columbia college english assessment/placement update
TRANSCRIPT
Columbia College English Assessment/Placement Update
Preliminary Data Summary
Fall 2008 – Spring 2011
Per AB3 and good practice, it is the responsibility of the College to monitor and review the accuracy of
its placements (cut score ranges) for placing students into coursework. Presumably, the more accurately
a student is placed, the greater chance he or she will have of being successful in the course. Thus, it is
also the responsibility of the College to make considered changes to those cut score ranges when there is
evidence that there is a need to do so. That is commonly called “cut score validation.”
The last cut score validation project for Columbia College English course placement was conducted and
reported in the 2008 “Accuplacer College Placement Validation Project.” This study also included an
“English Placement Consequential Validity” review whereby students and faculty evaluated the accuracy
of the placements via performance in the course. At that time, the cut scores were deemed reasonable
given the students’ performance and the feedback received from faculty. This report begins where the
2008 report ended analyzing the student placement and grade outcome data collected from Fall 2008 to
Spring 2011.
Background
Columbia College requires new students, or students who have been on an extended hiatus from college
coursework, to complete a placement exam in order to enroll in the appropriate English course for their
skill level. Testing is completed prior to the student’s orientation and advisement and before they register
in courses. The Datatel IT registration system blocks students from enrolling above their assessed course
level unless there is a manual override entered by a professional counselor or administrator who deems a
higher placement is appropriate for the student (multiple measures). Students may, however, self-select
to enroll in any English course that is below their assessed level.
Columbia uses the College Board’s online “Accuplacer” assessment system. For placement into English
courses, Accuplacer offers two test modules—sentence or writing skills, and reading. The College places
students into English via a score received on the sentence skills module only. Though the reading skills
module is available and students may choose to test in it, it is not required. Both test modules, sentence
and reading, have score ranges from between 0-120.
There are three levels of English courses: English 650 which is considered entry level or basic skills,
English 151 which is one level below transfer level English, and English 1A. The levels are determined
by the College and input into the Accuplacer system via a detailed set of decision rules. These decision
rules include several “background” questions that are weighted and add or subtract small increments to
the students total placement score. The background questions and decision rules can be updated or
altered by the College at any time. (Note: Though the background questions are currently in use, they
have never been statistically analyzed separately from the total scores.)
Data are taken from the Accuplacer test printouts and input into the district Datatel student database.
They are then extracted for analysis via the TSUM tables from a YCCD Crystal report. Likewise,
semester grades are obtained and downloaded through the Datatel system via the YCCD “Student
Demographic Detail” report from the same system. These term-driven downloads are then combined,
refined to isolate only the English scores and courses, and matched to each other via student IDs. When
the files are ready, they are then imported into the SPSS statistical application and Excel for analysis.
This study includes extracted data from Fall 2008 to the last term that had been graded and reported to
the California State Chancellor’s Office MIS system, Spring 2011.
We begin with summaries taken from the initial downloaded file contents, and then proceed through the
aggregating process to the final filtered and SPSS-ready numbers. They are included here as a means to
understand the data refinement process and to explain why the final Ns appear to be much lower than
would be expected given the number of terms included in the report.
Placement Distributions: Fall 2008 - Spring 2011
All Columbia College Sentence and Reading Skills TSUM test records,
July 15, 2008 to February 15, 2011:
All English Scores via the Accuplacer Assessment Battery
Sentence Skills Reading
N Valid 2301 1220
Mean 83.03 78.65
Median 84.00 81.00
Mode 84.04 81.35
Std. Deviation 20.274 20.903
Variance 411.022 436.954
Range 99 113
Percentiles 25 69 65
50 84 81
75 98 94
Included in the TSUM records were students’ retest data and as expected, average higher.
The following were removed:
All English Retests via the Accuplacer Assessment Battery
Sentence Skills Reading
N Valid 67 28
Mean 92.21 82.93
Median 93.00 85.00
Mode 93.33 84.00
Std. Deviation 16.052 18.599
Variance 257.653 345.921
Range 68 70
Percentiles 25 82 73
50 93 85
75 104 98
General Average Placement Levels1
Columbia College
Accuplacer Tests Taken
Fall 2008 – Spring 2011
Sentence
Skills
Reading
Skills Total
Test Score Mean: 83.03 78.65 81.51
Average Accuplacer Placement: English 151 English 151 Below
Transfer
Accuplacer Tests Completed: 2,301 1,220 3,521
Note: Average mean test scores did not contain records annotated as “retest,” previous course completion
advancements, and/or placements derived through outside institutions’ placement reports.
