case 1:17-cv-01427-tcb-wsd-bbm document 140 filed 03/27/18 … naacp 20180227... · 2018-03-31 ·...

1

UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF GEORGIA

ATLANTA DIVISION

NAACP, et al., ) ) Plaintiffs, ) )Case No: 1:17-CV-01427-

vs. )TCB-WSD-BBM ) BRIAN KEMP, in his official )capacity as Secretary of )State for the State of )Georgia, ) ) Defendant. )___________________________ )AUSTIN THOMPSON, et al., )CONSOLIDATED CASES ) Plaintiffs, ) )

vs. ) ) BRIAN KEMP, in his official )capacity as Secretary of )State for the State of )Georgia, ) ) Defendant. )

Deposition of Expert JOWEI CHEN, Ph.D

February 27, 2018

10:00 a.m.

BRYAN CAVE, LLP

One Atlantic Center, 14th Floor 1201 W. Peachtree Street, N.W.

Atlanta, GA 30039-3488

Marianne Vargas, CCR, CVR-M ______________________________________________________

VARGAS REPORTING SERVICES, INC. 5755 N. Hillbrooke Trace Johns Creek, GA 30005

678.458.4030

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

Case 1:17-cv-01427-TCB-WSD-BBM Document 140 Filed 03/27/18 of 166

2

VARGAS REPORTING SERVICES, INC.678.458.4030

APPEARANCES OF COUNSEL: ON BEHALF OF THE PLAINTIFFS GEORGIA STATE CONFERENCE OF THE NAACP, as an organization; LAVELLE LEMON, MARLON REID, LAURETHA CELESTE SIMS, PATRICIA SMITH, COLEY TYSON: JON M. GREENBAUM, ESQ. LAWYERS' COMMITTEE FOR CIVIL RIGHTS UNDER LAW 1401 New York Avenue, N.W. Suite 400 Washington, DC 20005-2124 202.662.8315 [email protected]

ON BEHALF OF THE PLAINTIFFS JAMAL BROOKS, an individual; AUSTIN THOMPSON, an individual; WAYNE SWANSON, an individual; DARRYL PAYTON, an individual; AUDRA CUNNINGHAM, an individual; SABRINA McKENZIE, an individual; JAMIDA ORANGE, an individual; ANDREA SNOW, an individual; SAMMY ARREY-MBI; LYNNE ANDERSON, an individual; and CORETTA JACKSON, an individual: ARIA C. BRANCH, ESQ. (via telephone) PERKINS COIE, LLP 700 13th Street, N.W. Suite 600 Washington, DC 20005-3960 202.654.6338 [email protected] ON BEHALF OF THE DEFENDANT THE STATE OF GEORGIA; and BRIAN KEMP, in his official capacity as Secretary of State for the State of Georgia: JOHN "JACK" PARK, JR., ESQ. STRICKLAND, BROCKINGTON, LEWIS, LLP Midtown Proscenium, Suite 2200 1170 Peachtree Street, N.E. Atlanta, Georgia 30309-7200 678.347.2208 [email protected]

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


3


INDEX OF EXAMINATIONS PAGE JOWEI CHEN

Examination by MR. PARK 4

Examination by MR. GREENBAUM 67

Examination by MR. PARK 69

INDEX OF EXHIBITS

DEFENDANT'S NO. DESCRIPTION PAGE 01 7Dec. 22, 2017 Original Report, Jowei Chen 02 51Jowei Chen, Reply Report

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


4


P R O C E E D I N G S

10:03 AM

MR. PARK: Would you please swear the

witness?

(The court reporter swore in

the witness.)

----------------------------

JOWEI CHEN, Ph.D,

having first been duly sworn, was examined and

testified as follows:

EXAMINATION

BY MR. PARK:

Q. Would you please state your name for the

record?

A. I'm Jowei Chen.

Q. And how do you spell your name?

A. J-O-W-E-I C-H-E-N.

Q. Thank you. And you're a Ph.D, right?

A. Yes, sir.

Q. Dr. Chen, have you been deposed before?

A. Yes, sir, I have.

Q. And so you know how they work?

A. Yes, sir.

Q. Just a couple of things. If I ask you a

question that you don't understand, would you please

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


5


feel free to ask me to rephrase it and I will attempt

to do so?

A. Yes, sir.

Q. And if you need to take a break at any

time, we'll be happy to accommodate you. I'd ask only

that you not take a break while a question is pending.

A. Yes, sir.

Q. What did you do to prepare for your

deposition?

A. I met with Plaintiffs' counsel last night,

yesterday, mostly yesterday evening. I reviewed my

two reports in this case, as well, and I reviewed the

rebuttal report written by Dr. Alford.

Q. When you prepared your report, did you

look at anything other than the electoral results and

the demographic data?

A. Which report are you asking about? Are

you asking about my original --

Q. Your original --

A. -- December report?

Q. Yes, sir.

A. Okay. We'll start there. In preparing my

December report, I certainly reviewed electoral data.

I also reviewed demographic data, census data. I also

reviewed various voter registration and voter history

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


6


turnout files. And I will just add that all of these

files were turned over in connection with my report.

So I had those files.

I also reviewed the deposition transcript

of Ms. Gina Wright. I believe there was another

document that was -- that outlined legislative

reapportionment criteria, I can't remember the precise

name of it, but it, too, was referenced in my original

expert report, and I reviewed that document, as well.

So those are all the documents I can

remember off the top of my head right now that I used

in connection for the original expert report. I

believe there were additional documents for my second

report.

Q. And do you recall any of those additional

documents that you reviewed for your second report?

A. Sure. There was an additional deposition

transcript that I reviewed and I referenced in my

second February report. If you'll just give me a

moment, I'll find -- I'll find the precise place where

I referenced it.

Q. Was it the deposition of Rob Strangia?

A. Yes, sir, that's correct.

So I reviewed Dr. Strangia's deposition.

I, again, reviewed Ms. Gina Wright's deposition, and

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


7


obviously I reviewed Dr. Alford's expert reports, as

well as the data files and computer codes turned over

in connection with Dr. Alford's rebuttal report.

I should add a few more things that I just

remembered. For my rebuttal report I, of course, also

considered the new election results of the

January 2018 Special House Election, and of course I

analyzed those, those election results in my

rebuttal -- sorry, in my reply report of February. So

those were certainly new documents that I didn't have

access to that I did not use in my original December

report.

Q. I'd like to hand you what's been marked as

Defendant's Exhibit 1, Dr. Chen. Can you tell me what

that is?

A. This exhibit is my original December 22nd,

2017 expert report in this case.

(Exhibit No. 01 was

marked/identified.)

BY MR. PARK:

Q. Does it appear to be a true and correct

copy?

A. Yes, sir.

Q. Turning your attention to Page 1 of this

report, you identify a number of cases. In Missouri

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


8


National Association for the Advancement of Colored

People vs. Ferguson-Florissant School District and St.

Louis County Board of Election Commissioners, which

side did you support?

A. I didn't support other side, sir.

Q. What was the nature of your expert report

in that case?

A. I was engaged by the Defendants to work on

a rebuttal report.

Q. Was a rebuttal report filed?

A. Yes, sir. I believe that report was

filed, so I worked on that rebuttal report. That's

what I was hired to do.

Q. What about René Romo, et al. vs. Ken

Detzner?

A. I was engaged by the Plaintiffs.

Q. What about League of Women Voters of

Florida and others vs. Detzner?

A. The Plaintiffs, sir.

Q. What about Raleigh Wake Citizens

Association vs. Wake County Board of Elections?

A. The Plaintiffs.

Q. What about Corrine Brown vs. Detzner?

A. I was hired by the Defendant Intervenors.

Q. And what was the nature of your work in

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


9


René Romo? Did you prepare --

A. In Romo --

Q. -- an expert report?

A. Sure. In Romo vs. Detzner I prepared a

few different expert reports.

Q. Did you testify?

A. I did not testify at trial.

Q. Were you deposed?

A. Yes, sir, I was.

Q. What about in League of Women Voters?

A. In The League of Women Voters vs. Detzner,

I believe that was the challenge to the Florida Senate

map, and, again, I was hired by Plaintiffs. I wrote

an expert report. I can't remember right now off the

top of my head if I was deposed or not, but I recall

that the case was somehow settled -- settled before

trial. It was somehow resolved. I'm not sure exactly

the nature of how it was resolved.

Q. What about Raleigh Wake Citizens

Association?

A. The Plaintiff's counsel, League of Women

Voters -- sorry, Plaintiff's counsel was the Southern

Coalition for Social Justice. They hired me. I wrote

an expert report, and I believe there was no

deposition. I testified at trial in that case.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


10


That's correct, right. I was not deposed. I

testified at trial.

Q. And then in Corrine Brown you worked for

the Defendant Intervenors; is that right?

A. Correct. Defendant Intervenors. I wrote

an expert report, and I was not deposed. I don't know

whether or not -- I don't believe that ever went to

trial. I'm not -- I couldn't tell you off the top of

my head.

Q. Okay. What about City of Greensboro and

others vs. Guilford County Board of Elections?

A. Okay. In the Greensboro case I was hired

by Plaintiffs, I was deposed, and I -- the case went

to trial in, I believe, February of 2017, and I

testified at that trial.

Q. Okay. What about Common Cause and others

vs. Robert Rucho?

A. Rucho, yes, sir. Sure, Common Cause vs.

Rucho, I wrote an expert report. I was engaged by

Plaintiffs. The deposition was, I believe, sometime

in the spring of 2017, I think in April of 2017. And

that case went to trial in October, just this past

October, and I testified at trial, as well.

Q. And then League of Women Voters of

Pennsylvania vs. Commonwealth of Pennsylvania?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


11


A. Sure. The Plaintiff's counsel hired me in

that case to write an expert report, as well as a

supplemental report later on. I was not deposed, and

I testified at trial this past December in that case.

Q. Have you ever worked with Maptitude?

A. I'm familiar with the software. And I've

used it just a little bit just to become familiar with

it, but I don't use it in my regular research process.

Q. In your familiarity with Maptitude, do you

know whether the Pending Changes box reflects a user's

choice?

A. Whether the -- I'm not sure. If you could

clarify for me what you mean by the "pending box."

Q. Do you know anything about the Pending

Changes box that are discussed in some of the

depositions?

A. I'm not exactly sure what you're referring

to.

MR. PARK: Can we go off just a second?

(An off-the-record

discussion was held.)

BY MR. PARK:

Q. Do you recall that in her deposition Gina

Wright answered a question about the Pending Changes

box by saying, "When I work on a plan, I use the

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


12


Pending Changes box, which is a feature in Maptitude

that shows how your numbers change as you select what

geography you're selecting"?

A. I don't recall reading that portion of Ms.

Wright's deposition, but I take your word for it.

Q. Okay. But you can't tell me what -- When

you've used Maptitude, you haven't relied on the

Pending Changes box at all?

A. Again, I would just reiterate that I don't

use this -- I don't use Maptitude as part of my normal

research process. I accept that you're describing a

feature of Maptitude, and I don't use it very commonly

as part of my normal research process, so I couldn't

give you first-hand information about exactly what

that box looks like.

Q. I’d like to, if you would, on page -- if

you would turn to Page 4 of your report.

A. Yes, sir.

Q. And you describe your methodology. You

say that you used ecological inference?

A. Yes, sir.

Q. And you use a procedure known as Maximum

Likelihood Estimation, combined with Duncan and Davis'

1953 Method of Bounds?

A. Yes, sir.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


13


Q. That's not a Bayesian approach, is it?

A. It's a little bit different.

Q. Is the procedure you're using like what

King describes?

A. It is exactly what Gary King describes.

It is -- In fact, I used his software package, his

computer code. So it is a package or set of computer

codes, code files, called EI, and it's a program for

our programming language. And professor King makes

that code, that computer code readily available, and

that is what I used for all of my ecological inference

calculations in this report, as well as in my February

response report.

Q. When you reviewed the results, how many

runs of data did you do?

A. If I could ask you to clarify your

question.

Q. Well, when you put your data in and you

run it through the software, that would be one

iteration, right?

A. You're saying if I analyze one election

and run it once, that would be one iteration. That's

what you want to refer to by "iteration"?

Q. Well, in order to come up with your

estimates in, say, Table 2, for example, in the 2012

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


14


House election, EI on Page 6 of your report, just

turning to it, you report 98.2 percent of -- And is

that a single run through the software, or is it --

how many -- how many runs through the software do you

do to get your --

A. Sure.

Q. -- confidence level?

A. Each coefficient that you see here on

Table 2, it represents one run of the software. And,

again, that software was the EI package that I

referred to a moment ago.

Q. Given that you say each of these

coefficients represents one run of the software, is

that -- did I understand that correctly?

A. Yes, sir, one run of the software.

Q. You don't burn any, do you?

A. If I could ask you to define what you mean

by "burn."

Q. Do you discard any of your runs?

A. I do not run the software and then discard

the run that I just did. That would not be part of my

normal research practice, or any social scientist's

normal research practice.

Q. If you turn back to Page 2 of your --

Page 3 of your report, in the middle of the paragraph

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


15


that starts "finally," there's a sentence toward the

end: "I found that the Legislature's primary map

drawer for the 2015 Plan had access only to racial

data, but not partisan data, at the sub-precinct

level;" is that right?

A. Let me just orient myself to where you

are. You're on Page 2?

Q. Page 2, Paragraph "finally," then the line

starts "account," and the sentence starts toward the

end of it. So it's about seven or ten lines down.

A. You're on --

Q. On Page 3.

A. Okay. You're on Page 3.

MR. PARK: Let's go off.

(An off-the-record

discussion was held.)

MR. PARK: Let's go back on the record.

BY MR. PARK:

Q. On Page 3 in the paragraph that starts

"finally," you have the sentence that says, "I found

that the Legislature's primary map drawer for the 2015

Plan had access only to racial data, but not partisan

data, at the sub-precinct level."

A. Yes, sir, I see that sentence.

Q. Who do you understand to be the

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


16


Legislature's primary map drawer?

A. When I wrote this original report -- and

you're referring to me -- you're referring me to a

sentence here from my December report -- I had, at

that time, only reviewed the deposition transcript of

Ms. Gina Wright. And so I was basing this sentence on

Ms. Gina Wright -- Ms. Wright's deposition transcript

regarding her process in drawing the map.

Q. And that was based on your understanding

of Ms. Wright's testimony?

A. It was -- Yes, sir. It was based on my

reading of Ms. Wright's deposition.

Q. I’d like you to turn to Page 5, pages 5

and 6 of your report. There are tables that address

your ecological inference and ecological regression

estimates for House District 105 and 111; is that

right?

A. Yes, sir.

Q. Looking at the ecological regression

estimates for Black for each of the three elections in

Table 1, you report 100 percent with 100 percent

confidence limits.

A. Yes, sir.

Q. Does that mask a result that was over

100 percent?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


17


A. How we report ecological regression

estimates is that if the estimate is over 100 percent,

then you report 100 percent, or you round it down to

100 percent, because that is logically the maximum --

the maximum coefficient that can explain the data. So

I wouldn't say that it masks, but certainly that is --

that is what we do when the regression coefficient is

over 100 percent.

Q. Do you recall how high the regression

coefficient was for the 2012 House election?

A. You're asking for the Table 1 results here

on the first row?

Q. Correct.

A. I don't recall the precise number, but

obviously, as I explained in my report, the regression

estimate would have been over 100 percent, and

logically reported as 100 percent.

Q. And in Tables 1 and 2 you're looking

exclusively at any part Black and non-Black; is that

right?

A. In Tables 1 and 2 when I conduct the EI

and the ER estimates, you're asking about those

tables?

Q. Yes.

A. And your question was whether Black

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


18


includes any part Black; is that right?

Q. Well, does Black include any part Black?

A. Okay. I'll just explain. For the purpose

of creating Tables 1 and 2, the data that I used to

identify the racial composition of voters came from

the voter registration files and the turnout files.

Those files are not as detailed as the Census Bureau's

breakdown of individuals' racial composition. So we

don't know, for example, because voters are not given

the opportunity to identify -- to identify as

precisely their racial composition.

So here the identification of Black versus

non-Black is anybody who checked the box for

African-American as opposed to checking one of the

other boxes.

Q. And everyone who checked the box other

than African-American is in the non-Black?

A. That's correct, sir.

Q. On Page 7 of your report --

A. Yes, sir.

Q. -- in the Paragraph No. 2 you say,

"Non-African-American voters within House District 105

as drawn under the 2012 Plan became somewhat more

likely to favor a Black Democratic House candidate in

November 2016."

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


19



Q. And is that -- Do you draw that conclusion

from Table 1?

A. If you could just give me a moment to

orient myself. (Reviews document.)

Yes, sir. I was describing the results

that were reported in the table.

Q. And Table 2 doesn't really show that, does

it? You can't draw the same conclusion about House

District 111 from Table 2?

A. What specifically are you asking me about

Table 2's results?

Q. That non-African-American voters within

House District 111, as drawn under the 2012 Plan,

became somewhat more likely to favor a Black

Democratic House candidate in November of 2016

compared to previous elections.

A. (No response.)

Q. If you substitute 111 for 105 with respect

to -- and ask that question with respect to Table 2,

it doesn't show that, does it?

A. I did not draw that conclusion regarding

111.

I just want to be as complete as possible,

and if you'll just give me a moment --

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


20


Q. Sure.

A. -- let me just review this section of my

report and --

Q. Please take your time.

A. -- make sure that I didn't answer your

question in my report.

(Reviews document.) I just wanted to make

sure that I didn't directly say anything about it in

the text of my report, and it appears that there's

nothing in the text of my report that directly answers

your question. I simply reported the numbers that I

do for Table 2 in House District 111.

Certainly what those numbers show to me is

that there is a bit of an increase in

non-African-American support for the Democratic

candidate between 2014 to 2016, but obviously not such

an increase in 2012 compared to 2016.

So I just wanted to answer that using

Table 2 of my report; answer your questions as

completely as possible.

Q. On Page 7 when you talk about the increase

in the African-American population in House District

105 and you illustrate that in Table 3 -- this is your

text -- which shows that African-Americans comprise

35.2 of the election day turnout in November 2012,

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


21


35.7 percent in November 2014, and 37.0 percent in

November 2016, counting only voters who reside within

the 2012 Plan's House District 105 boundaries; is that

right?

A. Yes, sir, I see that.

Q. And the African-Americans, they weren't a

majority in the district, were they?

A. You're asking me if those numbers that you

just read out comprise a majority of the district.

Q. Correct.

A. Well, obviously the numbers that you just

read out are under 50 percent.

Q. And the same is true for the

African-American population in House District 111; is

that right?

A. It's addressed in that next sentence, and

I -- as I stated in the report, the numbers were 36.1

percent -- the comparable numbers were 36.1 percent in

2012, 37.6 percent in 2014, and 40.3 percent in 2016.

And obviously, again, I confirmed that those numbers

are under 50 percent.

Q. I'd like to turn your attention to Page 10

of your report, your initial report.

A. Yes, sir.

Q. In the first paragraph, talking about

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


22


House District 105, the prediction that non-Black

support for a Black Democratic candidate increased

from 21.4 percent in November 2012 to 25.2 percent in

November 2016.

A. Yes, sir I see that.

Q. And you state, "This increase is partially

attributable to an increase in the Hispanic and Asian

shares of the non-Black population of the electorate

in House District 105 as illustrated in Table 7..."


Q. "...as Hispanic and Asian voters were more

likely than non-Hispanic White voters to support Black

Democratic candidates." It's the last sentence --

A. Yes, sir --

Q. -- in that paragraph.

A. -- I see that sentence. I see that

sentence you're referring to.

Q. Have you done any ecological inference or

ecological regression runs to see about the political

cohesion of Hispanic voters in House District 105?

A. No, sir, I did not use EI or ER to

specifically answer that question.

Q. Did you try?

A. Did I try doing what?

Q. Using EI or ER to see about the cohesion

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


23


of Hispanic voters in House District 105.

A. No, sir, I did not.

Q. What do you rely on, then, to say that

they're more likely to vote Democrat than non-Hispanic

Whites?

A. Okay. You're asking me what the basis of

that sentence is in the paragraph that we were just

reading?

Q. Yes, Dr. Chen.

A. Sure. Okay. I am purely relying on my

general knowledge of political behavior in the South.

If you would just give me a moment, let me

just make sure I haven't -- I've given you a complete

answer. So if you'll just give me a moment. (Reviews

document.)

Okay. I want to give as complete an

answer as I can. So, again, the basis for my answer

is my general knowledge -- for my answer for how I

came to the statement that I -- that we were just

talking about in that final sentence of that paragraph

is my knowledge of political behavior in the South

generally.

What I saw here specifically, and why I

reference Table 7 is because, as I stated in that

sentence and as I can see here in Table 7, there was a

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


24


noticeable increase in the Asian and Hispanic shares

of the non-Black portion of the electorate in House

District 105 between 2012 to 2016.

Now, logically I combined that information

with what we just discussed a moment ago, which is

that I can see in Table 1 that there was a noticeable

increase between 2012 to 2016 in the ecological

inference estimate of the percentage of non-Blacks

supporting the Democratic candidate.

So I put those two pieces of information

together and came to the conclusion that I stated in

that final sentence that these Hispanic and Asian

voters were more likely than non-Hispanic White voters

to support Black Democratic candidates. That confirms

my general intuition and knowledge about political

behavior in the South.

Q. Turning back to Table 1, what part of the

non-Black estimate would be attributable to Hispanic

and Asian voters? How do I see that in Table 1?

A. Okay. You're asking me in Table 1, what

part of the EI estimate for non-Black voter support

for Democratic candidates would be attributable to --

Q. Hispanics or Asians.

A. -- to Asian or Hispanic voters? And my

answer is that I, in Table 1, analyzed all non-Blacks

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


25


as a group.

Q. And that could include White Democrats,

couldn't it?

A. Oh. You're asking me if non-Blacks

include White Democratic voters?

Q. Yes, sir.

A. Obviously they're always -- Anytime you

have a group that includes non-Blacks, there are going

to be some number of White Democratic voters.

Q. And the census has a category of Others,

doesn't it?

A. You're asking me if the census allows a

respondent to check a box for Other; is that right?

Q. Yes, sir.

A. I think that's right. I'll take your word

for it. I couldn't just tell you off the top of my

head what that Other category specifies, but I'll take

your word for it.

Q. Well, if you'll turn to Table 7, the last

box down there are Others or Unknown, correct, for

turnout on Page 14?

A. Okay. You're asking me now about Table 7,

and you're asking me what the Other or Unknown line

represents?

Q. Right.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


26


A. I just want to be clear. This is not --

This is not census data. This is voter registration

and voter history turnout data. So what I'm reporting

here in Table 7 are the racial identifications of

individuals who turned out in these various elections.

So this is not census data, but that

bottom line indeed represents people who checked Other

or did not check a box at all.

Q. Well, going back to Table 1, would the

Other be included in the non-Black?

A. Yes, sir. Non-Blacks includes everybody

who did not check the box for African-American.

Q. And we don't really have any way of

knowing how they voted, right?

A. Who are you referring to?

Q. The Others and the Unknowns.

A. Okay. You're asking me if I can -- if I'm

able to draw any conclusions about who specifically

the Others are voting for, right?

Q. Yes, sir.

A. Okay. And the answer is that generally I

-- let me turn off my phone.

Okay. Your question is whether we can say

anything, or whether I can say anything about how the

Others or Unknowns voted, and my answer is that I

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


27


analyzed them as part of my EI estimates by grouping

voters into Black versus non-Black, so in fact we can

say something about those -- about those individuals.

Q. But none of the voting behavior of

Hispanics or Asians or Others or Unknown is reflected

in an ecological inference or ecological reflection

other than on Table 1; is that right?

A. They are included in my analysis, and the

behavior of those individuals is, in fact, reflected

in my analysis in Table 1.

Q. But they don't -- none of them has an

independent ecological inference or an ecological

regression.

A. What exactly, sir, do you mean by

"independent ecological inference" or --

Q. Well, you didn't --

A. -- "ecological regression"?

Q. Well, you told us you didn't do an

ecological reference or ecological regression for

Hispanic voters in -- run for Hispanic voters in House

District 105, correct?

A. I included all Hispanic voters in my

analysis, so my results do, in fact, account for the

behavior of Hispanic voters.

Q. But you didn't try to do Hispanics alone,

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


28


correct?

A. What do you mean by "try to do Hispanics

alone"? I just want to understand your question.

Q. If there were enough Hispanics, you

could -- and they were properly distributed, could you

try to identify the voting behavior of Hispanics as a

group with using ecological inference or ecological

regression?

A. My answer is that my ecological inference

and ecological regression analyses do, in fact,

account for the behavior of Hispanics.

Q. In the non-Black portion of Table 1?

A. That is correct.

Q. And that's the only place it's accounted

for in this report?

A. That's not the only place where I've

looked at Hispanic voters, but I think you're -- you

know, I think you're asking in the context of

conducting ecological inference estimates here. I

think you're trying to ask whether what I reported in

Table 1 and Table 2 represents the approach that I

used, and it is. That is exactly how I conducted EI

and ER analysis.

Q. If you'll look at Table 8 of your report.

A. Yes, sir.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


29


Q. The demographics of House District 111 as

reflected in the turnout show a pretty small Hispanic

and Asian turnout; is that right?

A. They show specifically that between

2012 to 2016 within the old boundaries, within the

2012 Plan's boundaries for House District 111, that

Hispanics increased from 1.8 percent to 2.4 percent,

and that Asians increased from 0.9 percent to

1.3 percent. So I'd say that comparatively those are

quite significant increases, but the point of this

table is to report exactly numerically what those

increases are.

Q. But the demographics of Henry County are

different from the demographics of Gwinnett County,

right?

A. I accept that.

Q. In Henry County it's more African-American

than White, isn't that right, than Gwinnett County?

A. I did not specifically study the

comparative demographics of Henry and Gwinnett

counties as a whole, so I can't answer that question

accurately.

Q. Well, it's even --

A. I'll take your word for it.

Q. It's even reflected in the difference

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


30


between Tables 7 and 8, isn't it?

A. Well, what I specifically studied in

Table 7 and Table 8 were the demographics of the

election turnout in House Districts 105 and 111

specifically, so that is what I can report in these

tables.

MR. PARK: Let's take a break for a couple

of minutes.

(Off the record at 10:54 AM)

--------------

(On the record at 10:59 AM)

BY MR. PARK:

Q. When you say you rely -- for your

conclusion about the behavior of Asian and Hispanic

voters in House District 105, you're purely relying on

your general knowledge of political behavior in the

South, or something like that. Is that what you said?

A. I gave a bit of an a longer answer than

that, so I'm happy to revisit that answer again, if it

would help you out. Would you like me to do that?

Q. When you talk about "the South," what do

you mean?

A. I mean the South of the United States.

Q. Romo vs. Detzner was in Florida, right?

A. Yes, sir.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


31


Q. And League of Women Voters vs. Detzner was

also in Florida?

A. Yes, sir.

Q. And Raleigh Wake was in North Carolina?

A. Yes, sir.

Q. And Corrine Brown was in Florida?

A. Yes, sir.

Q. The City of Greensboro was in North

Carolina?

A. Yes, sir, it was.

Q. And Rucho was in North Carolina?

A. Yes, sir.

Q. Have you done any work in Georgia?

A. Have I -- You're asking if I've worked on

a redistricting case --

Q. Yes, sir.

A. -- in Georgia as an expert witness before,

and the answer is no, outside of this case. No.

Obviously my work as an expert witness is not the

extent of my expertise as a political scientist. I

was a political scientist long before I became an

expert witness in any redistricting cases.

Q. Well, what goes to your general knowledge

of political behavior of Asians and Hispanics in

Georgia? Where do you draw your conclusion -- What do

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


32


you use to draw that conclusion?

A. Okay. You're asking me what is the basis

of me saying in this sentence that Hispanic and Asian

voters were more likely than non-Hispanic White voters

to support Black Democratic candidates, right?

Q. In Georgia.

A. Okay. So I think you're asking me the

same question that you asked me some time ago, so I'm

going to revisit my answer, because I think this is

responsive to your question.

So the basis for my answer is, as I said,

I looked at Table 7 and I looked at Table 1 and I

noticed a couple of things. The first thing that I

noticed is that in Table 7 I saw that the percentage

of the district that's Asian-Hispanic has increased in

House District 105 between 2012 to 2016 within the

boundaries -- within the boundaries of the 2012

districts. And then I saw the EI numbers that I

reported and discussed some time ago in Table 1.

I put those two things together, and that

confirmed -- that confirmed my general knowledge about

voting behavior of various minority groups, which is

that Hispanics and Asians don't vote quite the same as

non-Hispanic Whites and Asian voters, in general. And

it confirmed that these demographic changes had, in

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


33


fact, been part of the partisan -- the results

reported in Table 1, which were that there has been a

slight change in the percentage of non-Whites that

support Democratic candidates.

Q. How many times have you been to Georgia?

A. How many times have I been in the state of

Georgia?

Q. Yes, sir.

A. Well, too many to count. I’ll tell you

that there's no way I could possibly count. I grew up

about 1 mile from the border, the northern border of

Georgia, just outside of Chattanooga, Tennessee. So

Atlanta was the closest metropolitan -- large

metropolitan area to where we lived, so certainly my

family and I would've gone on trips down to Atlanta,

since that's where all the large museums, supermarkets

and other cultural amenities were. We certainly

would've gone down to Atlanta, I would say, at least

once every one or two months when I was growing up.

So I grew up for 18 years just outside of

Chattanooga, and so it's easily in the hundreds the

number of times that I've been into the state of

Georgia. Actually, I'm pretty sure it's in the

thousands, because I recall that when I was in

elementary school our school bus actually had to make

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


34


a detour into Georgia to pick up some students every

day. So every day I would have traveled into the

state of Georgia.

But obviously, as I said, I grew up very

close to this state.

Q. How many times to Gwinnett County or

Lawrenceville?

A. How many times have I been in Gwinnett

County?

Q. Yes.

A. Well, I love going to see the Gwinnett

Braves. They have a really nice baseball stadium out

there, and so I've been out there at least a few times

whenever I'm in conferences in the area. So I really

couldn't tell you off the top of my head.

Q. What about McDonough in Henry County?

A. How many times have I been to McDonough or

Henry County?

Q. Yes.

A. I recall going to Henry County when I was

-- when I was in high school. One time there was a

school trip down there. I really couldn't tell you

off the top of my head, but I recall being to Henry

County one time.

Q. Are you saying that Hispanic and Asian

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


35


voters behave the same way in Georgia that they do in

Florida and North Carolina?

A. That's not a question I analyzed.

Q. When you were doing the work in North

Carolina, were you looking -- did you look at the

behavior of Hispanic voters?

A. I'm just going to have to ask you to be

more specific. Which work in North Carolina are you

referring to?

Q. City of Greensboro or Common Cause vs.

Rucho?

A. Okay. So let's take those one at a time.

You're asking me in the expert work that I did in

Greensboro --

Q. Yes, sir.

A. -- whether I looked at the voting behavior

of Hispanic voters?

Q. Yes.

A. There certainly are a significant number

of Hispanic voters within the city limits of

Greensboro, and I had to analyze the partisan behavior

of those voters as part of my general analysis of

partisan voting patterns in Greensboro. So Hispanic

voters were certainly part of my analysis in

Greensboro.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


36


Q. Were they reflected in the non-Black

portion of an ecological regression chart, or were

they reflected in a Hispanic portion of the chart?

A. What exactly are you referring to? You're

asking about my report here in this case?

Q. In Rucho.

A. Oh, in Rucho?

Q. No. In Greensboro. In Greensboro.

A. Okay. You're asking me about Greensboro.

And what specifically -- you're asking me --

Q. Looking at Table 1 you’ve got Black and

non-Black. In Greensboro, did you -- were your

ecological inference runs based on Black and

non-Black?

A. And my answer is that I did not do an

ecological inference run in my expert report in

Greensboro.

Q. How did you analyze voting behavior, then,

in Greensboro?

A. How did I analyze voting behavior in my

Greensboro expert report?

Q. Yes, sir.

A. I looked at past election results.

Q. But you didn't use ecological inference or

ecological regression to make any inferences about the

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


37


voting behavior of various blocks?

A. You're asking about my expert report in

Greensboro, right?

Q. Yes.

A. And the answer is no, I did not use

ecological inference.

Q. What about Rucho? Did you do ecological

inference and ecological regression to determine the

behavior of blocks of voters?

A. I did not do so in my expert report for

the Common Cause vs. Rucho case.

Q. What about in Romo? Did you do -- was the

behavior of Hispanic voters at issue in Romo vs.

Detzner?

A. I can't tell you if that was an issue that

was litigated. I can tell you what's in my expert

report, though.

Q. What was in -- What was the general nature

of your expert report? Did you do ecological

inference/ecological regression analysis?

A. I did not, sir. My report in Romo --

You're asking about Romo vs. Detzner, right?

Q. Yes, sir.

A. In that report I did not present an

ecological inference analysis.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


38


Q. When you speak about your general

knowledge of political behavior in the South, are you

relying on any publications?

A. I am relying upon my education and

training as a political scientist, which included

courses on political behavior in the United States.

Certainly in those courses we would have read quite a

few publications.

Q. Can you remember any of the publications

that address the voting behavior of Hispanic voters?

A. I couldn't tell you off the top of my

head, but I'm certainly confident in generally saying

that that is a topic that is covered in political

behavior.

Q. What about Asian voters?

A. Same answer there. I couldn't tell you a

publication off the top of my head, but I'm confident

in saying that that is a commonly-studied topic in

political behavior.

Q. When you talk about the Reapportionment

Office and its treatment of sub-precinct behavior, as

you do on Page 33 of your report and Page 34 of your

report --

A. Yes, sir.

Q. -- you're relying on your understanding of

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


39


Ms. Wright's deposition; is that correct?

A. I am relying on Ms. Wright's deposition,

but in this section of the report I'm also drawing on

my own expert knowledge of census data, census

block-level data, precinct election data, and the

relationship between those different geographic units.

Q. Well, do you recall in Mr. Strangia's

deposition him saying that the software allocated

voters into split precincts by the voting age

population of the splits? Do you recall whether he

testified that way?

A. I recall that Mr. Strangia stated in his

deposition that his office would have estimated, using

voting age population, the partisanship of census

blocks within split -- or within any precinct. And

specifically what was important for me for my

understanding of its process was that Mr. Strangia

confirmed Ms. Wright's testimony that the

Reapportionment Office had no way to distinguish the

relative partisanship of different census blocks

within split precincts. It struck me that that is --

that made sense to me, that they had that limitation.

And I read that part of both of their

depositions, and they seemed to be saying the same

thing. And I wanted to review those portions of the

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


40


depositions to understand their process and what data

they had available at the block level within split

precincts, and that made sense to me.

Q. In essence, they used voting age

population as the basis for an estimate of who is in

each part of the split, right?

A. Well, I just want to make sure you

understand what I had gleaned from Mr. Strangia's

deposition, as well as from Ms. Wright's deposition.

It's not that they were estimating the relative

partisanship of census blocks within a split precinct.

They had no way -- Both of them testified they had no

way to distinguish the partisanship, the relative

partisan meanings of various census blocks within a

split precinct; therefore, they simply assumed, or

their software assumed that the partisanship of all

census blocks within a split precinct is exactly the

same, is perfectly uniform throughout all of the

census blocks within a split precinct. That's not

really estimating the partisanship of different census

blocks within a split precinct. That's simply

assuming uniformity.

And maybe that's what you're referring to,

but I just wanted to make sure that I was clear that

that is information that I gleaned from those two

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


41


depositions.

Q. Well, do you recall the portion of

Mr. Strangia's testimony or deposition where he said

the lowest level of geography for election results is

the precinct level of geography, okay? So that needs

to be allocated down to the block level of geography

based on voting age population?

A. I recall him discussing that topic. If

you'd like to put that deposition transcript in front

of me, I would be happy to confirm that he said that,

but I don't have that transcript in front of me here.

Q. Well, do you know whether -- do you recall

whether he said whether other states use a similar

formula?

A. I recall there was some discussion about

them. As I said, I generally recall the discussion.

I just can't confirm the precise words that he used

because I don't have the same transcript that you have

in front of you.

Q. Did you recall that Ms. Wright testified

that she had political data at the block level also?

(Reading) "It's an estimate. When we bring in our

file from the Secretary of State's Office, it is

completely accurate to the precinct. But when we

allocate that data from them to the block level, it

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


42


estimates that figure to the blocks based on the

percentage and proportions of the population"?

MR. GREENBAUM: Objection. Asked and

answered.

Go ahead. Answer the question one more

time.

BY MR. PARK:

Q. Do you recall her testifying to that?

A. My answer is the same, which is that I

can't confirm the precise words because I don't have

the same deposition transcript that you have in front

of you. I don't have a copy of it here with me right

now in front of me, but I obviously remember them

discussing these general topics in describing their

process.

And if you'd like to put the transcript in

front of me, I would be happy to help you understand

exactly what they're talking about.

Q. You would be happy to help me understand

what they're talking about?

A. I would be happy to explain to you the

process that Ms. Wright and Mr. Strangia are

describing, if you'd like to put the transcript in

front of me.

MR. GREENBAUM: And I think, in fact, you

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


43


already have described the process.

BY MR. PARK:

Q You talk about some of the precinct

changes in your report with respect to House District

105, and one change that was made in 2015 was Harbins

C was added to House District 105; is that right?

A. Yes, sir.

Q. Is Harbins C a rural area?

A. If you'll just give me a moment and let me

review my report. (Reviews document.)

Q. If you'll look at Page 24 of your report.

A. Yes, sir.

Q. In the bottom paragraph you say that House

District 105, a precinct Harbins C, was added to it?

A. Yes, sir.

Q. Do you know whether Harbins C is rural or

urban?

A. I’m just going to have to ask you to

define what you mean by those terms.

Q. Do you know whether it's Republican?

A. Do I know whether Harbins C is a

Republican precinct?

Q. Mainly Republican precinct.

A. Let me see if that's reported in my

report. So if you'll just give me a second to look up

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


44


-- to orient myself in my report.

(Reviews document.) I can't find the

place in my report where I specifically reported the

partisan composition of that particular precinct, but

I'll take your word for it.

Q. On Page 25 you talk about moving a portion

of Lawrenceville M out of 105 into 104?

A. Yes, sir.

Q. And that portion contains a BVAP of

45.6 percent?


Q. Does that 45.6 percent BVAP mean that the

non-BVAP population would be some 54.4 percent?

A. You're asking me if --

Q. Isn't that a majority non-BVAP move?

A. Okay. You're asking me of this portion of

the Lawrenceville M precinct is more than 50 percent

non-African-American.

Q. Yes, sir.

A. And I see -- I think I understand the math

you're asking me, and I confirm that, sure, it's a

little bit over 50 percent non-African-American. I

think what I refer to in this portion of my report is

that clearly this portion of the Lawrenceville M

precinct is more heavily African-American than the

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


45


rest of the district. I think that's all I was

referring to here. But I confirm your math.

Q. If you turn to Table 15 --

A. Yes, sir.

Q. -- this shows that by the 2010 Census,

voting age population, the Black population,

proportion of the population inside House District 105

boundaries was 29.8 percent; is that right?

A. If I could just ask you to repeat that

question before. I want to make sure I follow.

Q. All right. We're looking at 2010 Census

VAP by race within and outside House District 105 for

the 2015 Plan in split precincts; is that right?

A. Yes, sir. You're looking at the bottom

table.

Q. And in the totals by race for 2016

election turnout within, it's 29.8 percent; is that

right?

A. You're asking about the number --

Q. The total by race --

A. -- that was reported --

Q. -- for the Black proportion.

A. Sure. You're asking about the number that

I reported in the bottom row where I'm looking at --

where I'm looking at the House District 105 boundaries

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


46


in the 2015 Plan, and I confirm that it is, in fact,

29.8 percent that I reported as a Black proportion.

Q. And outside it was 30.4 percent Black

proportion?

