the relative importance of supply and demand factors in determining preschool attendance rob...

THE RELATIVE IMPORTANCE OF SUPPLY AND DEMAND FACTORS IN DETERMINING PRESCHOOL ATTENDANCE

Rob Wassmer, Chairperson and Professor,Public Policy and Administration, Sacramento State

Summary

IntroductionLiterature ReviewRegression ModelData DescriptionRegression ResultsPolicy ImplicationsConclusion

04/19/23

2

Introduction

… But today, fewer than 3 in 10 four year-olds are enrolled in a high-quality preschool program. Most middle-class parents can’t afford a few hundred bucks a week for a private preschool. And for poor kids who need help the most, this lack of access to preschool education can shadow them for the rest of their lives. So tonight, I propose working with states to make high-quality preschool available to every single child in America… (President Obama’s State of the Union Address, 2/12/2013)

04/19/23

3

Introduction

04/19/23

4

http://greatergreatereducation.org/post/18902/daycare-is-education-part-2-early-childhood-care-matters

Introduction

04/19/23

5

For 2009; http://earlyed.newamerica.net/blogposts/2010/2009_snapshot_of_preschool_programs_state_by_state_shows_losses_and_gains-31358

Introduction

04/19/23

6

All programs analyzed were high quality Proposed is universal, analyzed programs were

targeted to low-income and disadvantagedLynch, R. G. (2004). Exceptional Returns: Economic, Fiscal, and Social Benefits of Investment in Early Childhood Development. Washington, DC: Economic Policy Institute

Introduction

04/19/23

7

From CHIS survey data used here

Mimics what found elsewhere Latino

Difference Income

Difference

All

25.6

% (

21.9

to 2

9.4

)

Lati

no 1

7.3

% (

13.1

to 2

1.6

)

Whit

e 3

3.6

% (

27.5

to 3

9.7

)

Afr

ican A

meri

can 3

8.1

% (

17.5

to 5

8.7

)

Tw

o o

r M

ore

Race

s 38.2

% (

25.4

to 5

1.0

)

Asi

an 3

9.2

% (

24.0

to 5

4.5

)

Zero

to 1

00%

FPVL

12.5

% (

7.4

to 1

7.6

)

100 t

o 2

00%

FPVL

21.6

% (

13.3

to 2

9.1

)

200 t

o 3

00%

FPVL

17.0

% (

9.2

to 2

4.9

)

300 t

o 4

00%

FPVL

25.7

% (

15.9

to 3

5.5

)

400%

Plu

s FP

VL

49.6

% (

42.1

to 5

7.1

)

0

5

10

15

20

25

30

35

40

45

50

55

60

65

All Race/ Ethnicity Income

Figure 1: Preschool Attendance for 10 or more Hours a Week for Californians Aged 3-5 in 2011-12 (From California Health Interview Survey; 95 Percent Confidence Interval in Parentheses)

Introduction

04/19/23

8

Strong Start for Children Act November 2013, multi-year $100B federal subsidies for

high quality state preschool programs Serve three- to five-year-olds, teachers hold Bachelor’s

degrees, maximum class size, specific staff to student ratio, full-day, curriculum aligned with state standards, pay in line with K-12, regular state monitoring, comprehensive services offered, teacher professional development, meet Head Start standards, and health/safety standards.

California State Assembly commissioned this study

Literature Review

04/19/23

9

Economic model of parent(s)’ demand for child’s preschool education and available supply determining observed attendance

Table A1 offers summary of previous regression studies Demand factors

Income Clearly a normal good

Substitute availability Greater use if grandparent not available, mother/father/partner not

present Preferences

Race/ethnicity, English proficiency, employment, maternal education Supply of providers (price)

None did an adequate job of accounting for this Greater use if lower average tuition price in county, preschool

statewide funding, urban residence

Regression Model

04/19/23

10

Preschool Hoursi = f(Incomei, Substitute Availabilityi, Preferencesi, Supply of Providersi, Controlsi) Where i = 1, 2, 3, …1,928 households with

children aged three to five in CHIS survey Preschool Hoursi = f(Incomei, Familyi,

Parentsi, Race/Ethnicityi, Childi, Neighborhoodi, Child Care Facilitiesi)

Data Description

04/19/23

11

June 2011 to Jan 2013 telephone survey of 7,344 CA households containing newborns to age 18

Data from 1,928 respondents with child age 3-5 Questions used

Regular childcare of more than 10 hours a week in private preschool/nursery school or public Head Start/state preschool (yes for 499)