The next step was to further refine the data to include only those students who tested and enrolled in the
course of placement. After eliminating the retests, placements accepted from outside institutions, etc.,
the data were matched by the closest course completion date to the date of the Sentence Skills placement.
By matching to test and academic years, duplicate course completion records were reduced as much as
possible. (Many students had grades from two and three English courses during the study period.) This
reduced the data set to from 2,301 to 1,070. (See: Appendices, Table A1.)
1 Refer to the “Columbia College Placement Test Score Interpretation”, 2011-12 Columbia College Catalog.
Data were filtered again to include only the three English courses listed on the Columbia College
“Placement Test Score Interpretation” guide. (See: Footnote 1 and Appendices, Table A2.) The records
that remained numbered 712. (See: Appendices, Table A3.)
Because there were reading test scores in the TSUM data, the decision was made to conduct an similar
evaluation and a like process was conducted in matching and filtering for them. The result was 408
records which was sufficient for statistical analysis.
Cut Score Analysis – English
Because the primary Accuplacer instrument used for English placement is the Sentence Skills the first
analysis begins with the filtered 842 records derived from the original 4,301 records found in the TSUM.
English course grades included those from English 649, 650, 249, 250, 151, and 1A.
The Research Assumption:
Given an assessment through the Accuplacer Sentence Skills placement instrument and a review and
interview by a professional, the College would place students into appropriate levels of English
coursework for their skill levels. Students would then be more “likely” to succeed in those courses given
no external or uncontrolled circumstances, e.g., illness, life events, etc., that might interfere with their
performance.
An Overview of All Average English Grades Received (4pt_Grade Scale)
Sorted by the Accuplacer Writing Placement Received (the primary module used for English placement)
Fall 2008- Spring 2011
Note: The following data are highlighted to reflect Accuplacer Sentence Skills tests taken during 2008-
2011 where students enrolled in, and completed courses at the placement levels.
Accuplacer
Test Taken
CC Course
Completed:
English Placement Levels
ENGL-650 ENGL-151 ENGL-1A Total
N Mean N Mean N Mean N Mean
CACC-SENT ENGL 249,
649, 250, 650 37 2.14 72 1.60 4 a 1.50 113 1.77
ENGL 50, 151 4 a 1.25 402 2.16 14 2.21 420 2.16
ENGL 1A 36 1.69 273 2.34 309 2.27
Total 41 2.05 510 2.05 291 2.32 842 2.14
CACC-READ ENGL 249,
649, 250, 650 21 1.71 21 2.29 1 a 2.00 43 2.00
ENGL 50, 151 23 2.48 235 2.29 31 2.65 289 2.34
ENGL 1A 97 2.55 142 2.63 239 2.59
Total 44 2.11 353 2.36 174 2.63 571 2.42 a Cells with < 10 records - provided as information only.
Data were filtered to include only those students who tested and enrolled in their placement course, and
who received a grade during the study period of Fall 2008 to Spring 2010.
Average English Grades Received (4pt_Grade Scale)
by Accuplacer Placement Levels and Courses Completed at Placement Level
Fall 2008- Spring 2011
Accuplacer Sentence Skills Placements
Descriptives of Sentence Placements and Grade Outcomes after Filtering (N: 712)