A. Now you're asking about the right column,

and I looked at the portions of the House District 105

split precincts outside of the district boundaries

themselves, outside of the House District 105

boundaries, and I reported that the Black proportion

was 30.4 percent. So, yes, I reported that number.

Q. Do you know why there was a lower turnout

within House District 105 boundaries than without in

2016?

A. Do I know why there was a lower turnout

within than without? I'm not sure that I've actually

reported that there was a lower turnout.

Q. A lower percentage of turnout?

A. I --

Q. Why would the minority population turn out

at a slightly lower rate within?

A. Okay. You had asked me a moment ago

whether -- why there was a lower turnout, and I'm not

sure that I reported there was, in fact, a lower

turnout. I think what you're trying to ask is

specifically about the Black proportion of the

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


47


electorate, right?

Q. Correct.

A. Okay. I appreciate that. So you're

asking me why the total VAP within the House District

105 boundaries in these split precincts is

29.8 percent, and why the portion outside is

30.4 percent. Have I got your question right?

Q. Correct.

A. Okay. Sure. What I did in calculating

those numbers was I aggregated up across these three

split precincts, the three split portions that were

within House District 105 boundaries, and then I

aggregated up in the right column the House District

105 portions of these three split precincts outside.

Now, the aggregate racial composition of

those three precincts put together means absolutely

nothing. It's absolutely irrelevant for the purposes

of analyzing how these three individual precincts were

split.

And when you take these three split

precincts and aggregate up the racial composition,

that tells you a number that is pretty meaningless.

It's telling you something about an arbitrary three

chosen precincts, but you're not even analyzing one at

a time. You're just reporting something about the

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


48


total racial composition of these three chosen

precincts, the ones that were chosen to be split.

So it's a pretty arbitrary choice of three

precincts, and that is not the basis on which I

analyze split precincts in my report.

Q. With respect to House District 111, do you

know where the then incumbent Strickland lived?


head.

Q. One of the precincts -- On Page 361, one

of the split precincts is Tussahaw.


Q. Do you know whether Tussahaw performs

pretty strongly Republican?


head.

Q. Would that be a reason to put Tussahaw in

the district of a Republican, or that portion of the

Tussahaw precinct into the district of a Republican?

A. You're asking me if Tussahaw's strong

Republican performance would be a reason to put that

split portion of Tussahaw into House District 111?

That's your question?

Q. Correct.

A. Okay. And my answer is that I accept your

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


49


representation that there may be a portion of Tussahaw

that is strongly performing for Republicans. I accept

that because you've told me that. I think that's what

you're telling -- you're asking me to accept.

My answer to your question, though, is

that from my reading of Ms. Wright's deposition,

Mr. Strangia's deposition, it would be impossible for

the Legislative Reapportionment Office to know the

relative partisan performance of one split precinct of

the Tussahaw precinct as opposed to another split

portion of that same precinct, because they both

testified that they lacked the granular data necessary

to compare the relative partisan performance of two

split portions of the same precinct.

I also know from their depositions that

the only data they would've had to distinguish between

the split portions of a precinct like Tussahaw would

have been census demographic data, including racial

data.

So from the perspective of the Legislative

Reapportionment Office, it would not have been

possible to do what I think you're suggesting or

asking about.

Q. Just completely impossible? Is that your

testimony?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


50


A. It's my testimony, given my understanding

of the data available to the Reapportionment Office.

Now, I wouldn't go so far as to say it's completely

impossible, because what I found in my report actually

does make it possible to do a version of what you're

suggesting.

My report concluded that it is, in fact,

possible in House District 111 to use racial data to

effectively get at the kind of partisan consideration

that I think you are asking about.

So it certainly would have been possible

to use the available racial data that Ms. Wright and

Mr. Strangia had to effectively accomplish the

partisan manipulation of a splitting of the precinct

that I think you are suggesting in your question.

Q. Do you know whether there are White

Democrats in Henry County?

A. Do I know whether there are any White

Democrats at all?

Q. Yes, sir.

A. In Henry County? I assume there are.

Q. Do you know whether there's a significant

block of White Democrats in Henry County?

A. Well, I analyzed that question

specifically with respect to 111. And obviously, as

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


51


I've discussed earlier, those specific numbers are

reported in Table 2 of my report. So I was able to

answer that question only in the context of House

District 111, but I was able to answer it a bit more

precisely.

Q. Let's take a look at your reply report.

Let's mark this as Defendant's 2. Is that a copy of

your reply report?

A. Yes, it is.

Q. Does it appear to be a true and correct

copy?

A. Yes, sir, it does.

(Exhibit No. 02 was

marked/identified.)

BY MR. PARK:

Q. If you would turn to Page 15.

A. Yes, sir.

Q. In the paragraph that starts "moreover,"

you say, "The key issue I sought to analyze was

whether Black voters generally support Democratic or

Republican candidates."

A. I just want to try and find where you are

on the page here.

Q. The fourth line or the fifth line in the

middle, "The key issue"?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


52


A. "The key issue I sought to analyze." Yes,

sir, I see where you are. Go ahead.

Q. And you go on to say that the Democratic

candidate for November 2014 House election was

Caucasian, and that House election exhibited

substantially the same level of racially-polarized

voting as the other two HD 111 House elections I

analyzed in my original expert report."

A. Yes, sir, I see where I said that.

Q. And so do I correctly understand that

Black voters voted for that White Democratic candidate

in November 2014 in the House District 111 election?

Is that what you're saying?

A. If I could ask you to repeat the question.

Q. Are you saying that Black voters in House

District 111 in the November 2014 election supported

the White Democratic candidate?

A. Okay. You're basically asking me what the

race of the Democratic candidate was and whether or

not Black supporters supported that candidate at a

relatively-high rate, and the answer is yes. I, of

course, said here that the Democratic candidate is --

Q. You said he was Caucasian, right?

A. Yes, sir. I agree with you. The

candidate is -- the candidate in 2014 was Caucasian,

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


53


and I presented EI numbers for Black support of the

Democratic candidate that were in line with what I had

reported for 2012 and 2016.

Q. In his report Dr. Alford suggested that

Black voters generally supported Democratic candidates

in other races, and I know you disagree with his

analysis, but is that -- do you think that's correct,

that Black voters support Democratic candidates in,

say, U.S. Senate elections, without regard to whether

the candidate is White or Black?

A. Well, I analyzed the question with respect

to the House district elections, because those were

the most probative elections. And as we've just been

talking about, clearly that includes one election in

2014 where Blacks supported a Caucasian Democratic

candidate. That's the extent of my analysis.

Q. Is that the extent of your -- That's the

extent of your analysis? Are you aware of other

consistent instances of that happening?

A. Okay. You're asking more generally --

Q. Yes, sir.

A. -- beyond my expert report here?

Q. Yes, sir.

A. Sure. I'm happy to confirm that indeed

there are examples where Black voters have voted in

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


54


favor of a White Democratic candidate. I confirm that

as a general -- as a general matter.

Q. Do you think that would include Jim

Barksdale when he ran for U.S. Senate --

A. I did not --

Q. -- here in Georgia?

A. I did not analyze the results of that

election beyond looking at the data and analysis in

Dr. Alford's reports. But it's not one that in my

report, for my expert report, that I analyzed.

Q. What about Stephen Oppenheimer in 2012 for

Public Service Commission District 3?

A. I have the same answer there. I --

Q. That it's not in your report?

A. Sure. I have the same answer. I also

don't have Dr. Alford's report in front of me right

now, so I can't tell you the race of those various

candidates. But I take your word for it that that was

a White Democrat, and I'm just giving you the same

answer that I did not analyze it in my expert report.

Q. In his rebuttal report, again on Page 15,

Dr. Alford talks about the portions of the split

precincts inside --

MR. GREENBAUM: Are we talking about

Dr. Alford's report now, or Dr. Chen's report?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


55


MR. PARK: Dr. Chen's report.

MR. GREENBAUM: The rebuttal report?

MR. PARK: Yes, sir.

MR. GREENBAUM: Exhibit 2?

MR. PARK: Yes, sir.

MR. GREENBAUM: Okay. Thank you. Sorry

about the lack of clarity.

BY MR. PARK:

Q. You say, "Dr. Alford's observation about

the aggregate racial composition of these split

precinct portions is irrelevant."

A. Yes, sir.

Q. Why is it irrelevant?

A. Why is it relevant for my --

Q. Why is it irrelevant?

A. Sure. Why is it irrelevant for my

analysis? And I'm happy to explain that in some

detail.

I will start my answer by saying that I

explained why it's irrelevant in the latter half of

Page 15, Page 16, and into Page 17, but I'll give you

a briefer version here.

It's relevant for a number of different

reasons. When we're understanding split precincts

precisely in the context of this particular case,

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


56


there are -- were a number of important issues that I

noted. First, because I noted that there was strong

racially-polarized voting, it was clear that even just

a small change in the racial composition of a district

could have significant political consequences.

Now, specifically with respect to split

precincts, this means that it's important to analyze

these split precincts one at a time. The broader

reason why those split precincts need to be analyzed

one at a time rather than in the aggregate is because

each individual decision to split a precinct is one

that I sought to scrutinize.

I sought to scrutinize each individual

decision to split an individual precinct because there

was nothing that compelled, there was nothing that

forced the precincts to be split by the Legislature.

It wasn't something that the reapportionment

guidelines require, the splitting of precincts. It

was something that the Legislature chose to do.

And so it was specifically a deviation

from the redistricting guidelines whenever a precinct

was split, and so I analyzed those, one at a time.

So the point is that to properly analyze

them, it's necessary to go precinct by precinct rather

than to aggregate together the arbitrary set of

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


57


precincts that were chosen to be split by the

Reapportionment Office. Instead, it was necessary to

go individual precinct by individual precinct and

analyze the racial composition of those individual

split precincts.

Another important consideration for me in

determining that the relevant consideration was how

each individual precinct was split rather than how the

aggregate composition of those split precincts looked

was simply that I, based on Ms. Wright's and

Mr. Strangia's deposition testimony, knew that the

only available data within a particular split precinct

available to them was census demographic and racial

data and not partisan data.

So it was important for me for all of

those reasons to analyze how each individual precinct

was split rather than the aggregate composition of all

of those split precincts.

So those are the various reasons why I

said that focusing on the aggregate of those split

precincts was irrelevant and beside the point of how

to properly analyze split precincts. And I laid those

explanations, those reasons out in detail in my

response report.

Q. When we determine the winner of an

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


58


election in House District 105 and House District 111,

we aggregate the votes, don't we?

A. I agree with you that that is what the

Secretary of State does when determining the winner of

an election contest after the district lines have

already been drawn and are put into place.

Q. And a vote from a split precinct counts

just as much as a vote from a whole precinct, doesn't

it?

A. That is indeed how the Secretary of State

would put together election results. I affirm your

description of how votes are counted.

Q. On Page 17 of your reply report you say

that the aggregate racial composition is more heavily

African-American inside because of -- than outside is

entirely caused by two precincts, Lawrenceville D in

105 and Hickory Flat in 111. Is that -- Am I reading

that right?

A. I see where you're referring to. You're

referring to that second full paragraph on Page 17. I

do see where you're referring to.

Q. And you talk about racial segregation in

the next sentence. Is that de jure segregation?

A. Is the racial segregation of these two

precincts de jure segregation?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


59


Q. Yes, sir.

A. I really couldn't tell you, because I did

not conduct an analysis of the causes of racial

segregation in precinct Lawrenceville D and precinct

Hickory Flat.

Q. But you say it's racially segregated,

right? Why do you say it's racially segregated?

MR. GREENBAUM: Objection. Argumentative.

THE WITNESS: I am merely describing the

racial geography of those precincts that I saw

using census data, using census data and using

the election -- or the voter registration data

that I analyzed in my reports. I made -- I was

not attempting at all, and I apologize if this

portion of the report misled you, to say

anything further than an empirical statement of

the geographic patterns.

BY MR. PARK:

Q. Of residency, right?

A. The geography of race in these two

precincts. That's all I was talking about.

Q. Where people live, right, and who lives

there?

A. Sure. What we're talking about with

geography is data that measures where people live.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


60


Q. And the drafter put these two precincts,

or these two more heavily African-American portions of

precincts into 105 and 111, right? Mrs. Wright did

that?

A. Maybe we can take those precincts one at

the time.

Q. Okay.

A. I just want to make sure I understand your

question properly.

Q. Somehow Lawrenceville D ended up in -- or

a portion of Lawrenceville D ended up in House

District 105, right?

A. Indeed. It was a split precinct.

Q. And Ms. Wright put it in Lawrenceville D,

didn't she?

A. I assume the Reapportionment Office did.

I couldn't -- I couldn't speak for her, but I accept

-- I accept what you're saying.

Q. And the Reapportionment Office would have

put Hickory Flat in 111, right?

A. I agree that it was a split precinct.

Q. On Page 13 of your reply report you, in

Table 1 -- and you talk about Democratic candidates'

share of the two-party vote in House District 111 in

January of 2018.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


61



Q. Was that a -- Was that a jungle primary?

A. If I could ask you to define that term.

Q. How many candidates were involved in that

election? Do you recall?

A. I believe there were four.

Q. Two Democrats and two Republicans; is that

right?

A. I believe that's correct.

Q. And, again, on the ecological regression

for Black in January of 2018, you got a result that

was greater than 100 percent and rounded it down; is

that right?

A. That's correct. You're talking about the

right portion of Table 1. Same answer as when we

discussed that issue earlier in my original report.

Q. How do you get results that are over

100 percent? Do you know how that happens?

A. You're asking why does ecological

regression produce estimates for Black voting behavior

that are over 100 percent, right?

Q. For any behavior --

A. Any behavior.

Q. -- but in this case it's consistent for

Black voting behavior.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


62


A. Sure. So you're asking how an estimate

over 100 percent occurs. And the answer, just to

explain a little bit of the statistical details behind

this, is that when you conducted an ecological

regression estimate, you were fitting a regression

line to the data. So it's estimating one parameter

for Blacks and one parameter for non-Blacks that will

be applied to all precincts within a particular

district, and in this case, 111.

And when you conduct a regression

estimate, a linear regression estimate, what you're

doing is fitting all of the data to one line, to two

parameters, one regarding Blacks and one regarding

non-Blacks. And what the regression analysis does by

fitting that line is producing simply a straight line,

a straight line between the racial composition of a

precinct and its partisan voting patterns.

So when you produce that straight line,

sometimes regression analysis will result in estimates

at the very extremes that are obviously outside the

bounds of what is logically possible. In other words,

outside of 100 percent and zero percent. That is

something that does not happen with ecological

inference, but it does happen with ecological

regression.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


63


And so the convention has been that when

you produce an estimate using regression analysis that

results in, say, a logically-impossible conclusion,

say, a particular racial group voting at over

100 percent or under zero percent for a particular

candidate, then we simply round that number. We

report that number as zero or 100 percent.

So that's -- that's how it happens.

That's just a bit of the statistical background behind

how ecological regression works.

Q. You criticize the plan drafters for

splitting municipalities; is that right?

A. I don't criticize the plan drawers for

anything. I simply analyzed what the plan drawers

did. Again, I think what you're asking --

Q. You note --

A. -- your question --

Q. You note --

A. -- is whether I analyzed --

Q. You note --

A. -- split municipalities.

Q. You note that the drafters split

municipalities, correct?

A. That is a part of what I analyzed in my

report, yes, sir.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


64


Q. Did the guidelines, to the best of your

recollection, say anything about municipalities?

A. I think I've listed all of the guidelines

stated in my original report, so if you'll allow me a

second just to review that section of my report.

Q. Sure.

A. (Reviews document.) The answer is that

the redistricting guidelines are silent on the issue

of municipalities and only talk about avoiding a

splitting of precincts and counties.

Q. Correct. And do you know what the General

Assembly used as their allowable deviation in drawing

these plans, allowable population deviation?

A. I'm not sure I'm aware of that number.

Q. Do you know whether they went plus/minus

one percent?

A. I'm not sure I know.

Q. Do you think you can split more counties

if you went plus/minus one than if you went plus/minus

five?

A. Would you split more counties in a

district --

Q. Are you most likely to?

A. Are you more likely to split more counties

in a plus or minus one percent deviation world than if

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


65


you allowed plus or minus five percent?

Q. Correct.

A. And my answer is that, in general, that

doesn't make a difference because in the end typically

a plan will, particularly with larger counties that

have to be split into multiple districts, which

obviously applies to 105 and 111, a districting plan

in general only needs to split -- split a county only

when necessary to achieve equal population, which

specifically means that at most two districts might

need to be involved in the splitting of -- in being

split up.

In other words, it's possible to comply

with any reasonable equal population threshold in

larger counties, such as Gwinnett and Henry, by only

involving two districts and splitting up that -- and

being split -- I’m sorry -- in only involving two

districts in being split across counties.

So my answer is that, in general, it

shouldn't matter.

Q. If you'd look at Page 14 of your reply

report.

A. Yes, sir.

Q. In Paragraph 4 in the first paragraph you

say toward the end of the fifth line from the bottom,

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


66


"In my original report I found that in both House

District 105 and House District 111, 98 to 99 percent

Black voters support the Democratic state legislative

candidate in each election."


Q. And, "Meanwhile 75 to 85 percent of

non-Black support Republican candidates" --


Q. -- "constituting a level of Black voting

sufficient to defeat the Democratic candidate

supported by Black voters"?

A. Yes, sir.

Q. And, again, in House District 105 and 111,

the Black voters are a minority when compared to the

non-Black voters, right?

A. I can affirm that indeed it is less than

50 percent.

Q. So if it's less than 50 percent, then

everyone who could vote voted. Is that still

White-Black vote -- or non-Black voting defeating them

or is it just their minority status?

A. I want to make sure I understand the

question. I think there are a couple of things going

on here. You're asking me about a hypothetical world

in which every Black person shows up and votes?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


67


Q. If every Black showed up to vote in House

District 105 and voted for the Democrat, you still

need crossover from some of the non-Blacks to win,

right, because you're the minority?

A. And you're -- again, you're asking me

about a hypothetical world in which every Black person

votes. And are you telling me anything about how the

non-Black voters turn out? I just want understand

your scenario as accurately as I can here.

Q. Well, if you take every Black voter who

votes in House District 105, you have --

I'll withdraw the question?

MR. PARK: Let's take a break.

(Off the record at 12:05 PM)

--------------

(On the record at 12:07 PM)

MR. PARK: Dr. Chen, I believe that's all

the questions I have at this time.

MR. GREENBAUM: And I only have a couple

of questions.

EXAMINATION

BY MR. GREENBAUM:

Q. Dr. Chen, during Mr. Park's examination,

you said that for each election you run King's code

once, correct?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


68


A. Yes, sir, that is exactly what I do.

Q. What is entailed in -- When you run King's

code once, what does that mean?

A. Well, I call the software, and the

software, with its standard -- with its standard

inputs and its standard parameters, are run once.

Now, the software certainly does a substantial number

of what are called "burn-ins" as part of how the

software normally operates, but I physically only run

the software once.

Q. Just really quickly, what is a burn-in?

A. A burn-in involves runs that, at the

beginning of the algorithm, of the EI algorithm, that

are not kept and not analyzed. And typically the

software will do that by default. And certainly the

EI package that I was describing earlier to Mr. Park

does that. So a burn-in is simply a large number of

iterations at the beginning, certainly at least

several thousand or so, probably more, as part of any

EI software, as part of any EI algorithm, that are

conducted at the beginning and not kept, not used in

the analysis and not reported.

That doesn't mean that I am intentionally

throwing away any results. I obviously explained to

Mr. Park this morning that I reported all of my

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


69


analyses. But the EI software itself, the EI

algorithm, includes that as part of -- includes those

burn-ins as a standard part of how the algorithm runs.

Q. So each run contains -- Each time you do a

run, you run the code, the software goes through

several thousand iterations?

A. Oh, absolutely. I couldn't tell you the

exact number off the top of my head, but that is a

standard parameter in any EI software that one would

use in this sort of racially-polarized voting

analysis. It's a given that that is how ecological

inference is conducted.

MR. GREENBAUM: No further questions.

EXAMINATION

BY MR. PARK:

Q. Just so I'm clear, what is the trade name

of the software?

A. Sure. There's not a trade name. It's not

a trademarked software. It's open source software,

and it's by Professor Gary King. The software is just

called EI. It's runn in the R programming language.

It's one that's open source. It's freely available to

anybody on the Internet. And it's -- obviously it's

the open-source computer code that Professor King and

various collaborators developed as part of his

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


70


development and research on ecological inference back

in the early 2000's.

But I would just add to that that in the

computer code that I turned over in connection with my

original expert report, the layout, the installation

of that package, of that computer code, as well as the

use -- my use of that EI software, is all in the

computer code, and it was laid out precisely how I

uploaded the computer code, as well as how I used it

and the standard parameters that were used in my EI

analysis. So all of that is clearly laid out in the

computer code that I turned over in connection with my

expert reports.

Q. And, again, you call the package EI?

A. Yeah. It's just called EI. The letters

E-I. I know it's not the best name and not the most

creative name, but that is -- that's literally what

it's called. It just stands for Ecological Inference.

MR. PARK: Thank you, Dr. Chen.

THE WITNESS: Thank you, sir.

MR. GREENBAUM: Aria, any questions?

MS. BRANCH: No. No, thank you.

MR. GREENBAUM: All right. We're done.

THE COURT REPORTER: Aria, would you like

a copy of the deposition?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


71


MS. BRANCH: Yes, we would like a copy.

THE COURT REPORTER: All right. And Jon,

are you reserving signature?

MR. GREENBAUM: Yes. He'll read and sign.

(Time noted: 12:12 p.m.)

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


72


DISCLOSURE STATE OF GEORGIA COUNTY OF FULTON

WITNESS: JOWEI CHEN, Ph.D

Pursuant to Article 10.B of the Rules and Regulations of the Board of Court Reporting of the Judicial Council of Georgia, I make the following disclosure:

I am a Georgia Certified Court Reporter.

I am not disqualified for a relationship of interest under the provisions of O.C.G.A. 9-11-28(c).

I am a representative of Vargas Reporting Services, Inc.

Vargas Reporting Services, Inc. was contacted by the offices of STRICKLAND, BROCKINGTON & LEWIS, LLP to provide court reporting services for this proceeding.

Vargas Reporting Services, Inc. will not be taking this proceeding under any contract that is prohibited by Georgia law.

_________________________

Marianne Vargas, CCR, CVR-MCertified Court ReporterCertificate Number B-1832

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


73


C E R T I F I C A T E

STATE OF GEORGIA:

COUNTY OF FULTON:

I hereby certify that the foregoing proceeding

was taken down as stated in the caption, and the

colloquies and questions and answers were reduced to

typewriting under my direction; that the foregoing

transcript is a true and correct record of the

evidence given upon said hearing. I further certify

that I am not of kin or counsel to the parties in the

case, and am not in the regular employ of counsel for

any of the said parties, nor am I in any way

interested in the outcome of the case.

This certification is expressly withdrawn and

denied upon the alteration, disassembly, and/or

photocopying of the foregoing proceedings, including

exhibits, unless such is done by the undersigned

certified court reporter and the signature and

original seal is attached thereto.

This day, March 4, 2018.

___________________________

Marianne Vargas, CCR, CVR-MCertified Court ReporterCertificate Number B-1832

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


74


ERRATA SHEET I, JOWEI CHEN, Ph.D, the witness herein, do hereby certify that I have read the transcript of February 27, 2018, of my deposition testimony, and the same is true and correct, to the best of my knowledge, with the exception of the following changes noted below, if any:

Page_____ Line_____ should read:_________________ Reason for change:_______________________________ Page_____ Line_____ should read:_________________ Reason for change:_______________________________ Page_____ Line_____ should read:_________________ Reason for change:_______________________________ Page_____ Line_____ should read:_________________ Reason for change:_______________________________ Page_____ Line_____ should read:_________________ Reason for change:_______________________________ Page_____ Line_____ should read:_________________ Reason for change:_______________________________ Page_____ Line_____ should read:_________________ Reason for change:_______________________________

_______________________ JOWEI CHEN, Ph.D

Sworn to and subscribed before me, the undersigned Notary Public, on this ______ day of _________________, _______. ________________________________________ Notary Public

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25


BY MR. GREENBAUM: [1] 67/21 BY MR. PARK: [11] 4/11 7/19 11/21 15/17 30/11 42/6 43/1 51/14 55/7 59/17 69/14 MR. GREENBAUM: [12] 42/2 42/24 54/23 55/1 55/3 55/5 59/7 67/18 69/12 70/20 70/22 71/3 MR. PARK: [11] 4/2 11/18 15/13 15/16 30/6 54/25 55/2 55/4 67/12 67/16 70/18 MS. BRANCH: [2] 70/21 70/25 THE COURT REPORTER: [2] 70/23 71/1 THE WITNESS: [2] 59/8 70/19

.

...as [1] 22/11

00.9 [1] 29/801 [1] 7/1801427 [1] 1/402 [1] 51/13

11.3 percent [1] 29/91.8 [1] 29/710 [1] 21/2210.B [1] 72/5100 [1] 62/22100 percent [15] 16/21 16/21 16/25 17/2 17/3 17/4 17/8 17/16 17/17 61/12 61/18 61/21 62/2 63/5 63/7104 [1] 44/7105 [36] 16/16 18/22 19/19 20/23 21/3 22/1 22/9 22/20 23/1 24/3 27/21 30/4 30/15 32/16 43/5 43/6 43/14 44/7 45/7 45/12 45/25 46/6 46/8 46/12 47/5 47/12 47/14 58/1 58/17 60/3 60/12 65/7 66/2 66/13 67/2 67/1110:00 [1] 1/1810:03 [1] 4/210:54 [1] 30/910:59 [1] 30/11111 [26] 16/16 19/10 19/14 19/19 19/23 20/12 21/14 29/1 29/6 30/4 48/6 48/22 50/8 50/25 51/4 52/7 52/12 52/16 58/1 58/17 60/3 60/20 60/24 62/9 65/7 66/13111, 98 to [1] 66/2

1170 [1] 2/211201 [1] 1/2012:05 [1] 67/1412:07 [1] 67/1612:12 [1] 71/513 [1] 60/2213th [1] 2/1414 [2] 25/21 65/211401 [1] 2/514th [1] 1/2015 [4] 45/3 51/16 54/21 55/2116 [1] 55/2117 [3] 55/21 58/13 58/2018 [1] 33/201832 [2] 72/22 73/251953 [1] 12/241:17-CV-01427 [1] 1/4

22's [1] 19/122.4 [1] 29/72000's [1] 70/220005-2124 [1] 2/620005-3960 [1] 2/152010 [2] 45/5 45/112012 [15] 13/25 17/10 18/23 19/14 20/17 20/25 21/3 21/19 22/3 24/3 24/7 29/6 32/17 53/3 54/112012 to 2016 [2] 29/5 32/162014 [8] 20/16 21/1 21/19 52/4 52/12 52/16 52/25 53/152015 [5] 15/3 15/21 43/5 45/13 46/12016 [14] 18/25 19/16 20/16 20/17 21/2 21/19 22/4 24/3 24/7 29/5 32/16 45/16 46/13 53/32017 [5] 3/11 7/17 10/14 10/21 10/212018 [6] 1/17 7/7 60/25 61/11 73/20 74/3202.654.6338 [1] 2/15202.662.8315 [1] 2/721.4 percent [1] 22/32124 [1] 2/622 [1] 3/112200 [1] 2/2022nd [1] 7/1624 [1] 43/1125 [1] 44/625.2 percent [1] 22/327 [2] 1/17 74/328 [1] 72/929.8 percent [4] 45/8 45/17 46/2 47/6

330.4 percent [3] 46/3 46/10 47/730005 [1] 1/24

30039-3488 [1] 1/2130309-7200 [1] 2/2133 [1] 38/2234 [1] 38/223488 [1] 1/2135.2 [1] 20/2535.7 percent [1] 21/136.1 [1] 21/1736.1 percent [1] 21/18361 [1] 48/1037.0 percent [1] 21/137.6 [1] 21/193960 [1] 2/15

440.3 percent [1] 21/19400 [1] 2/645.6 percent [2] 44/10 44/12

550 [1] 44/2250 percent [5] 21/12 21/21 44/17 66/17 66/1854.4 [1] 44/135755 [1] 1/24

6600 [1] 2/14678.347.2208 [1] 2/22678.458.4030 [1] 1/25

7700 [1] 2/147200 [1] 2/2175 [1] 66/6

885 [1] 66/6

99-11-28 [1] 72/998.2 percent [1] 14/299 [1] 66/2

Aa.m [1] 1/18able [3] 26/18 51/2 51/4about [73] 5/17 5/18 8/14 8/17 8/20 8/23 9/10 9/19 10/10 10/16 11/14 11/24 12/14 15/10 17/22 19/9 19/11 20/8 20/21 21/25 22/19 22/25 23/20 24/15 25/22 26/18 26/24 27/3 27/3 30/14 30/21 32/21 33/11 34/16 36/5 36/9 36/25 37/2 37/7 37/12 37/22 38/1 38/15 38/20 41/15 42/18 42/20 43/3 44/6 45/19 45/23 46/5 46/25 47/23 47/25 49/23

50/10 53/14 54/11 54/22 54/24 55/7 55/9 58/22 59/21 59/24 60/23 61/14 64/2 64/9 66/24 67/6 67/7abranch [1] 2/16absolutely [3] 47/16 47/17 69/7accept [7] 12/11 29/16 48/25 49/2 49/4 60/17 60/18access [3] 7/11 15/3 15/22accommodate [1] 5/5accomplish [1] 50/13account [3] 15/9 27/23 28/11accounted [1] 28/14accurate [1] 41/24accurately [2] 29/22 67/9achieve [1] 65/9across [2] 47/10 65/18actually [4] 33/23 33/25 46/15 50/4add [3] 6/1 7/4 70/3added [2] 43/6 43/14additional [3] 6/13 6/15 6/17address [2] 16/14 38/10addressed [1] 21/16Advancement [1] 8/1affirm [2] 58/11 66/16African [16] 18/14 18/17 18/22 19/13 20/15 20/22 20/24 21/6 21/14 26/12 29/17 44/18 44/22 44/25 58/15 60/2African-American [9] 18/14 18/17 20/22 21/14 26/12 29/17 44/25 58/15 60/2African-Americans [2] 20/24 21/6after [1] 58/5again [13] 6/25 9/13 12/9 14/10 21/20 23/17 30/19 54/21 61/10 63/15 66/13 67/5 70/14age [5] 39/9 39/14 40/4 41/7 45/6aggregate [10] 47/15 47/21 55/10 56/10 56/25 57/9 57/17 57/20 58/2 58/14aggregated [2] 47/10 47/13ago [5] 14/11 24/5 32/8 32/19 46/21agree [3] 52/24 58/3 60/21ahead [2] 42/5 52/2al [3] 1/3 1/10 8/14

Alford [3] 5/13 53/4 54/22Alford's [6] 7/1 7/3 54/9 54/16 54/25 55/9algorithm [5] 68/13 68/13 68/20 69/2 69/3all [26] 6/1 6/10 12/8 13/11 24/25 26/8 27/22 33/16 40/16 40/18 45/1 45/11 50/19 57/15 57/17 59/14 59/21 62/8 62/12 64/3 67/17 68/25 70/7 70/11 70/23 71/2allocate [1] 41/25allocated [2] 39/8 41/6allow [1] 64/4allowable [2] 64/12 64/13allowed [1] 65/1allows [1] 25/12alone [2] 27/25 28/3already [2] 43/1 58/6also [9] 5/24 5/24 6/4 7/5 31/2 39/3 41/21 49/15 54/15alteration [1] 73/15always [1] 25/7am [15] 4/2 23/10 30/9 30/11 38/4 39/2 58/17 59/9 68/23 72/7 72/8 72/10 73/10 73/11 73/12amenities [1] 33/17American [14] 18/14 18/17 18/22 19/13 20/15 20/22 21/14 26/12 29/17 44/18 44/22 44/25 58/15 60/2Americans [2] 20/24 21/6analyses [2] 28/10 69/1analysis [20] 27/8 27/10 27/23 28/23 35/22 35/24 37/20 37/25 53/7 53/16 53/18 54/8 55/17 59/3 62/14 62/19 63/2 68/22 69/11 70/11analyze [14] 13/21 35/21 36/18 36/20 48/5 51/19 52/1 54/7 54/20 56/7 56/23 57/4 57/16 57/22analyzed [15] 7/8 24/25 27/1 35/3 50/24 52/8 53/11 54/10 56/9 56/22 59/13 63/14 63/19 63/24 68/14analyzing [2] 47/18 47/24and/or [1] 73/15ANDERSON [1] 2/11ANDREA [1] 2/11another [3] 6/5 49/10 57/6


Aanswer [37] 20/5 20/18 20/19 22/22 23/14 23/17 23/17 23/18 24/25 26/21 26/25 28/9 29/21 30/18 30/19 31/18 32/9 32/11 36/15 37/5 38/16 42/5 42/9 48/25 49/5 51/3 51/4 52/21 54/13 54/15 54/20 55/19 61/15 62/2 64/7 65/3 65/19answered [2] 11/24 42/4answers [2] 20/10 73/6any [30] 5/4 6/15 14/16 14/19 14/22 17/19 18/1 18/2 22/18 26/13 26/18 31/13 31/22 36/25 38/3 38/9 39/15 50/18 61/22 61/23 65/14 68/19 68/20 68/24 69/9 70/21 72/14 73/12 73/12 74/4anybody [2] 18/13 69/23anything [9] 5/15 11/14 20/8 26/24 26/24 59/16 63/14 64/2 67/7Anytime [1] 25/7apologize [1] 59/14appear [2] 7/21 51/10APPEARANCES [1] 2/1appears [1] 20/9applied [1] 62/8applies [1] 65/7appreciate [1] 47/3approach [2] 13/1 28/21April [1] 10/21arbitrary [3] 47/23 48/3 56/25are [56] 5/17 5/17 6/10 11/15 15/7 16/14 18/7 18/9 19/11 21/12 21/21 25/8 25/20 26/4 26/15 26/19 27/8 29/9 29/12 29/13 34/25 35/8 35/19 36/4 38/2 42/22 50/10 50/15 50/16 50/18 50/21 51/1 51/22 52/2 52/15 53/18 53/25 54/24 56/1 57/19 58/6 58/12 61/17 61/21 62/20 64/8 64/23 64/24 66/14 66/23 67/7 68/6 68/8 68/14 68/20 71/3area [3] 33/14 34/14 43/8Argumentative [1] 59/8ARIA [3] 2/13 70/21 70/24

ARREY [1] 2/11ARREY-MBI [1] 2/11Article [1] 72/5as [83] Asian [13] 22/7 22/11 24/1 24/12 24/19 24/24 29/3 30/14 32/3 32/15 32/24 34/25 38/15Asian-Hispanic [1] 32/15Asians [5] 24/23 27/5 29/8 31/24 32/23ask [13] 4/24 5/1 5/5 13/16 14/17 19/20 28/20 35/7 43/18 45/9 46/24 52/14 61/3asked [3] 32/8 42/3 46/21asking [41] 5/17 5/18 17/11 17/22 19/11 21/8 23/6 24/20 25/4 25/12 25/22 25/23 26/17 28/18 31/14 32/2 32/7 35/13 36/5 36/9 36/10 37/2 37/22 44/14 44/16 44/21 45/19 45/23 46/5 47/4 48/20 49/4 49/23 50/10 52/18 53/20 61/19 62/1 63/15 66/24 67/5Assembly [1] 64/12Association [3] 8/1 8/21 9/20assume [2] 50/21 60/16assumed [2] 40/15 40/16assuming [1] 40/22at [60] 5/4 5/15 9/7 9/25 10/2 10/15 10/23 11/4 12/8 15/4 15/23 16/4 16/19 17/19 26/8 28/17 28/24 30/9 30/11 32/12 32/12 33/18 34/13 35/5 35/12 35/16 36/11 36/23 37/13 40/2 41/21 43/11 45/11 45/14 45/24 45/25 46/6 46/20 47/24 50/9 50/19 51/6 52/20 54/8 56/8 56/10 56/22 59/14 60/5 62/20 63/4 65/10 65/21 67/14 67/16 67/18 68/12 68/18 68/18 68/21ATLANTA [6] 1/2 1/21 2/21 33/13 33/15 33/18Atlantic [1] 1/20attached [1] 73/19attempt [1] 5/1attempting [1] 59/14attention [2] 7/24 21/22attributable [3] 22/7 24/18 24/22AUDRA [1] 2/10

AUSTIN [2] 1/10 2/9available [7] 13/10 40/2 50/2 50/12 57/12 57/13 69/22Avenue [1] 2/5avoiding [1] 64/9aware [2] 53/18 64/14away [1] 68/24

BB-1832 [2] 72/22 73/25back [5] 14/24 15/17 24/17 26/9 70/1background [1] 63/9Barksdale [1] 54/4baseball [1] 34/12based [6] 16/9 16/11 36/13 41/7 42/1 57/10basically [1] 52/18basing [1] 16/6basis [6] 23/6 23/17 32/2 32/11 40/5 48/4Bayesian [1] 13/1BBM [1] 1/5be [33] 5/5 7/21 13/19 13/22 14/21 15/25 19/24 24/18 24/22 25/9 26/1 26/10 35/7 39/24 41/6 41/10 42/17 42/19 42/21 44/13 48/2 48/17 48/21 49/1 49/7 51/10 56/9 56/16 57/1 62/8 65/6 65/11 72/14became [3] 18/23 19/15 31/21because [18] 17/4 18/9 23/24 32/9 33/24 41/18 42/10 49/3 49/11 50/4 53/12 56/2 56/10 56/14 58/15 59/2 65/4 67/4become [1] 11/7been [18] 4/9 4/20 7/13 17/16 33/1 33/2 33/5 33/6 33/22 34/8 34/13 34/17 49/18 49/21 50/11 53/13 58/6 63/1before [6] 4/20 9/16 31/17 31/21 45/10 74/22beginning [3] 68/13 68/18 68/21BEHALF [3] 2/2 2/9 2/17behave [1] 35/1behavior [30] 23/11 23/21 24/16 27/4 27/9 27/24 28/6 28/11 30/14 30/16 31/24 32/22 35/6 35/16 35/21 36/18 36/20 37/1 37/9 37/13 38/2 38/6 38/10 38/14 38/19 38/21 61/20 61/22 61/23 61/25behind [2] 62/3 63/9

being [4] 34/23 65/11 65/17 65/18believe [11] 6/5 6/13 8/11 9/12 9/24 10/7 10/14 10/20 61/6 61/9 67/17below [1] 74/4beside [1] 57/21best [3] 64/1 70/16 74/3between [9] 20/16 24/3 24/7 29/4 30/1 32/16 39/6 49/16 62/16beyond [2] 53/22 54/8bit [8] 11/7 13/2 20/14 30/18 44/22 51/4 62/3 63/9Black [61] 16/20 17/19 17/19 17/25 18/1 18/2 18/2 18/12 18/13 18/17 18/24 19/15 22/1 22/2 22/8 22/12 24/2 24/14 24/18 24/21 26/10 27/2 27/2 28/12 32/5 36/1 36/11 36/12 36/13 36/14 45/6 45/22 46/2 46/3 46/9 46/25 51/20 52/11 52/15 52/20 53/1 53/5 53/8 53/10 53/25 61/11 61/20 61/25 66/3 66/7 66/9 66/11 66/14 66/15 66/20 66/20 66/25 67/1 67/6 67/8 67/10Blacks [11] 24/8 24/25 25/4 25/8 26/11 53/15 62/7 62/7 62/13 62/14 67/3block [6] 39/5 40/2 41/6 41/21 41/25 50/23block-level [1] 39/5blocks [10] 37/1 37/9 39/15 39/20 40/11 40/14 40/17 40/19 40/21 42/1Board [4] 8/3 8/21 10/11 72/5border [2] 33/11 33/11both [4] 39/23 40/12 49/11 66/1bottom [5] 26/7 43/13 45/14 45/24 65/25boundaries [12] 21/3 29/5 29/6 32/17 32/17 45/8 45/25 46/7 46/9 46/12 47/5 47/12bounds [2] 12/24 62/21box [13] 11/10 11/13 11/15 11/25 12/1 12/8 12/15 18/13 18/16 25/13 25/20 26/8 26/12

boxes [1] 18/15BRANCH [1] 2/13Braves [1] 34/12break [4] 5/4 5/6 30/7 67/13breakdown [1] 18/8BRIAN [3] 1/6 1/13 2/18briefer [1] 55/22bring [1] 41/22broader [1] 56/8BROCKINGTON [2] 2/20 72/12BROOKS [1] 2/9Brown [3] 8/23 10/3 31/6BRYAN [1] 1/19Bureau's [1] 18/7burn [7] 14/16 14/18 68/8 68/11 68/12 68/17 69/3burn-in [3] 68/11 68/12 68/17burn-ins [2] 68/8 69/3bus [1] 33/25but [45] 6/8 9/15 11/8 12/5 12/6 15/4 15/22 17/6 17/14 20/16 25/17 26/6 27/4 27/11 27/25 28/17 29/10 29/13 34/4 34/23 36/24 38/12 38/17 39/3 40/24 41/11 41/24 42/13 44/4 45/2 47/24 51/4 53/7 54/9 54/18 55/21 59/6 60/17 61/24 62/24 68/9 69/1 69/8 70/3 70/17BVAP [4] 44/9 44/12 44/13 44/15