Number of hours in an average week last year Survey weights used

Data Description

04/19/23

12

Variable Name Description (Variable derived from in 2011-12 CHIS Public Use Data)

Mean

StandardDeviation

Min Max

DependentPreschool Hours Hours in Head Start, State

Preschool, Preschool, or Nursery School (CG2, CG3B, and CG3C)

7.387

13.993 0 63

IncomeBelow Poverty Line Federal poverty line used

(POVLL)0.225

0.417 0 1

Poverty Line to Three Times Poverty Line

Federal poverty line used (POVLL)

0.351

0.477 0 1

Own Home (SRTENR) 0.523

0.500 0 1

EducationCompleted High School For MKA [Most

Knowledgeable Adult] (CHEDUCA)

0.180

0.384 0 1

Completed Vocational School

For MKA (CHEDUCA) 0.032

0.176 0 1

Completed Some College For MKA (CHEDUCA) 0.116

0.320 0 1

Completed Associates Degree


0.260 0 1

Completed Bachelor’s Degree


0.423 0 1

Completed Less than Master ’s Degree


0.096 0 1

Completed Master’s Degree For MKA (CHEDUCA) 0.119

0.324 0 1

Completed Doctorate Degree


0.216 0 1

Data Description

04/19/23

13

FamilyMarried (FAMT4) 0.854 0.353 0 1Number of Household Members

(HHSIZE_P) 4.449 1.190 2 7

Parent(s)Age 30 to 49

For MKA (AGEGRP_A) 0.781 0.413 0 1

Age 50 Plus For MKA (AGEGRP_A) 0.057 0.232 0 1Does Not Speak English For MKA (CH18) 0.101 0.302 0 1Mother in U.S. Less than Four Years

(YRUSM) 0.023 0.151 0 1

Father in U.S. Less than Four Years

(YRUSF) 0.021 0.142 0 1

Mother Non-Citizen (CITIZ2_M) 0.023 0.151 0 1Father Non-Citizen (CITIZ2_F) 0.021 0.143 0 1Race/EthnicityLatino Considers Self Mexican Latino and identified

themselves as Mexican subtype (LATIN9TP)

0.377 0.485 0 1

Latino CA Department of Finance Classification (RACEDO_P)

0.460 0.498 0 1

African American CA Department of Finance Classification Non-Latino (RACEDO_P)

0.033 0.179 0 1

Native American Same as above 0.007 0.085 0 1Asian Same as above 0.095 0.293 0 1Two or More Races Same as above 0.072 0.258 0 1

Data Description

04/19/23

14

ChildChild Female (SRSEX) 0.493 0.500 0 1Child Age Four (SRAGE_P) 0.358 0.480 0 1Child Age Five (SRAGE_P) 0.342 0.474 0 1Parent Believes Child Health Poor

Assessment of general health (CA6)

0.005 0.068 0 1

Parent Believes Limited Ability Child

To do age-appropriate things(CA7)

0.042 0.201 0 1

NeighborhoodNeighborhood People Not Helpful

Strongly disagree or disagree with people in neighborhood willing to help (CG39)

0.026 0.161 0 1

Neighborhood Not Safe Some or none of the time feel safe in the neighborhood (CG42)

0.018 0.133 0 1

Second City Residence Census Claritas definition based on Census Tract (UR-TRACT)

0.202 0.402 0 1

Suburban Residence Same as above 0.202 0.402 0 1Town/Rural Residence Same as above 0.192 0.394 0 1Preschool FacilitiesPreschool Slots Per 3-5 Year Olds in 5 Mile Radius

GIS based upon 2010 Census and 2012 State of CA Preschool Data

0.152 0.068 0 0.443

Preschool Slots Per 3-5 Year Olds in 10 Mile Radius

Same as above 0.165 0.058 0 0.383

Preschool Slots Per 3-5 Year Olds in 20 Mile Radius

Same as above 0.177 0.052 0.017 0.319

Regression Results

04/19/23

15

Use STATA Two-Part Model (TPM) Logistic regression

One for 499 respondents, zero for 1,429 respondents OLS

Typical weekly hours in preschool for 499 respondents (mean=28.5)

Robust standard errors (heteroskedasticity and survey weights)

Also calculate conditional marginal effects at mean values for all statistically significant influences on overall preschool hours (comparable to Tobit)