Sentence Skills Placement Level
Mean Grade Averages (based on 4 pt. grade scale) Statistic Std. Error
Placed into
ENGL 650 level and
received a grade in
ENGL 650 level course 2
N: 37
Mean 1.77 .119
95% Confidence Interval for Mean
Lower Bound 1.53
Upper Bound 2.00
5% Trimmed Mean 1.74
Median 2.00
Variance 1.589
Std. Deviation 1.261
Minimum 0
Maximum 4
Range 4
Interquartile Range 2
Skewness -.071 .227
Kurtosis -.738 .451
Placed into
ENGL 50, 151
and
received a grade
N: 402
Mean 2.16 .066
95% Confidence Interval for Mean Lower Bound 2.03
Upper Bound 2.29
5% Trimmed Mean 2.17
Median 2.00
Variance 1.823
Std. Deviation 1.350
Minimum 0
Maximum 4
Range 4
Interquartile Range 2
Skewness -.352 .119
Kurtosis -1.019 .238
2 The “ENGL 650 level” includes the courses ENGL 249-250 and ENGL 649-650. (See: Appendices, Table A3.)
Descriptives of Sentence Placements and Grade Outcomes after Filtering (N: 712)
Placed into
ENGL 1A and
received a grade
N: 273
Mean 2.27 .086
95% Confidence Interval for Mean
Lower Bound 2.10
Upper Bound 2.43
5% Trimmed Mean 2.29
Median 3.00
Variance 2.280
Std. Deviation 1.510
Minimum 0
Maximum 4
Range 4
Interquartile Range 4
Skewness -.441 .139
Kurtosis -1.244 .276
Correlations for Sentence Skills Assessment
To grades in
courses:
Predictive Correlations
Accuplacer Sentence Skills Score
Pearson’s Kendall’s Spearman’s
N R p Tau_b p Rho p
English 649-650 37 .633* .000 .532* .000 .656* .000
English 151 402 .095 .057 .080* .033 .104* .036
English 1A 273 .078 .201 .066 .148 ,089 .143
* Indicates a statistically significant correlation that is not due to chance error (p < .05).
Accuplacer Reading Placements
Descriptives for Reading Placements and Grade Outcomes after Filtering (N: 398)
Reading Placement Level
Mean Grade Averages (based on 4 pt. grade scale)
Statistic Std. Error
Placed into
ENGL 650 level and
received a grade in
ENGL 650 level
N: 21
Mean 1.71 .302
95% Confidence Interval for Mean
Lower Bound 1.08
Upper Bound 2.34
5% Trimmed Mean 1.68
Median 2.00
Variance 1.914
Std. Deviation 1.384
Minimum 0
Maximum 4
Range 4
Interquartile Range 3
Skewness -.056 .501
Kurtosis -1.186 .972
Placed into
ENGL 151 level
and received a grade in
ENGL 151
N: 235
Mean 2.29 .088
95% Confidence Interval for Mean
Lower Bound 2.11
Upper Bound 2.46
5% Trimmed Mean 2.32
Median 3.00
Variance 1.837
Std. Deviation 1.355
Minimum 0
Maximum 4
Range 4
Interquartile Range 2
Skewness -.478 .159
Kurtosis -.933 .316
Reading Placement Level
Mean Grade Averages (based on 4 pt. grade scale)
N: 398 Statistic Std. Error
(continued)
Placed into
ENGL 1A level
and received a grade
in ENGL 1A
N: 142
Mean 2.63 .122
95% Confidence Interval for Mean
Lower Bound 2.39
Upper Bound 2.87
5% Trimmed Mean 2.70
Median 3.00
Variance 2.122
Std. Deviation 1.457
Minimum 0
Maximum 4
Range 4
Interquartile Range 2
Skewness -.806 .203
Kurtosis -.691 .404
Correlations for Reading Skills Assessment
To grades in
courses:
Predictive Correlations
Accuplacer Reading Score
Pearson’s Kendall’s Spearman’s
N R p Tau_b p Rho p
English 649-650 21 -.170 .460 -.155 .385 -.171 ,458
English 151 235 .136* .037 .110* .025 .145* ,027
English 1A 142 -.006 .944 .003 .962 -.006 .942
* Indicates a statistically significant correlation that is not due to chance error (p < .05).
Disproportionate Impact
Highlighted areas indicate percentages where placements are disproportionately higher in lower level
courses as compared to the overall and majority population. The rule of thumb: The percent distribution
of any student category should be within 20% of the majority group’s placements (80% of the placement
level). If the proportion is more than that, it has disproportional impact and requires that a plan to
monitor and address the situation be implemented and rectified if possible.
Note: We return to the earlier dataset of 1,070 students with an English placemen, and any English
course grade between the dates of July 15, 2008 and February 15, 2011.