CC-H-E-N [1] 4/17calculating [1] 47/9calculations [1] 13/12call [2] 68/4 70/14called [5] 13/8 68/8 69/21 70/15 70/18came [3] 18/5 23/19 24/11can [18] 6/10 7/14 11/19 17/5 23/17 23/25 24/6 26/17 26/23 26/24 27/2 30/5 37/16 38/9 60/5 64/18 66/16 67/9can't [10] 6/7 9/14 12/6 19/9 29/21 37/15 41/17 42/10 44/2 54/17candidate [20] 18/24 19/16 20/16 22/2 24/9 52/4 52/11 52/17 52/19 52/20 52/22 52/25 52/25 53/2 53/10 53/16 54/1 63/6 66/4 66/10


Ccandidates [11] 22/13 24/14 24/22 32/5 33/4 51/21 53/5 53/8 54/18 61/4 66/7candidates' [1] 60/23capacity [3] 1/6 1/13 2/18caption [1] 73/5Carolina [6] 31/4 31/9 31/11 35/2 35/5 35/8case [20] 1/4 5/12 7/17 8/7 9/16 9/25 10/12 10/13 10/22 11/2 11/4 31/15 31/18 36/5 37/11 55/25 61/24 62/9 73/11 73/13cases [3] 1/10 7/25 31/22category [2] 25/10 25/17Caucasian [4] 52/5 52/23 52/25 53/15Cause [4] 10/16 10/18 35/10 37/11caused [1] 58/16causes [1] 59/3CAVE [1] 1/19CCR [3] 1/22 72/21 73/24CELESTE [1] 2/3census [21] 5/24 18/7 25/10 25/12 26/2 26/6 39/4 39/4 39/14 39/20 40/11 40/14 40/17 40/19 40/20 45/5 45/11 49/18 57/13 59/11 59/11Center [1] 1/20certainly [14] 5/23 7/10 17/6 20/13 33/14 33/17 35/19 35/24 38/7 38/12 50/11 68/7 68/15 68/18Certificate [2] 72/22 73/25certification [1] 73/14certified [4] 72/7 72/21 73/18 73/24certify [3] 73/4 73/9 74/2challenge [1] 9/12change [11] 12/2 33/3 43/5 56/4 74/6 74/8 74/10 74/12 74/14 74/16 74/18changes [8] 11/10 11/15 11/24 12/1 12/8 32/25 43/4 74/4chart [2] 36/2 36/3Chattanooga [2] 33/12 33/21check [3] 25/13 26/8 26/12checked [3] 18/13 18/16 26/7checking [1] 18/14

CHEN [14] 1/16 3/11 3/12 4/8 4/15 4/20 7/14 23/9 67/17 67/23 70/19 72/4 74/2 74/21Chen's [2] 54/25 55/1choice [2] 11/11 48/3chose [1] 56/19chosen [4] 47/24 48/1 48/2 57/1Citizens [2] 8/20 9/19city [4] 10/10 31/8 35/10 35/20CIVIL [1] 2/5clarify [2] 11/13 13/16clarity [1] 55/7clear [4] 26/1 40/24 56/3 69/16clearly [3] 44/24 53/14 70/11close [1] 34/5closest [1] 33/13Coalition [1] 9/23code [13] 13/7 13/8 13/10 13/10 67/24 68/3 69/5 69/24 70/4 70/6 70/8 70/9 70/12codes [2] 7/2 13/8coefficient [4] 14/8 17/5 17/7 17/10coefficients [1] 14/13cohesion [2] 22/20 22/25COIE [1] 2/13COLEY [1] 2/3collaborators [1] 69/25colloquies [1] 73/6Colored [1] 8/1column [2] 46/5 47/13combined [2] 12/23 24/4come [1] 13/24Commission [1] 54/12Commissioners [1] 8/3COMMITTEE [1] 2/5Common [4] 10/16 10/18 35/10 37/11commonly [2] 12/12 38/18commonly-studied [1] 38/18Commonwealth [1] 10/25comparable [1] 21/18comparative [1] 29/20comparatively [1] 29/9compare [1] 49/13compared [3] 19/17 20/17 66/14compelled [1] 56/15

complete [3] 19/24 23/13 23/16completely [4] 20/20 41/24 49/24 50/3comply [1] 65/13composition [14] 18/5 18/8 18/11 44/4 47/15 47/21 48/1 55/10 56/4 57/4 57/9 57/17 58/14 62/16comprise [2] 20/24 21/9computer [10] 7/2 13/7 13/7 13/10 69/24 70/4 70/6 70/8 70/9 70/12concluded [1] 50/7conclusion [8] 19/2 19/9 19/22 24/11 30/14 31/25 32/1 63/3conclusions [1] 26/18conduct [3] 17/21 59/3 62/10conducted [4] 28/22 62/4 68/21 69/12conducting [1] 28/19CONFERENCE [1] 2/2conferences [1] 34/14confidence [2] 14/7 16/22confident [2] 38/12 38/17confirm [8] 41/10 41/17 42/10 44/21 45/2 46/1 53/24 54/1confirmed [5] 21/20 32/21 32/21 32/25 39/18confirms [1] 24/14connection [5] 6/2 6/12 7/3 70/4 70/12consequences [1] 56/5consideration [3] 50/9 57/6 57/7considered [1] 7/6consistent [2] 53/19 61/24CONSOLIDATED [1] 1/10constituting [1] 66/9contacted [1] 72/11contains [2] 44/9 69/4contest [1] 58/5context [3] 28/18 51/3 55/25contract [1] 72/14convention [1] 63/1copy [6] 7/22 42/12 51/7 51/11 70/25 71/1CORETTA [1] 2/12correct [25] 6/23 7/21 10/1 10/5 17/13 18/18 21/10 25/20 27/21 28/1 28/13 39/1 47/2 47/8 48/24 51/10

53/7 61/9 61/14 63/23 64/11 65/2 67/25 73/8 74/3correctly [2] 14/14 52/10Corrine [3] 8/23 10/3 31/6could [13] 11/12 13/16 14/17 19/4 25/2 28/5 28/5 33/10 45/9 52/14 56/5 61/3 66/19couldn't [14] 10/8 12/13 25/3 25/16 34/15 34/22 38/11 38/16 48/8 48/15 59/2 60/17 60/17 69/7Council [1] 72/6counsel [7] 2/1 5/10 9/21 9/22 11/1 73/10 73/11count [2] 33/9 33/10counted [1] 58/12counties [8] 29/21 64/10 64/18 64/21 64/24 65/5 65/15 65/18counting [1] 21/2counts [1] 58/7county [19] 8/3 8/21 10/11 29/13 29/14 29/17 29/18 34/6 34/9 34/16 34/18 34/20 34/24 50/17 50/21 50/23 65/8 72/2 73/3couple [5] 4/24 30/7 32/13 66/23 67/19course [3] 7/5 7/7 52/22courses [2] 38/6 38/7court [8] 1/1 4/5 72/5 72/7 72/12 72/21 73/18 73/24covered [1] 38/13creating [1] 18/4creative [1] 70/17Creek [1] 1/24criteria [1] 6/7criticize [2] 63/11 63/13crossover [1] 67/3cultural [1] 33/17CUNNINGHAM [1] 2/10CV [1] 1/4CVR [3] 1/22 72/21 73/24CVR-M [3] 1/22 72/21 73/24

DDARRYL [1] 2/10data [39] 5/16 5/23 5/24 5/24 7/2 13/15 13/18 15/4 15/4 15/22 15/23 17/5 18/4 26/2 26/3 26/6 39/4 39/5 39/5 40/1 41/21 41/25 49/12 49/16 49/18 49/19 50/2 50/8 50/12

54/8 57/12 57/14 57/14 59/11 59/11 59/12 59/25 62/6 62/12Davis' [1] 12/23day [5] 20/25 34/2 34/2 73/20 74/23DC [2] 2/6 2/15de [2] 58/23 58/25Dec [1] 3/11December [6] 5/20 5/23 7/11 7/16 11/4 16/4December 22nd [1] 7/16decision [2] 56/11 56/14default [1] 68/15defeat [1] 66/10defeating [1] 66/20Defendant [6] 1/8 1/15 2/17 8/24 10/4 10/5DEFENDANT'S [3] 3/9 7/14 51/7Defendants [1] 8/8define [3] 14/17 43/19 61/3Democrat [3] 23/4 54/19 67/2Democratic [26] 18/24 19/16 20/15 22/2 22/13 24/9 24/14 24/22 25/5 25/9 32/5 33/4 51/20 52/3 52/11 52/17 52/19 52/22 53/2 53/5 53/8 53/15 54/1 60/23 66/3 66/10Democrats [5] 25/2 50/17 50/19 50/23 61/7demographic [5] 5/16 5/24 32/25 49/18 57/13demographics [5] 29/1 29/13 29/14 29/20 30/3denied [1] 73/15deposed [7] 4/20 9/8 9/15 10/1 10/6 10/13 11/3deposition [28] 1/16 5/9 6/4 6/17 6/22 6/24 6/25 9/25 10/20 11/23 12/5 16/5 16/7 16/12 39/1 39/2 39/8 39/13 40/9 40/9 41/3 41/9 42/11 49/6 49/7 57/11 70/25 74/3depositions [5] 11/16 39/24 40/1 41/1 49/15describe [1] 12/19described [1] 43/1describes [2] 13/4 13/5describing [6] 12/11 19/6 42/14 42/23 59/9 68/16description [2] 3/10


Ddescription... [1] 58/12detail [2] 55/18 57/23detailed [1] 18/7details [1] 62/3determine [2] 37/8 57/25determining [2] 57/7 58/4detour [1] 34/1Detzner [9] 8/15 8/18 8/23 9/4 9/11 30/24 31/1 37/14 37/22developed [1] 69/25development [1] 70/1deviation [4] 56/20 64/12 64/13 64/25did [41] 5/8 5/14 7/11 8/4 9/1 9/6 9/7 13/15 14/14 14/21 19/22 22/21 22/23 22/24 23/2 26/8 26/12 29/19 35/5 35/13 36/12 36/15 36/18 36/20 37/5 37/7 37/10 37/12 37/19 37/21 37/24 41/20 47/9 54/5 54/7 54/20 59/2 60/3 60/16 63/15 64/1didn't [9] 7/10 8/5 20/5 20/8 27/16 27/18 27/25 36/24 60/15difference [2] 29/25 65/4different [7] 9/5 13/2 29/14 39/6 39/20 40/20 55/23direction [1] 73/7directly [2] 20/8 20/10disagree [1] 53/6disassembly [1] 73/15discard [2] 14/19 14/20disclosure [2] 72/1 72/6discussed [5] 11/15 24/5 32/19 51/1 61/16discussing [2] 41/8 42/14discussion [4] 11/21 15/16 41/15 41/16disqualified [1] 72/8distinguish [3] 39/19 40/13 49/16distributed [1] 28/5district [61] 1/1 1/1 8/2 16/16 18/22 19/10 19/14 20/12 20/22 21/3 21/7 21/9 21/14 22/1 22/9 22/20 23/1 24/3 27/21 29/1 29/6 30/15 32/15 32/16 43/4 43/6 43/14 45/1 45/7 45/12 45/25 46/6 46/7 46/8 46/12 47/4 47/12 47/13 48/6

48/18 48/19 48/22 50/8 51/4 52/12 52/16 53/12 54/12 56/4 58/1 58/1 58/5 60/12 60/24 62/9 64/22 66/2 66/2 66/13 67/2 67/11districting [1] 65/7districts [6] 30/4 32/18 65/6 65/10 65/16 65/18DIVISION [1] 1/2do [73] 4/16 5/2 5/8 6/15 8/13 11/9 11/14 11/23 13/15 14/4 14/5 14/16 14/19 14/20 15/25 17/7 17/9 19/2 20/12 23/3 24/19 27/14 27/18 27/23 27/25 28/2 28/2 28/10 30/20 30/21 31/25 31/25 35/1 36/15 37/7 37/10 37/12 37/19 38/22 39/7 39/10 41/2 41/12 41/12 42/8 43/16 43/20 43/21 46/11 46/14 48/6 48/13 49/22 50/5 50/16 50/18 50/22 52/10 53/7 54/3 56/19 58/21 59/7 61/5 61/17 61/18 64/11 64/15 64/18 68/1 68/15 69/4 74/2document [8] 6/6 6/9 19/5 20/7 23/15 43/10 44/2 64/7documents [4] 6/10 6/13 6/16 7/10does [17] 7/21 16/24 18/2 19/8 19/21 44/12 50/5 51/10 51/12 58/4 61/19 62/14 62/23 62/24 68/3 68/7 68/17doesn't [6] 19/8 19/21 25/11 58/8 65/4 68/23doing [3] 22/24 35/4 62/12don't [21] 4/25 10/6 10/7 11/8 12/4 12/9 12/10 12/12 14/16 17/14 18/9 26/13 27/11 32/23 41/11 41/18 42/10 42/12 54/16 58/2 63/13done [4] 22/18 31/13 70/23 73/17down [9] 15/10 17/3 25/20 33/15 33/18 34/22 41/6 61/12 73/5Dr [3] 7/3 54/16 54/25Dr. [15] 4/20 5/13 6/24 7/1 7/14 23/9 53/4 54/9 54/22 54/25 55/1 55/9 67/17 67/23 70/19Dr. Alford [3] 5/13 53/4 54/22Dr. Alford's [4] 7/1

54/9 54/25 55/9Dr. Chen [6] 4/20 7/14 23/9 67/17 67/23 70/19Dr. Chen's [1] 55/1Dr. Strangia's [1] 6/24drafter [1] 60/1drafters [2] 63/11 63/22draw [6] 19/2 19/9 19/22 26/18 31/25 32/1drawer [3] 15/3 15/21 16/1drawers [2] 63/13 63/14drawing [3] 16/8 39/3 64/12drawn [3] 18/23 19/14 58/6duly [1] 4/9Duncan [1] 12/23during [1] 67/23

EE-I [1] 70/16each [12] 14/8 14/12 16/20 40/6 56/11 56/13 57/8 57/16 66/4 67/24 69/4 69/4earlier [3] 51/1 61/16 68/16early [1] 70/2easily [1] 33/21ecological [42] 12/20 13/11 16/15 16/15 16/19 17/1 22/18 22/19 24/7 27/6 27/6 27/12 27/12 27/15 27/17 27/19 27/19 28/7 28/7 28/9 28/10 28/19 36/2 36/13 36/16 36/24 36/25 37/6 37/7 37/8 37/19 37/20 37/25 61/10 61/19 62/4 62/23 62/24 63/10 69/11 70/1 70/18education [1] 38/4effectively [2] 50/9 50/13EI [23] 13/8 14/1 14/10 17/21 22/21 22/25 24/21 27/1 28/22 32/18 53/1 68/13 68/16 68/20 68/20 69/1 69/1 69/9 69/21 70/7 70/10 70/14 70/15election [26] 7/6 7/7 7/8 8/3 13/21 14/1 17/10 20/25 30/4 36/23 39/5 41/4 45/17 52/4 52/5 52/12 52/16 53/14 54/8 58/1 58/5 58/11 59/12 61/5 66/4 67/24elections [9] 8/21 10/11 16/20 19/17

26/5 52/7 53/9 53/12 53/13electoral [2] 5/15 5/23electorate [3] 22/8 24/2 47/1elementary [1] 33/25empirical [1] 59/16employ [1] 73/11end [4] 15/2 15/10 65/4 65/25ended [2] 60/10 60/11engaged [3] 8/8 8/16 10/19enough [1] 28/4entailed [1] 68/2entirely [1] 58/16equal [2] 65/9 65/14ER [4] 17/22 22/21 22/25 28/23ERRATA [1] 74/1ESQ [3] 2/4 2/13 2/19essence [1] 40/4estimate [12] 17/2 17/16 24/8 24/18 24/21 40/5 41/22 62/1 62/5 62/11 62/11 63/2estimated [1] 39/13estimates [10] 13/25 16/16 16/20 17/2 17/22 27/1 28/19 42/1 61/20 62/19estimating [3] 40/10 40/20 62/6Estimation [1] 12/23et [3] 1/3 1/10 8/14even [4] 29/23 29/25 47/24 56/3evening [1] 5/11ever [2] 10/7 11/5every [7] 33/19 34/1 34/2 66/25 67/1 67/6 67/10everybody [1] 26/11everyone [2] 18/16 66/19evidence [1] 73/9exact [1] 69/8exactly [11] 9/17 11/17 12/14 13/5 27/14 28/22 29/11 36/4 40/17 42/18 68/1examination [4] 4/11 67/21 67/23 69/14EXAMINATIONS [1] 3/1examined [1] 4/9example [2] 13/25 18/9examples [1] 53/25exception [1] 74/4exclusively [1] 17/19exhibit [5] 7/14 7/16 7/18 51/13 55/4Exhibit 1 [1] 7/14exhibited [1] 52/5exhibits [2] 3/8 73/17expert [30] 1/16 6/9 6/12 7/1 7/17 8/6 9/3

9/5 9/14 9/24 10/6 10/19 11/2 31/17 31/19 31/22 35/13 36/16 36/21 37/2 37/10 37/16 37/19 39/4 52/8 53/22 54/10 54/20 70/5 70/13expertise [1] 31/20explain [5] 17/5 18/3 42/21 55/17 62/3explained [3] 17/15 55/20 68/24explanations [1] 57/23expressly [1] 73/14extent [4] 31/20 53/16 53/17 53/18extremes [1] 62/20

Ffact [10] 13/6 27/2 27/9 27/23 28/10 33/1 42/25 46/1 46/23 50/7familiar [2] 11/6 11/7familiarity [1] 11/9family [1] 33/15far [1] 50/3favor [3] 18/24 19/15 54/1feature [2] 12/1 12/12February [6] 1/17 6/19 7/9 10/14 13/12 74/3feel [1] 5/1Ferguson [1] 8/2Ferguson-Florissant [1] 8/2few [4] 7/4 9/5 34/13 38/8fifth [2] 51/24 65/25figure [1] 42/1file [1] 41/23filed [2] 8/10 8/12files [8] 6/1 6/2 6/3 7/2 13/8 18/6 18/6 18/7final [2] 23/20 24/12finally [3] 15/1 15/8 15/20find [4] 6/20 6/20 44/2 51/22first [7] 4/9 12/14 17/12 21/25 32/13 56/2 65/24first-hand [1] 12/14fitting [3] 62/5 62/12 62/15five [2] 64/20 65/1Flat [3] 58/17 59/5 60/20Floor [1] 1/20Florida [6] 8/18 9/12 30/24 31/2 31/6 35/2Florissant [1] 8/2focusing [1] 57/20follow [1] 45/10following [2] 72/6 74/4follows [1] 4/10


Fforced [1] 56/16foregoing [3] 73/4 73/7 73/16formula [1] 41/14found [4] 15/2 15/20 50/4 66/1four [1] 61/6fourth [1] 51/24free [1] 5/1freely [1] 69/22front [8] 41/9 41/11 41/19 42/11 42/13 42/17 42/24 54/16full [1] 58/20FULTON [2] 72/2 73/3further [3] 59/16 69/13 73/9

GGA [2] 1/21 1/24Gary [2] 13/5 69/20gave [1] 30/18general [17] 23/11 23/18 24/15 30/16 31/23 32/21 32/24 35/22 37/18 38/1 42/14 54/2 54/2 64/11 65/3 65/8 65/19generally [7] 23/22 26/21 38/12 41/16 51/20 53/5 53/20geographic [2] 39/6 59/17geography [7] 12/3 41/4 41/5 41/6 59/10 59/20 59/25GEORGIA [24] 1/1 1/7 1/14 2/2 2/17 2/18 2/21 31/13 31/17 31/25 32/6 33/5 33/7 33/12 33/23 34/1 34/3 35/1 54/6 72/2 72/6 72/7 72/15 73/2get [3] 14/5 50/9 61/17Gina [5] 6/5 6/25 11/23 16/6 16/7give [10] 6/19 12/14 19/4 19/25 23/12 23/14 23/16 43/9 43/25 55/21given [6] 14/12 18/9 23/13 50/1 69/11 73/9giving [1] 54/19gleaned [2] 40/8 40/25go [9] 11/19 15/14 15/17 42/5 50/3 52/2 52/3 56/24 57/3goes [2] 31/23 69/5going [8] 25/8 26/9 32/9 34/11 34/20 35/7 43/18 66/23gone [2] 33/15 33/18got [3] 36/11 47/7 61/11granular [1] 49/12greater [1] 61/12

GREENBAUM [1] 2/4Greensboro [16] 10/10 10/12 31/8 35/10 35/14 35/21 35/23 35/25 36/8 36/8 36/9 36/12 36/17 36/19 36/21 37/3grew [3] 33/10 33/20 34/4group [4] 25/1 25/8 28/7 63/4grouping [1] 27/1groups [1] 32/22growing [1] 33/19guidelines [5] 56/18 56/21 64/1 64/3 64/8Guilford [1] 10/11Gwinnett [7] 29/14 29/18 29/20 34/6 34/8 34/11 65/15

Hhad [18] 6/3 15/3 15/22 16/4 32/25 33/25 35/21 39/19 39/22 40/2 40/8 40/12 40/12 41/21 46/21 49/16 50/13 53/2half [1] 55/20hand [2] 7/13 12/14happen [2] 62/23 62/24happening [1] 53/19happens [2] 61/18 63/8happy [8] 5/5 30/19 41/10 42/17 42/19 42/21 53/24 55/17Harbins [5] 43/5 43/8 43/14 43/16 43/21has [5] 25/10 27/11 32/15 33/2 63/1have [44] 4/20 4/21 7/10 11/5 15/20 17/16 22/18 25/8 26/13 31/13 31/14 33/5 33/6 34/2 34/8 34/12 34/17 35/7 38/7 39/13 41/11 41/18 41/18 42/10 42/11 42/12 43/1 43/18 47/7 49/18 49/21 50/11 53/25 54/13 54/15 54/16 56/5 58/5 60/19 65/6 67/11 67/18 67/19 74/2haven't [2] 12/7 23/13having [1] 4/9HD [1] 52/7he [7] 39/10 41/3 41/10 41/13 41/17 52/23 54/4He'll [1] 71/4head [11] 6/11 9/15 10/9 25/17 34/15 34/23 38/12 38/17 48/9 48/16 69/8hearing [1] 73/9heavily [3] 44/25

58/14 60/2held [2] 11/21 15/16help [3] 30/20 42/17 42/19Henry [11] 29/13 29/17 29/20 34/16 34/18 34/20 34/23 50/17 50/21 50/23 65/15her [4] 11/23 16/8 42/8 60/17here [19] 14/8 16/4 17/11 18/12 23/23 23/25 26/4 28/19 36/5 41/11 42/12 45/2 51/23 52/22 53/22 54/6 55/22 66/24 67/9hereby [2] 73/4 74/2herein [1] 74/2Hickory [3] 58/17 59/5 60/20high [3] 17/9 34/21 52/21Hillbrooke [1] 1/24him [2] 39/8 41/8hired [6] 8/13 8/24 9/13 9/23 10/12 11/1his [11] 1/6 1/13 2/18 13/6 13/6 39/12 39/13 53/4 53/6 54/21 69/25Hispanic [30] 22/7 22/11 22/12 22/20 23/1 23/4 24/1 24/12 24/13 24/18 24/24 27/20 27/20 27/22 27/24 28/17 29/2 30/14 32/3 32/4 32/15 32/24 34/25 35/6 35/17 35/20 35/23 36/3 37/13 38/10Hispanics [10] 24/23 27/5 27/25 28/2 28/4 28/6 28/11 29/7 31/24 32/23history [2] 5/25 26/3House [55] 7/7 14/1 16/16 17/10 18/22 18/24 19/9 19/14 19/16 20/12 20/22 21/3 21/14 22/1 22/9 22/20 23/1 24/2 27/20 29/1 29/6 30/4 30/15 32/16 43/4 43/6 43/13 45/7 45/12 45/25 46/6 46/8 46/12 47/4 47/12 47/13 48/6 48/22 50/8 51/3 52/4 52/5 52/7 52/12 52/15 53/12 58/1 58/1 60/11 60/24 66/1 66/2 66/13 67/1 67/11how [40] 4/16 4/22 9/18 12/2 13/14 14/4 14/4 17/1 17/9 23/18 24/19 26/14 26/24 28/22 33/5 33/6 34/6 34/8 34/17 36/18 36/20 47/18 57/7 57/8 57/16 57/21 58/10 58/12 61/4 61/17

61/18 62/1 63/8 63/10 67/7 68/8 69/3 69/11 70/8 70/9hundreds [1] 33/21hypothetical [2] 66/24 67/6

II'd [4] 5/5 7/13 21/22 29/9I'll [9] 6/20 6/20 18/3 25/15 25/17 29/24 44/5 55/21 67/12I'm [27] 4/15 9/17 10/8 11/6 11/12 11/17 26/3 26/17 30/19 32/8 33/23 34/14 35/7 38/12 38/17 39/3 45/24 45/25 46/15 46/22 53/24 54/19 55/17 64/14 64/14 64/17 69/16I've [9] 11/6 23/13 28/16 31/14 33/22 34/13 46/15 51/1 64/3identification [1] 18/12identifications [1] 26/4identified [2] 7/19 51/14identify [5] 7/25 18/5 18/10 18/10 28/6if [52] 4/24 5/4 6/19 9/15 11/12 12/16 12/16 13/16 13/21 14/17 14/24 17/2 19/4 19/19 19/25 21/8 23/12 23/14 25/4 25/12 25/19 26/17 26/17 28/4 28/24 30/19 31/14 37/15 41/8 42/16 42/23 43/9 43/11 43/24 43/25 44/14 45/3 45/9 48/20 51/16 52/14 59/14 61/3 64/4 64/19 64/19 64/25 65/21 66/18 67/1 67/10 74/4illustrate [1] 20/23illustrated [1] 22/9important [5] 39/16 56/1 56/7 57/6 57/15impossible [4] 49/7 49/24 50/4 63/3in [327] INC [4] 1/23 72/10 72/11 72/14include [4] 18/2 25/2 25/5 54/3included [4] 26/10 27/8 27/22 38/5includes [6] 18/1 25/8 26/11 53/14 69/2 69/2including [2] 49/18 73/16increase [7] 20/14 20/17 20/21 22/6 22/7 24/1 24/7

increased [4] 22/2 29/7 29/8 32/15increases [2] 29/10 29/12incumbent [1] 48/7indeed [5] 26/7 53/24 58/10 60/13 66/16independent [2] 27/12 27/15INDEX [2] 3/1 3/8individual [19] 2/9 2/9 2/10 2/10 2/10 2/11 2/11 2/11 2/12 2/12 47/18 56/11 56/13 56/14 57/3 57/3 57/4 57/8 57/16individuals [3] 26/5 27/3 27/9individuals' [1] 18/8inference [22] 12/20 13/11 16/15 22/18 24/8 27/6 27/12 27/15 28/7 28/9 28/19 36/13 36/16 36/24 37/6 37/8 37/20 37/25 62/24 69/12 70/1 70/18inference/ecological [1] 37/20inferences [1] 36/25information [4] 12/14 24/4 24/10 40/25initial [1] 21/23inputs [1] 68/6ins [2] 68/8 69/3inside [3] 45/7 54/23 58/15installation [1] 70/5instances [1] 53/19Instead [1] 57/2intentionally [1] 68/23interest [1] 72/9interested [1] 73/13Internet [1] 69/23Intervenors [3] 8/24 10/4 10/5into [12] 27/2 33/22 34/1 34/2 39/9 44/7 48/19 48/22 55/21 58/6 60/3 65/6intuition [1] 24/15involved [2] 61/4 65/11involves [1] 68/12involving [2] 65/16 65/17irrelevant [7] 47/17 55/11 55/13 55/15 55/16 55/20 57/21is [162] isn't [3] 29/18 30/1 44/15issue [7] 37/13 37/15 51/19 51/25 52/1 61/16 64/8issues [1] 56/1it [84] it's [45] 13/2 13/8 15/10 21/16 22/13


Iit's... [40] 28/14 29/17 29/23 29/25 33/21 33/23 40/10 41/22 43/20 44/21 45/17 47/17 47/23 48/3 50/1 50/3 54/9 54/14 55/20 55/23 56/7 56/24 59/6 59/7 61/24 62/6 65/13 66/18 69/11 69/18 69/19 69/20 69/21 69/22 69/22 69/23 69/23 70/15 70/16 70/18iteration [3] 13/20 13/22 13/23iterations [2] 68/18 69/6its [6] 38/21 39/17 62/17 68/5 68/5 68/6itself [1] 69/1

JJ-O-W-E-I [1] 4/17JACK [1] 2/19JACKSON [1] 2/12JAMAL [1] 2/9JAMIDA [1] 2/11January [3] 7/7 60/25 61/11January 2018 [1] 7/7jgreenbaum [1] 2/7Jim [1] 54/3jjp [1] 2/22JOHN [1] 2/19Johns [1] 1/24JON [2] 2/4 71/2JOWEI [8] 1/16 3/11 3/12 4/8 4/15 72/4 74/2 74/21JR [1] 2/19Judicial [1] 72/6jungle [1] 61/2jure [2] 58/23 58/25just [59] 4/24 6/1 6/19 7/4 10/22 11/7 11/7 11/19 12/9 14/1 14/21 15/6 18/3 19/4 19/24 19/25 20/2 20/7 20/18 21/9 21/11 23/7 23/12 23/13 23/14 23/19 24/5 25/16 26/1 28/3 33/12 33/20 35/7 40/7 40/24 41/17 43/9 43/18 43/25 45/9 47/25 49/24 51/22 53/13 54/19 56/3 58/8 60/8 62/2 63/9 64/5 66/21 67/8 68/11 69/16 69/20 70/3 70/15 70/18Justice [1] 9/23

KKEMP [3] 1/6 1/13 2/18Ken [1] 8/14kept [2] 68/14 68/21key [3] 51/19 51/25

52/1kin [1] 73/10kind [1] 50/9King [5] 13/4 13/5 13/9 69/20 69/24King's [2] 67/24 68/2knew [1] 57/11know [25] 4/22 10/6 11/10 11/14 18/9 28/18 41/12 43/16 43/20 43/21 46/11 46/14 48/7 48/13 49/8 49/15 50/16 50/18 50/22 53/6 61/18 64/11 64/15 64/17 70/16knowing [1] 26/14knowledge [10] 23/11 23/18 23/21 24/15 30/16 31/23 32/21 38/2 39/4 74/3known [1] 12/22

Llack [1] 55/7lacked [1] 49/12laid [3] 57/22 70/8 70/11language [2] 13/9 69/21large [3] 33/13 33/16 68/17larger [2] 65/5 65/15last [3] 5/10 22/13 25/19later [1] 11/3latter [1] 55/20LAURETHA [1] 2/3LAVELLE [1] 2/2law [2] 2/5 72/15Lawrenceville [9] 34/7 44/7 44/17 44/24 58/16 59/4 60/10 60/11 60/14LAWYERS' [1] 2/5lawyerscommittee.org [1] 2/7layout [1] 70/5League [6] 8/17 9/10 9/11 9/21 10/24 31/1least [3] 33/18 34/13 68/18legislative [4] 6/6 49/8 49/20 66/3Legislature [2] 56/16 56/19Legislature's [3] 15/2 15/21 16/1LEMON [1] 2/2less [2] 66/16 66/18let [6] 15/6 20/2 23/12 26/22 43/9 43/24let's [7] 15/14 15/17 30/7 35/12 51/6 51/7 67/13letters [1] 70/15level [12] 14/7 15/5 15/23 39/5 40/2 41/4 41/5 41/6 41/21 41/25

52/6 66/9LEWIS [2] 2/20 72/12like [14] 7/13 12/15 12/16 13/3 16/13 21/22 30/17 30/20 41/9 42/16 42/23 49/17 70/24 71/1Likelihood [1] 12/23likely [8] 18/24 19/15 22/12 23/4 24/13 32/4 64/23 64/24limitation [1] 39/22limits [2] 16/22 35/20line [20] 15/8 25/23 26/7 51/24 51/24 53/2 62/6 62/12 62/15 62/15 62/16 62/18 65/25 74/5 74/7 74/9 74/11 74/13 74/15 74/17linear [1] 62/11lines [2] 15/10 58/5listed [1] 64/3literally [1] 70/17litigated [1] 37/16little [4] 11/7 13/2 44/22 62/3live [2] 59/22 59/25lived [2] 33/14 48/7lives [1] 59/22ll [1] 33/9LLP [4] 1/19 2/13 2/20 72/12logically [5] 17/4 17/17 24/4 62/21 63/3logically-impossible [1] 63/3long [1] 31/21longer [1] 30/18look [7] 5/15 28/24 35/5 43/11 43/25 51/6 65/21looked [7] 28/17 32/12 32/12 35/16 36/23 46/6 57/9looking [9] 16/19 17/18 35/5 36/11 45/11 45/14 45/24 45/25 54/8looks [1] 12/15Louis [1] 8/3love [1] 34/11lower [7] 46/11 46/14 46/16 46/17 46/20 46/22 46/23lowest [1] 41/4LYNNE [1] 2/11

Mmade [4] 39/22 40/3 43/5 59/13Mainly [1] 43/23majority [3] 21/7 21/9 44/15make [13] 20/5 20/7 23/13 33/25 36/25 40/7 40/24 45/10 50/5 60/8 65/4 66/22 72/6makes [1] 13/9manipulation [1]

50/14many [10] 13/14 14/4 14/4 33/5 33/6 33/9 34/6 34/8 34/17 61/4map [5] 9/13 15/2 15/21 16/1 16/8Maptitude [6] 11/5 11/9 12/1 12/7 12/10 12/12March [1] 73/20Marianne [3] 1/22 72/21 73/24mark [1] 51/7marked [3] 7/13 7/19 51/14marked/identified [2] 7/19 51/14MARLON [1] 2/3mask [1] 16/24masks [1] 17/6math [2] 44/20 45/2matter [2] 54/2 65/20maximum [3] 12/22 17/4 17/5may [1] 49/1maybe [2] 40/23 60/5MBI [1] 2/11McDonough [2] 34/16 34/17McKENZIE [1] 2/10me [67] 5/1 6/19 7/14 9/23 11/1 11/13 12/6 15/6 16/3 16/3 19/4 19/11 19/25 20/2 20/13 21/8 23/6 23/12 23/12 23/14 24/20 25/4 25/12 25/22 25/23 26/17 26/22 30/20 32/2 32/3 32/7 32/8 35/13 36/9 36/10 39/16 39/21 39/22 40/3 41/10 41/11 42/12 42/13 42/17 42/19 42/24 43/9 43/9 43/24 43/25 44/14 44/16 44/21 46/21 47/4 48/20 49/3 49/4 52/18 54/16 57/6 57/15 64/4 66/24 67/5 67/7 74/22mean [10] 11/13 14/17 27/14 28/2 30/22 30/23 43/19 44/12 68/3 68/23meaningless [1] 47/22meanings [1] 40/14means [3] 47/16 56/7 65/10Meanwhile [1] 66/6measures [1] 59/25merely [1] 59/9met [1] 5/10Method [1] 12/24methodology [1] 12/19metropolitan [2] 33/13 33/14middle [2] 14/25 51/25

Midtown [1] 2/20might [1] 65/10mile [1] 33/11minority [5] 32/22 46/19 66/14 66/21 67/4minus [5] 64/15 64/19 64/19 64/25 65/1minutes [1] 30/8misled [1] 59/15Missouri [1] 7/25moment [9] 6/20 14/11 19/4 19/25 23/12 23/14 24/5 43/9 46/21months [1] 33/19more [21] 7/4 18/23 19/15 22/11 23/4 24/13 29/17 32/4 35/8 42/5 44/17 44/25 51/4 53/20 58/14 60/2 64/18 64/21 64/24 64/24 68/19moreover [1] 51/18morning [1] 68/25most [4] 53/13 64/23 65/10 70/16mostly [1] 5/11move [1] 44/15moving [1] 44/6Mr. [12] 39/7 39/12 39/17 40/8 41/3 42/22 49/7 50/13 57/11 67/23 68/16 68/25Mr. Park [2] 68/16 68/25Mr. Park's [1] 67/23Mr. Strangia [4] 39/12 39/17 42/22 50/13Mr. Strangia's [5] 39/7 40/8 41/3 49/7 57/11Mrs. [1] 60/3Mrs. Wright [1] 60/3Ms [3] 12/4 16/7 16/12Ms. [15] 6/5 6/25 16/6 16/7 16/10 39/1 39/2 39/18 40/9 41/20 42/22 49/6 50/12 57/10 60/14Ms. Gina [4] 6/5 6/25 16/6 16/7Ms. Wright [4] 41/20 42/22 50/12 60/14Ms. Wright's [7] 16/10 39/1 39/2 39/18 40/9 49/6 57/10much [1] 58/8multiple [1] 65/6municipalities [5] 63/12 63/21 63/23 64/2 64/9museums [1] 33/16my [113] myself [3] 15/6 19/5 44/1


NN.W [3] 1/20 2/5 2/14NAACP [2] 1/3 2/2name [7] 4/13 4/16 6/8 69/16 69/18 70/16 70/17National [1] 8/1nature [4] 8/6 8/25 9/18 37/18necessary [4] 49/12 56/24 57/2 65/9need [4] 5/4 56/9 65/11 67/3needs [2] 41/5 65/8new [3] 2/5 7/6 7/10next [2] 21/16 58/23nice [1] 34/12night [1] 5/10no [20] 1/4 3/10 7/18 9/24 18/21 19/18 22/21 23/2 31/18 31/18 33/10 36/8 37/5 39/19 40/12 40/12 51/13 69/13 70/22 70/22non [39] 17/19 18/13 18/17 18/22 19/13 20/15 22/1 22/8 22/12 23/4 24/2 24/8 24/13 24/18 24/21 24/25 25/4 25/8 26/10 26/11 27/2 28/12 32/4 32/24 33/3 36/1 36/12 36/14 44/13 44/15 44/18 44/22 62/7 62/14 66/7 66/15 66/20 67/3 67/8non-African-American [5] 18/22 19/13 20/15 44/18 44/22non-Black [18] 17/19 18/13 18/17 22/1 22/8 24/2 24/18 24/21 26/10 27/2 28/12 36/1 36/12 36/14 66/7 66/15 66/20 67/8non-Blacks [8] 24/8 24/25 25/4 25/8 26/11 62/7 62/14 67/3non-BVAP [2] 44/13 44/15non-Hispanic [5] 22/12 23/4 24/13 32/4 32/24non-Whites [1] 33/3none [2] 27/4 27/11nor [1] 73/12normal [4] 12/10 12/13 14/22 14/23normally [1] 68/9North [6] 31/4 31/8 31/11 35/2 35/4 35/8northern [2] 1/1 33/11not [69] 5/6 7/11 9/7 9/15 9/17 10/1 10/6 10/7 10/8 11/3 11/12 11/17 13/1 14/20 14/21 15/4 15/22 18/7 18/9 19/22 20/16 22/21 23/2 26/1 26/2