Table 2 Four sets of Logit/OLS results with (1) no provider

measure, (2) providers within 5 miles, (3) providers within 10 miles, (4) providers within 20 miles Only show (4)

04/19/23

16

Provider Measure Within Twenty Miles

Explanatory VariableAttend or Not

7

Hours Attended

8Income

Below Poverty Line-54.1%-0.779*(0.451)

-3.090(2.859)

Poverty Line to Three Times Poverty Line1

-46.8%-0.632**(0.338)

-3.084(2.525)

Own Home0.053

(0.246)4.469**(1.886)

EducationCompleted High School

0.021(0.472)

4.567*(2.729)

Completed Vocational School

0.852(0.599)

0.386(4.848)

Completed Some College

0.113(0.507)

-0.072(3.414)

Completed Associate ’s Degree

0.385(0.594)

3.848(3.596)

Completed Bachelor’s Degree

154.5%0.934*(0.520)

4.916(3.239)

Completed Less than Master’s Degree

437.1%1.681**(0.961)

-0.709(3.853)

Completed Master’s Degree

305.5%1.400**(0.603)

3.388(3.069)

Completed Doctorate Degree2

468.6%1.738***(0.612)

4.452(3.159)


Explanatory Variable

Attend or Not

7

Hours Attended

8Family

Married-0.272(0.267)

-5.149**(2.140)

Number of Household Members

-0.130(0.093)

0.877(0.808)

Parent(s)

Age 30 to 490.019

(0.3334)-5.571**(2.300)

Age 50 Plus3 -0.339(0.683)

-1.560(4.007)

Does Not Speak English

-70.3%-1.213*(0.601)

1.991(5.220)

Mother in U.S. Less than Four Years

0.025(0.560)

11.626*(7.240)

Father in U.S. Less than Four Years

661.4%2.030***(0.712)

-0.114(7.130)

Mother Non-Citizen-0.293(0.294)

3.969*(2.148)

Father Non-Citizen0.0005(0.319)

1.130(2.211)

Race/EthnicityLatino Considers Self Mexican

0.468(0.418)

-2.270(3.061)

Latino-0.157(0.455)

1.038(3.086)

African American219.6%1.162**(0.572)

3.910(3.211)

Native American508.6%1.806**(0.664)

-2.543(7.308)

Asian0.349

(0.324)-4.075*(2.362)

Two or More Races4

0.398(0.382)

-2.870(2.695)



Attend or Not7

Hours Attended

8Child

Child Female-0.232(0.201)

-0.499(1.292)

Child Age Four120.1%0.789***(0.209)

1.369(1.524)

Child Age Five5

-57.2%-0.848***(0.291)

-2.374(1.950)

Parent Believes Child Health Poor

1.639(2.100)

11.558*(6.850)

Parent Believes Limited Ability Child

-0.559(0.552)

-2.692(2.845)

Neighborhood

Neighborhood People Not Helpful

-79.0%-1.562**(0.776)

2.654(3.871)

Neighborhood Not Safe

-1.172(0.802)

-15.958***(4.019)

Second City Residence

-0.193(0.282)

-1.516(2.607)

Suburban Residence

-40.8%-0.525***(0.227)

-4.692**(1.896)

Town/Rural Residence6

-0.282(0.262)

-5.554**(2.369)

Preschool FacilitiesPreschool Slots Per 3-5 Year Old in 20 Mile Radius

5,267.8%3.983*(2.268)

-13.761(14.758)

R-Squared(Pseudo for Logit)

0.218 0.2181Excluded category is More than Three Times Poverty Line; 2Excluded category is Less than High School; 3 Excluded category is Primary Caregiver Less than Age 30; 4Excluded category is White; 5Excluded category is Child Age Six; 6Excluded category is Urban.

Table 2: Logit (Preschool Attend or Not) and OLS (Preschool Hours Attended) Regression Results for Various Measures of Provider Availability (estimated with STATA Two-Part Model)

04/19/23

17

Provider Measure Within Five Miles Provider Measure Within

Ten Miles



Attend or Not11

Hours Attended

12

Attend or Not13

Hours Attended

14

Attend or Not15

Hours Attended

16

NeighborhoodSecond City Residence -0.595

(0.733)-2.173(4.907)

-1.032(0.989)

3.979(5.698)

-1.188(1.154)

-3.172(7.801)

Suburban Residence -0.968(0.751)

-10.382(7.111)

-1.603(1.017)

-14.344*(8.322)

-1.922(1.265)

-13.836*(8.999)