Gender by Sentence Skills Placement Level
Placement Level
Gender Engl-650 Engl-151 Engl-1A Transfer Gender Total
Count % Count % Count % Count %
Female 18 3.0% 349 58.3% 232 38.7% 599 100.0%
Male 25 5.4% 220 47.4% 219 47.2% 464 100.0%
Unk/Unrpt 0 .0% 4 57.1% 3 42.9% 7 100.0%
Placement
Subtotal 43 4.0% 573 53.6% 454 42.4% 1,070 100.0%
Though no gender reached the threshold of 20% higher proportions in their placements compared to the
total, or would trigger a plan for action, it is noteworthy that a higher proportion of females placed in
English 151 than did males. Though the numbers of unknown/unreported are too small to determine
impact, they are included here to explain the missing 7 records and to bring the table’s total to 1,070
records.
Data Regarding Disproportionate Impact (continued)
Ethnicity by Sentence Skills Placement Level
Placement Level
Engl-650/250 Engl-151 Engl-1A Transfer Ethnicity Total
Ethnicity Count % Count % Count % Count %
1. Asian 0 .0% 6 66.7% 3 33.3% 9 100.0%
2. African Am. 1 7.7% 10 76.9% 2 15.4% 13 100.0%
3. Filipino 0 .0% 4 80.0% 1 20.0% 5 100.0%
4. Hispanic 6 5.5% 68 61.8% 36 32.7% 110 100.0%
5. Native Am. 3 14.3% 11 52.4% 7 33.3% 21 100.0%
6. Pacific Isl. 0 .0% 3 100.0% 0 .0% 3 100.0%
7. Other Non-White 1 3.8% 15 57.7% 10 38.5% 26 100.0%
8. White 23 3.1% 371 50.1% 346 46.8% 740 100.0%
9. Unk-Unrpt 9 6.3% 85 59.4% 49 34.3% 143 100.0%
Placement Total 43 4.0% 573 53.6% 454 42.4% 1,070 100.0%
The proportion of Hispanic students who placed into English 151 are noticeably higher than that of the
overall and the majority group.
While the numbers and proportions of African Americans, Native Americans, and Pacific Islanders had
larger proportions of students who placed into lower than transfer level coursework, the numbers
underlying those percentages are very small and therefore not highlighted. That does not mean that the
information is not worthy of notice or further exploration.
Disproportionate Impact (continued)
Age Levels by Sentence Skills Placement
Placement Level
Ages Engl-650 Engl-151 Engl-1A Age Group Total
Count % Count % Count % Count %
00-17 1 1.1% 42 45.7% 49 53.3% 92 100.0%
18-24 31 4.1% 407 54.1% 314 41.8% 752 100.0%
25-34 6 4.7% 71 55.0% 52 40.3% 129 100.0%
35-44 3 6.5% 20 43.5% 23 50.0% 46 100.0%
45-54 1 2.4% 29 70.7% 11 26.8% 41 100.0%
55-64 1 12.5% 3 37.5% 4 50.0% 8 100.0%
65-00 0 .0% 1 50.0% 1 50.0% 2 100.0%
Placement Subtotal 43 4.0% 573 53.6% 454 42.4% 1,070 100.0%
The students aged 24 and younger had the highest proportion of placements into transfer level English
1A compared to other age groups. The 45-54 year old group had the highest proportion overall of
students placing into English 151.
Primary Language by Sentence Skills Placement Levels
Placement Level
Engl-650 Engl-151 Engl-1A Language Total
Language Count % Count % Count % Count %
English 35 3.4% 552 53.5% 445 43.1% 1032 100.0%
Not English 8 28.6% 16 57.1% 4 14.3% 28 100.0%
Unk/Unrpt 0 .0% 5 50.0% 5 50.0% 10 100.0%
Placement Subtotal 43 4.0% 573 53.6% 455 42.5% 1,070 100.0%
With only 28 students indicating primary languages other than English, the percentages are questionable
as a true representation for this population; however, it may be worth noting that over half of the 28, or
57.1%, placed into English 151. This is contrary to the assumption that this group’s placements would be
lower than the English 151 level.