26/6 26/8 26/12 28/16 29/19 31/19 35/3 36/15 37/5 37/10 37/21 37/24 40/10 40/19 46/15 46/22 47/24 48/4 49/21 52/20 54/5 54/7 54/9 54/14 54/20 57/14 59/3 59/14 62/23 64/14 64/17 68/14 68/14 68/21 68/21 68/22 69/18 69/18 70/16 70/16 72/8 72/14 73/10 73/11Notary [2] 74/23 74/25note [4] 63/16 63/18 63/20 63/22noted [4] 56/2 56/2 71/5 74/4nothing [4] 20/10 47/17 56/15 56/15noticeable [2] 24/1 24/6noticed [2] 32/13 32/14November [10] 18/25 19/16 20/25 21/1 21/2 22/3 22/4 52/4 52/12 52/16November 2012 [1] 22/3November 2014 [2] 52/12 52/16November 2016 [3] 18/25 21/2 22/4now [12] 6/11 9/14 24/4 25/22 42/13 46/5 47/15 50/3 54/17 54/25 56/6 68/7number [19] 7/25 17/14 25/9 33/22 35/19 45/19 45/23 46/10 47/22 55/23 56/1 63/6 63/7 64/14 68/7 68/17 69/8 72/22 73/25numbers [12] 12/2 20/11 20/13 21/8 21/11 21/17 21/18 21/20 32/18 47/10 51/1 53/1numerically [1] 29/11

OO.C.G.A [1] 72/9Objection [2] 42/3 59/8observation [1] 55/9obviously [14] 7/1 17/15 20/16 21/11 21/20 25/7 31/19 34/4 42/13 50/25 62/20 65/7 68/24 69/23occurs [1] 62/2October [2] 10/22 10/23off [18] 6/11 9/14 10/8 11/19 11/20

15/14 15/15 25/16 26/22 30/9 34/15 34/23 38/11 38/17 48/8 48/15 67/14 69/8off-the-record [2] 11/20 15/15office [10] 38/21 39/13 39/19 41/23 49/8 49/21 50/2 57/2 60/16 60/19offices [1] 72/12official [3] 1/6 1/13 2/18Oh [3] 25/4 36/7 69/7okay [29] 5/22 10/10 10/12 10/16 12/6 15/13 18/3 23/6 23/10 23/16 24/20 25/22 26/17 26/21 26/23 32/2 32/7 35/12 36/9 41/5 44/16 46/21 47/3 47/9 48/25 52/18 53/20 55/6 60/7old [1] 29/5on [58] 2/2 2/9 2/17 8/8 8/12 11/3 11/25 12/7 12/16 14/1 14/8 15/7 15/11 15/12 15/13 15/17 15/19 16/6 16/9 16/11 17/12 18/19 20/21 23/3 23/10 25/21 27/7 30/11 30/15 31/14 33/15 36/13 38/3 38/6 38/22 38/25 39/2 39/3 41/7 42/1 44/6 48/4 48/10 51/23 52/3 54/21 57/10 57/20 58/13 58/20 60/22 61/10 64/8 66/24 67/16 69/23 70/1 74/23once [6] 13/22 33/19 67/25 68/3 68/6 68/10one [35] 1/20 13/19 13/21 13/22 14/9 14/13 14/15 18/14 33/19 34/21 34/24 35/12 42/5 43/5 47/24 48/10 48/10 49/9 53/14 54/9 56/8 56/10 56/11 56/22 60/5 62/6 62/7 62/12 62/13 62/13 64/16 64/19 64/25 69/9 69/22one percent [1] 64/16ones [1] 48/2only [17] 5/5 15/3 15/22 16/5 21/2 28/14 28/16 49/16 51/3 57/12 64/9 65/8 65/8 65/15 65/17 67/19 68/9open [3] 69/19 69/22 69/24open-source [1] 69/24operates [1] 68/9Oppenheimer [1]

54/11opportunity [1] 18/10opposed [2] 18/14 49/10or [55] 9/15 10/7 13/7 14/3 14/22 15/10 17/3 22/18 22/21 22/25 24/23 24/24 25/20 25/23 26/8 26/24 26/25 27/5 27/5 27/5 27/6 27/12 27/15 27/19 28/7 30/17 33/19 34/6 34/17 35/10 36/2 36/24 39/15 40/15 41/3 43/16 48/18 49/22 51/20 51/24 52/19 53/10 54/25 59/12 60/2 60/10 63/5 63/7 64/25 65/1 66/20 66/21 68/19 73/10 73/15ORANGE [1] 2/11order [1] 13/24organization [1] 2/2orient [3] 15/6 19/5 44/1original [14] 3/11 5/18 5/19 6/8 6/12 7/11 7/16 16/2 52/8 61/16 64/4 66/1 70/5 73/19other [17] 5/15 8/5 18/15 18/16 25/13 25/17 25/23 26/7 26/10 27/7 33/17 41/13 52/7 53/6 53/18 62/21 65/13others [9] 8/18 10/11 10/16 25/10 25/20 26/16 26/19 26/25 27/5our [3] 13/9 33/25 41/22out [12] 21/9 21/12 26/5 30/20 34/12 34/13 44/7 46/19 57/23 67/8 70/8 70/11outcome [1] 73/13outlined [1] 6/6outside [12] 31/18 33/12 33/20 45/12 46/3 46/7 46/8 47/6 47/14 58/15 62/20 62/22over [13] 6/2 7/2 16/24 17/2 17/8 17/16 44/22 61/17 61/21 62/2 63/4 70/4 70/12own [1] 39/4

Pp.m [1] 71/5package [6] 13/6 13/7 14/10 68/16 70/6 70/14page [40] 3/2 3/10 7/24 12/16 12/17 14/1 14/24 14/25 15/7 15/8

15/12 15/13 15/19 16/13 18/19 20/21 21/22 25/21 38/22 38/22 43/11 44/6 48/10 51/16 51/23 54/21 55/21 55/21 55/21 58/13 58/20 60/22 65/21 74/5 74/7 74/9 74/11 74/13 74/15 74/17Page 1 [1] 7/24Page 10 [1] 21/22Page 13 [1] 60/22Page 14 [2] 25/21 65/21Page 15 [3] 51/16 54/21 55/21Page 16 [1] 55/21Page 17 [3] 55/21 58/13 58/20Page 2 [2] 15/7 15/8Page 2 of [1] 14/24Page 24 [1] 43/11Page 25 [1] 44/6Page 3 [2] 14/25 15/12Page 33 [1] 38/22Page 34 [1] 38/22Page 361 [1] 48/10Page 4 [1] 12/17Page 5 [1] 16/13Page 6 [1] 14/1Page 7 [2] 18/19 20/21pages [1] 16/13paragraph [13] 14/25 15/8 15/19 18/21 21/25 22/15 23/7 23/20 43/13 51/18 58/20 65/24 65/24Paragraph 4 [1] 65/24parameter [3] 62/6 62/7 69/9parameters [3] 62/13 68/6 70/10PARK [3] 2/19 68/16 68/25Park's [1] 67/23part [21] 12/10 12/13 14/21 17/19 18/1 18/2 24/17 24/21 27/1 33/1 35/22 35/24 39/23 40/6 63/24 68/8 68/19 68/20 69/2 69/3 69/25partially [1] 22/6particular [6] 44/4 55/25 57/12 62/8 63/4 63/5particularly [1] 65/5parties [2] 73/10 73/12partisan [13] 15/4 15/22 33/1 35/21 35/23 40/14 44/4 49/9 49/13 50/9 50/14 57/14 62/17partisanship [6] 39/14 39/20 40/11 40/13 40/16 40/20


Pparty [1] 60/24past [3] 10/22 11/4 36/23PATRICIA [1] 2/3patterns [3] 35/23 59/17 62/17PAYTON [1] 2/10Peachtree [2] 1/20 2/21pending [7] 5/6 11/10 11/13 11/14 11/24 12/1 12/8Pennsylvania [2] 10/25 10/25people [4] 8/2 26/7 59/22 59/25percent [52] 14/2 16/21 16/21 16/25 17/2 17/3 17/4 17/8 17/16 17/17 21/1 21/1 21/12 21/18 21/18 21/19 21/19 21/21 22/3 22/3 29/7 29/7 29/8 29/9 44/10 44/12 44/13 44/17 44/22 45/8 45/17 46/2 46/3 46/10 47/6 47/7 61/12 61/18 61/21 62/2 62/22 62/22 63/5 63/5 63/7 64/16 64/25 65/1 66/2 66/6 66/17 66/18percentage [5] 24/8 32/14 33/3 42/2 46/17perfectly [1] 40/18performance [3] 48/21 49/9 49/13performing [1] 49/2performs [1] 48/13PERKINS [1] 2/13perkinscoie.com [1] 2/16person [2] 66/25 67/6perspective [1] 49/20Ph.D [6] 1/16 4/8 4/18 72/4 74/2 74/21phone [1] 26/22photocopying [1] 73/16physically [1] 68/9pick [1] 34/1pieces [1] 24/10place [5] 6/20 28/14 28/16 44/3 58/6Plaintiff's [3] 9/21 9/22 11/1Plaintiffs [10] 1/4 1/11 2/2 2/9 8/16 8/19 8/22 9/13 10/13 10/20Plaintiffs' [1] 5/10plan [12] 11/25 15/3 15/22 18/23 19/14 45/13 46/1 63/11 63/13 63/14 65/5 65/7Plan's [2] 21/3 29/6plans [1] 64/13please [4] 4/3 4/13 4/25 20/4

plus [5] 64/15 64/19 64/19 64/25 65/1plus/minus [3] 64/15 64/19 64/19PM [2] 67/14 67/16point [3] 29/10 56/23 57/21polarized [3] 52/6 56/3 69/10political [15] 22/19 23/11 23/21 24/15 30/16 31/20 31/21 31/24 38/2 38/5 38/6 38/13 38/19 41/21 56/5population [16] 20/22 21/14 22/8 39/10 39/14 40/5 41/7 42/2 44/13 45/6 45/6 45/7 46/19 64/13 65/9 65/14portion [19] 12/4 24/2 28/12 36/2 36/3 41/2 44/6 44/9 44/16 44/23 44/24 47/6 48/18 48/22 49/1 49/11 59/15 60/11 61/15portions [9] 39/25 46/6 47/11 47/14 49/14 49/17 54/22 55/11 60/2possible [8] 19/24 20/20 49/22 50/5 50/8 50/11 62/21 65/13possibly [1] 33/10practice [2] 14/22 14/23precinct [44] 15/4 15/23 38/21 39/5 39/15 40/11 40/15 40/17 40/19 40/21 41/5 41/24 43/3 43/14 43/22 43/23 44/4 44/17 44/25 48/19 49/9 49/10 49/11 49/14 49/17 50/14 55/11 56/11 56/14 56/21 56/24 56/24 57/3 57/3 57/8 57/12 57/16 58/7 58/8 59/4 59/4 60/13 60/21 62/17precincts [39] 39/9 39/21 40/3 45/13 46/7 47/5 47/11 47/14 47/16 47/18 47/21 47/24 48/2 48/4 48/5 48/10 48/11 54/23 55/24 56/7 56/8 56/9 56/16 56/18 57/1 57/5 57/9 57/18 57/21 57/22 58/16 58/25 59/10 59/21 60/1 60/3 60/5 62/8 64/10precise [5] 6/7 6/20 17/14 41/17 42/10precisely [4] 18/11 51/5 55/25 70/8prediction [1] 22/1

prepare [2] 5/8 9/1prepared [2] 5/14 9/4preparing [1] 5/22present [1] 37/24presented [1] 53/1pretty [5] 29/2 33/23 47/22 48/3 48/14previous [1] 19/17primary [4] 15/2 15/21 16/1 61/2probably [1] 68/19probative [1] 53/13procedure [2] 12/22 13/3proceeding [3] 72/13 72/14 73/4proceedings [1] 73/16process [9] 11/8 12/11 12/13 16/8 39/17 40/1 42/15 42/22 43/1produce [3] 61/20 62/18 63/2producing [1] 62/15professor [3] 13/9 69/20 69/24program [1] 13/8programming [2] 13/9 69/21prohibited [1] 72/15properly [4] 28/5 56/23 57/22 60/9proportion [6] 45/7 45/22 46/2 46/4 46/9 46/25proportions [1] 42/2Proscenium [1] 2/20provide [1] 72/12provisions [1] 72/9Public [3] 54/12 74/23 74/25publication [1] 38/17publications [3] 38/3 38/8 38/9purely [2] 23/10 30/15purpose [1] 18/3purposes [1] 47/17Pursuant [1] 72/5put [14] 13/18 24/10 32/20 41/9 42/16 42/23 47/16 48/17 48/21 58/6 58/11 60/1 60/14 60/20

Qquestion [29] 4/25 5/6 11/24 13/17 17/25 19/20 20/6 20/11 22/22 26/23 28/3 29/21 32/8 32/10 35/3 42/5 45/10 47/7 48/23 49/5 50/15 50/24 51/3 52/14 53/11 60/9 63/17 66/23 67/12questions [6] 20/19 67/18 67/20 69/13 70/21 73/6quickly [1] 68/11

quite [3] 29/10 32/23 38/7

Rrace [6] 45/12 45/16 45/20 52/19 54/17 59/20races [1] 53/6racial [23] 15/3 15/22 18/5 18/8 18/11 26/4 47/15 47/21 48/1 49/18 50/8 50/12 55/10 56/4 57/4 57/13 58/14 58/22 58/24 59/3 59/10 62/16 63/4racially [5] 52/6 56/3 59/6 59/7 69/10racially-polarized [3] 52/6 56/3 69/10Raleigh [3] 8/20 9/19 31/4ran [1] 54/4rate [2] 46/20 52/21rather [4] 56/10 56/24 57/8 57/17read [13] 21/9 21/12 38/7 39/23 71/4 74/2 74/5 74/7 74/9 74/11 74/13 74/15 74/17readily [1] 13/10reading [6] 12/4 16/12 23/8 41/22 49/6 58/17really [8] 19/8 26/13 34/12 34/14 34/22 40/20 59/2 68/11reapportionment [10] 6/7 38/20 39/19 49/8 49/21 50/2 56/17 57/2 60/16 60/19reason [10] 48/17 48/21 56/9 74/6 74/8 74/10 74/12 74/14 74/16 74/18reasonable [1] 65/14reasons [4] 55/24 57/16 57/19 57/23rebuttal [9] 5/13 7/3 7/5 7/9 8/9 8/10 8/12 54/21 55/2recall [20] 6/15 9/15 11/23 12/4 17/9 17/14 33/24 34/20 34/23 39/7 39/10 39/12 41/2 41/8 41/12 41/15 41/16 41/20 42/8 61/5recollection [1] 64/2record [9] 4/14 11/20 15/15 15/17 30/9 30/11 67/14 67/16 73/8redistricting [4] 31/15 31/22 56/21 64/8reduced [1] 73/6refer [2] 13/23 44/23reference [2] 23/24 27/19referenced [3] 6/8 6/18 6/21

referred [1] 14/11referring [12] 11/17 16/3 16/3 22/17 26/15 35/9 36/4 40/23 45/2 58/19 58/20 58/21reflected [6] 27/5 27/9 29/2 29/25 36/1 36/3reflection [1] 27/6reflects [1] 11/10regard [1] 53/9regarding [4] 16/8 19/22 62/13 62/13registration [4] 5/25 18/6 26/2 59/12regression [27] 16/15 16/19 17/1 17/7 17/9 17/15 22/19 27/13 27/17 27/19 28/8 28/10 36/2 36/25 37/8 37/20 61/10 61/20 62/5 62/5 62/10 62/11 62/14 62/19 62/25 63/2 63/10regular [2] 11/8 73/11Regulations [1] 72/5REID [1] 2/3reiterate [1] 12/9relationship [2] 39/6 72/8relative [5] 39/20 40/10 40/13 49/9 49/13relatively [1] 52/21relatively-high [1] 52/21relevant [3] 55/14 55/23 57/7relied [1] 12/7rely [2] 23/3 30/13relying [6] 23/10 30/15 38/3 38/4 38/25 39/2remember [5] 6/7 6/11 9/14 38/9 42/13remembered [1] 7/5René [2] 8/14 9/1repeat [2] 45/9 52/14rephrase [1] 5/1reply [7] 3/12 7/9 51/6 51/8 58/13 60/22 65/21report [107] reported [19] 17/17 19/7 20/11 28/20 32/19 33/2 43/24 44/3 45/21 45/24 46/2 46/9 46/10 46/16 46/23 51/2 53/3 68/22 68/25reporter [5] 4/5 72/7 72/21 73/18 73/24reporting [8] 1/23 26/3 47/25 72/5 72/10 72/11 72/12 72/14reports [6] 5/12 7/1 9/5 54/9 59/13 70/13representation [1] 49/1representative [1]


Rrepresentative... [1] 72/10represents [5] 14/9 14/13 25/24 26/7 28/21Republican [9] 43/20 43/22 43/23 48/14 48/18 48/19 48/21 51/21 66/7Republicans [2] 49/2 61/7require [1] 56/18research [6] 11/8 12/11 12/13 14/22 14/23 70/1reserving [1] 71/3reside [1] 21/2residency [1] 59/19resolved [2] 9/17 9/18respect [7] 19/19 19/20 43/4 48/6 50/25 53/11 56/6respondent [1] 25/13response [3] 13/13 19/18 57/24responsive [1] 32/10rest [1] 45/1result [3] 16/24 61/11 62/19results [16] 5/15 7/6 7/8 13/14 17/11 19/6 19/12 27/23 33/1 36/23 41/4 54/7 58/11 61/17 63/3 68/24review [4] 20/2 39/25 43/10 64/5reviewed [14] 5/11 5/12 5/23 5/24 5/25 6/4 6/9 6/16 6/18 6/24 6/25 7/1 13/14 16/5Reviews [6] 19/5 20/7 23/14 43/10 44/2 64/7revisit [2] 30/19 32/9right [54] 4/18 6/11 9/14 10/1 10/4 13/20 15/5 16/17 17/20 18/1 21/4 21/15 25/13 25/15 25/25 26/14 26/19 27/7 29/3 29/15 29/18 30/24 32/5 37/3 37/22 40/6 42/12 43/6 45/8 45/11 45/13 45/18 46/5 47/1 47/7 47/13 52/23 54/16 58/18 59/7 59/19 59/22 60/3 60/12 60/20 61/8 61/13 61/15 61/21 63/12 66/15 67/4 70/23 71/2RIGHTS [1] 2/5Rob [1] 6/22Robert [1] 10/17Romo [9] 8/14 9/1 9/2 9/4 30/24 37/12 37/13 37/21 37/22round [2] 17/3 63/6rounded [1] 61/12

row [2] 17/12 45/24Rucho [9] 10/17 10/18 10/19 31/11 35/11 36/6 36/7 37/7 37/11Rules [1] 72/5run [17] 13/19 13/22 14/3 14/9 14/13 14/15 14/20 14/21 27/20 36/16 67/24 68/2 68/6 68/9 69/4 69/5 69/5runn [1] 69/21runs [7] 13/15 14/4 14/19 22/19 36/13 68/12 69/3rural [2] 43/8 43/16

SSABRINA [1] 2/10said [14] 30/17 32/11 34/4 41/3 41/10 41/13 41/16 52/9 52/22 52/23 57/20 67/24 73/9 73/12said whether [1] 41/13same [19] 19/9 21/13 32/8 32/23 35/1 38/16 39/24 40/18 41/18 42/9 42/11 49/11 49/14 52/6 54/13 54/15 54/19 61/15 74/3SAMMY [1] 2/11saw [4] 23/23 32/14 32/18 59/10say [27] 12/20 13/25 14/12 17/6 18/21 20/8 23/3 26/23 26/24 27/3 29/9 30/13 33/18 43/13 50/3 51/19 52/3 53/9 55/9 58/13 59/6 59/7 59/15 63/3 63/4 64/2 65/25saying [12] 11/25 13/21 32/3 34/25 38/12 38/18 39/8 39/24 52/13 52/15 55/19 60/18says [1] 15/20sbllaw.net [1] 2/22scenario [1] 67/9school [5] 8/2 33/25 33/25 34/21 34/22scientist [3] 31/20 31/21 38/5scientist's [1] 14/22scrutinize [2] 56/12 56/13seal [1] 73/19second [7] 6/13 6/16 6/19 11/19 43/25 58/20 64/5Secretary [6] 1/6 1/13 2/18 41/23 58/4 58/10section [3] 20/2 39/3 64/5see [25] 14/8 15/24 19/1 21/5 22/5 22/10

22/16 22/16 22/19 22/25 23/25 24/6 24/19 34/11 43/24 44/11 44/20 48/12 52/2 52/9 58/19 58/21 61/1 66/5 66/8seemed [1] 39/24segregated [2] 59/6 59/7segregation [5] 58/22 58/23 58/24 58/25 59/4select [1] 12/2selecting [1] 12/3Senate [3] 9/12 53/9 54/4sense [2] 39/22 40/3sentence [19] 15/1 15/9 15/20 15/24 16/4 16/6 19/1 21/16 22/10 22/13 22/16 22/17 23/7 23/20 23/25 24/12 32/3 44/11 58/23Service [1] 54/12services [5] 1/23 72/10 72/11 72/12 72/14set [2] 13/7 56/25settled [2] 9/16 9/16seven [1] 15/10several [2] 68/19 69/6share [1] 60/24shares [2] 22/8 24/1she [2] 41/21 60/15SHEET [1] 74/1should [8] 7/4 74/5 74/7 74/9 74/11 74/13 74/15 74/17shouldn't [1] 65/20show [5] 19/8 19/21 20/13 29/2 29/4showed [1] 67/1shows [4] 12/2 20/24 45/5 66/25side [2] 8/4 8/5sign [1] 71/4signature [2] 71/3 73/18significant [4] 29/10 35/19 50/22 56/5silent [1] 64/8similar [1] 41/13simply [8] 20/11 40/15 40/21 57/10 62/15 63/6 63/14 68/17SIMS [1] 2/3since [1] 33/16single [1] 14/3sir [80] slight [1] 33/3slightly [1] 46/20small [2] 29/2 56/4SMITH [1] 2/3SNOW [1] 2/11so [68] 4/22 5/2 6/3 6/10 6/24 7/9 8/12 12/13 13/7 15/10 16/6

17/5 18/8 18/12 20/18 23/14 23/17 24/10 26/3 26/6 27/2 27/23 29/9 29/21 30/5 30/19 32/7 32/8 32/11 33/12 33/14 33/20 33/21 34/2 34/13 34/14 35/12 35/23 37/10 41/5 43/25 46/10 47/3 48/3 49/20 50/3 50/11 51/2 52/10 54/17 56/20 56/22 56/23 57/15 57/19 62/1 62/6 62/18 63/1 63/8 64/4 65/19 66/18 68/17 68/19 69/4 69/16 70/11social [2] 9/23 14/22software [27] 11/6 13/6 13/19 14/3 14/4 14/9 14/10 14/13 14/15 14/20 39/8 40/16 68/4 68/5 68/7 68/9 68/10 68/15 68/20 69/1 69/5 69/9 69/17 69/19 69/19 69/20 70/7some [10] 11/15 25/9 32/8 32/19 34/1 41/15 43/3 44/13 55/17 67/3somehow [3] 9/16 9/17 60/10something [7] 27/3 30/17 47/23 47/25 56/17 56/19 62/23sometime [1] 10/20sometimes [1] 62/19somewhat [2] 18/23 19/15sorry [4] 7/9 9/22 55/6 65/17sort [1] 69/10sought [4] 51/19 52/1 56/12 56/13source [3] 69/19 69/22 69/24South [7] 23/11 23/21 24/16 30/17 30/21 30/23 38/2Southern [1] 9/22speak [2] 38/1 60/17Special [1] 7/7specific [2] 35/8 51/1specifically [16] 19/11 22/22 23/23 26/18 29/4 29/19 30/2 30/5 36/10 39/16 44/3 46/25 50/25 56/6 56/20 65/10specifies [1] 25/17spell [1] 4/16split [59] 39/9 39/15 39/21 40/2 40/6 40/11 40/15 40/17 40/19 40/21 45/13 46/7 47/5 47/11 47/11 47/14 47/19 47/20 48/2 48/5 48/11 48/22 49/9 49/10 49/14 49/17 54/22 55/10 55/24

56/6 56/8 56/9 56/11 56/14 56/16 56/22 57/1 57/5 57/8 57/9 57/12 57/17 57/18 57/20 57/22 58/7 60/13 60/21 63/21 63/22 64/18 64/21 64/24 65/6 65/8 65/8 65/12 65/17 65/18splits [1] 39/10splitting [6] 50/14 56/18 63/12 64/10 65/11 65/16splitting of [1] 65/11spring [1] 10/21St [1] 8/2stadium [1] 34/12standard [6] 68/5 68/5 68/6 69/3 69/9 70/10stands [1] 70/18start [2] 5/22 55/19starts [5] 15/1 15/9 15/9 15/19 51/18state [19] 1/7 1/7 1/14 1/14 2/2 2/17 2/18 2/18 4/13 22/6 33/6 33/22 34/3 34/5 58/4 58/10 66/3 72/2 73/2State's [1] 41/23stated [6] 21/17 23/24 24/11 39/12 64/4 73/5statement [2] 23/19 59/16states [4] 1/1 30/23 38/6 41/13statistical [2] 62/3 63/9status [1] 66/21Stephen [1] 54/11still [2] 66/19 67/2straight [3] 62/15 62/16 62/18Strangia [5] 6/22 39/12 39/17 42/22 50/13Strangia's [6] 6/24 39/7 40/8 41/3 49/7 57/11Street [3] 1/20 2/14 2/21STRICKLAND [3] 2/20 48/7 72/12strong [2] 48/20 56/2strongly [2] 48/14 49/2struck [1] 39/21students [1] 34/1studied [2] 30/2 38/18study [1] 29/19sub [3] 15/4 15/23 38/21sub-precinct [3] 15/4 15/23 38/21subscribed [1] 74/22substantial [1] 68/7substantially [1]


Ssubstantially... [1] 52/6substitute [1] 19/19such [3] 20/16 65/15 73/17sufficient [1] 66/10suggested [1] 53/4suggesting [3] 49/22 50/6 50/15Suite [3] 2/6 2/14 2/20supermarkets [1] 33/16supplemental [1] 11/3support [14] 8/4 8/5 20/15 22/2 22/12 24/14 24/21 32/5 33/4 51/20 53/1 53/8 66/3 66/7supported [5] 52/16 52/20 53/5 53/15 66/11supporters [1] 52/20supporting [1] 24/9sure [33] 6/17 9/4 9/17 10/18 11/1 11/12 11/17 14/6 20/1 20/5 20/8 23/10 23/13 33/23 40/7 40/24 44/21 45/10 45/23 46/15 46/23 47/9 53/24 54/15 55/16 59/24 60/8 62/1 64/6 64/14 64/17 66/22 69/18sure I [1] 45/10SWANSON [1] 2/10swear [1] 4/3swore [1] 4/5sworn [2] 4/9 74/22

Ttable [45] 13/25 14/9 16/21 17/11 19/3 19/7 19/8 19/10 19/12 19/20 20/12 20/19 20/23 22/9 23/24 23/25 24/6 24/17 24/19 24/20 24/25 25/19 25/22 26/4 26/9 27/7 27/10 28/12 28/21 28/21 28/24 29/11 30/3 30/3 32/12 32/12 32/14 32/19 33/2 36/11 45/3 45/15 51/2 60/23 61/15tables [7] 16/14 17/18 17/21 17/23 18/4 30/1 30/6take [16] 5/4 5/6 12/5 20/4 25/15 25/17 29/24 30/7 35/12 44/5 47/20 51/6 54/18 60/5 67/10 67/13taken [1] 73/5taking [1] 72/14talk [8] 20/21 30/21 38/20 43/3 44/6 58/22

60/23 64/9talking [9] 21/25 23/20 42/18 42/20 53/14 54/24 59/21 59/24 61/14talks [1] 54/22TCB [1] 1/5TCB-WSD-BBM [1] 1/5telephone [1] 2/13tell [16] 7/14 10/8 12/6 25/16 33/9 34/15 34/22 37/15 37/16 38/11 38/16 48/8 48/15 54/17 59/2 69/7telling [3] 47/23 49/4 67/7tells [1] 47/22ten [1] 15/10Tennessee [1] 33/12term [1] 61/3terms [1] 43/19testified [10] 4/10 9/25 10/2 10/15 10/23 11/4 39/11 40/12 41/20 49/12testify [2] 9/6 9/7testifying [1] 42/8testimony [7] 16/10 39/18 41/3 49/25 50/1 57/11 74/3text [3] 20/9 20/10 20/24than [25] 5/15 18/17 22/12 23/4 24/13 27/7 29/18 29/18 30/18 32/4 44/17 44/25 46/12 46/15 56/10 56/25 57/8 57/17 58/15 59/16 61/12 64/19 64/25 66/16 66/18thank [5] 4/18 55/6 70/19 70/20 70/22that [391] that's [31] 6/23 8/12 10/1 13/1 13/22 18/18 25/15 28/14 28/16 32/15 33/16 35/3 40/19 40/21 40/23 43/24 45/1 48/23 49/3 53/7 53/16 53/17 59/21 61/9 61/14 63/8 63/8 63/9 67/17 69/22 70/17their [8] 18/11 39/23 40/1 40/16 42/14 49/15 64/12 66/21them [9] 27/1 27/11 40/12 41/16 41/25 42/13 56/24 57/13 66/20themselves [1] 46/8then [12] 10/3 10/24 14/20 15/8 17/3 23/3 32/18 36/18 47/12 48/7 63/6 66/18there [38] 5/22 6/5 6/13 6/17 9/24 16/14 20/14 23/25 24/6 25/8

25/20 28/4 33/2 34/13 34/13 34/21 34/22 35/19 38/16 41/15 46/11 46/14 46/16 46/22 46/23 49/1 50/16 50/18 50/21 53/25 54/13 56/1 56/2 56/14 56/15 59/23 61/6 66/23there's [5] 15/1 20/9 33/10 50/22 69/18therefore [1] 40/15thereto [1] 73/19these [20] 6/1 14/12 24/12 26/5 30/5 32/25 42/14 47/5 47/10 47/14 47/18 47/20 48/1 55/10 56/8 58/24 59/20 60/1 60/2 64/13they [25] 4/22 9/23 21/6 21/7 26/14 27/8 27/11 28/5 29/4 34/12 35/1 36/1 36/3 39/22 39/24 40/2 40/4 40/10 40/12 40/12 40/15 49/11 49/12 49/16 64/15they're [4] 23/4 25/7 42/18 42/20thing [2] 32/13 39/25things [5] 4/24 7/4 32/13 32/20 66/23think [22] 10/21 25/15 28/17 28/18 28/20 32/7 32/9 42/25 44/20 44/23 45/1 46/24 49/3 49/22 50/10 50/15 53/7 54/3 63/15 64/3 64/18 66/23this [44] 5/12 7/16 7/17 7/24 10/22 11/4 12/10 13/12 16/2 16/6 20/2 20/23 22/6 26/1 26/2 26/2 26/6 28/15 29/10 31/18 32/3 32/9 34/5 36/5 39/3 44/16 44/23 44/24 45/5 51/7 55/25 56/7 59/14 61/24 62/4 62/9 67/18 68/25 69/10 72/12 72/14 73/14 73/20 74/23THOMPSON [2] 1/10 2/9those [43] 6/3 6/10 6/15 7/8 7/8 7/10 17/22 18/7 20/13 21/8 21/20 24/10 27/3 27/3 27/9 29/9 29/11 32/20 35/12 35/22 38/7 39/6 39/25 40/25 43/19 47/10 47/16 51/1 53/12 54/17 56/9 56/22 57/4 57/9 57/16 57/18 57/19 57/20 57/22 57/23 59/10 60/5 69/2though [2] 37/17 49/5

thousand [2] 68/19 69/6thousands [1] 33/24three [10] 16/20 47/10 47/11 47/14 47/16 47/18 47/20 47/23 48/1 48/3threshold [1] 65/14through [4] 13/19 14/3 14/4 69/5throughout [1] 40/18throwing [1] 68/24time [17] 5/5 16/5 20/4 32/8 32/19 34/21 34/24 35/12 42/6 47/25 56/8 56/10 56/22 60/6 67/18 69/4 71/5times [7] 33/5 33/6 33/22 34/6 34/8 34/13 34/17together [5] 24/11 32/20 47/16 56/25 58/11told [2] 27/18 49/3too [2] 6/8 33/9top [11] 6/11 9/15 10/8 25/16 34/15 34/23 38/11 38/17 48/8 48/15 69/8topic [3] 38/13 38/18 41/8topics [1] 42/14total [3] 45/20 47/4 48/1totals [1] 45/16toward [3] 15/1 15/9 65/25Trace [1] 1/24trade [2] 69/16 69/18trademarked [1] 69/19training [1] 38/5transcript [12] 6/4 6/18 16/5 16/7 41/9 41/11 41/18 42/11 42/16 42/23 73/8 74/2traveled [1] 34/2treatment [1] 38/21trial [10] 9/7 9/17 9/25 10/2 10/8 10/14 10/15 10/22 10/23 11/4trip [1] 34/22trips [1] 33/15true [5] 7/21 21/13 51/10 73/8 74/3try [6] 22/23 22/24 27/25 28/2 28/6 51/22trying [2] 28/20 46/24turn [10] 12/17 14/24 16/13 21/22 25/19 26/22 45/3 46/19 51/16 67/8turned [5] 6/2 7/2 26/5 70/4 70/12turning [3] 7/24 14/2 24/17turnout [15] 6/1 18/6

20/25 25/21 26/3 29/2 29/3 30/4 45/17 46/11 46/14 46/16 46/17 46/22 46/24Tussahaw [8] 48/11 48/13 48/17 48/19 48/22 49/1 49/10 49/17Tussahaw's [1] 48/20two [19] 5/12 24/10 32/20 33/19 40/25 49/13 52/7 58/16 58/24 59/20 60/1 60/2 60/24 61/7 61/7 62/12 65/10 65/16 65/17two-party [1] 60/24typewriting [1] 73/7typically [2] 65/4 68/14TYSON [1] 2/3

UU.S [2] 53/9 54/4under [9] 2/5 18/23 19/14 21/12 21/21 63/5 72/9 72/14 73/7undersigned [2] 73/17 74/23understand [13] 4/25 14/14 15/25 28/3 40/1 40/8 42/17 42/19 44/20 52/10 60/8 66/22 67/8understanding [5] 16/9 38/25 39/17 50/1 55/24uniform [1] 40/18uniformity [1] 40/22UNITED [3] 1/1 30/23 38/6units [1] 39/6Unknown [3] 25/20 25/23 27/5Unknowns [2] 26/16 26/25unless [1] 73/17up [16] 13/24 33/10 33/19 33/20 34/1 34/4 43/25 47/10 47/13 47/21 60/10 60/11 65/12 65/16 66/25 67/1up that [1] 65/16uploaded [1] 70/9upon [3] 38/4 73/9 73/15urban [1] 43/17us [1] 27/18use [17] 7/11 11/8 11/25 12/10 12/10 12/12 12/22 22/21 32/1 36/24 37/5 41/13 50/8 50/12 69/10 70/7 70/7used [14] 6/11 11/7 12/7 12/20 13/6 13/11 18/4 28/22 40/4 41/17 64/12 68/21 70/9 70/10


Uuser's [1] 11/10using [9] 13/3 20/18 22/25 28/7 39/13 59/11 59/11 59/11 63/2

VVAP [2] 45/12 47/4Vargas [7] 1/22 1/23 72/10 72/11 72/14 72/21 73/24various [8] 5/25 26/5 32/22 37/1 40/14 54/17 57/19 69/25ve [1] 36/11version [2] 50/5 55/22versus [2] 18/12 27/2very [3] 12/12 34/4 62/20via [1] 2/13vote [8] 23/4 32/23 58/7 58/8 60/24 66/19 66/20 67/1voted [6] 26/14 26/25 52/11 53/25 66/19 67/2voter [8] 5/25 5/25 18/6 24/21 26/2 26/3 59/12 67/10voters [53] 8/17 9/10 9/11 9/22 10/24 18/5 18/9 18/22 19/13 21/2 22/11 22/12 22/20 23/1 24/13 24/13 24/19 24/24 25/5 25/9 27/2 27/20 27/20 27/22 27/24 28/17 30/15 31/1 32/4 32/4 32/24 35/1 35/6 35/17 35/20 35/22 35/24 37/9 37/13 38/10 38/15 39/9 51/20 52/11 52/15 53/5 53/8 53/25 66/3 66/11 66/14 66/15 67/8votes [5] 58/2 58/12 66/25 67/7 67/11voting [24] 26/19 27/4 28/6 32/22 35/16 35/23 36/18 36/20 37/1 38/10 39/9 39/14 40/4 41/7 45/6 52/7 56/3 61/20 61/25 62/17 63/4 66/9 66/20 69/10

WWake [4] 8/20 8/21 9/19 31/4want [11] 13/23 19/24 23/16 26/1 28/3 40/7 45/10 51/22 60/8 66/22 67/8wanted [4] 20/7 20/18 39/25 40/24was [117] Washington [2] 2/6 2/15

wasn't [1] 56/17way [8] 26/13 33/10 35/1 39/11 39/19 40/12 40/13 73/12WAYNE [1] 2/9we [24] 11/19 17/1 17/7 18/8 23/7 23/19 24/5 26/13 26/23 27/2 33/14 33/17 38/7 41/22 41/24 54/24 57/25 58/2 58/2 60/5 61/15 63/6 63/6 71/1we'll [2] 5/5 5/22we're [4] 45/11 55/24 59/24 70/23we've [1] 53/13well [30] 5/12 6/9 7/2 10/23 11/2 13/12 13/18 13/24 18/2 21/11 25/19 26/9 27/16 27/18 29/23 30/2 31/23 33/9 34/11 39/7 40/7 40/9 41/2 41/12 50/24 53/11 67/10 68/4 70/6 70/9went [6] 10/7 10/13 10/22 64/15 64/19 64/19were [37] 6/2 6/13 7/10 9/8 19/7 21/7 21/17 21/18 22/11 23/7 23/19 24/13 28/4 28/5 30/3 32/4 33/2 33/17 35/4 35/5 35/24 36/1 36/2 36/12 40/10 47/11 47/18 48/2 53/2 53/12 56/1 57/1 61/4 61/6 62/5 70/10 73/6weren't [1] 21/6what [90] what's [2] 7/13 37/16when [33] 5/14 11/25 12/6 13/14 13/18 16/2 17/7 17/21 20/21 30/13 30/21 33/19 33/24 34/20 34/21 35/4 38/1 38/20 41/22 41/24 47/20 54/4 55/24 57/25 58/4 61/15 62/4 62/10 62/18 63/1 65/9 66/14 68/2whenever [2] 34/14 56/21where [20] 6/20 15/6 28/16 31/25 33/14 33/16 41/3 44/3 45/24 45/25 48/7 51/22 52/2 52/9 53/15 53/25 58/19 58/21 59/22 59/25whether [25] 10/7 11/10 11/12 17/25 26/23 26/24 28/20 35/16 39/10 41/12 41/13 41/13 43/16 43/20 43/21 46/22 48/13 50/16 50/18 50/22 51/20 52/19 53/9 63/19 64/15

which [15] 5/17 8/3 12/1 20/24 24/5 32/22 33/2 35/8 38/5 42/9 48/4 65/6 65/9 66/25 67/6while [1] 5/6White [16] 22/12 24/13 25/2 25/5 25/9 29/18 32/4 50/16 50/18 50/23 52/11 52/17 53/10 54/1 54/19 66/20White-Black [1] 66/20Whites [3] 23/5 32/24 33/3who [13] 15/25 18/13 18/16 21/2 26/5 26/7 26/12 26/15 26/18 40/5 59/22 66/19 67/10whole [2] 29/21 58/8why [16] 23/23 46/11 46/14 46/19 46/22 47/4 47/6 55/13 55/14 55/15 55/16 55/20 56/9 57/19 59/7 61/19will [8] 5/1 6/1 55/19 62/7 62/19 65/5 68/15 72/14win [1] 67/3winner [2] 57/25 58/4withdraw [1] 67/12withdrawn [1] 73/14within [26] 18/22 19/13 21/2 29/5 29/5 32/16 32/17 35/20 39/15 39/15 39/21 40/2 40/11 40/14 40/17 40/19 40/21 45/12 45/17 46/12 46/15 46/20 47/4 47/12 57/12 62/8without [3] 46/12 46/15 53/9witness [7] 4/4 4/6 31/17 31/19 31/22 72/4 74/2Women [6] 8/17 9/10 9/11 9/21 10/24 31/1word [6] 12/5 25/15 25/18 29/24 44/5 54/18words [4] 41/17 42/10 62/21 65/13work [9] 4/22 8/8 8/25 11/25 31/13 31/19 35/4 35/8 35/13worked [4] 8/12 10/3 11/5 31/14works [1] 63/10world [3] 64/25 66/24 67/6would [41] 4/3 4/13 4/25 12/9 12/16 12/17 13/19 13/22 14/21 17/16 23/12 24/18 24/22 26/9 30/20 30/20 33/18 34/2 38/7 39/13 41/10 42/17

42/19 42/21 44/13 46/19 48/17 48/21 49/7 49/17 49/21 50/11 51/16 54/3 58/11 60/19 64/21 69/9 70/3 70/24 71/1would've [3] 33/15 33/18 49/16wouldn't [2] 17/6 50/3Wright [9] 6/5 11/24 16/6 16/7 41/20 42/22 50/12 60/3 60/14Wright's [11] 6/25 12/5 16/7 16/10 16/12 39/1 39/2 39/18 40/9 49/6 57/10write [1] 11/2written [1] 5/13wrote [5] 9/13 9/23 10/5 10/19 16/2WSD [1] 1/5

YYeah [1] 70/15years [1] 33/20yes [83] yesterday [2] 5/11 5/11York [1] 2/5you [241] you'd [4] 41/9 42/16 42/23 65/21you'll [9] 6/19 19/25 23/14 25/19 28/24 43/9 43/11 43/25 64/4you're [70] 4/18 11/17 12/3 12/11 13/3 13/21 15/7 15/11 15/13 16/3 16/3 17/11 17/18 17/22 21/8 22/17 23/6 24/20 25/4 25/12 25/22 25/23 26/17 28/17 28/18 28/20 30/15 31/14 32/2 32/7 35/13 36/4 36/9 36/10 37/2 37/22 38/25 40/23 44/14 44/16 44/21 45/14 45/19 45/23 46/5 46/24 47/3 47/24 47/25 48/20 49/4 49/4 49/22 50/5 52/13 52/18 53/20 58/19 58/19 58/21 60/18 61/14 61/19 62/1 62/11 63/15 66/24 67/4 67/5 67/5you've [2] 12/7 49/3your [76]

Zzero [3] 62/22 63/5 63/7zero percent [1] 63/5


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Filed 12/22/17 Page 1 of 44

EXPERT REPORT OF JOWEI CHEN. Ph.D.

I am an Associate Professor in the Department of Political Science at the University of

Michigan, Ann Arbor. I am also a Faculty Associate at the Center for Political Studies of the

Institute for Social Research at the University of Michigan as well as a Research Associate at the

Spatial Social Science Laboratory at Stanford University. In 2007,1 received a M.S. in Statistics

from Stanford University, and in 2009,1 received a Ph.D. in political science from Stanford

University. I have published academic papers on political geography and districting in top

political science journals, including The American Journal ofPolitical Science and The

American Political Science Review, and Election Law Journal. My academic areas of expertise

include spatial statistics, redistricting, racial politics, legislatures, elections, and political

geography. I have unique expertise in the use of geographic information systems (GIS) data to

study questions related to political geography and redistricting.

1 have provided expert reports in the following redistricting court cases: Missouri

National Association for the Advancement of Colored People v. Ferguson-Florissant School

District and St. Louis County Board of Election Commissioners (E.D. Mo. 2014); Rene Romo et

al. V. Ken Detzner et al. (Fla. 2d Judicial Cir. Leon Cnty. 2013); The League of Women Voters

of Florida et al. v. Ken Detzner et al. (Fla. 2d Judicial Cir. Leon Cnty. 2012); Raleigh Wake

Citizens Association et al. v. Wake County Board ofElections (E.D.N.C. 2015); Corrine Brown

et al. V. Ken Detzner et al. (N.D. Fla. 2015); City of Greensboro et al. v. Guilford County Board

ofElections, (M.D.N.C. 2015); Common Cause et al. v. Robert A. Rucho et al. (M.D.N.C.

2016); League ofWomen Voters of Pennsylvania et al. v. Commonwealth of Pennsylvania et al.

(No. 261 M.D. 2017). I have testified at trial in the following cases: Raleigh Wake Citizens

Association et al. v. Wake County Board ofElections (E.D.N.C. 2015); City of Greensboro et al.

V. Guilford County Board ofElections (M.D.N.C. 2015); Common Cause et al. v. Robert A.

Rucho et al. (M.D.N.C. 2016); League of Women Voters ofPennsylvania et al. v.

Commonwealth ofPennsylvania et al. (No. 261 M.D. 2017). 1 am being compensated $250 per

hour for my work in this case.

Research Questions and Summary of Findings:

The attorneys for the plaintiffs in this case have asked me to analyze House Districts 105

and 111 in the 2012 Georgia House districting plan, as created by Act No. 277 (S.B. 513) of

DEFENDANT'SEXi^lBIT

/



2012, and in the 2015 Georgia House districting plans, as created by Act No. 251 (2015 Ga. L.

1413) (H.B. 566) of2015. Specifically, I was asked to analyze:

1) Whether there is racially polarized voting within HD 105 and HD 111 under the two plans;

2) What the partisan results of the House races in HD 105 and HD 111 would have been in

November 2016 if these two House races had been held using the boundaries of the 2012 House

districting plan (Act No. 277); and

3) Whether race predominated in the drawing of HD 105 and HD 111 under the 2015 plan.

I answered these questions by analyzing individual-level voter registration files,

individual-level voter turnout history files, and precinct-level election results for Georgia's state

house elections held in November 2012, 2014, and 2016.1 also analyzed 2010 Census data

describing the racial and ethnic breakdowns of Georgia's precincts and Census blocks, as well as

shapefiles depicting the district boundaries within the 2012 and 2016 Plans.

In Georgia, residents are asked to select their racial identification when they register to

vote. However, voters are not given the opportunity to select a partisan affiliation. Therefore, the

publicly available voter registration list in Georgia contains information on the racial

identification of each individual voter, along with the precinct and the House district in which

each voter resides. I therefore analyzed this data in order to identify the number of voters of each

racial identity residing within each precinct and within HD 105 and HD 111, as drawn by both

the 2012 Plan and the 2015 Plans. Because Georgia voters are not asked to identify their partisan

affiliation, it is not possible to obtain or analyze data regarding voter partisanship or election

results at the sub-precinct level. In Georgia, election results are available only at the precinct

level.

By analyzing these precinct-level data, I concluded that voters in both HD 105 and HD

111 exhibit significant racially polarized voting. In both districts, virtually all Black voters

supported Democratic House candidates in the 2012, 2014, and 2016 House elections, while 75-

85% of non-Black voters supported Republican candidates. Thus, race is an extremely strong

proxy for partisanship in both districts.

Next, I estimated the hypothetical outcomes of the November 2016 House elections,

assuming they had been held under the old 2012 Plan boundaries for HD 105 and HD 111.1

found that, under the 2012 Plan boundaries, a Black Democratic candidate would have defeated a


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Filedl2/22/17 Page 3 of 44

White Republican candidate in November 2016, winning approximately 50.3%-54.4% of the

vote in the two districts.

Finally, I analyzed the motivations for the redrawing of HD 105 and HD 111 in the 2015

plan. First, I found that the 2015 plan decreased the African-American share of the turnout

electorate by 4.0 percentage points in HD 105 and by 2.9 percentage points in HD 111. Overall,

in HD 105 and HD 111, the 2015 Plan generally decreased compliance with traditional

districting principles and with the principles set forth in the "2011-2012 Guidelines for the House

Legislative and Congressional Reapportionment Committee" (Hereinafter: "Redistricting

Guidelines"). Given that race and partisanship are highly correlated within these two districts, I

also sought to analyze whether partisan considerations, rather than racial considerations, could

account for the drawing of the new district boundaries in the 2015 plan. I found that the

Legislature's primary map-drawer for the 2015 Plan had access only to racial data, but not

partisan data, at the sub-precinct level. Yet strikingly, I also found that the 2015 Plan splits three

precincts in HD 105 and five precincts in HD 111 in ways that consistently decreased the

African-American share of the population in both districts. These two findings demonstrate that

racial considerations, not partisanship, predominated in the drawing of the 2015 Plan boundaries

within these eight split precincts.

This report proceeds as follows. First, I describe my analysis of racially polarized voting

in HD 105 and HD 111. Second, I illustrate how increasing racial minority proportions caused a

pro-Democratic shift during 2012 to 2016 within the 2012 Plan's boundaries for HD 105 and

111. Third, I produce vote estimates of hypothetical November 2016 House elections held using

the previous 2012 Plan boundaries. Fourth, I describe how the 2015 Plan made a series of

changes to the boundaries of HD 105 and HD 111 that altered the racial composition of these

districts by subordinating traditional districting principles, including principles set forth in the

Redistricting Guidelines. Finally, I describe how the 2015 Plan's boundaries for HD 105 and HD

111 within split precincts appears to have been driven by racial considerations.

Racially Polarized Voting Analysis

To analyze whether there was racially polarized voting within HD 105 and HD 111 under

the 2012 and 2015 Plans, I first calculated precinct-level racial breakdowns of the turnout

electorate - the set of registered voters who cast ballots - within the boundaries of HD 105 and



HD 111 during the November 2012, 2014, and 2016 general elections. I then compared these

precinct-level racial breakdowns to the precinct-level House election results for HD 105 and 111

during these tliree elections.

To estimate the partisan voting patterns of each racial group within each district, I use

ecological inference (El), a commonly-used and widely-accepted statistical technique for

estimating different racial groups' political behavior when racial breakdowns of such behavior is

not directly reported in publicly-available data. El uses a procedure known as maximum

likelihood estimation, combined with Duncan and Davis' (1953) method of bounds, to estimate

the level support for a particular party's candidate among members of different racial groups

across the different precincts contained within a district. The key advantage of El is that it uses

observed election results and racial data jBrom all precincts within the district and estimates any

differences across precincts in a particular racial group's voting behavior.

Table 1 reports the El estimates of each racial group's tendency to support Democratic

candidates during the November 2012, 2014, and 2016 House elections in HD 105, while Table

2 reports the El estimates for HD 111. It is clear that both HD 105 and HD 111 exhibited

significantly racially polarized voting during each of these three elections. In HD 105,

approximately 98-99% of Black voters supported the Democratic candidate during the three

elections, whereas only 19-25% of non-Black voters supported the Democratic candidate.

HD 111 exhibited a similar pattern of racially polarized voting during each of the three

elections. Approximately 98-99% of Black voters supported the Democratic candidate during the

2012, 2014, and 2016 elections, whereas only 16-18% of non-Black voters supported the

Democratic candidate.


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Fi

led 12

/22/

17 Page 5 of 44

Table 1:

Ecological Inf

eren

ce and Ecological Re

gres

sion

Estimates of De

mocr

atic

Candidates' Share of Tw

o-Pa

rty Vote Among Among

Blacks and Non-Blacks in House District 105

Ecol

ogic

al Inference Estimates

Ecological Regression Estimates:

Black

Non-Black

Black

Non-Black

2012 House Election

99.0%

[98.0%, 99.7%]

21.4%

[21.0%, 22.0%]

100%

[100%, 100%]

7.4%

[0.4%, 10.3%]

2014 House Election

97.9%

[93.3%, 99.6%]

19.2%

[18.2%, 21.7%]

100%

[100%, 100%]

6.6%

[4.0%, 9.0%]

2016 House Election

99.3%

[98.8%, 99.7%]

25.2%

[25.0%, 25.4%]

100%

[100%, 100%]

10.2%

[7.8%, 12.6%]

[95% Confidence Intervals listed in br

acke

ts]


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Filedl2/22/17

Page 6 of 44

Table 2:

Ecological Inf

eren

ce and Ecological Regression Estimates of Democratic Can

dida

tes'

Share of Two-Party Vote Among Among

Blacks and Non-Blacks in House District 111:

Ecol

ogic

al Inference Estimates

Ecological Regression Estimates:

Black

Non-Black

Black

Non-Black

2012 House Election

98.2%

[90,1%, 99.8%]

18.1%

[17.2%, 20.4%]

100%

[100%, 100%]

8.0%

[6.0%, 10.0%]

2014 House Election

98.6 %

[94.9%, 99.8%]

15.7%

[14.9%, 17.9%]

100%

[100%, 100%]

7.4%

[5.4%, 9.5%]

2016 House Election

99.3%

[98.6%, 99.8%]

17.8%

[17.5%, 18.1%]

100%

[100%, 100%]

7.6%

[5.3%, 10.0%]

[95% Con

fide

nce Intervals listed in br

acke

ts]



Demographic and Partisan Changes in HD 105 and 111 under the 2012 Plan

Having found that HD 105 and HD 111 both exhibited racially polarized voting in the

2012, 2014, and 2016 House elections, I next analyzed the racial composition and partisan

performance of the two districts, as drawn by the 2012 Plan. Overall, my analysis revealed three

findings:

1) The African-American share of the turnout electorate increased noticeably from

November 2012 to November 2016 in both HD 105 and HD 111 under the 2012 Plan.

2) Non-African-American voters within HD 105, as drawn under the 2012 Plan, became

somewhat more likely to favor a Black Democratic House candidate in November 2016,

compared to previous elections.

3) As a result of these racial and partisan shifts. Democratic House candidates' vote share

significantly increased among voters residing within the 2012 Plan boundaries for HD 105 and

HD 111 from November 2012 to November 2016.

Below, I describe and illustrate these three findings in greater detail:

First, both HD 105 and HD 111, as drawn under the 2012 Plan, became more heavily

African-American from 2014 to 2016. This increasing African-American share of the electorate

within the 2012 Plan's HD 105 boundaries is illustrated in Table 3, which shows that African-

Americans comprised 35.2% of the Election Day turnout in November 2012, 35.7% in

November 2014, and 37,0% by November 2016 (counting only voters who reside within the

2012 Plan's HD 105 boundaries). Table 4 illustrates an even more significant increase in

African-American share of the electorate within the 2012 Plan's HD 111 boundaries: African-

Americans comprised 36.1% of the Election Day turnout in November 2012, 37.6% in

November 2014, and 40.3% by November 2016 (counting only voters who reside within the

2012 Plan's HD 111 boundaries). During all three elections, voters in both districts exhibited

significant racially polarized voting patterns, with African-Americans favoring Democratic

House candidates at a rate of around 98-99%. Thus, it is clear that this demographic pattern of

increasing African-American population within the 2012 Plan's HD 105 and HD 111 boundaries

would have caused a substantial increase in Democratic vote share by the November 2016 House

elections in both districts.



led 12

/22/

17 Page 8 of 44

Table 3:

HD 105 Precinct-Level Voter Turnout by Race under the 2012 and 2015 Pla

ns

2012 Election Turnout




Within HD 105

Within HD 105

Within HD 105

Within HD 105

Boundaries (2012 Plan)




Pet:

Precinct Name:

Black

Non-Black

Black

Non-Black

Black

Non-Black

Black

Non-Black

001

Harbins A

329'

1651'

060

Lawrenceville D

897'

705'

474'

379'

938'

893'

1088'

933'

071

Lawrenceville F

1083

1067

734

706

1143

1203

1143

1203

078

Baycreek K

488

1520

387

1009

681

1677

681

1677

080

Baycreek C

1149

2169

794

1492

1463

2394

1463

2394

091

Baycreek D

585

2130

433

1470

744

2204

744

2204

134

Baycreek F

845

1366

587

956

1105

1563

1105

1563

144

Lawrenceville M

1233'

1152'

865'

681'

1490'

1525'

146

Baycreek H

333'

1997'

246'

1396'

468'

2096'

468'

2096'

147

Baycreek I

973

1869

660

1246

1173

2110

1173

2110

151

Harbins C

249

1367

Tota

ls by Race:

7,586

13,9

755,180

9,335

9,199

15,6

608,

443

17,1

98(35.2%)

(64.8%)

(35.7%)

(64.3%)

(37.0%)

(63.0%)

(32.9%)

(67.1%)

Totals:

21,5

6114,515

24,945

25,6

41

* Indicates tha

t th

e pr

ecin

ct was spl

it int

o multiple districts, in

clud

ing House District 10

5. Only those vot

ers re

sidi

ng within

HD 105 are included in thi

s table's tu

rnou

t nu

mber

s. In particular, no

te that the HD 105

por

tion

of La

wrenceville D was different

unde

r the 2012 plan than under the 2015 pla

n. The

refo

re, the November 2016 tur

nout

numbers for

Lawrenceville D wit

hin HD 105 are

different under th

e 2012 plan and un

der th

e 2015 plan



led 12

/22/

17 Page 9 of 44

Table 4:

HD 111 Pre

cinc

t-Le

ve! Vo

ter Turnout by Race under the

2012 and 2015 Plans





Within HD 111

Within HD 111

Within HD 111

Within HD 111





Precinct Name:

Black

Non-Black

Black

Non-Black

Black

Non-Black

Black

Non-Black

26 - Tussahaw

25'

498'

29 - Lowes

1539

2516

1148

1669

2146

2823

2146

2823

31 - North Hampton

505'

1208'

395'

823'

659'

1240'

32 - Mount Carmel

949

1004

831

725

1517

1196

1023'

803'

34 - Wesley Lakes

1505

1436

993

947

1653

1463

1653

1463

35 - McDonough

534

611

38 - Hic

kory

Flat

795'

583'

40 - Stockbridge West

1570

777

1093

429

1690

750

41 - Sta

geco

ach

703

1298

505

849

735

1253

48 - Uni

ty Grove

328

2203

228

1596

386

2567

386

2567

50 - Pates Creek

991

2186

738

1515

1265

2096

1265

2096

51 - Oakland

591

1786

509

1355

911

1850

911

1850

53 - Fli

ppen

»

oooo

1106'

57 - Dutchtown

351

1372

273

973

425

1339

425

1339

59 - Grove Park

445

1559

61 - McDonough Cen

tral

288'

731'

185'

562'

348'

775'

302

759'

Totals by Race:

9,320

16,5

176,898

11,443

11,735

17,352

10,7

9418,057

(36.1%)

(63.9%)

(37.6%)

(62.4%)

(40.3%]

(59.7%)

(37.4%)

(62.6%)

Totals:

25,837

18,3

4129,087

28,851

* Ind

icat

es that the pr

ecin

ct was split into mu

ltip

le districts, i

ncluding House Dis

tric

t 111. Onl

y th

ose voters residing within HD 111

are included in this tab

le's

turnout numbers



Second, the non-African-American portion of the electorate in HD 105 exhibited a

noticeable increase in its support for a Black Democratic candidate in November 2016, compared

to earlier elections. This increase in Democratic support is illustrated by the Ecological Inference

estimates in Table 1, which predict that non-Blacks support for a Black Democratic candidate

increased from 21.4% in November 2012 to 25.2% in November 2016. This increase is partially

attributable to an increase in the Hispanic and Asian shares of the non-Black portion of the

electorate in HD 105, illustrated in Table 7, as Hispanic and Asian voters were more likely than

non-Hispanic white voters to support Black Democratic candidates.

As a result of these two demographic shifts within the 2012 Plan's HD 105 and HD 111-

the increase in African-Americans and other Democratic-supporting minority populations — both

districts would have exhibited a substantial increase in Democratic vote share in the November

2016 House elections, if not for the 2015 Plan's redrawing of the two districts' boundaries.

This pro-Democratic shift within the 2012 Plan's HD 105 and HD 111 boundaries is

clearly seen in Tables 5 and 6. These Tables show the actual precinct-level House election vote

counts for only those precincts that were assigned to HD 105 or HD 111 and whose district

boundaries were identical under both the 2012 and the 2015 Plans. In other words, these

precincts are the ones whose House district assignments were unaffected by the 2015 Plan

redistricting. There were seven such precincts in HD 105 (Table 5) and six such precincts in HD

111 (Table 6).

Table 5 illustrates that all seven precincts in HD 105 that were unaffected by the 2015

Plan substantially increased their Democratic vote shares in House elections from November

2012 to 2016; in fact, all seven precincts increased their Democratic vote share from November

2014 to 2016. For example, voters in Precinct 71 (Lawrenceville F) supported Democrat Renita

Hamilton at a 66.7% rate in November 2012 and 2014, but by November 2016, the precinct's

support for the Democratic candidate (Donna McLeod) increased to 70.2%. Overall, all seven

precincts increased their respective Democratic vote shares by a margin of 3.5 to 9.6 percentage

points between November 2012 and 2016.

10



led 12/22/17 Page 11 of 44

Table 5:

House Ele

ctio

n Results in

Pre

cinc

ts in which HD 105 Boundaries Remained Unchanged from the 20

12 Pla

n to

the 2015 Pl

an

2012 Ele

ctio

n Results (HD 105)

2014 Election Results (HD 105)

2016 Ele

ctio

n Results (HD 105)

Precinct Name:

Renita Hamilton

Joyc

e Ch

andl

erRenita Hamilton

Joyce Ch

andl

erDonna McLeod

Joyc

e Ch

andl

er(Black

(White

(Black

(White

(Bla

ck(White

Democrat)

Repu

blic

an)

Democrat)

Republican)

Democrat)

Repu

blic

an)

71-Lawrenceville F

1369

682

937

467

1571

667

78-B

aycr

eek K

683

1224

519

851

1018

1236

80-Baycreek C

1496

1662

1017

1229

1999

1701

91-B

aycr

eek D

898

1712

604

1262

1137

1702

134-Baycreek F

1039

1086

727

788

1495

1060

146-Baycreek H

542

1703

398

1215

817

1665

147-Baycreek I

1299

1407

878

988

1712

1449

2012 Ele

ctio

n Re

sult

s (HD 105)

2014 Ele

ctio

n Results (ED) 105)

2016 Election Results (ED) 105)

Precinct Name:

Democratic Candidate Vote Share



71-Lawrenceville F

66.7%

66.7%

70.2%

78-B

aycr

eek K

35.8%

37.9%

45.2%

80-Baycreek C

47.4%

45.3%

54.0%

91-Baycreek D

34.4%

32.4%

40.0%

134-Baycreek F

48.9%

48.0%

58.5%

146-Baycreek H

24.1%

24.7%

32.9%

147-Baycreek I

48.0%

47.1%

54.2%

Only

includes pr

ecin

cts in whi

ch the

boundaries of HD 105

did not cha

nge from the

2012 to th

e 2015 Plan.

11



led 12/22/17 Page 12 of 44

Table 6:

House Election Results in

Pre

cinc

ts in wh

ich HD 111 Bou

ndar

ies Remained Unchanged from the

2012 Plan to

the

2015 Pla

n

2012 Ele

ctio

n Vote Counts (HD 111)

2014 Ele

ctio


2016 Ele

ctio


Precinct Name:

Brian Strickland

Brian Strickland

Brian Strickland

Bill Blackmon

(White

Jim Nichols

(White

Darryl Payton

(White

(Black Democrat)

Republican)

(White Democrat)

Repu

blic

an)

(Black Democrat)

Republican)

29 - Lowes

1934

2083

1356

1439

2602

2283

34 - Wesley Lakes

1798

1029

1176

718

1991

1054

48 - Uni

ty Grove

513

1911

361

1428

621

2264

50 - Pates Creek

1307

1739

954

1238

1613

1628

51 - Oakland

780

1445

644

1165

1124

1532

57 - Dutchtown

520

1128

373

842

614

1095

2012 Ele

ctio

n Results (HD 111)

2014 Ele

ctio

n Results (HD 111)

2016 Ele

ctio

n Results (HD 111)

Precinct Name:




29 - Lowes

48.1%

48.5%

53.3%

34 - Wesley Lakes

63.6%

62.1%

65.4%

48 - Uni

ty Grove

21.2%

20.2%

21.5%

50 - Pates Creek

42.9%

43.5%

49.8%

51 - Oakland

35.1%

35.6%

42.3%

57 - Dutchtown

31.6%

30.7%

35.9%

Only includes precincts in which the

bou

ndar

ies of HD 105 did

not

cha

nge from the 2012 to the 2015 Plan.

12



Table 6 illustrates a similar pro-Democratic pattern for the six precincts in HD 111 that

were unaffected by the 2015 Plan: All six precincts substantially increased their Democratic vote

shares in House elections from November 2012 to 2016; all six precincts also increased their

Democratic vote share from November 2014 to 2016. For example, voters in Precinct 29

(Lowes) supported the Democratic House candidate at a 48.1% rate in November 2012 and a

48.5% rate in 2014, but by November 2016, the precinct's support for the Democratic candidate

(Darryl Payton) increased to 53.3%. Overall, all six precincts increased their respective

Democratic vote shares by a margin of 0.3 to 6.9 percentage points between November 2012 and

2016.

Overall, these two Tables illustrate that HD 105 and 111, as drawn by the 2012 Plan,

would have exhibited a substantial increase in Democratic vote share in the November 2016

House elections, if not for the 2015 Plan's redrawing of the two districts' boundaries. Among the

seven unaffected precincts in HD 105, the Black Democratic candidate's vote share increased by

7.1 percentage points from November 2012 to November 2016. Among the six unaffected

precincts in HD 111, the Black Democratic candidate's vote share increased by 4.2 percentage

points from November 2012 to November 2016. These changes in the unaffected portions ofHD

105 and 111 are attributable primarily to the increasing minority proportions of the electorate

within the two districts' boundaries under the 2012 Plan.

Given that voting patterns in these two districts are highly racially polarized, it is not

surprising that these partisan shifts coincided with racial shifts in the composition of the

electorate. Tables 7 and 8 show how the partisan shifts within the 2012 Plan's boundaries for HD

105 and 111 are clearly attributable to the increasing African-American, Hispanic, and Asian

proportions of the electorate within the two districts' boundaries under the 2012 Plan. For HD

105, Table 7 lists the racial breakdown of voters residing in the 2012 Plan's HD 105 boundaries

who turned out to vote in November 2012, 2014, and 2016. The final column then lists the racial

breakdown of November 2016 voters who turned out and who resided within the new HD 105

boundaries, as drawn by the 2015 Plan. Table 8 shows the analogous calculations for HD 111, as

drawn by the 2012 and the 2015 Plans.

13



ledl

2/22

/17

Page 14 of 44

Table 7:

HD 105 District-Wide Turnout by Race under the

2012 and 2015 Pla

ns

Raci

al Group:


Within HD 105



Within HD 105



Within HD 105



Within HD 105


White

10,8

85

(50.5%)

7,468

(51.5%)

10,8

00(43.4%)

12,5

54(49%)

Blac

k (n

on-H

ispa

nic)

7,586

(35.2%)

5,180

(35.7%)

9,199

(37%)

8443

(32.9%)

Hisp

anic

747

(3.5%)

372

(2.6%)

1397

(5.6%)

1178

(4.6%)

Asian or Pacific Islander

348

(1.6%)

175

(1.2%)

603

(2.4%)

552

(2.2%)

American Indian or

Alaskan Native

10

(0%)

7

(0%)

17

(0.1%)

22

(0.1%)

Other or Unknown

1,985

(9.2%)

1,313

(9.0%)

2,843

(11.4%)

2,892

(11.3%)

Total Turnout:

21,5

6114,515

24,859

25,641

* Indicates that th

e precinct was spl

it int

o mu

ltip

le districts, i

ncluding House District 105. Onl

y those vo

ters

res

idin

g within HD 105

are included in th

is tab

le's

tur

nout

numbers. In particular, not

e th

at the HD 105

por

tion

of La

wren

cevi

lle D was different und

er the

2012 pla

n than under the 201

5 pl

an. Therefore, th

e November 2016 tur

nout

numbers for

Law

renc

evil

le D wit

hin HD 105 are dif

fere

ntunder the 2012 plan an

d under th

e 2015 pla

n.

14


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Filed 12

/22/17

Page 15 of 44

Table 8:

HD 111 District-Wide Turnout by Race under the 2012 and 2015 Pla

ns

Raci

al Group:


Within HD 111



Within HD 111



Within HD 111



Within HD 111


White

13,3

49(51.7%)

9,422

(51.4%)

13,2

51(45.6%)

13,8

36(48.0%)

Black (non-Hispanic)

9,320

(36.1%)

6,898

(37.6%)

11,735

(40.3%)

10,794

(37.4%)

Hisp

anic

463

(1.8%)

259

(1.4%)

692

(2.4%)

679

(2.4%)

Asian or Pacific Islander

235

(0.9%)

113

(0.6%)

379

(1.3%)

436

(1.5%)

American Indian or

Alaskan Native

1

(0%)

0

(0%)

10

(0%)

8

(0%)

Other or Unknown

2,469

(9.6%)

1,649

(9.0%)

3,020

(10.4%)

3,098

(10.7%)

Total Turnout:

25,837

18,3

4129,087

28,8

51

* Indicates tha

t th

e precinct was sp

lit into mul

tipl

e di

stri

cts,

including House Dis

tric

t 11

1. Onl

y those voters res

idin

g wi

thin

HD 111

are included in this table's turnout numbers.

15



Together, these Tables illustrate a similar pattern in both districts. From November 2012

to 2016, each racial group's share of the total election-day turnout increased significantly within

the 2012 Plan's HD 105 boundaries: African-Americans increased from 35.2% to 37%,

Hispanics increased from 3.5% to 5.6%, and Asians increased from 1.6% to 2.4%. Similarly,

within the 2012 Plan's HD 111 boundaries, each racial group's share of the total election-day

turnout also increased significantly from November 2012 to 2016: African-Americans increased

from 36.1% to 40.3%, Hispanics increased from 1.8% to 2.4%, and Asians increased from 0.9%

to 1.3%. These increases in racial minority population explain why every single precinct that

remained within HD 105 and 111 in both the 2012 Plan and the 2015 Plan exhibited a noticeable

increase in Democratic candidate vote share in the 2016 House elections, as compared to the

2012 House elections.

Moreover, Tables 7 and 8 also illustrate how this trend of increasing racial minority

populations in HD 105 and 111 was successfully reversed by the 2015 Plan's redrawing of the

two districts. Within the 2012 Plan's boundaries for HD 105, the November 2016 turnout

electorate consisted of 37% African-American voters and 5.6% Hispanic voters. But within the

2015 Plan's new boundaries for HD 105, the November 2016 turnout electorate consisted of only

32.9% African-Americans and 4.6% Hispanics. A similar reversal occurred in HD 111: Within

the 2012 Plan's boundaries for HD 111, the November 2016 turnout electorate consisted of

40.3% African-American voters. But within the 2015 Plan's new boundaries for HD 111, the

November 2016 turnout electorate consisted of only 37.4% African-Americans. By redrawing

the boundaries of HD 105 and 111, the 2015 Plan decreased the racial minority proportions of

the electorate, thus reversing the demographic changes that had occurred within the 2012 Plan's

boundaries for HD 105 and 111 in recent years.

Estimate of November 2016 Election Results Under 2012 Plan Boundaries

Next, I estimated what the partisan results of the House races in HD 105 and HD 111

would have been in November 2016 if these two House races had been held using the boundaries

of the 2012 House districting plan (Act No. 277). Specifically, I assumed that the set of election-

day voters would have been exactly the same as the voters who actually turned out in November

2016. In total, I identified a total of 23,696 voters in HD 105 and 29,087 who satisfy the

following two criteria:

16



1. The voter cast a ballot in the November 2016 General Election.

2. As of November 2016, the voter resided within the 2012 Plan boundaries of HD 105 or

HD 111.

I then used this set of voters, along with their respective racial identifications on their

voter registrations, to construct estimates of hypothetical November 2016 House election

outcomes within the borders of HD 105 and HD 111 of the 2012 Plan. Specifically, I use

Ecological Inference (El) to derive predicted voting patterns by race and to estimate the rate at

which voters cast ballots in House elections. Because the November 2016 elections included a

US presidential race, and because turnout levels differ significantly between presidential and

non-presidential elections, 1 use the November 2012 House election results and precinct-level

turnout counts by race in order to derive precinct-level El estimates about racial voting patterns. I

then apply these racial voting estimates to precinct-level turnout counts by racial group in

November 2016 in order to estimate how many votes would have been cast for each party's

candidate in each precinct.

The November 2012 House elections featured a Black Democratic candidate and a White

Republican candidate in both the HD 105 (Renita Hamilton and Joyce Chandler) and HD 111

(Brian Strickland and Bill Blackmon) races, which were held using the boundaries of the 2012

Plan. Thus the El estimates derived using the results of this election give us reliable predictions

regarding the racial voting patterns within each precinct in a House election featuring a Black

Democratic and White Republican candidate during a presidential election year.

Table 9 reports the El estimates for HD 105, while Table 10 reports the El estimates for

HD 111. The first row of Table 9 reports, for example, that African-American voters in Gwinnett

County's Precinct 60 ("Lawrenceville D") who turn out to vote exhibit a roll-off rate of 8.3%,

meaning that 91.7% of those who turn out are expected to cast a vote in their House race; Among

those who cast a vote, 98.2% are expected to vote for a Black Democratic candidate. Meanwhile,

only 44.7% of non-Black voters in Precinct 60 would have supported a Black Democratic

candidate.

17



led 12

/22/

17 Page 18 of 44

Table 9:

El-B

ased

Est

imat

es of Hy

poth

etic

al 2016 Ele

ctio

n Re

sult

s Within the

Boundaries of HD 105 from the

2012 House Pla

n

El Estimates of Voter

El Estimates within HD 105

El-Based Estimates of Democratic Votes in

Roll-Off within HD 105

Boundaries from the 2012 House

November 2016 within HD 105 Boundaries from

Boundaries from the 2012

Plan

the 2012 House Plan

House Plan

Precinct Name:

El

El

El Estimates of

El Estimates of

El-based Estimate of


Estimates of

Estimates of

Black Voter

Non-Black

Nov. 2016 Votes for a


Black Voter

Non-Black

Support for a

Voter Support

Black Democratic

White Republican

Roll-Off:

Voter Roll-

Off:

Black

Democratic

Candidate:

for a Black

Democratic

Candidate:

Candidate:

Candidate:

60-Lawrenceville D'

8.3%

3.5%

98.2%

44.7%

1230

492

71-Lawrenceville F

6.2%

3.0%

98.2%

35.9%

1472

768

78-B

aycr

eek K

8.6%

3.9%

97.4%

17.0%

880

1354

80-B

aycr

eek C

7.5%

3.4%

97.9%

21.8%

1828

1838

91 -Baycreek D

6.6%

3.1%

98.1%

17.5%

1056

1774

134-Baycreek F

5.6%

2.8%

98.8%

19.5%

1316

1245

144-Lawrenceville M'

6.9%

3.2%

98.2%

39.6%

1940

913

146-Baycreek H '

6.8%

3.1%

98.2%

12.3%

677

1789

147-Baycreek 1

7.5%

3.3%

97.9%

23.1%

1534

1590

Total Estimated Votes


for a Black Democratic

for a White

Candidate:

Republican Candidate:

11,9

3311,763

* Indicates that th

e precinct was spl

it into mu

ltip

le dis

tric

ts, in

clud

ing House District 105. Only those vot

ers residing within HD 105

are in

clud

ed in the El

est

imat

es and est

imat

ed vote to

tals

reported in

this Ta

ble.

Note that an extremely sm

all po

rtio

n of pr

ecin

ct"Baycreek G" al

so lies within HD 105

from the 2012 Plan. However, th

ere were no registered voters within th

is portion, so Baycreek

G is not lis

ted on thi

s Table.

18



led 12

/22/

17 Page 19 of 44

Table 10:

El-Based Est

imat

es of Hypothetical 2016 Ele

ctio

n Re

sult

s Within the Boundaries of HD 111 from the 2012 House Plan

El Estimates of Voter

Roll-Off within HD 111


El Estimates within HD 111

Boundaries from the 2012 House

Plan

El-Based Estimates of Democratic Votes in

November 2016 within HD 111 Boundaries from

the 2012 House Plan

Precinct Name:

El

El

Estimates of Estimates of

Black Voter

Non-Black

Roll-Off:

Voter Roll-

Off:

El Estimates of

El Estimates of

Black Voter

Non-Black

Support for

a

Vote

r Support

Black

for a Black

Democratic

Democratic

Candidate:

Candidate:





Black Democratic

White Rep

ubli

can

Candidate:

Candidate:

29 - Lowes

1.0%

0.9%

98.5%

17.3%

2603

2366

31 - Nor

th Ham

pton

'3.1%

3.9%

98.3%

12.3%

803

1096

32 - Mount Carmel

1.4%

1.9%

98.7%

23.2%

1776

937

34 - Wesley Lakes

3.2%

4.6%

98.8%

26.2%

2023

1093

40 - Sto

ckbr

idge

West

4.6%

5.5%

99.2%

41.2%

1988

452

41 - Sta

geco

ach

3.5%

5.3%

98.7%

20.0%

982

1006

48 - Unity Grove

3.4%

4.3%

98.8%

9.5%

628

2325

50 - Pates Creek

3.0%

4.6%

98.6%

17.2%

1617

1744

51 - Oakland

4.6%

7.0%

98.4%

13.5%

1156

1605

57 - Dutchtown

3.5%

4.6%

98.4%

14.3%

611

1153

61 - McDonough

Central*

3.6%

5.1%

98.4%

13.1%

447

676



for a Black Democratic

for a White

Candidate:

Republican Candidate:

14,6

34

14,4

53* Indicates that th

e pr

ecin

ct was spl

it int

o multip

of the 2012 Plan are included in the El estimates

e di

stri

cts,

including House District 111. Only th

ose voters res

idin

g wi

thin

HD 111

and estimated vote tot

als reported in th

is Table

.

19



From these El estimates, I am able to predict that Precinct 60 would have produced a

total of 1,230 votes for a Black Democratic candidate and 492 votes for a White Republican

candidate. Note that Precinct 60 was split by the 2012 Plan into HD 104 and HD 105, and this El

analysis considers only voters who resided, as of November 2016, within the boundaries of HD

105 from the 2012 Plan.

Applying this methodology to all precincts within the two districts, I find that both HD

105 and HD 111 would have been won by a Black Democratic candidate in November 2016 if

these House races had been held using the boundaries of the 2012 Plan. Specifically, as Table 9

illustrates, voters in HD 105 would have favored the Democratic over the Republican candidate

by 11,933 to 11,763 votes. Meanwhile, voters in HD 111 would have favored the Democratic

over the Republican candidate by 14,634 to 14,453 votes, as illustrated in Table 10.

Yet even these El estimates likely under-estimate the number of Democratic voters

residing within both districts as of November 2016. The El estimates used in Tables 9 and 10 are

based on voting patterns observed in the November 2012 House elections. From 2014 to 2016,

these voting patterns shifted noticeably in a pro-Democratic direction, due to increases in racial

minority proportions in HD 105 and 111, as described earlier in this report. This pro-Democratic

shift in the two districts suggests that the use of El-based predictions likely under-estimates the

true Democratic vote share in a hypothetical November 2016 election held using the 2012 Plan's

boundaries.