Appendices
Table A1. Descriptive Mean Score Statistics by Placement Level 3
Descriptives – English Placement Sentence Skills Mean Scores by Placement Level
N: 712 Placement Level Statistic Std. Error
MEAN SCORE
Engl-650 Mean 41.63 .905
95% Confidence
Interval for Mean
Lower Bound 39.80
Upper Bound 43.45
5% Trimmed Mean 41.83
Median 42.00
Variance 35.192
Std. Deviation 5.932
Minimum 29
Maximum 50
Range 21
Interquartile Range 9
Skewness -.452 .361
Kurtosis -.464 .709
Engl-151 Mean 75.36 .452
95% Confidence
Interval for Mean
Lower Bound 74.47
Upper Bound 76.25
5% Trimmed Mean 75.73
Median 77.00
Variance 117.140
Std. Deviation 10.823
Minimum 51
Maximum 91
Range 40
Interquartile Range 16
Skewness -.508 .102
3 Table A1 represents only average (mean) scores by placement level given a Sentence Skills placement test taken between
Fall 2008-Spring 2011. It is provided as a means to evaluation the average score within each placement level, and to highlight
any clusters with possible anomalies.
Descriptives – English Placement Sentence Skills Mean Scores by Placement Level
N: 712 Placement Level Statistic Std. Error
Kurtosis -.752 .204
Engl-1A Mean 104.44 .380
95% Confidence
Interval for Mean
Lower Bound 103.69
Upper Bound 105.19
5% Trimmed Mean 104.30
Median 104.00
Variance 65.611
Std. Deviation 8.100
Minimum 92
Maximum 120
Range 28
Interquartile Range 14
Skewness .179 .115
Kurtosis -1.089 .229
Descriptives – English Placement Reading Skills Mean Scores by Placement Level Descriptives
Placement Level Statistic
Std.
Error
English 650 Mean 37.32 1.057
95% Confidence
Interval for Mean
Lower Bound 35.19
Upper Bound 39.45
5% Trimmed Mean 37.22
Median 37.00
Variance 49.152
Std. Deviation 7.011
Minimum 27
Maximum 50
Range 23
Interquartile Range 12
Skewness .046 .357
Kurtosis -1.268 .702
English 151 Mean 75.47 .581
95% Confidence
Interval for Mean
Lower Bound 74.33
Upper Bound 76.62
5% Trimmed Mean 75.89
Median 77.00
Variance 119.017
Std. Deviation 10.909
Minimum 51
Maximum 91
Range 40
Interquartile Range 18
Skewness -.461 .130
Kurtosis -.772 .259
English 1A Mean 101.86 .522
95% Confidence
Interval for Mean
Lower Bound 100.83
Upper Bound 102.89
5% Trimmed Mean 101.52
Median 101.00
Variance 47.449
Std. Deviation 6.888
Minimum 92
Maximum 120
Range 28
Interquartile Range 10
Skewness .610 .184
Kurtosis -.223 .366
Table A2. Distribution of all English courses completed by placed students, 2008-2011.
YSTU_COURSE_NAME
Frequency Percent Valid Percent Cumulative Percent
Valid CENGL-125 5 .5 .5 3.3
CENGL-249 3 .3 .3 88.1
CENGL-250 8 .7 .7 88.9
CENGL-606 2 .2 .2 89.7
CENGL-649 68 6.4 6.4 96.1
CENGL-650 34 3.2 3.2 99.3
CENGL-47 6 .6 .6 89.4
CENGL-50 1 .1 .1 89.5
CENGL-151 419 39.2 39.2 42.4
CENGL-1A 309 28.9 28.9 71.6
CENGL-1B 142 13.3 13.3 84.9
CENGL-1C 32 3.0 3.0 87.9
CENGL-81 8 .7 .7 100.0
CENGL-10 19 1.8 1.8 1.8
CENGL-11 11 1.0 1.0 2.8
CENGL-17 1 .1 .1 42.5
CENGL-18 2 .2 .2 42.7
Total 1,070 100.0 100.0
Table A3. English courses completed by students with placement scores and enrollments in
placement courses during the study period of 2008-2011.
Distribution of Students Placements after Filtering
Sentence
Skills Percent
Reading
Skills Percent
Valid CENGL-249, 250, 649, 650 37 5.2 21 5.3
CENGL-151 402 56.5 235 59.0
CENGL-1A 273 38.3 142 35.7
Total 712 100.0 398 100.0
* N records for courses ENGL 249, 250, and 50 were insufficient for data analysis but are provided here
as a means to account for the student course completions given placements. Summaries were aggregated
to include these courses in the ENGL 650 placement level.