Hence, a more realistic method of estimating hypothetical November 2016 election

outcomes involves using actual precinct-level House election outcomes from November 2016 for

those precincts that were not removed from HD 105 or 111 by the 2015 Plan; for precincts that

were removed, the same El predicted results are used.

20



ledl

2/22/1

7 Page 21 of 44

Table 11:

Combined Est

imat

es of Hypothetical 2016 Election Re

sult

sWithin the Boundaries of HD 105 from the 2012 House Plan

Estimates of November 2016 Votes for a Black

Democratic Candidate within the Boundaries of HD 105

of the 2012 House Plan

Estimates of November 2016 Votes for a White

Republican Candidate wit

hin th

e Boundaries of HD 105


Precinct Name:

Actual Nov. 2016 Votes for

Dem. Donna McLeod (in

precincts where HD 105 was

not altered by the 2015 Pla

n):

Bl-based Estimate of Nov.

2016 Votes for a Black

Democratic Candidate (i

nprecincts where HD 105

was

alte

red by the 2015 Plan):


Rep. Joyce Cha

ndle

r (i

npr

ecin

cts where HD 105

was

not al

tere

d by the 2015 Pla

n):

El-based Estimate of Nov.



nprecincts wh

ere HD 105

was

alte


60-Lawrenceville D '

1230

492

71-Lawrenceville F

1432

484

78-B

aycr

eek K

1018

1236

80-B

aycr

eek C

1999

1701

91-Baycreek D

1137

1702

134-

Bayc

reek

F1495

1060

144-Lawrenceville M'

1940

913

146-Baycreek H'

817

1665

147-Baycreek 1

1712

1449

Combined Total Estimated Votes for a Black Democratic

Candidate:

Combined Tot

al Estimated Votes for

a White Rep

ubli

can

Candidate:

12,7

80(54.4%)

10,702

(45.6%)

* Indicat

es tha

t the pr

ecin

ct was spl

it into multiple districts, i

ncluding House District 111. Onl

y th

ose voters residing within HD 111

of the 2012 Plan are in

clud

ed in the El est

imat

es and

estimated vote to

tals

reported in thi

s Table.

21


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Fl

ledl

2/22/1

7 Page 22 of 44

Table 12:

Combined Est

imat

es of Hyp

othe

tica

l 2016 Ele

ctio

n Re

sult

sWithin the Boundaries of HD 111 from the 2012 House Plan

Estimates of November 2016 Votes for a Black

Democratic Candidate within the Boundaries of HD 111


Estimates of November 2016 Votes for a White

Republican Candidate wit

hin th

e Boundaries of HD 111


Precinct Name:


Dem. Dar

ryl Payton (in

prec

inct

s where HD 111 was

not al

tere

d by the

2015 Pla

n):




npr

ecin

cts wh

ere HD 111

was

alte



Rep. Bri

an Strickland (i

nprecincts wh

ere HD 111

was

not altered by the

2015 Pla

n):


2016 Votes for a White

Repu

blic

an Candidate (in

prec

inct

s wh

ere HD 111

was

altered by the 2015 Pla

n):

29 - Lowes

2602

2283

31 - Nor

th Ham

pton

'803

1096

32 - Mount Carmel

1776

937

34 - Wesley Lakes

1991

1054

40 - Sto

ckbr

idge

West

1988

452

41 - Sta

geco

ach

982

1006

48 - Unity Grove

621

2264

50 - Pates Creek

1613

1628

51 - Oakland

1124

1532

57 - Dutchtown

614

1095

61 - McDonough

Cent

ral*

447

676

Combined Total Estimated Votes for a Black Democratic

Candidate:

Combined Total Estimated Votes for a White Rep

ubli

can

Candidate:

14,5

61

(50.9%)

14,023

(49.1%)

* Indicates tha

t the pr

ecin

ct was spl

it into mu

ltip

le dis

tric

ts, i

ncluding House Dis

tric

t 11

1. Only those voters residing within HD 111

of th

e 2012 Pla

n are in

clud

ed in the El est

imat

es and

est

imat

ed vot

e totals reported in thi

s Ta

ble.

22



Table 11 illustrates this method for HD 105. Under the 2012 Plan, HD 105 contained

nine precincts, including three split precincts. Of these nine precincts, seven were unaffected by

the 2015 Plan: The same portions of these seven precincts assigned to HD 105 under the 2012

Plan were again assigned to HD 105 under the 2015 Plan. The remaining two precincts were

affected by the 2015 redistricting: Precinct 144 (Lawrenceville M) was completely removed

from HD 105, while the borders ofHD 105 were altered within Precinct 60 (Lawrenceville D).

Moreover, all of Lawrenceville M and portions of Lawrenceville D were reassigned to HD 104

in the 2015 Plan. In November 2016, HD 104 featured an uncontested House race with no

Democratic candidate; therefore, no meaningful election results from November 2016 are

available for these two reassigned precincts.

For the seven precincts in HD 105 unaffected by the 2015 Plan, Table 11 simply counts

the number of House election votes received by the Black Democratic candidate (Donna

McLeod) and the White Republican candidate (Joyce Chandler) in November 2016, with no El

estimates used. For the remaining two precincts affected by the 2015 Plan - Precincts 60 and 144

- Table 11 uses the same El-based predictions derived previously in Table 9.

Table 11 sums together these actual election results and El-based predictions for the nine

precincts within the 2012 Plan boundaries of HD 105. In total, a Black Democratic candidate is

expected to receive 12,780 votes, whereas a White Republican would receive 10,702 votes in a

November 2016 House election held using the 2012 Plan boundaries for HD 105. The Black

Democratic candidate's predicted vote share of 54.4% accounts for the increases in African-

American and other minority populations that occurred in HD 105 during 2012-2016, thus

producing an even more Democratic-leaning prediction than the Table 9 estimates, which solely

relied upon El predictions about racial voting patterns from 2012 election data.

Table 12 uses this identical methodology for HD 111, yielding a similar prediction. The

2012 Plan boundaries for HD 111 contained all or split portions of 11 different precincts. Five of

the precincts were affected by the 2015 Plan redistricting, while the House district assignments

for the remaining six precincts were unaffected by the 2015 Plan. For the six unaffected

precincts. Table 12 uses the actual election results from the November 2016 between Darryl

Payton (Democrat) and Brian Strickland (Republican). For the remaining five precincts that

affected by the 2015 Plan's redrawing of the HD 111, Table 12 uses the El-based predictions

23



derived previously in Table 10 to estimate the number of votes that would have been cast in

favor of a Black Democrat and a White Republican candidate.

In total. Table 12 predicts a Black Democratic candidate would receive 14,561 votes,

whereas a White Republican would receive 14,023 votes in a November 2016 House election

held using the 2012 Plan boundaries for HD 111. Once again, this prediction, by using actual

November 2016 election results in precincts unaffected by the 2015 Plan redistricting, yields a

somewhat higher Democratic vote share than the purely El-based predictions in Table 10.

Changes Made to HD 105 and HD 111 by the 2015 Plan

In analyzing the sequence of changes made to HD 105 and HD 111 by the 2015 Plan, I

first reviewed the November 20, 2017 deposition testimony of Ms. Gina Wright, the

Legislature's primary map-drawer for the 2015 Plan. In this section, I describe how the 2015

Plan made a series of changes to the boundaries of HD 105 and HD 111 that altered the racial

composition of these districts by subordinating traditional districting principles, including

principles set forth in the Redistricting Guidelines. I first describe the series of changes the 2015

Plan made to HD 105 and 111 and how these changes noticeably decreased the African-

American population in both districts. I then describe how these racial changes to the district

boundaries subordinated traditional districting principles.

Changes to HD 105: As constructed by the 2012 Plan, HD 105 had a Black Voting Age

Population (BVAP) of 33.1%, based on 2010 Census data. The 2015 Plan made a series of three

changes to the boundaries of HD 105, and all three of these changes uniformly had the effect of

decreasing the African-American population of HD 105. This sequence of changes to HD 105

was described by Ms. Wright in her deposition in this case, and 1 relied upon pp. 214-219 from

the transcript of her deposition in analyzing her she redrew the HD 105 boundaries.

First, the 2015 Plan added two precincts to the eastern portion of HD 105, and both of

these added precincts are comprised of overwhelmingly White population, thus decreasing the

overall BVAP of HD 105. The two precincts added to HD 105 were Harbins C (11.2% BVAP)

and a portion of Harbins A (containing a 14.5% BVAP). Because of the overwhelmingly White

populations in Harbins A and Harbins C, these additions to HD 105 each served to decrease the

African-American and Hispanic proportions of HD 105.

24



To offset the population gain caused by adding Harbins A and Harbins C, the 2015 Plan

then removed more heavily African-American population from the northwestern end of district

in the Lawrenceville area. Specifically, at the northwestern end of district, the 2015 Plan

removed the portion of the Lawrenceville M precinct that had previously been assigned to HD

105 under the 2012 Plan. This portion of Lawrenceville contains a BVAP of 45.6%. To partially

offset the population loss caused by the removal of Lawrenceville M, the 2015 Plan slightly

expanded the portions of the adjacent Lawrenceville D precinct assigned to HD 105. These two

alterations to HD 105 had the overall effect of decreasing the Voting Age Population by 3,713 of

which 35.1% consisted of African-Americans. Hence, these two alterations combined to decrease

the overall BVAP of HD 105. Altogether, combined with the Harbins A and Harbins C additions

to the district, these changes decreased the BVAP of HD 105 from 33.1% under the 2012 Plan to

31.0% under the 2015 Plan.

Changes to HD 111: As constructed by the 2012 Plan, HD 111 had a Black Voting Age

Population (BVAP) of 33.9%, based on 2010 Census data. The 2015 Plan made a similar series

of changes in HD 111, having the overall effect of decreasing the BVAP of the district by 2.2%.

I relied upon pp. 183 and pp. 224-230 of the transcript of Ms. Wright's deposition in analyzing

her she redrew the 111 105 boundaries.

First, the 2015 Plan removed the northernmost precincts in HD 111, which includes areas

with substantial black population. Most notably, the 2015 Plan removed the Stockbridge West

precinct from the northern portion of the district. Stockbridge West's BVAP of 59.8% is the

highest of any precinct in District 111 in the 2012 plan. Having removed Stockbridge West, the

2015 Plan was then also forced to remove the Stagecoach precinct (28.3% BVAP) because this

precinct would not have been contiguous with the rest of HD 111 if it remained in the district.

Overall, then, the removal of Stockbridge West and Stagecoach constituted the removal of a

northern portion of HD 111 consisting of 45.2% BVAP. In the northern portion of HD 111, the

population loss created by the removal of Stockbridge West and Stagecoach were partially offset

by the addition of part of the Flippen precinct. This portion of Flippen contains a BVAP of

38.7%, which is lower than the combined BVAP of the removed Stockbridge West and

Stagecoach precincts. Thus, replacing Stockbridge West and Stagecoach with this portion of

Flippen served to decrease the overall BVAP ofHD 111.

25



Additionally, the 2015 Plan further removed portions of the North Hampton and Mount

Carmel precincts from the western side of HD 111. The removed portions of these two precincts

had a combined BVAP of 32.7%. To offset this population loss created by the removal of the

northern and western precincts in HD 111, the 2015 Plan expanded the southeastern portion of

HD 111 into a predominantly White portion of Henry County. Specifically, the entirety of the

Grove Park precinct (21.4% BVAP) and a portion of the Tussahaw precinct (4.8% BVAP) were

added to HD 111, and these additions substantially decreased the overall BVAP of HD 111.

Finally, the 2015 Plan made alterations to the HD 111 boundaries within the split precinct

of McDonough Central that again had the effect of decreasing the district's overall African-

American population. The redrawing of the HD 111 boundaries within the McDonough Central

precinct caused a noticeable decrease in the BVAP of the portion of McDonough Central

belonging to HD 111. The BVAP of the portion of the McDonough Central precinct lying within

HD 111 decreased from 30.2% under the 2012 Plan to 26.9% under the 2015 Plan.

The 2015 Plan and Traditional Districting Principles: In evaluating the 2015 Plan's

compliance with traditional districting principles, I first examined the principles set forth in the

"2011-2012 Guidelines for the House Legislative and Congressional Reapportionment

Committee" (Hereinafter: "Redistricting Guidelines"). These principles include population

equality, geographic compactness, and avoiding the splitting of precincts and counties. Tables 13

and 14 describe various characteristics of HD 105 and HD 111 under both the 2012 and the 2012

Plans.

First, I found that the 2015 Plan creates more significant population deviations in both

HD 105 and HD 111, relative to the 2012 Plan. Georgia's total population, as of the 2010

Census, is 9,687,653, so the ideal district population for a House district is 53,820.29. HD 105's

population deviation increased in magnitude from 102 under the 2012 Plan to 278 under the

2015 Plan. Similarly, HD 111's population deviation increased from 376 under the 2012 Plan to

472 under the 2015 Plan. Although these deviations are not especially large, it is nevertheless

clear that the 2015 Plan had a slight deleterious effect on compliance with population equality in

both districts.

26



led 12

/22/17

Page 27 of 44

Table 13:

Characteristics of House District 105:

2012 Plan:

2015 Plan:

Split Pr

ecin

cts:

Lawrenceville D

Lawrenceville M

Baycreek H

Harbins A

Lawrenceville D

Baycreek H

Split Municipalities:

Grayson (1

05,1

06,1

14)

Lawrenceville (

101,102,104,105,106,107)

Snel

lvil

le (9

3,10

5,10

6,10

7)

Grayson (1

05,1

06,1

14)

Lawrencevil l

e (101,102,104,105,106,107)

Snel

lvil

le (9

3,10

5,10

6,10

7)

Reock Compactness:

0.3213

0.3175

Popp

er-P

olsb

y Compactness:

0.2354

0.2280

Total Po

pula

tion

(2010 Cen

sus)

:53,718

53,542

Tota

l Voting Age Pop

ulat

ion (2010

Cens

us):

36,580

36,449

Any Par

t Bl

ack Voting Age

Popu

lati

on (2010 Cen

sus)

:12

,100

(33.

1%)

11,313(31.0%)

November 2016 Total Turnout:

24,945

26,6

41

November 2016 African-American

Turnout:

9,199 (36.9%)

8,44

3 (32.9%)

* Geo

rgia

's sta

tewi

de pop

ulat

ion as of the 2010 Cen

sus was 9,6

87,6

53, s

o the id

eal di

stri

ct pop

ulat

ion is

53,820.29.

27



led 12/22/17 Page 28 of 44

Table 14:

Characteristics of House District 111:

2012 Plan:

2015 Plan:

Split Precincts:

McDonough Cen

tral

Nort

h Hampton

Flip

pen

Hickory Fl

atMcDonough Central

Mount Carmel

Tussahaw

Split Mu

nici

pali

ties

:Lo

cust

Grove (1

11,1

30)

McDonough (1

09,1

10,1

11)

Stockbridge (

76,78,90,109,111)

Locu

st Grove (111,130)

McDonough (10

9,11

0,11

1)Stockbridge (

76,7

8,90

,109

,111

)

Reock Compactness:

0.3160

0.3153

Popp

er-P

olsb

y Compactness:

0.1317

0.1234

Tota

l Po

pula

tion

(2010 Cen

sus)

:54,197

54,293

Tota

l Vo

ting

Age Pop

ulat

ion (2010

Census):

38,545

38,235

Any Part Black Vot

ing Age

Popu

lati

on (2010 Census):

13,068 (33.9%)

12,1

03(3

1.7%

)

November 2016 Total Turnout:

29,087

28,851

November 2016 African-American

Turnout:

11,7

35 (40.3%)

10,7

94 (37.4%)

Georgia's statewide population as of th

e 2010 Cen

sus was 9,687,653, so th

e ideal district population is

53,

820.

29.

28



Next, I then evaluated how the 2015 Plan affected the geographic compactness of the two

districts. Once again, I found that the 2015 Plan worsened compliance with the Redistricting

Guidelines in both districts. I measured the geographic compactness of the two districts under

both plans using the Reock score and the Popper-Polsby score. Both of these compactness

measures are commonly used used by scholars of redistricting, and with measures, higher scores

indicating greater compactness, while lower scores indicate more non-compact districts.

First, I calculated the average Reock score of both districts under both plans. The Reock

score for each individual district is calculated as the ratio of the district's area to the area of the

smallest bounding circle that can be drawn to completely contain the district. HD 105's Reock

score slightly worsened from 0.3213 under the 2012 Plan to 0.3175 under the 2015 Plan.

Similarly, HD 111's Reock score slightly worsened from 0.3160 under the 2012 Plan to 0.3153

under the 2015 Plan.

Second, I calculate the average Popper-Polsby score of both districts under both plans.

The Popper-Polsby score for each individual district is calculated as the ratio of the district's area

to the area of a hypothetical circle whose circumference is identical to the length of the district's

perimeter. HD 105's Popper-Polsby score slightly worsened from 0.2354 under the 2012 Plan to

0.2280 under the 2015 Plan. Similarly, HD 111's Popper-Polsby score slightly worsened from

0.1317 under the 2012 Plan to 0.1234 under the 2015 Plan. Hence, it is clear that using either the

Reock or the Popper-Polsby measure, the 2015 Plan had a slight deleterious effect on the

geographic compactness of both districts.

Next, I evaluated how well the two plans avoided splitting precincts in HD 105 and HD

111.1 found that HD 105 split two precincts (Lawrencevilie D, and Lawrenceville M) under the

2012 Plan and two precincts (Harbins A, and Lawrenceville D) under the 2015 Plan. Thus, the

2015 Plan neither increased nor decreased the total number of precincts split by HD 105.

In HD 111, however, the 2015 substantially increased the number of split precincts from

two to five. Under the 2012 Plan, HD 111 split only two precincts (North Hampton and

McDonough Central). The 2015 Plan removed North Hampton from HD 111, also added partial

fragments of three new precincts (Tussahaw, Hickory Flat, and Flippen), while removing a

portion of Mount Carmel from the district. These numerous changes resulted in HD 111 splitting

five precincts (McDonough Central, Mount Carmel, Tussahaw, Hickory Flat, and Flippen) under

29



the 2015 Plan. Hence, the 2015 Plan severely worsened HD 111 with respect to the traditional

districting criterion of avoiding precinct splits.

I further evaluated how well the two plans followed county boundaries in the areas

surrounding HD 105 and HD 111. Overall, 1 found that the 2015 Plan significantly deviates from

the Redistricting Guidelines' principle of avoiding county splits in both Gwinnett County and in

Henry County. In general, it is relatively straightforward to draw a legislative plan with equally

populated, contiguous districts while having, at most, one district in each county that crosses the

county's borders. In both Gwinnett and Henry Counties, the 2015 Plan has far more districts

crossing county borders than is necessary.

In Gwinnett County, there are six districts - HD 81, 93, 94, 95, 103, and 114 - that cross

the borders of Gwinnett County. That is, each of these six districts partially lies within Gwinnett

and partially lies within another adjacent county. Similarly, Henry County has seven districts -

HD 73, 76, 78, 90, 109, 110, and 130 - that cross the borders of Henry County. Each of these

seven districts partially lies within Henry County and partially lies within another adjacent

county. Overall, such a large number of cross-county districts involving Gwinnett and Henry

Counties indicates a failure of the 2015 Plan to follow county boundaries in the drawing of the

House districts in these two counties.

A further measure of any districting plan's adherence to county boundaries is the number

of districts that lie fully within any given county's boundaries. Gwinnett County has a 2010

Census population of 842,046, so with an ideal House district population of 53,830, Gwinnett

can geographically accommodate up to 15 full House districts within its borders. However, under

the 2015 plan, Gwinnett County contains only 12 House districts lying fiilly within its borders.

The remaining six districts in Gwinnett County - HD 81, 93, 94, 95, 103, and 114 - spill over

into adjacent counties. Hence, it is clear that the 2015 Plan created districts in Gwinnett County

that generally failed to respect county boundaries.

Yet more striking is the failure of the 2015 Plan to draw districts respecting the

boundaries of Henry County. Henry County has a 2010 Census population of209,053, so with an

ideal House district population of 53,830, Henry County can geographically accommodate up to

3 full House districts within its borders. However, the 2015 Plan creates only one district (HD

111) that lies completely within the boundaries of Henry County. The remaining seven districts

in Henry County - HD 73, 76, 78, 90, 109, 110, and 130 - spill over into adjacent counties.

30



Hence, it is clear that the 2015 Plan created districts in Henry County that generally failed to

respect county boundaries.

Finally, although the avoiding of municipal splits is not specifically mentioned in the

Redistricting Guidelines, I analyzed HD 105 and HD 111 with respect to its adherence to

municipal boundaries because avoiding municipal splits is a traditional districting principle.

Specifically, I found that the 2015 Plan splits up municipalities within HD 105 and HD 111 into

significantly more fragmented pieces than was necessary.

HD 105 contains three municipalities that are split up into several districts by the 2015

Plan. Most notably the city of Lawrenceville contains a population of 28,546, so can be fully

contained within a single House district. Yet, the 2015 Plan splits of Lawrenceville into portions

of seven different House districts (HD 101, 102, 104, 105, 106, 107). This fragmentation of the

city of Lawrenceville is especially noteworthy because of the city's significant African-

American and Hispanic population (32.0% African-American and 22.3% Hispanic population).

As described earlier in this section, the 2015 Plan altered the boundaries of HD 105 in

Lawrenceville's precincts in ways that decreased the BVAP of the district's population in the

Lawrenceville area.

Similarly, the 2015 Plan splits the city of Grayson into three different House districts,

while the city of Snellville is split into four different House districts. Given the small populations

of both cities (2,666 for Grayson and 18,242 for Snellville), each of these two cities could have

been fully accommodated within a single House district. Overall, then, the 2015 Plan's splitting

of Lawrenceville, Grayson, and Snellville into several districts indicates a significant failure to

respect municipal boundaries in the drawing of HD 105.

In HD 111, the 2015 Plan split three different municipalities: Locust Grove, McDonough,

and Stockbridge. Most notably, the 2015 Plan splits the city of McDonough into three different

districts (HD 109, 110, and 111), even though McDonough's population of22,084 could easily

have been accommodated within a single House district. This decision to split McDonough, a

majority-African-American city, is notable: As described earlier in this section, the 2015 Plan

altered the boundaries of HD 111 in McDonough's precincts in ways that incorporated less-

heavily African-American portions of McDonough into HD 111 while excluding some more

heavily African-American portions of McDonough from HD 111.

31


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Filedl2/22/17 Page 32 of 44

Equally notable is the precise way in which the 2015 Plan split the city of Stockbridge,

which also has a majority African-American population, into five districts (HD 76, 78, 90, 109,

111). Stockbridge's 2010 Census population of 25,636 could easily have been accommodated

within a single House district. But the 2015 Plan's decision to split up Stockbridge into five

separate districts had important racial consequences: As described earlier in this section, the

2015 Plan removed one of Stockbridge's precincts containing a majority African-American

population (the Stockbridge West precinct) while redrawing HD Ill's boundaries in the northern

portion of the district in a manner that excluded much of the Stockbridge's African-American

population from HD 111. Yet the 2015 Plan did not simply re-assign the majority-African-

American population of Stockbridge into another adjacent district. Rather, the heavily African-

American population of Stockbridge was scattered into portions of four other districts (HD 76,

78, 90, 109) by the 2015 Plan, thus minimizing the influence that Stockbridge voters would have

within any single House district.

Overall, I find that the 2015 Plan redrew the boundaries of HD 105 and HD 111 through

a series of changes that consistently decreased the African-American proportion of both districts.

Moreover, I find that the redrawing of the two districts' boundaries along these racial lines came

at the sacrifice of adherence to traditional districting principles: The new 2015 Plan worsened the

population equality and geographic compactness of HD 105 and HD 111. The 2015 Plan also

significantly increased the number of split precincts in HD 111 from two to five. Moreover,

districts around the areas of HD 105 and HD 111 split up municipalities and ignored county

boundaries far more than necessary.

Analysis of Racial Breakdowns of Split Precincts

This section discusses my evaluation of the racial composition of split precincts in HD

105 and 111, as drawn by the 2015 Plan. When evaluating precincts split by House districts in

Georgia, it is important to recognize what sub-precinct-level data is available and unavailable to

the Legislature's map-drawer when drawing any districting plan. In Georgia, all partisan data,

including election results, are available only at the precinct level. Election returns are not

reported by the Secretai7 of State at the sub-precinct level, nor does the Secretary of State

provide any estimated sub-precinct level election returns. In other words, it is impossible for a

32



map-drawer to gain detailed knowledge of whether one split portion of a precinct is more heavily

Democratic or Republican-leaning than another split portion of the same precinct.

On the other hand, racial data is indeed available at the sub-precinct level from two

sources: In Georgia, residents are asked to select their racial identification when they register to

vote. Therefore, publicly available voter registration lists in Georgia contain information on the

racial identification of each individual voter, along with the voter's residential address. Hence, it

is indeed possible to calculate how the racial composition of one split portion of a precincts

differs from another split portion of the same precinct. Furthermore, census block-level racial

data, as provided by the US Census Bureau in Redistricting Data Summary File 1, enables map-

drawers to easily calculate the 2010 Census racial composition of any split precinct in Georgia.

In evaluating the racial composition of split precincts in the 2015 Plan, I first reviewed

the November 20, 2017 deposition testimony of Ms. Gina Wright, the Legislature's primary

map-drawer for the Act 251 of 2015 State House Plan. Ms. Wright testified that I) The

Legislative and Congressional Reapportionment Office obtained precinct-level political data,

including election results and registration counts, from the Georgia Secretary of State; 2) The

Reapportionment Office analyzed no partisan data at any lower level of geography than the

precinct level; and 3) The Reapportionment Office simply assumed that all census blocks (and

other sub-precinct geographies) within any individual precinct had the same percentage of

Democrats and Republicans with respect to any available partisan measure (pp. 111-113).

Based on Ms. Wright's deposition testimony, as well as my own experience and expertise

using redistricting and Census data, it is thus apparent that partisan considerations cannot explain

the particular ways in which the Reapportionment Office drew boundaries within the precincts

that were split by HD 105 and HD 111 of the 2015 Plan. Therefore, having eliminated partisan

considerations as an explanation for the particular ways in which precincts were split, I evaluated

whether the particular ways in which the HD 105 and HD 111 split precincts followed any

apparent racial patterns.

In the 2015 Plan, HD 105 splits three precincts, while HD 111 splits five precincts. In

other words, there are three precincts partially assigned to HD 105 and partially assigned to

another neighboring district, and there are five precincts partially assigned to HD 111 and

partially assigned to another district. Hence, there are eight total split precincts. Among these

eight split precincts, I found a striking racial pattern in the way particular ways these precincts

33



were split: Six of the eight precincts were split such that the respective portions of the precincts

assigned to HD 105 or HD 111 had a lower African-American share of the Voting Age

Population (YAP), while the portions of the precincts not assigned to HD 105 or HD 111 had a

higher African-American proportion. Given that the Reapportionment Office did not analyze

partisan data at the sub-precinct level, this striking racial pattern suggests that racial

considerations explain the particular ways in which these precincts were split.

For each of the eight split precincts created by HD 105 and HD 111,1 analyzed the Any

Part African-American (including multi-racial African-Americans) share of the Total Voting Age

Population (hereinafter: BVAP). These BVAP calculations were produced using census block-

level racial breakdowns reported in the 2010 US Census Redistricting Data Summary File 1.1

overlaid a 2010 Census block shapefile onto a shapefile of the 2015 Plan districts and a shapefile

of Georgia's 2,756 precincts (as of 2014), and I calculated the population within each split

fragment within the eight split precincts. My findings regarding the racial breakdowns of each

split precinct, as detailed in Tables 15 and 16, are as follows:

Split Precincts Involving HD 105: The boundaries of HD 105 in the 2015 Plan produce

three split precincts. These three split precincts, along with the 2010 Census racial breakdown of

the split fragments comprising each precinct, are listed in Table 15.

Precinct 001 ("Harbins A") is split between HD 104 and HD 105. The portion of Harbins

A contained within HD 105 has a 14.5% BVAP. By contrast, the portion of Harbins A assigned

to HD 104 has a 33.2% BVAP.

Precinct 60 ("Lawrenceville D") is split between HD 104 and HD 105. The portion of

Lawrenceville D contained within HD 105 has a 45.6% BVAP, while the portion of the precinct

assigned to HD 104 has a 30.4% BVAP.

Precinct 146 ("Baycreek H") is split between FID 105 and HD 114. The portion of

Baycreek H contained within HD 105 has a 13.5% BVAP, while the portion of the precinct


Overall, two out of the three precincts split by HD 105 (Harbins A and Baycreek H) were

split in such a manner that the portion of the precinct within HD 105 has a lower BVAP than the

portion of the precinct outside of HD 105.

34



led 12/22/17 Page 35 of 44

Table 15:

Vote

r Turnout by Race Wit

hin and Outside of HD 105 of the

2015 Plan In

Spl

it Precincts

2016 Election Turnout Within

2016 Election Turnout Outside of

HD 105 Boundaries (2015 Plan)


Pet:

Precinct Name:

Black

Non-Black

Black

Proportion

Black

Non-Black

Black

Proportion

001

Harbins A

329

1651

16.6%

602

1139

34.6%

060

Lawrenceville D

1088

933

53.8%

167

224

42.7%

146

Baycreek H

468

2096

18.3%

309

684

31.1%

Tota

ls by Race:

28.7%

34.5%

* Includes only precincts tha

t ar

e sp

lit in

to mul

tipl

e di

stri

cts.

2010 Census YAP by Race Within and Outside of HD 105 of the

2015 Pla

nIn Spl

it Pre

cinc

ts





Pet:

Precinct Name:

Black

Total YAP

Black

Proportion

Black

Total YAP

Black

Proportion

001

Harbins A

313

2161

14.5%

845

2542

33.2%

060

Lawrenceville D

2428

5320

45.6%

405

1334

30.4%

146

Baycreek H

419

3112

13.5%

343

1357

25.3%

Totals by Race:

3160

10592

29.8%

1,59

35,233

30.4%

* Includes only precincts tha

t ar

e sp

lit in

to mul

tipl

e di

stri

cts.

Sou

rce:

2010 US Cen

sus Redistricting Da

ta Summary Fil

e 1

35



Moreover, none of these three precincts are split in a manner that was necessitated by the

principles set forth in the "2011-2012 Guidelines for the House Legislative and Congressional

Reapportionment Committee" (Hereinafter: "Redistricting Guidelines"). First, the Redistricting

Guidelines call for plans to preserve existing precinct boundaries. HD 105's splitting of three

precincts is more than was necessary to achieve population equality in the district. Second, none

of these three precincts included non-contiguous fragments, so the splitting of these precincts

was not necessary for preserving the geographic contiguity of HD 105 or any neighboring

district. Third, the decision to split each of these three precincts, rather than assign the entire

precinct to HD 105 or to another district, decreased the geographic compactness of the entire

district, as measured by Reock score. Finally, none of these three precincts crosses a county

boundary; thus, the splitting of these three precincts was not necessary for the preservation of

county boundaries.

In sum, the splitting of three precincts in HD 105 appears to have followed the racial

consideration of excluding more heavily-African-American portions of these precincts from HD

105 while including less heavily-African-American portions in HD 105. This racial consideration

in splitting the precincts also required the subordination of the Redistricting Guidelines,

particularly the preservation of precinct boundaries.

Split Precincts Involving HD 111: The boundaries of HD 111 in the 2015 Plan produce

five split precincts. These five split precincts, along with the racial breakdown of the split

fragments comprising each precinct, are listed in Table 16.

Precinct 26 ("Tussahaw") is split between HD 110 and HD 111. The portion ofTussahaw

contained within HD 111 has a 4.8% BVAP, while the portion of the precinct assigned to HD

110hasa6.9%BVAP.

Precinct 32 ("Mount Carmel") is split between HD 073 and HD 111. The portion of

Mount Carmel contained within HD 111 has a 43.7% BVAP, while the portion of the precinct


Precinct 38 ("Hickory Flat") is split between HD 109 and HD 111. The portion of

Hickory Flat contained within HD 111 has a 50.6% BVAP, while the portion of the precinct


36



led 12/22/17 Page 37 of 44

Table 16;

Voter Turnout by Race Within and Outside of HD. Ill of the 2015 Plan





Precinct Name:

Black

Non-Black

Black

Proportion

Black

Non-Black

Black

Proportion

26 - Tussahaw

25'

498'

4.8%

98

1554

5.9%

32 - Mount Carmel

1023'

803'

56.0%

494

393

55.7%

38 - Hic

kory

Flat

795'

583'

57.7%

699

1173

37.3%

53 - Flippen

884'

1106'

44.4%

379

452

45.6%

61 - McDonough Central

302'

759'

28.5%

649

765

45.9%

Tota

ls by Rac

e:3029

3749

44.7%

2319

4337

34.8%

* Includes only precincts that are sp

lit into multiple di

stri

cts.

2010 Census YAP by Race Within and Outside of HD 111 of the

2015 Pla

nIn Each Spl

it Precinct





Precinct Name:

Black

Total YAP

Black

Proportion

Black

Total YAP

Black

Proportion

26 - Tussahaw

38

789

4.8%

171

2461

6.9%

32 - Mount Carmel

699

1601

43.7%

381

842

45.2%

38 - Hickory Flat

995

1967

50.6%

785

2338

33.6%

53 - Fli

ppen

961

2478

38.8%

520

1264

41.1%

61 - McDonough Cen

tral

404

1504

26.9%

950

1895

50.1%

Totals by Race:

3097

8339

37.1%

2807

8800

31.9%

* Includes only precincts that are split in

to mul

tipl

e districts. Sou

rce:

2010 US Cen

sus Re

dist

rict

ing Da

ta Summary Fil

e

37



Precinct 53 ("Flippen") is split between HD 109 and HD 111. The portion of Flippen

contained within HD 111 has a 38.8% BVAP, while the portion of the precinct assigned to HD

109 has a 41.1% BVAP.

Precinct 61 ("McDonough Central") is split between HD 109 and HD 111. The portion of

McDonough Central contained within HD 111 has a 26.9% BVAP, while the portion of the

precinct assigned to HD 109 has a 50.1% BVAP.

Overall, four out of the five precincts split by HD 111 (Tussahaw, Mount Carmel,

Flippen, and McDonough Central) were split in such a manner that the portion of the precinct

within HD 111 has a lower BVAP than the portion of the precinct outside of HD 111.

Moreover, none of these five precincts are split in a manner that was necessitated by the

principles set forth in the Redistricting Guidelines. First, HD 111's splitting of five precincts is

more than was necessary to achieve population equality in the district. Second, none of these five

precincts included non-contiguous fragments, so the splitting of these precincts was not

necessary for preserving the geographic contiguity of HD 111 or any neighboring district. Third,

the decision to split each of these five precincts, rather than assign the entire precinct to HD 111

or to another neighboring district, decreased the geographic compactness of the entire district, as

measured by Reock score. Finally, none of these five precincts crosses a county boundary; thus,

the splitting of these five precincts was not necessary for the preservation of county boundaries.

In sum, the splitting of five precincts in HD 111 appears to have followed the racial

consideration of excluding more heavily-African-American portions of these precincts from HD

five while including less heavily-African-American portions in HD five. This racial

consideration in splitting the precincts also required the subordination of the Redistricting

Guidelines, particularly the preservation of precinct boundaries.

38


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Filed 12122117 Page 39 of 44

I declare under penalty of perjury the foregoing is tme and correct. Executed this 22nd day of

December 2017.

Signed:

Jowei Chen

39


Case l:17-cv-01427-TCB-WSD-BBM Document 63-1 Filed 12122117 Page 40 of 44

Jowei Chen

Curriculum Vitae

Department of Political ScienceUniversity of Michigan5700 Haven Hall

505 South State Street

Ann Arbor, MI 48109-1045Phone: 917-861-7712, Email: [email protected]: http://www.umich.edu/~iowei

Academic Positions:

Associate Professor (2015-present), Assistant Professor (2009-2015), Department of PoliticalScience, University of Michigan.Faculty Associate, Center for Political Studies, University of Michigan, 2009 - Present.W. Glenn Campbell and Rita Ricardo-Campbell National Fellow, Hoover Institution, StanfordUniversity, 2013.Principal Investigator and Senior Research Fellow, Center for Governance and Public PolicyResearch, Willamette University, 2013 - Present.

Education:

Ph.D., Political Science, Stanford University (June 2009)M.S., Statistics, Stanford University (January 2007)B.A., Ethics, Politics, and Economics, Yale University (May 2004)

Publications:

Chen, Jowei and Neil Malhotra. 2007. "The Law of k/n: The Effect of Chamber Size onGovernment Spending in Bicameral Legislatures."

American Political Science Review. 101(4J: 657-676.

Chen, Jowei, 2010. "The Effect of Electoral Geography on Pork Barreling in BicameralLegislatures."

American Journal of Political Science, 54(21: 301-322.

Chen, Jowei, 2013. "Voter Partisanship and the Effect of Distributive Spending on PoliticalParticipation."

American Journal of Political Science. 57(11: 200-217.

Chen, Jowei and Jonathan Rodden, 2013. "Unintentional Gerrymandering: Political Geographyand Electoral Bias in Legislatures"

Quarterly Journal of Political Science, 8(31: 239-269.

Chen, Jowei, 2014. "Split Delegation Bias: The Geographic Targeting of Pork Barrel Earmarksin Bicameral Legislatures."

Revise and Resubmit. State Politics and Policy Quarterly.

40


Casel:17-cv-01427-TCB-WSD-BBM Document63-1 Filedl2/22/17 Page41 of44

Bradley, Katharine and Jowei Chen, 2014. "Participation Without Representation? SeniorOpinion, Legislative Behavior, and Federal Health Reform."

Journal ofHealth Politics. Policy and Law, 39(2L 263-293.

Chen, Jowei and Tim Johnson, 2015. "Federal Employee Unionization and Presidential Controlof the Bureaucracy: Estimating and Explaining Ideological Change in Executive Agencies."

Journal of Theoretical Politics^ Volume 27. No. 1; 151-174.

Bonica, Adam, Jowei Chen, and Tim Johnson, 2015. "Senate Gate-Keeping, PresidentialStaffing of'Inferior Offices' and the Ideological Composition of Appointments to the PublicBureaucracy,"

Quarterly Journal of Political Science. Volume 10. No. 1: 5-40.

Chen, Jowei and Jonathan Rodden, 2015, "Cutting through the Thicket: RedistrictingSimulations and the Detection of Partisan Gerrymanders,"

Election Law Journal. Volume 14. Number 4: 331-345.

Chen, Jowei and David Cottrell, 2016. "Evaluating Partisan Gains from CongressionalGerrymandering: Using Computer Simulations to Estimate the Effect of Gerrymandering in theU.S. House."

Electoral Studies. Volume 44 (December 2016'!; 329-340.

Chen, Jowei, 2016, "Analysis of Computer-Simulated Districting Maps for the Wisconsin StateAssembly."

Forthcoming 2017. Election Law Journal.

Research Grants:

Principal Investigator. National Science Foundation Grant SES-1459459. September 2015 -August 2017 ($165,008). "The Political Control of U.S. Federal Agencies and BureaucraticPolitical Behavior."

"Economic Disparity and Federal Investments in Detroit," (with Brian Min) 2011. GrahamInstitute, University of Michigan ($30,000).

"The Partisan Effect of OSHA Enforcement on Workplace Injuries," (with Connor Raso) 2009.John M. Olin Law and Economics Research Grant ($4,410).

Invited Talks:

September, 2011. University of Virginia, American Politics Workshop.October 2011. Massachusetts Institute of Technology, American Politics Conference.January 2012. University of Chicago, Political Economy/American Politics Seminar.February 2012. Harvard University, Positive Political Economy Seminar.

41



September 2012. Emory University, Political Institutions and Methodology Colloquium.November 2012. University of Wisconsin, Madison, American Politics Workshop.September 2013. Stanford University, Graduate School of Business, Political EconomyWorkshop.February 2014. Princeton University, Center for the Study of Democratic Politics Workshop.November 2014. Yale University, American Politics and Public Policy Workshop.December 2014. American Constitution Society for Law & Policy Conference: Building theEvidence to Win Voting Rights Cases.February 2015. University of Rochester, American Politics Working Group.March 2015. Harvard University, Voting Rights Act Workshop.May 2015. Harvard University, Conference on Political Geography.October 2015. George Washington University School of Law, Conference on RedistrictingReform.

September 2016. Harvard University Center for Governmental and International Studies, VotingRights Institute.March 2017. Duke University, Redistricting Reform: Mapping our Future Conference.October 2017. Willamette University College of Law.October 2017. University of Wisconsin, Metric Geometry and Gerrymandering Group.

Conference Service:

Section Chair, 2017 APSA (Chicago, IL), Political Methodology SectionDiscussant, 2014 Political Methodology Conference (University of Georgia)Section Chair, 2012 MPSA (Chicago, IL), Political Geography Section.Discussant, 2011 MPSA (Chicago, IL) "Presidential-Congressional Interaction."Discussant, 2008 APSA (Boston, MA) "Congressional Appropriations."Chair and Discussant, 2008 MPSA (Chicago, IL) "Distributive Politics: Parties and Pork.

Conference Presentations and Working Papers:

"Ideological Representation of Geographic Constituencies in the U.S. Bureaucracy," (with TimJohnson). 2017 APSA.

"Incentives for Political versus Technical Expertise in the Public Bureaucracy," (with TimJohnson). 2016 APSA.

"Black Electoral Geography and Congressional Districting: The Effect of Racial Redistricting onPartisan Gerrymandering". 2016 Annual Meeting of the Society for Political Methodology (RiceUniversity)

"Racial Gerrymandering and Electoral Geography." Working Paper, 2016.

"Does Deserved Spending Win More Votes? Evidence from Individual-Level DisasterAssistance," (with Andrew Healy). 2014 APSA.

42



"The Geographic Link Between Votes and Seats: How the Geographic Distribution of PartisansDetermines the Electoral Responsiveness and Bias of Legislative Elections," (with DavidCottrell). 2014 APS A.

"Gerrymandering for Money: Drawing districts with respect to donors rather than voters." 2014MPSA.

"Constituent Age and Legislator Responsiveness: The Effect of Constituent Opinion on the Votefor Federal Health Reform." (with Katharine Bradley) 2012 MPSA.

"Voter Partisanship and the Mobilizing Effect of Presidential Advertising." (with Kyle Dropp)2012 MPSA.

"Recency Bias in Retrospective Voting: The Effect of Distributive Benefits on VotingBehavior." (with Andrew Feher) 2012 MPSA.

"Estimating the Political Ideologies of Appointed Public Bureaucrats," (with Adam Bonica andTim Johnson) 2012 Annual Meeting of the Society for Political Methodology (University ofNorth Carolina)

"Tobler's Law, Urbanization, and Electoral Bias in Florida." (with Jonathan Rodden) 2010Annual Meeting of the Society for Political Methodology (University of Iowa)

"Unionization and Presidential Control of the Bureaucracy" (with Tim Johnson) 2011 MPSA.

"Estimating Bureaucratic Ideal Points with Federal Campaign Contributions" 2010 APSA.(Washington, DC).

"The Effect of Electoral Geography on Pork Spending in Bicameral Legislatures," VanderbiltUniversity Conference on Bicameralism, 2009.

"When Do Government Benefits Influence Voters' Behavior? The Effect of FEMA Disaster

Awards on US Presidential Votes," 2009 APSA (Toronto, Canada).

"Are Poor Voters Easier to Buy Off?" 2009 APSA (Toronto, Canada).

"Credit Sharing Among Legislators: Electoral Geography's Effect on Pork Barreling inLegislatures," 2008 APSA (Boston, MA).

"Buying Votes with Public Funds in the US Presidential Election," Poster Presentation at the2008 Annual Meeting of the Society for Political Methodology (University of Michigan).

"The Effect of Electoral Geography on Pork Spending in Bicameral Legislatures," 2008 MPSA.

"Legislative Free-Riding and Spending on Pure Public Goods," 2007 MPSA (Chicago, IL).

43



"Free Riding in Multi-Member Legislatures," (with Neil Malhotra) 2007 MPSA (Chicago, IL).

"The Effect of Legislature Size, Bicameralism, and Geography on Government Spending:Evidence from the American States," (with Neil Malhotra) 2006 APSA (Philadelphia, PA).

Reviewer Service:

American Journal of Political Science

American Political Science Review

Journal of Politics

Quarterly Journal of Political ScienceAmerican Politics Research

Legislative Studies QuarterlyState Politics and Policy QuarterlyJournal of Public PolicyJournal of Empirical Legal StudiesPolitical Behavior

Political Research QuarterlyPolitical AnalysisPublic Choice

Applied Geography

44



REPLY REPORT OF JO WEI CHEN, Ph.D.

In response to my December 22, 2017 expert report in this case, Defendants' counsel

submitted a January 31, 2018 rebuttal report authored by Dr. John Alford. This reply report

addresses the issues raised by Dr. Alford's rebuttal report. 1 also point out several empirical

findings from my original report that Dr. Alford did not dispute and did not discuss.

First, I explain how, contrary to the claims in the text of Dr. Alford's rebuttal report, the

computer code and data files turned over in connection with his report reveal that Dr. Alford did

not analyze any election results from the 2012 and 2016 state house elections in HD 105 and HD

111. Instead, Dr. Alford's own computer code and underlying data files reveal that he actually

analyzed the 2012 and 2016 US Presidential election results in Gwinnett and Henry Counties.

Nor did Dr. Alford even attempt to analyze any election results from the 2014 state house

elections in HD 105 and HD 111. Overall, then, Dr. Alford effectively failed to analyze any of

the state house elections held in HD 105 and HD 111 during 2012, 2014, and 2016. This failure

is significant because these six elections are the most probative elections when conducting

racially polarized voting analysis of house election results in HD 105 and HD 111. Thus, the fact

that Dr. Alford's computer code files reveal that he failed to analyze any of these six state house

elections illustrates that Dr. Alford has not provided any evidence to directly dispute my

Ecological Inference (El) and Ecological Regression (ER) calculations for the 2012, 2014, and

2016 state house elections in HD 105 and HD 111, as reported in my original December 2017

expert report.

Second, I explain that in conducting his analyses of non-legislative elections (such as

Sheriff, Coroner, and County Commission Chair contests), Dr. Alford's computer code reveals a

fundamental and significant mistake that renders every one of his Ecological Inference and

Ecological Regression calculations invalid for the purposes of evaluating racially polarized

voting in HD 105 and HD 111. Specifically, Dr. Alford did not even attempt to account for split

precincts that are divided across multiple house districts. Instead, in calculating the racial

composition of HD 105 and HD 111 precincts. Dr. Alford simply treated all voters in split

precincts as if they all resided within one of the two challenged districts. This mistake caused Dr.

Alford to significantly miscalculate the racial composition of voters in multiple precincts, thus

rendering his Ecological Inference and Ecological Regression calculations invalid.

DEFENDANTS

EXHIBIT



Third, I present in this report my racially polarized voting analysis of the January 2018

special house election in HD 111, an election that occurred after I submitted my original

December 22, 2017 expert report. The results from this analysis indicate that voters in HD 111

continued to exhibit substantially the same level of racially polarized voting as they had in

previous elections in the district.

Fourth, I address Dr. Alford's claim in his rebuttal report that voters respond to the

partisanship of a candidate, not to the race of the candidate. I explain that this claim is irrelevant

to the racially polarized voting analysis in my original expert report.

Fifth, I address Dr. Alford's claim in his rebuttal report that when combined together, the

portions of split precincts outside of HD 105 and HD 111 are not very different in their aggregate

racial composition from the portions of these precincts inside of the two challenged districts.

Finally, in this report, 1 note several empirical findings from my original December 22,

2017 expert report that were neither addressed nor disputed by Dr. Alford's rebuttal report.

1. Dr. Afford Failed to Analyze Any State House Elections from 2012,2014, or 2016

In Tables 1 and 2 of his rebuttal report (p. 7), Dr. Afford purports to present his

Ecological Inference and Ecological Regression analysis of several state and local election

contests within HD 105 and HD 111 during November 2012 and November 2016. In particular.

Dr. Afford purports to have analyzed the 2012 and 2016 state house elections, as well as other

non-legislative contests, in HD 105 and HD 111. In connection with his report. Dr. Afford also

turned over his computer code (Stata .do files and .R code files) and data files (.dta and .xlsx

files) used to create the results in Tables I and 2 of his report.

Examination of Dr. Alford's computer code and data files reveals that Dr. Afford did not

analyze election results from the HD 105 and HD 111 state house elections in 2012 and 2016.

Instead, Dr. Alford's computer code and data files reveal that he analyzed the results of the 2012

US Presidential election contests. Although the text of Dr. Alford's rebuttal report states that he

intended to analyze the 2012 and 2016 state house election results, his computer code files

illustrate clearly that he did not analyze these state house elections. Below, I provide a complete

listing of every computer code file in which Dr. Afford performed El and ER calculations, and I

detail the actual election contests each computer code file analyzed.



For both HD 105 and HD 111, the number of computer code files tumed over by Dr.

Alford perfectly matches the number of 2012 and 2016 election contests for which he reported

El and ER results in Tables 1 and 2. Therefore, it is apparent that Dr. Alford turned over the

complete set of all computer code files he used to produce the El and ER calculations in his

rebuttal report. However, Dr. Alford labeled the elections incorrectly in Tables 1 and 2.1 explain

these mistakes below, and I identify below the actual election analyzed by each one of Dr.

Alford's computer code files.

Dr. Alford's 2012 Election Results for HD 105: In rows 2-4 of Table 1 of his rebuttal

report (p. 7), Dr. Alford lists three 2012 elections for which he claims to have analyzed voting

behavior in HD 105: The 2012 state house election for HD 105, the Public Service Commission

(District 3) election, and the Clerk of Superior Court election.

However, Dr. Alford's computer code and data files indicate that he did not analyze the

HD 105 state house election results, although he did analyze the results from the Public Service

Commission and Clerk of Superior Court elections. Specifically, the following computer code

files reveal that Dr. Alford actually analyzed the three election contests listed below, with

precinct-level election results imported from the data files listed below:

1) Dr. Alford's computer code file titled "elec2012_HD105_contest_l.do" (reprinted in

Appendix F of this report) imports and analyzes precinct-level election results from the 2012 US

Presidential election between Barack Obama, Mitt Romney, and Gary Johnson. Dr. Alford's

computer code imports these results from a data file called "HD 105/Rdata_2012_contest_l.dta"

(reprinted in Appendix E of this report) and attempts to conduct El and ER calculations using the

data in this file. I visually inspected the precinct-level election data contained in this data file and

verified that they perfectly match the Gwinnett County precinct-level results of the Obama-

Romney 2012 US Presidential election, as reported on the Georgia Secretary of State website.'

Furthermore, another data file tumed over by Dr. Alford in connection with his rebuttal report,

called "HD 105/controlfile_2012_processed.dta" (reprinted in Appendix A of this report), clearly

states that Dr. Alford's computer code was analyzing the results from the 2012 US Presidential

election, not a state house election.

Downloaded from: http://results.enr.clarityelections.com/GA/Gwinnett/42345/112372/en/md_data.html?cid=5&



2) Dr. Alford's computer code file titled "elec2012_HD105_contest_2.do" (reprinted in

Appendix G of this report) imports and analyzes precinct-level election results from the 2012

Public Service Commission - District 3 election between Stephen Oppenheimer and Chuck

Eaton. Dr. Alford's computer code imports these election results from a data file called

"HD 105/Rdata_2012_contest_2.dta" and attempts to conduct El and ER calculations using the

data in this file.

3) Dr. Alford's computer code file titled "elec2012_HD105_contest_3.do" (reprinted in

Appendix H of this report) imports and analyzes precinct-level election results from the 2012

Clerk of Superior Court election between Richard Alexander and Brian Whiteside. Dr. Alford's

computer code imports these election results from a data file called "HD

105/Rdata_2012_contest_3.dta" and attempts to conduct El and ER calculations using the data in

this file.

I executed these three aforementioned computer code files on my own computer, and I

found that these files indeed produce numerical results close to the ones reported in the upper

half of Dr. Alford's Table 1 in his rebuttal report. However, Dr. Alford labeled the first election

contest he analyzed in Table 1 incorrectly. Dr. Alford incorrectly labeled the second row of

Table 1 if he had analyzed the November 2012 state house election in HD 105. As detailed

above, Dr. Alford's computer code and data files reveal that he did not analyze the November

2012 HD 105 state house election. Instead, he analyzed the results of the November 2012 US

Presidential election in Gwinnett County.

Dr. Alford's 2016 Election ResultsforHD 105: Dr. Alford made the same mistake in

describing his 2016 election results for HD 105. In rows 7-9 of Table 1 of his rebuttal report (p.

7), Dr. Alford lists three 2016 elections for which he claims to have analyzed voting behavior in

HD 105: The 2016 state legislative election for HD 105, the US Senator election, and the County

Commission Chair election.

However, Dr. Alford's computer code and data files reveal that once again, he did not

analyze the HD 105 state house election results, although he did analyze the results of the other

two elections. Specifically, the following computer code files reveal that Dr. Alford actually

analyzed the three election contests listed below, with precinct-level election results imported

from the data files listed below:



1) Dr. Alford's computer code file titled "elec2016_HD105_contest_2.do" imports and

analyzes precinct-level election results from the 2016 US Presidential election between Donald

Trump and Hillary Clinton. Dr. Alford's computer code imports these results from a data file

called "HD 105/Rdata_2016_contest_2.dta" and attempts to conduct El and ER calculations

using the data in this file. I visually inspected the precinct-level election data contained in this

data file and verified that they perfectly match the Gwinnett County precinct-level results of the

Trump-Clinton 2016 US Presidential election, as reported on the Georgia Secretary of State

website.^ Furthermore, another data file turned over by Dr. Alford in connection with his rebuttal

report, called "HD 105/controlfile_2016_processed.dta" (reprinted in Appendix B of this report),

clearly states that Dr. Alford's computer code was analyzing the results from the 2016 US

Presidential election, not a state house election.

2) Dr. Alford's computer code file titled "elec2016_HD105_contest_l.do" imports and

analyzes precinct-level election results from the 2016 US Senator election. Dr. Alford's computer

code imports these election results from a data file called "HD 105/Rdata_2016_contest_l.dta"

and attempts to conduct El and ER calculations using the data in this file.


analyzes precinct-level election results from the 2016 County Commission Chair election. Dr.

Alford's computer code imports these election results from a data file called


data in this file.

I executed these three aforementioned computer code files on my own computer, and I

found that these files indeed produce numerical results close to the ones reported in the lower

half of Dr. Alford's Table 1 in his rebuttal report. However, Dr. Alford labeled the first of the

2016 election contests he analyzed in Table 1 incorrectly. Dr. Alford incorrectly labeled the

seventh row of Table 1 as if he had analyzed the November 2016 state house election in HD 105.

As detailed above. Dr. Alford's computer code and data files reveal that he did not analyze the

November 2016 HD 105 state house election. Instead, he analyzed the results of the November

2016 US Presidential election in Gwinnett County.

Downloaded from: http://results.enr.clarityelections.coni/GA/63991/184321/en/md_data.html?cid=5000&



Dr. Alford's 2012 Election Results for HD 111: Dr. Alford made the same mistake in

describing his 2012 election results for HD 111, In rows 2-5 of Table 1 of his rebuttal report (p.

7), Dr. Alford lists four 2012 elections for which he claims to have analyzed voting behavior in

HD 111: The 2012 state house election for HD 111, the Public Service Commission-District 3

election, the Flint Circuit District Attorney election, and the County Commission Chair election.

However, Dr. Alford's computer code and data files indicate that once again, he did not

analyze the HD 111 state house election results, although he did analyze the results from the

other three elections. Specifically, the following computer code files reveal that Dr. Alford

actually analyzed the four election contests listed below, with precinct-level election results

imported from the data files listed below:

1) Dr. Alford's computer code file titled "elec2012_HDl 1 l_contest_l.do" imports and

analyzes precinct-level election results from the 2012 US Presidential election between Barack

Obama, Mitt Romney, and Gary Johnson. Dr. Alford's computer code imports these results from

a data file called "HD 111/Rdata_2016_contest_l.dta" and attempts to conduct El and ER

calculations using the data in this file. I visually inspected the precinct-level election data

contained in this data file and verified that they perfectly match the Henry County precinct-level

results of the Obama-Romney 2012 US Presidential election, as reported on the Georgia

Secretary of State website.^ Furthermore, another data file turned over by Dr. Alford in

connection with his rebuttal report, called "HD 11 l/controlfile_2012_processed.dta" (reprinted in

Appendix C of this report), clearly states that Dr. Alford's computer code was analyzing the

results from the 2012 US Presidential election, not a state house election.

2) Dr. Alford's computer code file titled "elec2012_HDl 1 l_contest_2.do" imports and

analyzes precinct-level election results from the 2012 Public Service Commission-District 3

election. Dr. Alford's computer code imports these election results from a data file called "HD


this file.


analyzes precinct-level election results from the 2012 Flint Circuit District Attorney election. Dr.


Downloaded from: http://results.enr.clarltyelections.eom/GA/Henry/42353/l 12380/en/md_data.html?cid=5&




data in this file.





data in this file.

I executed these four aforementioned computer code files on my own computer, and I

found that these files indeed produce numerical results reasonably close to the ones reported in

the upper half of Dr. Alford's Table 2 in his rebuttal report (p. 7). However, Dr. Alford labeled

the first of the 2012 election contests he analyzed in Table 2 incorrectly. Dr. Alford incorrectly

labeled the second row of Table 2 as if he had analyzed the November 2012 state house election

in HD 111. As detailed above. Dr. Alford's computer code and data files reveal that he did not

analyze the November 2012 HD 111 state house election. Instead, he analyzed the results of the

November 2012 US Presidential election in Henry County.

Dr. Alford's 2016 Election Results forHD 111: Dr. Alford made the same mistake in

describing his 2016 election results for HD 111. In rows 8-13 of Table 2 of his rebuttal report (p.

7), Dr. Alford lists six 2016 elections for which he claims to have analyzed voting behavior in

HD 105: The 2016 state house election for HD 111, the US Senator election, the Clerk of

Superior Court election, the Sheriff election, the Coroner election, and the County Commission

Chair election.

However, Dr. Alford's computer code and data files indicate that once again, he did not

analyze the HD 111 state house election results, although he did analyze the results from the

other five elections. Specifically, the following computer code files reveal that Dr. Alford

actually analyzed the six election contests listed below, with precinct-level election results

imported from the data files listed below:

1) Dr. Alford's computer code file titled "elec2016_HDl 1 l_contest_l.do" imports and

analyzes precinct-level election results from the 2016 US Presidential election between Donald

Trump and Hillary Clinton. Dr. Alford's computer code imports these election results from a data

file called "HD 111/Rdata_2016_contest_l.dta" and attempts to conduct El and ER calculations



using the data in this file. I visually inspected the precinct-level election data contained in this

data file and verified that they perfectly match the Henry County precinct-level results of the

Trump-Clinton 2016 US Presidential election, as reported on the Georgia Secretary of State

website."* Furthermore, another data file turned over by Dr. Alford in connection with his rebuttal

report, called "HD 11 l/controlfile_2016_processed.dta" (reprinted in Appendix D of this report),

clearly states that Dr. Alford's computer code was analyzing the results from the 2016 US

Presidential election, not a state house election.


analyzes precinct-level election results fi'om the 2016 US Senator election. Dr. Alford's computer

code imports these election results from a data file called "HD 111/Rdata_2016_contest_2.dta"



analyzes precinct-level election results from the 2016 Clerk of Superior Court election. Dr.

Alford's computer code imports these election results from a data file called "HD


this file.


analyzes precinct-level election results fi om the 2016 Sheriff election. Dr. Alford's computer




analyzes precinct-level election results from the 2016 Coroner election. Dr. Alford's computer







data in this file.

Downloaded from: http://results.enr.clarityelections.eom/GA/Henry/64067/l 83247/en/md_data.html?cid=5&

8



I executed these six aforementioned computer code files on my own computer, and I

found that these files indeed produce numerical results close to the ones reported in the lower

half of Dr. Alford's Table 2 in his rebuttal report. However, Dr. Alford labeled the first of the

2016 election contests he analyzed in Table 2 incorrectly. Dr. Alford incorrectly labeled the

eighth row of Table 2 as if he had analyzed the November 2016 state house election in HD 111.

As detailed above. Dr. Alford's computer code and data files reveal that he did not analyze the

November 2016 HD 111 state house election. Instead, he analyzed the results of the November

2016 US Presidential election in Henry County.

In addition to the failure of Dr. Alford's computer code to actually analyze any election

results from the 2012 and 2016 state house elections. Dr. Alford did not even claim to analyze

any election results from the 2014 state house elections in HD 105 and HD 111. Overall, then.

Dr. Alford effectively failed to analyze any of the state house elections held in HD 105 and HD

111 during 2012, 2014, and 2016. This failure is significant because these six elections are the

most probative elections for the purpose of conducting racially polarized voting analysis of

house election results in HD 105 and HD 111. These six elections are the most probative

because: 1) These are the actual election contests held to elect state house representatives for the

two challenged districts; and 2) These are the most recent election contests held during a regular

November election.

Thus, the fact that Dr. Alford's computer code files reveal that he failed to analyze any of

these six state house elections illustrates that Dr. Alford has not provided any evidence to

directly dispute my Ecological Inference and Ecological Regression calculations for the 2012,

2014, and 2016 state house elections in HD 105 and HD 111, as reported in my original expert

report of December 22,2017.

2) Dr. Alford Failed to Account for Voters in Split Precincts in HD 105 and HD 111

For every election analyzed in Tables 1 and 2 of his rebuttal report (p. 7), Dr. Alford used

incorrect data on the racial breakdown of voters within HD 105 and HD 111 because he failed to

account for split precincts within the two districts. In the 2015 Plan, HD 105 splits three

precincts out of ten total precincts covering the district, while HD 111 splits five precincts out of

13 total precincts covering the district. Similarly, in the 2012 Plan, HD 105 splits three precincts

out of ten total precincts covering the district, while HD 111 splits 2 precincts out of 11 total



precincts covering the district. Because such a significant percentage of the precincts covering

HD 105 and HD 111 are split, one must carefully determine which voters within each precinct

reside within and outside of the borders of HD 105 and HD 111. Failing to account for district

borders within each split precinct would lead to significantly inaccurate data regarding the racial

composition of voters within precincts and within each district.

1 explained and illustrated in my original expert report how correctly accounting for split

precincts is a simple and straightforward task for political scientists using voter registration files

and turnout history files. Yet Dr. Alford failed to account for the split precincts covering HD 105

and HD 111, and he did not rely upon a Georgia voter registration file and turnout history file,

which would have allowed him to easily calculate the racial composition of voters within each

district's portion of each split precinct. Instead, in each one of his computer code files performing

El and ER analyses (e.g.," elec2012_HD105_contest_l.do", which is reprinted in Appendix F),

Dr. Alford simply used data on the racial composition of all voters within each precinct,

regardless of whether these voters resided within or outside of HD 105 and HD 111. In

producing his Table 1 and 2 results. Dr. Alford incorrectly analyzed all voters within all

precincts that contain any portion of HD 105 and HD 111, without sorting out which voters

actually resided within the boundaries of the two challenged districts.

Dr. Alford made this same mistake in every single one of his computer code files

analyzing various election contests from the 2012 and 2016 elections. Dr. Alford's computer

code failed to account for split precincts and thus used incorrect voter race counts in all sixteen

of his computer code files analyzing these 2012 and 2016 elections.^

Perhaps in response to this data problem, each one of these sixteen computer code files -

e.g., "elec2012_HD105_contest_l.do" (reprinted in Appendix F) - acknowledges that Dr.

Alford's data contain mismatches between the number of voters who tumed out within some

precincts and the number of ballots actually cast in various election contests within those same

precincts. To address these mismatches, lines 77-78 of Dr. Alford's computer code file titled

"elec2012_HD105_contest_l.do" alters Dr. Alford's voter count numbers using an adjustment

^ The complete list of Dr. Alford's computer code files containing this mistake is as follows;"elec2012_HD 105_contest_l .do", "elec2012_HD 105_contest_2.do", "elec2012_HD 105_contest_3 .do","elec2016_HD 105_contest_l .do", "elec2016_HD 105_contest_2.do", "elec2016_HD 105_contest_3 .do","elec2012_HDl 1 l_contest_l.do", "elec2012_HDl 1 l_contest_2.do", "elec2012_HDl 1 l_contest_3.do","elec2012_HDl 1 l_contest_4.do"," elec2016_HDl 1 l_contest_l.do"," elec2016_HDl 1 l_contest_2.do",elec2016_HDl 1 l_contest_3.do", " elec2016_HDlll_contest_4.do", "elec2016_HDl 1 l_contest_5.do",'elec2016 HDlll contest 6.do"

10



that Dr. Alford describes as follows: "so few precincts I cant afford to drop but have to allocate

the the [sic] cast vote so give it to others based on proportionality" (see Appendix F, lines 77-78).

Specifically, this line of Dr. Alford's computer code appears to adjust the number of non-Black

voters within each precinct so as to artificially force the voter counts to match the ballots counts

from the election data being analyzed. Every single one of these sixteen computer code files

mentioned in Footnote 5 makes a similar artificial numerical adjustment to the precinct-level

racial data prior to conducting Dr. Alford's El and ER analysis.

At this point, it is apparent that Dr. Alford's computer code is not relying upon the actual

number of Black and non-Black voters residing within the actual boundaries of HD 105 and HD

111. Instead, Dr. Alford appears to be using data that were adjusted or transformed and were not

based on any attempt to account for the actual district boundaries of HD 105 and HD 111 within

split precincts. Hence, it is clear that Dr. Alford's 2012 and 2016 election results in Tables 1 and

2 are not analyzing data on the actual voters within HD 105 and HD 111 Thus, Dr. Alford's El

and ER analyses reported in these Tables are invalid because the racial numbers used for these

analyses are not describing the actual set of voters within the two districts.

To examine the magnitude and substantive impact of Dr. Alford's data accuracy

problems, 1 examined how Dr. Alford's precinct-level racial voter counts differed from my

counts, which were based only on the voters actually residing within the geographic boundaries

of HD 105 and HD 111, with no numerical alterations like the ones performed by Dr. Alford's

computer code. 1 found that Dr. Alford's racial voter counts were substantially inaccurate in

multiple precincts:

For example, Henry County's Precinct 38 ("Hickory Flat") is split between HD 111 and

HD 109 in the 2015 House Plan. The portion of this split precinct within HD 111 contained 795

Black and 583 non-Black voters who participated in the November 2016 election, so Blacks

comprised 57.7% of the electorate within the HD 111 portion of Hickory Flat, as detailed in

Table 16 of my original expert report (p. 37).

But Dr. Alford's computer code failed to account for the split portions of this precinct, in

addition to the other adjustments Dr. Alford made to the racial data, as described earlier. As a

result. Dr. Alford's racially polarized voting analysis for all 2016 elections used his inaccurate

data, which claimed that HD 111 contained 1,495 Black and 1,768 non-Black voters in Precinct

38 ("Hickory Flat") in November 2016, implying that Blacks comprised only 45.8% of total

11



voters (far below the reality of 57.7%). Thus, Dr. Alford's inaccurate data led him to

significantly miscalculate both the number of voters, as well as the Black share of voters within

this Precinct residing within HD 111.

Dr. Alford produced similarly faulty data for Henry County's Precinct 61 ("McDonough

Central"), which is also split between HD 111 and HD 109 in the 2015 House Plan. In reality,

the portion of this split precinct within HD 111 contained 302 Black and 759 non-Black voters

who participated in the November 2016 election, so Blacks comprised 28.5% of the electorate

within the HD 111 portion of the McDonough Central precinct, as detailed in Table 16 of my

original expert report (p. 37). But Dr. Alford's racially polarized voting analysis for 2016

elections used his inaccurate data, which claimed that HD 111 contained 949 Black and 1,528

non-Black voters in Precinct 61 ("McDonough Central"), implying that Blacks comprised 38.3%

of total November 2016 voters (far above the reality of 28.5%). Once again. Dr. Alford's

inaccurate data thus led him to significantly miscalculate both the number of voters, as well as

the Black share of voters within this Precinct residing within HD 111. Altogether, data

inaccuracies such as these are sufficiently substantial to have caused Dr. Alford to reach faulty

and invalid estimates in his racially polarized voting analyses.

3) Racially Polarized Voting Analysis of the January 2018 HD 111 Special Election

In this section, I present a racially polarized voting analysis of the January 2018 special

house election in HD 111, an election that occurred after I submitted my original December 22,

2017 expert report. The results from this analysis indicate that voters in HD 111 continued to

exhibit substantially the same level of racially polarized voting as they had in previous elections

in the district.

1 conducted this racially polarized voting analysis in exactly the same way as in my

original expert report: I first calculated precinct-level racial breakdowns of the turnout electorate

- the set of registered voters who cast ballots - within the boundaries of HD 111 during the

January 2018 special election. 1 then compared these precinct-level racial breakdowns to the

precinct-level partisan results for HD 111 in this special house election. I calculated both

Ecological Inference and Ecological Regression estimates, and Table 1 presents both sets of

results.

12



Table 1:

Ecological Inference and Ecological Regression Estimates of Democratic Candidates' Shareof Two-Party Vote Among Blacks and Non-Blacks in House District 111

Ecological Inference Estimates Ecological Regression Estimates:

Black Non-Black Black Non-Black

January 2018 SpecialElection, House District 111

98.7%

[97.1%, 99.5%1

14.2%

[13.8%, 15.1%]100%

[100%, 100%]

4.3%

[1.8%, 6.7%1

[95% Confidence Intervals listed in brackets]

Table 1 illustrates that in the January 2018 special election, voters in HD 111 continued

to exhibit significantly racially polarized voting, just as they had in state house elections during

2012, 2014, and 2016. In the January 2018 special election, approximately 98.7% of Black

voters supported one of the two Democratic candidates, whereas only 14.2% of non-Black voters

supported one of the Democratic candidates. These voting patterns are evidence of very

significant racially polarized voting, and these patterns indicate a continuation of the same

racially polarized voting patterns described in my original expert report in state house elections

during 2012, 2014, and 2016.

In his rebuttal report. Dr. Alford also presented an El and ER analysis of the January

2018 special house election for HD 111, as reported in the bottom row of Table 2 of his rebuttal

report. Dr. Alford's ER estimates are almost identical to mine, but his El estimates were slightly

different: Dr. Alford reported that 89.1% of Black voters and 18.8% of non-Black preferred a

Democratic candidate.

Although Dr. Alford's estimates substantially confirm my conclusion that there was

strong racially polarized voting in the January 2018 special election, I nevertheless examined Dr.

Alford's computer code and underlying data files to understand why he reached slightly different

estimates. Examination of his underlying precinct-level data files (named:

"datafor2018spechdl 1 l_4cand.dta" and "datafor2018spechdl 1 l_2cand.dta") immediately

revealed that once again. Dr. Alford's analysis suffered from one of the same data inaccuracy

problems that plagued his earlier calculations described above.

Specifically, it is apparent from these two precinct-level data files that once again, just as

he did in the "elec2012_HD105_contest_l.do" computer code described earlier. Dr. Alford made

13



his own set of numerical adjustments to the number of non-Black voters in each precinct in HD

111, as reported in these data files in the column named "allothturn". The result of Dr. Alford's

adjustment of the non-Black voter counts is that when he conducted his Ecological Inference

calculations using a computer code file titled "bo_4candSE.r", his calculations were based upon

artificially modified data on the number of non-Black voters within each precinct in HD 111.

Among the data files that Dr. Alford turned over was an Excel file titled "data for 2018

spec hdl 11 .xlsx", which correctly lists the actual number of black and non-Black voters that

participated in the January 2018 election within each precinct, as reported by the Georgia

Secretary of State. However, Dr. Alford's Ecological Inference calculations did not use these

official non-Black voter counts. Instead, Dr. Alford used his modified count of non-Black voters.

Regardless of Dr. Alford's motivations for making these numerical adjustments to the racial data,

it is clear from Dr. Alford's computer code that his analysis of the January 2018 special election

was based on data that did not accurately describe the actual racial composition of voters who

participated in the election. This use of inaccurate data helps explain why Dr. Alford arrived at

Ecological Inference estimates slightly different from my El estimates described in Table 1.

4) Racially Polarized Voting Analysis and the Racial Identity of Candidates

In my original December 22, 2017 report, 1 conducted a racially polarized voting (RPV)

analysis of black and non-black voters in state legislative election contests in HD 105 and HD

111 during November 2012,2014, and 2016. The standard approach taken by political scientists

when analyzing RPV in election contests such as these involves asking two questions: 1)

Whether Black voters usually vote for the same candidates in legislative races; and 2) Whether

non-Black voters exhibit sufficient bloc voting to defeat candidates supported by Black voters. In

my original report, 1 found that in both HD 105 and HD 111, 98-99% Black voters support the

Democratic state legislative candidate in each election. Meanwhile, 75-85% of non-Black voters

support Republican candidates, constituting a level of bloc voting sufficient to defeat the

Democratic candidate supported by Black voters.

In response, Dr. Alford's rebuttal report attempts to argue that "voters are responding to

the party of the candidate, and not the race of the candidate." Dr. Alford states that Black voters

support both white and Black Democratic candidates, whereas white voters largely oppose both

white and Black Democratic candidates.

14



These claims by Dr, Alford are irrelevant to the empirical analysis in my original report. I

simply sought to analyze whether Black voters generally support the same candidate in a

legislative contest and whether non-Black voters generally oppose the candidate preferred by

Black voters. Answering these two questions did not require me to analyze the effect of the

candidate's racial identity on Black or non-Black voter behavior. Therefore, Dr. Alford's claims

are irrelevant to the racially polarized voting analysis in my original report.

Moreover, in this case, the Legislature's primary map-drawer, Ms. Gina Wright, testified

in her November 2017 deposition that the 2015 Plan was drawn to enhance the likelihood that

Republican house candidates would defeat Democratic candidates (pp. 22, 24-26, 28-29). Given

this admission of partisan intent, in the context of this case, the key issue I sought to analyze is

whether Black voters generally support Democratic or Republican candidates and whether non-

black voters oppose the Black-preferred candidate, not the race of the candidates. Indeed, the

Democratic candidate for the November 2014 HD 111 house election was Caucasian, and this

house election exhibited substantially the same level of racially polarized voting as the other two

HD 111 house elections I analyzed in my original expert report.

5) The Predominance of Race in the Splitting of Precincts

On pages 8-9 of his rebuttal report, Dr. Alford argues that the 2015 Plan does not split

precincts in HD 105 and HD 111 on the basis of race. Dr. Alford's evidence for this argument is

his observation that the portions of the split precincts inside of HD 105 and HD 111 are not

extremely different in their aggregate racial composition from the portions of the split precincts

outside these two districts.

Dr. Alford's observation about the aggregate racial composition of these split precinct

portions is irrelevant, and Dr. Alford's misunderstands how a district's split precincts are properly

analyzed:

First, the fact that virtually all Black voters in HD 105 and HD 111 support Democratic

candidates, while the vast majority of non-Blacks oppose Democratic candidates, implies that a

voter's race is a statistically strong proxy for partisanship. Therefore, changing the racial

composition of a district by even only a small amount can guarantee a noticeable change in the

partisan performance of the district.

15



Second, the "2011-2012 Guidelines for the House Legislative and Congressional

Reapportionment Committee" (Hereinafter: "Redistricting Guidelines") require that districts

follow precinct boundaries. The fact that HD 105 and HD 111 in 2015 Plan split a total of eight

precincts constitutes a significant violation this traditional districting principle. Thus, as none of

these eight split precincts was compelled by the Redistricting Guidelines, it is important to

scrutinize how the choice of splitting each individual precinct was made by the Reapportionment

Office.

Third, the Legislature's primary map-drawer, Ms. Gina Wright, testified in her November

20, 2017 deposition that in producing the 2015 Plan, the Legislative and Congressional

Reapportionment Office analyzed no partisan data at any lower level of geography than the

precinct level; instead, the Reapportionment Office simply assumed that all census block within

any single precinct contain the same proportion of Republican and Democratic voters (pp. 111-

113). In his December 18, 2017 deposition, Mr. Robert Strangia, a GIS specialist at the

Reapportionment Office, confirmed Ms. Wright's statement regarding this assumption. Mr.

Strangia's deposition testimony confirmed that the Reapportionment Office did not possess any

data allowing it to distinguish different census blocks within the same precinct with respect to

their relative partisan balance. Instead, according to Mr. Strangia, the Reapportionment Office

simply assumed that all census blocks within the same precinct contain the same proportion of

Republican and Democratic voters (pp. 24-27).

Therefore, the Reapportionment Office could not possibly have used partisan data in

deciding how to split any individual precinct into two different districts. Instead, the

Reapportionment Office only had access to census block-level data, including population and

racial data, in deciding how to split any individual precinct.

Fourth, because the Reapportionment Office had only racial data at the sub-precinct level,

rather than any meaningful partisan data allowing it to distinguish among different blocks within

a precinct, and because the Redistricting Guidelines do not call for any precincts to be split, each

of the eight precincts split by the 2015 Plan should be analyzed individually. The

Reapportionment Office was not required to split apart any particular set of precincts; therefore,

each specific instance that an individual precinct was split should be individually examined as a

potential deviation from the Redistricting Guidelines. Given that each individual split precinct

could not have been split on the basis of any sub-precinct partisan data, it is therefore logical to

16



analyze whether each of these precincts were split on the basis of the racial sub-precinct-level

data that was available to the Reapportionment Office.

When each of the eight split precincts is analyzed individually, there is strong evidence

that race was a predominant factor in how these eight precincts were split: As described in my

original report, six of the eight precincts were split such that the respective portions of the

precincts assigned to HD 105 or HD 111 had a lower African-American share of the Voting Age

Population (YAP), while the portions of the precincts not assigned to HD 105 or HD 111 had a

higher African-American proportion. Given that the Reapportionment Office did not analyze

partisan data at the sub-precinct level, this striking racial pattern suggests that racial

considerations, not partisanship, explain the particular ways in which these eight precincts were

split.

Dr. Alford observes that the portions of the split precincts inside of HD 105 and HD 111

have an aggregate racial composition that is more heavily African-American than the portions of

the split precincts outside these two districts. But this fact is entirely caused by two precincts -

Precinct 60 ("Lawrenceville D") in HD 105 and Precinct 38 ("Hickory Flat") in HD 111 - with

geographic patterns of racial segregation that made inevitable the outcome observed by Dr.

Alford.

Precinct 60 ("Lawrenceville D") lies at the northwestern comer of HD 105 in the 2015

Plan. The southern half of Lawrenceville D is more heavily African-American than the northem

half of the precinct. Furthermore, since Lawrenceville D is at the northwestem comer of HD 105,

it is not possible to incorporate the northem half of Lawrenceville D into HD 105 without also

including the southem half of Lawrenceville D; otherwise, the geographic contiguity of HD 105

would be violated. Therefore, any decision to split Lawrenceville D would inevitably involve

including the southem, more heavily African-American portion of the precinct in HD 105 while

excluding the northem, less heavily African-American portion of the precinct.

Finally, it is important to note that the northem portion of Lawrenceville D, which was

excluded from HD 105, still has a significantly higher African-American proportion than the

overall composition of HD 105 under the 2015 Plan: The November 2016 tumout electorate

inside the excluded portion of the Lawrenceville D precinct was 42.7% African-American, while

the 2016 tumout electorate within the entire HD 105 district was only 32.9% African-American.

It is thus clear that the decision to split the Lawrenceville D precinct and exclude the northern

17



portion of the precinct from HD 105 had the effect of decreasing HD 105's African-American

proportion. Therefore, the particular manner in which the Lawrenceville D precinct was split, by

excluding the northern, heavily-African-American portion of the precinct, appears to support the

map drawer's overall strategy of splitting precincts so as to minimize the African-American

population ofHD 105.

A similar explanation applies to Precinct 38 ("Hickory Flat") in HD 111. The western

portion of Hickory Flat is more heavily African-American than the eastern portion of the

precinct. The Hickory Flat precinct connects to the remainder of HD 111 through the western

portion of the precinct. Thus, it is not possible to include the eastern portion of Hickory Flat into

HD 111 without also including the western portion of Hickory Flat; otherwise, the geographic

contiguity of HD 111 would be violated. Therefore, any decision to split Hickory Flat would

inevitably involve including the western, more heavily African-American portion of the precinct

in HD 111 while excluding the eastern, less heavily African-American portion of the precinct.

6) Findings Not Rebutted by Dr. Alford's Report

In my report, I analyze results from the six most probative elections for the purpose of

examining racially polarized voting within the two districts challenged in this case: The 2012,

2014, and 2016 general house elections in HD 105 and in HD 111. Of those six elections. Dr.

Alford did not purport to analyze the two house elections held in November 2014. For the

remaining four house elections held in 2012 and 2016, Dr. Alford's report claimed that he

analyzed the election results, but his computer code and data files clearly indicate that he instead

analyzed the 2012 and 2016 US Presidential elections in Gwinnett and Henry Counties. Dr.

Alford's effective failure to analyze any of these six elections is significant because these six

elections are the most probative elections when conducting a racially polarized voting analysis of

house election results in HD 105 and HD 111.

Dr. Alford did analyze one state house election contest: The 2018 special election for HD

111. Although Dr. Alford's Ecological Inference calculations suffered from the data inaccuracy

problems described earlier in this report, his conclusions nevertheless confirm the broader

pattern of significant racially polarized voting in HD 111 legislative elections.

My original December 22, 2017 expert report included several additional empirical

findings that were not disputed and were not mentioned by Dr. Alford's report. First, I found that

18



the 2014 House elections exhibited significant racially polarized voting, with over 98% of Black

voters supporting Democratic House candidates, while 75-85% of non-Black voters supported

Republican candidates. These findings were neither addressed nor disputed by Dr. Alford.

Second, I estimated the hypothetical outcomes of the November 2016 House elections,

assuming they had been held under the old 2012 Plan boundaries for HD 105 and HD 111.1

found that, under the 2012 Plan boundaries, a Black Democratic candidate would have defeated a

White Republican candidate in November 2016, winning approximately 50.3%-54.4% of the

vote in the two districts. These findings were neither addressed nor disputed by Dr. Alford.

Third, I found that demographic and partisan changes in HD 105 and HD 111 between

2012 and 2016 explain why a Democratic candidate would have defeated a Republican candidate

in November 2016 under the old 2012 Plan boundaries for HD 105 and HD 111. Specifically, my

original expert report found that the African-American share of the turnout electorate increased

noticeably from November 2012 to November 2016 in both HD 105 and HD 111 under the 2012

Plan. Meanwhile, non-African-American voters within HD 105, as drawn under the 2012 Plan,

became somewhat more likely to favor a Black Democratic House candidate in November 2016,

compared to previous elections. As a result of these collective racial and partisan shifts,

Democratic House candidates' vote share significantly increased among voters residing within

the 2012 Plan boundaries for HD 105 and HD 111 from November 2012 to November 2016. All

of these findings were neither addressed nor disputed by Dr. Alford.

Fourth, I analyzed the motivations for the redrawing of HD 105 and HD 111 in the 2015

plan. First, I found that the 2015 plan decreased the African-American share of the turnout

electorate by 4.0 percentage points in HD 105 and by 2.9 percentage points in HD 111. Overall,

in HD 105 and HD 111,1 found that the 2015 Plan generally decreased compliance with

traditional districting principles and with the principles set forth in the "2011-2012 Guidelines

for the House Legislative and Congressional Reapportionment Committee". Specifically, 1 found

that the 2015 Plan worsened compliance with the districting principles of equal population,

geographic compactness (as measured by Reock and Popper-Polsby scores), following precinct

boundaries, and avoiding county splits, while generally splitting more municipalities than

necessary. These findings were neither addressed nor disputed by Dr. Alford.

Finally, given that race and partisanship are highly correlated within these two districts, I

also analyzed in my original expert report whether partisan considerations, rather than racial

19



considerations, could account for the drawing of the new district boundaries in the 2015 plan. I

found that the 2015 Plan made a series of precinct-level switches that noticeably decreased the

African-American population in both HD 105 and HD 111 while subordinating traditional

districting principles and violating the 2011-2012 Redistricting Guidelines. These findings were

neither addressed nor disputed by Dr. Alford.

Furthermore, I found that the Legislature's primary map-drawer for the 2015 Plan had

access only to racial data, but not partisan data, at the sub-precinct level. Yet strikingly, I also

found that the 2015 Plan splits three precincts in HD 105 and five precincts in HD 111 in ways

that consistently decreased the African-American share of the population in both districts. The

fact that six of eight split precincts were split in ways that decreased the African-American share

of the population in HD 105 and HD 111 was neither addressed nor disputed by Dr. Alford.

I declare under penalty of perjury the foregoing is true and correct. Executed this 12th day of

February 2018.

Signed:

Jowei Chen

20



led 02

/12/

18 Page 21 of 37

Appendix A:

This appendix is a printout of Dr. Alf

ord'

s data fil

e: "HD 105/controlfile_2012_processed.dta"

list

ing th

e actual November 2012 ele

ctio

ns ana

lyze

d by Dr. Alford's computer code for HD 105 in his January 31,2018 rebuttal report.

election

contestnumber

contest

candidate

party

candorder

numcand

H1

Joutcome

race

ElectionlD

OfficelD

2012 GE

1President of the United States

MITT ROMNEY (R)

R1

3NA

NA

NA

11

2012

1

2012 GE


BARACK OBAMA

(l)(D)

D2

3NA

NA

NA

11

2012

1

2012 GE


GARY JOHNSON (L)

L3

3NA

NA

NA

11

2012

1

2012 GE

2

Public Ser

vice

Commission,

Dist

rict

3 - Metro-Atlanta

CHUCK EATON (R)

R1

3NA

NA

NA

11

2012

2

2012 GE

2

Public Ser

vice

Commission,

Dist

rict

3 - Metro-Atlanta

STEPHEN

OPPENHEIMER (D)

D2

3NA

NA

NA

11

2012

2

2012 GE

2

Public Ser

vice

Commission,

Dist

rict

3 - Metro-Atlanta

BRAD PLOEGER (L)

L3

3NA

NA

NA

11

2012

2

2012 GE

3Cl

erk Su

peri

or Court

RICH ALEXANDER

(l)(R)

R1

2NA

NA

NA

11

2012

3

2012 GE

3Cl

erk Superior Court

BRIAN WHITESIDE

(D)

D2

2NA

NA

NA

11

2012

3

21



led 02/12/18 Page 22 of 37

Appendix B:

This app

endi

x is

a printout of

Dr. Alford's da

ta file: "HD 105/controlfile_2016_processed.dta"

listing the ac

tual

November 201

6 elections an

alyz

ed by Dr. Al

ford's computer code for

HD 105 in his Ja

nuar

y 31,2018 rebuttal rep

ort.

election

contestnumber

contest

candidate

party

candorder

numcand

H1

Joutcome

race

ElectionID

OfficelD

2016 GE

1United States Senator, Isa

kson

JOHNNY ISAKSON

(!)

R1

3NA

NA

NA

11

2016

1

2016 GE

1United States Se

nato

r, Isa

kson

JIM BARKSDALE

D2

3NA

NA

NA

11

2016

1

2016 GE

1United States Senator, Isa

kson

ALLEN BUCKLEY

L3

3NA

NA

NA

12016

1

2016 GE


DONALD J.

TRUMP

R1

3NA

NA

NA

11

2016

2

2016 GE


HILLARY CLINTON

D2

3NA

NA

NA

11

2016

2

2016 GE


GARY JOHNSON

L3

3NA

NA

NA

11

2016

2

2016 GE

3CC Chair

C. NASH (l

)(R)

R1

2NA

NA

NA

11

2016

3

2016 GE

3CC Chair

J. SHEALEY (D)

D2

2NA

NA

NA

11

2016

3

22



led 02

/12/

18 Page 23 of 37

Appendix C:

This appendix is a printout of Dr. Alf

ord'

s da

ta file:

"HD

lll\

\con

trol

fiIe

_201

2jpr

oces

sed.

dta"

listing th

e actual November 2012 elections analyzed by Dr. Alf

ord'

s computer code for HD 111 in his January 31,2018 reb

utta

l re

port

.

election

contestnumber

contest

candidate

party

candorder

numcand

notes

outcome

race

ElectionID

OfficelD

election

contestnumber

2012 GE

1President of the United

States

MITT ROMNEY

(R)

R1

3NA

11

2012

12012 GE

1

2012 GE


States

BARACK OBAMA

(l)D

D2

3NA

11

2012

12012 GE

1

2012 GE


States

GARY JOHNSON

(L)

L3

3NA

11

2012

12012 GE

1

2012 GE

2Public Service

Commission, Di

stri

ct 3

- Metro-Atlanta

CHUCK EATON

(l)R

R1

3NA

11

2012

22012 GE

2

2012 GE

2Public Service

Commission, Dis

tric

t 3

- Metro-Atlanta

STEPHEN

OPPENHEIMER

(D)

D2

3NA

11

2012

22012 GE

2

2012 GE

2Public Service

Commission, Di

stri

ct 3

- Metro-Atlanta

BRAD PLOEGER

(L)

L3

3NA

11

2012

22012 GE

2

2012 GE

3Di

stri

ct Attorney, Fl

int

Circuit

JIM WRIGHT (R)

R1

2NA

11

2012

32012 GE

3

2012 GE

3District Attorney, Fl

int

Circuit

DARIUS

PATT

ILLO

(D)

D2

2NA

11

2012

32012 GE

3

2012 GE

4CHAIR COUNTY

COMMISSION

TOMMY SMITH

(R)

R1

2NA

11

2012

42012 GE

4

2012 GE

4CHAIR COUNTY

COMMISSION

CARLOTTA

HARRELL (D)

D2

2NA

11

2012

42012 GE

4

23



led 02

/12/

18 Page 24 of 37

Appendix D:

This app

endi

x is a printout of Dr. Alf

ord'

s data file: "HD lll/controlfile_2016_processed.dta"

listing th

e actual November 2016 elections ana

lyze

d by Dr. Alf

ord'

s computer code for HD 111 in his January 31,2018 reb

utta

l re

port

.

election

contestnumber

contest

candidate

party

candorder

numcand

notes

outcome

race

ElectionID

OfficelD

2016 GE


DONALD J. TRUMP

R1

3UNITY...

11

2016

1

2016 GE


HILLARY CLINTON

D2

31

12016

1

2016 GE


GARY JOHNSON

L3

31

12016

1

2016 GE

2United States Se

nato

r, Isa

kson

JOHNNY ISAKSON (1)

R1

31

12016

2

2016 GE

2Un

ited

States Senator, Isa

kson

JIM BARKSDALE

D2

31

12016

2

2016 GE

2Un

ited

States Senator, Isa

kson

ALLEN BUCKLEY

L3

31

12016

2

2016 GE

3CLERK, SUPERIOR COURT

B. HARRISON (1)

RR

12

11

2016

3

2016 GE

3CLERK, SUPERIOR COURT

S. HILL (D)

D2

21

12016

3

2016 GE

4SHERIFF

R. MCBRAYER (1

) RR

12

11

2016

4

2016 GE

4SHERIFF

J. M. ECKLER (D)

D2

21

12016

4

2016 GE

5CORONER

D. CLEVELAND (1

) R

R1

21

12016

5

2016 GE

5CORONER

T. BROWN (D)

D2

21

12016

5

2016 GE

6CHAIR COUNTY COMMISSION

J, WOOD (R)

R1

21

12016

6

2016 GE

6CHAIR COUNTY COMMISSION

C. HARRELL(D)

D2

21

12016

6

24


Case l:17-cv-01427-TCB-WSD-BBM Document 94-1 Filed 02

/12/

18 Page 25 of 37

Appendix £:

This

appendix is a printout of

Dr. Alford's data fil

e: "HD 105

/Rda

ta_2

012_

cont

est_

l.dt

a"co

ntai

ning

the pre

cinc

t-le

vel el

ecti

ons analyzed to pr

oduc

e Table 1 of Dr. Alford's January 31

,201

8 re

butt

al report.

precinct

candl

cand2

cand3

biktrn

allothturn

novote

OfficelD

ElectlonID

cands

60

535

1619

14

1211

960

31

2012

3

71

657

1472

17

1079

1070

31

2012

3

78

1221

766

13

483

1521

41

2012

3

80

1658

1616

29

1143

2173

13

12012

3

91

1702

977

32

585

2126

01

2012

3

134

1065

1128

13

841

1366

11

2012

3

144

720

1837

27

1360

1225

11

2012

3

146

2258

904

26

537

2657

61

2012

3

147

1391

1418

29

973

1868

31

2012

3

25



led 02/12/18 Page 26 of 37

Appendix F:

Dr. Alf

ord'

s computer file titled "el

ec20

12_H

D105

_con

test

_l.d

o";

Used to cr

eate

results for

the

first

2012 ele

ctio

n co

ntes

t analyzed in Table 1 of Dr. Alf

ord'

s January 31,2018 rebuttal re

port

.

1 /* th

is fi

le is pa

rt of a sequence of f

iles tha

t do one el

ecti

on eac

h fo

r th

e St

ate of

Georgia, HD105 exo

geno

us ele

ctio

ns. Ea

ch of t

hese do

file

s2

gets

called fr

om a run

file

{ru

n_ex

og.d

o)3 4

75 6

r set the

hom

e directory a

nd the contest and year gl

obal

s7 8

NOTE: Di

rect

ory of th

e correct R fi

les a

re har

d coded below, as is

the

location wh

ere those

9 files wi

ll be lo

ckin

g fo

r tempdata.

10

1112 7

1314

glob

al mys

tart

'"C:/Users/randy/Dropbox//"'

1516

glob

al dat

apat

h ""

_red

istr

icti

ng-s

hare

d-ri

ck-r

andy

\Gwi

nnet

t - Georgia\_final//"'

17

18

glob

al ana

lpat

h *"

cons

ulti

ng\S

tate

of Georgia\HD105//"'

19

20

2122

glob

al yea

r=20

122324

glob

al contest=1

2526

glob

al num

cans

=327

2829

/* Bri

ng in th

e demographics, do transformations, a

nd sav

e as

a sor

ted stata file 7

30

3132

impo

rt ex

cel "

${mystart}${datapath}${year}\demographics_${year}.xlsx", sh

eet(

"She

et1"

) fir

stro

w cl

ear

3334

35

duplicates li

st pre

cinc

t name I*

check for dupllicate precincts 7

36373 8

gen blktm=bf+bm

39

gen allothtum=nvf+nvm+of+om+uu+wf+wm+apf+apm+hsf+hsm

40

gen to

ttm=

tota

l4142

keep

if pre

cinc

t_nu

mber

==60

| III

26



led 02

/12/

18 Page 27 of 37

Appendix F:

Dr. Alf

ord'

s computer fil

e ti

tled

"elec2012_HD105_contest_l.do";

Used to cr

eate

results for

the fir

st 2012 ele

ctio

n contest an

alyz

ed in Table 1 of Dr. Alf

ord'

s January 31,2018 reb

utta

l re

port

.

43

prec

inct

_num

ber=

=711

III

44

precinct_number==78| II

I45

prec

inct

_num

ber=

=80|

III

46

precinct_number==911 II

I47

prec

inct

_num

ber=

=134

|///

48

precinct_number==144| I

II49

prec

inct

_nun

iber

==14

6| II

I50

precinct_number==147

5152

rena

me precinct_number precinct

53

sort pre

cinc

t54

5556

save

"${

myst

art}

${an

alpa

th}c

lemo

grap

hics

_${y

ear}

.dta

", re

plac

e57

58

/* bring in th

e eiection dat

a for t

he contest in

que

stio

n, merge wit

h demographics, bu

ild no

vote va

riab

le and save for use

in R •/

59

60

61

impo

rt ex

cel "

${my

star

t}${

data

path

}${y

ear}

\con

test

_${y

ear}

_${c

onte

st}.

xlsx

", sh

eet(

"She

et1"

) fir

stro

w clear

62

63

capt

ure dr

op to

tal

64

rena

me pre

cinc

t precinct_name2

65

rena

me pre

cinc

tid precinct

6667

sort precinct

68

merge 1:

1 precinct usi

ng "${mystart}${anaipath}\demographics_${year}.dta"

6970

keep

if _me

rge=

=37172

r cre

ate no vot

e */

7374

egen

cas

t=ro

wtot

al(c

and*

)75

gen no

vote

=tot

ai-c

ast

7677

repl

ace aliothturn=aliothtum+abs(novote) if

novote<0

/* so fe

w precincts

I cant af

ford

to drop but

hav

e to

ali

ocat

e th

e78

the ca

st vot

e so giv

e it to ot

hers

based on

proportionality */

79

replace no

vote

=0 if

novote<0

80

8182

/* add some variables the R fi

le needs to have under cer

tain

names */

8384

gen Of

fice

lD=$

{con

test

}

27



led 02

/12/18 Page 28 of 37

Appendix F:

Dr. Alf

ord'

s computer fil

e ti

tled


Used to cr

eate

results for

the

fir

st 2012 ele

ctio

n co

ntes

t an

alyz


ord'

s January 31,2018 rebuttal report.

85

gen ElectionlD=${year}

86

gen cands=${numcans}

87

88

keep precinct bi

ktrn

aliothtum ElectlonID OfficelD can

d* nov

ote

89

sort precinct

90

9192

save "${

myst

art}

${an

alpa

th}\

Rdat

a_${

year

}_co

ntes

t_${

cont

est}

.dta

", rep

lace

/* th

is Is t

he fi

le R wil

l re

ad - ac

tual

y we wi

ll mak

e a tem

p file

93

copy

of

It to

read so the

y wi

ll all

have th

e same name */

94

95

r cre

ate th

e tempdata tha

t R wil

l actually read (the file above Is f

or reference - cha

nge th

e names of th

e ca

ndid

ate variables to

V's

*/9697

98

local nu

mcan

s=${

numc

ans}

99

1 GO

fore

ach kk of nu

mlls

t 1/'numcans' {

101

rename cand'kk' V'kk'

102

}103

104

save "C

:\Us

ers\

rand

y\Dr

opbo

x\co

nsul

tlng

\Gen

erlc

Temp

Data

\tem

pdat

a.dt

a", r

eplace

105

106

107

108

r run the r fi

le - which sav

es out res

ults

as "R

resu

lts"

, which ar

e then read In

and

saved und

er a different name In the co

rrec

t directory

109

110

Note

that there Is

a kk va

riab

le which not nee

ded here - I le

ave

It In fo

r ease of tr

ansi

tion

thi

s to

a loop

111

*/112

113

local kk

=${c

onte

st}

114

loca

l nu

mcan

s=${

numc

ans}

115

local ye

ar=$

(yea

r}116

117

capture er

ase "C:\Users\randy\Desktop\GenerlcResultsFolder\Rresults.dta"

118

If'n

umca

ns-=

2 {119

shell "

C:\P

rogr

am Flles\R\R-3.3.2\bln\x64\R.exe" CMD BATCH "C

:\\U

sers

\\ra

ndy\

\Dro

pbox

\\co

nsul

tlng

\\AL

L El R FIL

ES\\

bo_2

cand

.r"

120

121

preserve

122

drop _all

123

use "C:\Users\randy\Desktop\GenerlcResultsFolder\Rresults.dta", c

lear

124

save

"C:\Users\randy\Desktop\HD105 Res

ults

\exo

g_'y

ear'

_HD 1

05_res'kk'.dta", rep

lace

125

restore

126

28



led 02

/12/18 Page 29 of 37

Appendix F:

Dr. Alf

ord'

s computer file ti

tled


Used to cr

eate

results for

the

first

2012 ele

ctio

n co

ntes


ord'


port

.

127

C:\U

sers

\ran

dy\D

ropb

ox\c

onsu

ltin

g\AL

L El R FIL

ES128

local kk

='kk

'+1

129

}130

else if

'num

cans

-=3 {

131

shell "C:

\Pro

gram

Files\R\R-3.3.2\bin\x64\R.exe" CMD BATCH "C:

\\Us

ers\

\ran

dy\\

Drop

box\

\con

sult

lng\

\ALL

El R FIL

ES\\

bo_3

cand

.r"

132

133

pres

erve

134

drop _a

ll135


lear

136

save "C:\Users\randy\Desktop\HD105 Results\exog_'year'_HD105_res'kk'.dta", replace

137

restore

138

139

140

loca

l kk='kk'+1

141

}142

siss

if 'num

c3ns

'~~4

143

shell "

C:\Progr

am Files\R\R-3.3.2\bin\x64\R.exe" CMD BATCH "C:

\\Us

ers\

\ran

dy\\

Drop

box\

\con

sult

lng\

\ALL

El R FIL

ES\\

bo_4

cand

.r"

144

145

preserve

146

drop_all

147

use "C

:\Us

ers\

rand

y\De

skto

p\Ge

neri

cRes

ults

Fold

ertR

resu

lts.

dta"

, clear

148

save "C:

\Use

rs\r

andy

\Des

ktop

\HD1

05 Res

ults

\exo

g_"y

ear'

_HD1

05_r

es'k

k'.d

ta",

rep

lace

149

rest

ore

150

151

152

local kk

='kk

'+1

153

}154

0ls0 if

*nLinic3ns'"~5

155

shell "

C:\P

rogr

am Flles\R\R-3.3.2\bin\x64\R.exe" CMD BATCH "C:

\\Us

ers\

\ran

dy\\

Drop

box\

\con

sult

lng\

\ALL

El R FILES\\bo_5cand.r"

156

157

pres

erve

158

drop_all

159

use "C

:\Us

ers\

rand

y\De

skto

p\Ge

nerl

cRes

ults

Fold

er\R

resu

lts.

dta"

, cle

ar160

save "C:\Users\randy\Desktop\HD105 Results\exog_'year'_HD105_res'kk'.dta", rep

lace

161

rest

ore

162

163

loca

l kk='kk'+1

164

}165

166

29



led 02/12/18 Page 30 of 37

Appendix G:

Dr. Alf

ord'

s computer file ti

tled

"elec2012_HD105_contest_2.do";

Used to create results for the

second 2012 election contest analyzed in Ta

ble 1 of

Dr. Alford's January 31,2018 rebuttal report.

1 2 /* th

is fi

le is pa

rt of a sequence of

file

s tha

t do on

e election eac

h for t

he State of G

eorgia, HD105 exogenous el

ecti

ons.

Each of

these do fi

les

3 gets

call

ed from a ru

n fi

le (r

un_e

xog.

do)

4 5 */

6 7 /* set th

e home directory and

the

contest and year gl

obal

s

9 NOTE: Di

rect

ory of th

e co

rrec

t R fi

les a

re har

d co

ded below, as is

the

loc

atio

n where those

10

file

s will be locking fo

r tempdata.

1112

13

*!14

15

glob

al mys

tart

"'C:

/Use

rs/r

andy

/Dro

pbox

/r'

1617

global dat

apat

h *"

_red

istr

icti

ng-s

hare

d-ri

ck-r

andy

\Gwi

nnet

t - Geo

rgia

\_fi

nal/

/"'

1819

global analpath "'

cons

ulti

ng\S

tate

of Georgia\HD1 O

S/f"

20

2122

23

glob

al year=2012

24

25

global contest=2

2627

global numcans=3

2829

30

r Bri

ng in t

he dem

ogra

phic

s, do tr

ansf

orma

tion

s, and sav

e as

a sorted st

ata file *!

3132

33

import excel "$

{mys

tart

}${d

atap

ath}

${ye

ar}\

demo

grap

hlcs

_${y

ear}

.xls

x", s

heet("Sheet1") firs

trow

clear

34

3536

dupl

icat

es li

st pre

cinc

t_na

me /*

che

ck fo

r dupllicate precincts */

37

3839

gen bl

ktrn

=bf+

bm40

gen a!

loth

tum=

nvf+

nvm+

of+o

m+uu

+wf+

wm+a

pf+a

pm+h

sf+h

sm41

gen

tott

rn=t

otal

42

30



led 02

/12/

18 Page 31 of 37

Appendix G:

Dr. Alf

ord'

s computer fil

e ti

tled


Used to cr

eate

results for

the second 2012 election co

ntes

t an

alyz

ed in Table I of Dr. Alf

ord'

s January 31,2018 rebuttal report.

43

keep

if pre

cinc

t_nu

mber

==60

| III

44

preclnct_number==711 II

I45


I46


IA1

precinct_number==911 II

I48

precinct_number==134|///

49


50

preclnct_number==146| I

II51

precinct number==147

52

53

rena

me precinct_number pre

cinc

t54

sort pre

cinc

t555657

save "${

myst

art}

${an

alpa

th}d

emog

raph

ics_

${ye

ar}.

dta"

, re

plac

e58

59

r bring in the election data for th

e contest in

question, mer

ge with demographics, bu

ild no

vote variable and sav

e fo

r use in

R *

/60

6162

import ex

cel "

${my

star

t}${

data

path

}${y

ear}

\con

test

_${y

ear}

_${c

onte

st}.

xlsx

", sh

eet(

"She

et1"

) fir

stro

w clear

63

64

capt

ure drop total

65

rena

me pre

cinc

t pr

ecin

ct_n

ame2

66

rena

me precinctid precinct

67

68

sort precinct

69

merg

e 1:1 pr

ecin

ct usi

ng "${

myst

art}

${an

alpa

th}\

demo

grap

hics

_${y

ear}

.dta

"70

7172

73

/* cre

ate no vot

e */

74

75

egen cast=rowtotal(cand*)

76

gen novote=total-cast

77

drop

if novote<0

7879

r add some var

iabl

es the

R fil

e needs to have under certain names *1

80

81

gen Of

fice

lD=$

{con

test

}82

gen El

ecti

onlD

=${y

ear}

83

gen cands=${numcans}

84

31



led 02

/12/

18 Page 32 of 37

Appendix G:

Dr. Alf

ord'

s computer fil

e ti

tled


Used to cr

eate

results for the sec

ond 2012 election co

ntes


ord'


utta

l report.

85

keep pre

cinc

t bi

ktrn

allothturn ElectlonlD OfficelD ca

nd* novote

86

sort precinct

87

8889

save "$

{mys

tart

}${a

nalp

ath}

\Rda

ta_$

{yea

r}_c

onte

st_$

{con

test

}.dt

a", r

epla

ce !*

this

is the

file R wi

ll read ~ ac

tualy we wi

ll mak

e a te

mp fi

le90

copy

of

it to rea

d so the

y will al

l ha

ve the

same name */

9192

r cre

ate th

e te

mpda

ta that R will a

ctually read (the fi

le above is f

or ref

eren

ce ~ change the na

mes of th

e ca

ndid

ate va

riab

les to

Vs *!

9394

95

loca

l numcans=${numcans}

9697

fore

ach kk

of nu

mlis

t 1/

'num

cans

' {98


99

}100

101

save

"C:\Users\randy\Dropbox\consulting\GenericTempData\tempdata.dta", rep

lace

102

103

104

105

r run

the

r fi

le - which saves out res

ults

as "Rresults", which ar

e then read

in and saved under a different name in

the

cor

rect

directory

106

107

Note tha

t the

re is a kk variable which not needed here - I le

ave

It in

for

ease of tr

ansition this t

o a loop

108

•/109

110

local kk

=${c

onte

st}

111

loca


112

loca

l year=${year}

113

114

capt

ure er

ase "C

:\Us

ers\

rand

y\De

skto

p\Ge

neri

cRes

ults

Fold

er\R

resu

lts.

dta"

115

if'numcans'==2{

116

shell "

C:\P

rogr

am Fil

es\R

\R-3

.3.2

\bin

\x64

\R.e

xe" CMD BATCH "C:\\Users\\randy\\Dropbox\\consulting\\ALL El R FIL

ES\\

bo_2

cand

.r"

117

118

preserve

119

drop _all

120

use "C

;\Us

ers\

rand

y\De

skto

p\Ge

neri

cRes

ults

Fold

er\R

resu

lts.

dta"

, clear

121

save

"C:\Users\randy\Desktop\HD105 Res

ults

\exo

g_'y

ear'

_HD1

05_r

es'k

k'.d

ta",

replace

122

rest

ore

123

124

C:\Users\randy\Dropbox\consulting\ALL El R FILES

125

local kk

='kk

'+1

126

}

32



led 02

/12/

18 Page 33 of 37

Appendix G:

Dr. Alford's computer fi

le titled "e

Iec2

012_

HD10

5_co

ntes

t_2.

do";

Used to cr

eate

results for the second 2012 ele

ctio

n co

ntes

t analyzed in Table 1 of

Dr. Alf

ord'

s Ja

nuar

y 31,2018 rebuttal report.

12 7

else

If' numcans'==3 {

128

shell "

C:\Program Fil

es\R

\R-3

.3.2

\bln

\x64

\R.e

xe" CMD BATCH "C

:\\Users\\randy\\Dropbox\\consulting\\ALL El

R FILES\\bo_3cancl.r"

130

pres

erve

131

drop

all

132

use "C:\Users\randy\Desktop\GenericResultsFolder\Rresults.dta", c

lear

133

save "C:\Users\randy\Desktop\HD105 Results\excg_'year'_HD105_res'kk'.dta", re

place

134

restore

135

136

137

local kk

='kk

'+1

138

}139

else

If' numcans-=4 {

140

shell "

C:\Program Fil

es\R

\R-3

.3.2

\bln

\x64

\R.e

xe" CMD BATCH "C

:\\Users\\randy\\Dropbox\\consulting\\ALL El R FILES\\bo_4cand.r"

142

pres

erve

143

drop

al

l144

use "C:\Users\randy\Desktop\GeneiicResultsFolder\Rresults.dta", c

lear

145

save "C:\Users\randy\Desktop\HD105 Results\exogJyear'_HD105_res'kk'.dta", r

eplace

146

restore

147

148

149

loca

l kk

='kk

'+1

150

}15

1 el

se If'

numc

ans'

==5 {

152

shel

l "C:\Program Fll

es\R

\R-3

.3.2

\bin

\x64

\R.e

xe" CMD BATCH "C

:\\Users\\randy\\Dropbox\\consulting\\ALL El

R FIL

ES\\

bo_5

cand

.r"

154

preserve

155

drop_all

156

use "C:\Users\randy\Desktop\GenericResultsFolder\Rresults.dta", c

lear

157

save

"C:

\Use

rs\r

andy

\Des

ktop

\HD1

05 Results\exogJyear'_HD105_res"kk'.dta", r

eplace

158

restore

159

160

local kk

='kk

'+1

161

}162

163

33



led 02

/12/

18 Page 34 of 37

Appendix H:

Dr. Alf

ord'

s computer file ti

tled


Used to cr

eate

results for

the

third 2012 ele

ctio

n co

ntes


ord'


port

.

1 2 r this f

ile i

s pa

rt of a sequence of

fiies t

hat do one election each for the

State of Ge

orgi

a, HD105 exogenous elections. Each of t

hese do fi

les

3 gets

called fr

om a run

file

(run

_exo

g.do

)4 5

•/6 7

/* set the

home directory and the

contest and yea

r globals

8 9 NOTE: Di

rectory of

the

cor

rect

R fi

les a

re har

d co

ded below, as is

the

location wh

ere th

ose

10

fiies will be lo

ckin

g for tempdata.

1112

13 7

14

15

glob

al mys

tart

'"C:/Users/randy/Dropbox//"'

1617

global dat

apat

h ■"

_red

istric

ting-

shar

ed-ri

ck-ra

ndy\

Gwi

nnet

t - G

eorg

ia\_

final

//"'

18 19

globa

l ana

lpath

"'co

nsul

ting\

Stat

e of

Geo

rgia

\HD1

05//"

'20 21 22 23

glo

bal y

ear=

2012

24 25

glob

al c

onte

st=3

26 27

glob

al n

umca

ns=2

28 29 30

r Br

ing in

the

dem

ogra

phics

, do

trans

fom

natio

ns, a

nd s

ave

as a

sor

ted

stata

file

731 32 33

im

port

exce

l "${

mys

tart}

${da

tapa

th}$

{yea

r}\de

mog

raph

ics_$

{yea

r}.xls

x", s

heet

("She

et1"

) firs

trow

clear

34 35 36

dupl

icate

s lis

t pre

cinct_

nam

e /*

chec

k fo

r dup

liicat

e pr

ecinc

ts 7

37 38 39

gen

blkt

rn=b

f+bm

40

gen

allo

thtu

m=n

vf+n

vm+o

f+om

+uu+

wf+w

m+a

pf+a

pm+h

sf+h

sm41

ge

n to

ttm=t

otal

42

34



led 02

/12/

18 Page 35 of 37

Appendix H:

Dr. Alf

ord'

s computer file ti

tled


Used to cr

eate

results for

the thi

rd 2012 ele

ctio

n co

ntes

t an

alyz


ord'


port

.

43

keep if

precinct_numbep==60| I

II44

prec

inct

_num

ber=

=711

III

45

prec

lnct

_num

ber=

=78|

III

46

prec

inct

_num

ber=

=80|

III

47

prec

inct

_num

bep=

=91|

///

48

preclnct_number==134|///

49

prec

inct

_num

bep=

=144

| III

50


51

prec

inct

number==147

52

53

rena

me pre

cinc

t_nu

mber

precinct

54

sort

pre

cinc

t555657

save

"${

myst

art}

${an

alpa

th}d

emog

raph

ics_

${ye

ar}.

dta"

, replace

5859

/* bring in th

e election data for t

he con

test

in que

stio

n, mer

ge with demographics, bu

ild no

vote variable and save for us

e in

R *

/60

6162

import exc

ei "$

{mys

tart

}${d

atap

ath}

${ye

ar}\

cont

est_

${ye

ar}_

${co

ntes

t}.x

lsx"

, she

et("

Shee

t1")

firs

trow

cle

ar63

64

capture drop tot

al65

rena

me precinct precinct_name2

66

rena

me pre

cinc

tid precinct

67

68

sort pre

cinc

t69

merg

e 1:

1 pr

ecin

ct usi

ng "${

myst

art}

${an

alpa

th}\

demo

grap

hics

_${y

ear}

.dta

"70

7172

73

/* create no vot

e */

74

75

egen cast=rowtotal(cand*)

76

gen novote=total-cast

77

drop

If novote<0

7879

r add some variables the

R file needs to have under certain names */

80

81

gen Of

fice

lD=$

{con

test

}82

gen Ei

ecti

onlD

=${y

ear}

83

gen ca

nds=

${nu

mcan

s}84

35



led 02/12/18 Page 36 of 37

Appendix H:

Dr. Alf

ord'

s computer fil

e ti

tled

"eIec2012_HD105_contest_3.do";

Used to cr

eate

results for

the thi

rd 2012 ele

ctio

n contest an

alyz


ord'


utta

l report.

85

keep

pre

cinc

t bi

ktrn

allothturn El

ecti

onlD

Office!D ca

nd* novote

86

sort precinct

87

8889

save "$

{mys

tart

}${a

na!p

ath}

\Rda

ta_$

{yea

r}_c

onte

st_$

{con

test

}.dt

a", r

epia

ce /*

this

is the fi

le R wi

ll read ~ ac

tual

y we wi

ll mak

e a te

mp fi

le90

copy of

it to

rea

d so the

y wi

ll all

have th

e same name */

9192

/* cre

ate th

e te

mpda

ta that R wi

ll actually read (t

he file

above is f

or reference - ch

ange

the

names of t

he can

dida

te var

iabl

es to V's */

9394

95

loca


9697

fore

ach kk

of nu

mlis

t 1/"numcans' {

98


99

}100

101

save

"C:

\Use

rs\r

andy

\Dro

pbox

\con

sult

ing\

Gene

ricT

empD

ataU

empd

ata.

dta"

, replace

102

103

104

105

/* run

the r file - whi

ch sav

es out results as "Rresuits", which ar

e then rea

d in and

saved under a different name in

the cor

rect

directory

106

107

Note tha

t the

re is a kk variable which not nee

ded he

re -1 lea

ve it

in for ease of tr

ansition thi

s to a loop

108

*/109

110

loca

l kk

=${c

onte

st}

111

loca

l nu

mcan

s=${

numc

ans}

112

loca

l year=${year}

113

114

capture er

ase "C

:\Us

ers\

rand

y\De

skto

p\Ge

neri

cRes

ults

Fold

er\R

resu

its.

dta"

115

if'numcans'==2 {

116

shell "

C:\Program Fil

es\R

\R-3

.3.2

\bin

\x64

\R.e

xe" CMD BATCH "C

:\\U

sers

\\ra

ndy\

\Dro

pbox

\\co

nsul

ting

\\AL

L El R FILES\\bo_2cand.r"

117

118

pres

erve

119

drop_all

120

use "C:\Users\randy\Desktop\GenericResultsFoldertRresults.dta", c

lear

121

save

"C:\Users\randy\Desktop\HD105 Re

sult

s\ex

og_'

year

'_HD

105_

res'

kk'.

dta"

, re

plac

e122

rest

ore

123

124

125

}126

else

if'

numc

ans'

==3 {

36



led 02/12/18 Page 37 of 37

Appendix H:

Dr. Alf

ord'

s computer file ti

tled

"eI

ec20

12_H

D105

_con

test

_3.d

o";

Used to create results for the thi

rd 2012 election contest an

alyz

ed in Ta

ble 1 of Dr. Alford's January 31,2018 reb

utta

l report.

127

shel

l "C:\Program Fil

es\R

\R-3

.3.2

\bin

\x64

\R.e

xe" CMD BATCH "Ci

WUse

rsWr

andy

WDro

pbox

Wcon

sult

ingW

ALL El R FIL

ESWb

o 3cand.r"

128

129

pres

erve

130

drop_all

131


lear

132

save "C:

\Use

rs\r

andy

\Des

ktop

\HD1

05 Res

ults

\exo

g_'y

ear'

_HD1

05_r

es'k

k'.d

ta",

replace

133

rest

ore

134

135

136

137

}13 8

else

If'

numcans'==4 {

139

shell "

C:\Program Files\R\R-3.3.2\bin\x64\R.exe" CMD BATCH "C

;\\Users\\randy\\Dropbox\\consulting\\ALL El

R FIL

ESWb

o 4c

and.

r"140

141

preserve

142

drop

_all

143

use "C

:\Us

ers\

rand

y\De

skto

p\Ge

nerl

cRes

ults

Fold

ertR

resu

lts.

dta"

, clear

144

save

"C:

\Use

rs\r

andy

\Des

ktop

\HD1

05 Results\exogJyear'_HD105_res'kk'.dta", r

eplace

145

rest

ore

146

147

148

149

}15 0

else

if" num

cans

'==5

{15

1 shell "

C:\Program Flles\R\R-3.3.2\bln\x64\R.exe" CMD BATCH "C

:WUs

ersW

rand

yWDr

opbo

xWco

nsul

tlng

\\AL

L El R FIL

ESWb

o_5c

and.

r"152

153

preserve

154

drop _all

155

use "C:\Users\randy\Desktop\GenerlcResultsFoldertRresults.dta", c

lear

156

save "C:

\Use

rs\r

andy

\Des

ktop

\HD1

05 Results\exogJyear'_HD105_res'kk'.dta", re

plac

e157

restore

158

159

160

}161

162

37


case 1:17-cv-01427-tcb-wsd-bbm document 140 filed 03/27/18 … naacp 20180227... · 2018-03-31 ·...

Documents