Town/Rural Residence -0.191(0.689)

-12.669***(4.595)

-0.705(0.868)

-16.878***(4.970)

-0.373(1.076)

-18.335***(6.609)

Preschool FacilitiesPreschool Slots Per 3-5 Year Olds in Given Miles

0.723(3.090)

-16.108(18.652)

-0.730(3.965)

-39.839*(20.743)

0.773(4.741)

-38.432^

(25.904) Preschool Slots X Second City Residence

1.922(3.723)

-21.815(27.137)

4.422(5.197)

-37.464(32.568)

5.217(5.938)

5.929(40.947)

Preschool Slots X Suburban Residence

2.488(4.127)

30.695(37.255)

5.608(5.353)

48.488(40.831)

7.099(6.424)

44.468(42.845)

Preschool Slots X Town/Rural Residence

2.923(4.4452)

55.081**(24.813)

5.730(4.891)

69.533***(23.638)

3.262(6.058)

75.596**(35.430)

R-Squared(Pseudo for Logit

0.214 0.230 0.216 0.240 0.217 0.226

Table 3: Provider Interaction Results (Other Explanatory Variables Included but Results not Recorded)

04/19/23

18

Explanatory VariableNo Provider

Measure

Provider Measure Within Five Miles

Provider Measure Within

Ten Miles

Provider Measure Within Twenty

MilesIncomeBelow Poverty Line -4.206* -4.091* -4.090* -4.071*Poverty Line to Three Times Poverty Line

-3.516** -3.433** -3.419** -3.413**

EducationCompleted Bachelor’s Degree 5.228** 5.089** 5.197** 5.118**Completed Less than Master’s Degree 7.703* 7.218* 7.546* 7.390*Completed Master’s Degree 7.057*** 6.629*** 7.014*** 6.892***Completed Doctorate Degree 8.971*** 8.501*** 8.675*** 8.626***FamilyMarried -2.290* -2.285* -2.264* -2.198*Parent(s)Does Not Speak English -4.921* -5.046* -4.9995* -5.049*Father in U.S. Less than Four Years 9.122*** 8.949*** 9.109*** 9.065***Race/EthnicityAfrican American 5.900** 6.152** 6.050** 5.944**Native American 7.308** 7.016** 7.155** 7.603**ChildChild Age Four 3.878*** 3.826*** 3.808*** 3.792***Child Age Five -4.264*** -4.169*** -4.269*** -4.247***NeighborhoodNeighborhood People Not Helpful -6.861** -6.658** -6.682** -6.489*Neighborhood Not Safe -8.179** -8.101** -8.296** -8.276**Suburban Residence -3.479*** -3.298*** -3.359*** -3.241***

Table 4: Conditional Marginal Effects Calculated at Mean Values for all Statistically Significant Influences on Overall Value Preschool Hours(from a mean of 7.387)

~Derived using STATA’s margins, dy/dx atmean command. These indicate the expected change in hours of preschool attended for a one-unit change in the respective variable, holding all other explanatory variables constant.

Policy Implications

04/19/23

19

Different demand and supply influences found important to parent(s)’ decision to (1) send child to preschool and (2) hours to attend

Demand factors and attendance 50% greater: Household income > three times federal poverty level 70% smaller: MKA does not speak English 80% smaller: MKA believes neighborhood people not helpful 150 to 500% greater: MKA holds Bachelor’s to doctorate degree 200 and 400% greater: African and Native American compared to

White Latino affect disappears

Above findings could help guide outreach/subsidy efforts

Policy Implications

04/19/23

20

Demand factors and hours (mean = 28.5) attended 4.0 hours less: Asian 4.5 hours more: Own home 5 hours less: Married (as compared to unmarried) parents 5.5 hours less: Age 30 to 39 (as compared to less than age

30)

Above findings could help guide outreach/subsidy efforts Note cumulative affect of being Asian, renting,

married and age 30 to 39 is 19 hours less

Policy Implications

04/19/23

21

Supply factors and attendance 40% smaller: Suburban (as compared to urban) residence Supply of preschool slots per 3-5 year olds within

10 mile radius is 0.165 20 mile radius is 0.177

Raise supply of preschool slots one standard deviation (0.06) Within 10 mile radius, increase likelihood of attendance by 75% Within 20 miles radius, increases likelihood of attendance by 300%

Supply of preschool slots influence nearly as great as education influences noted earlier

Policy Implications

04/19/23

22

Supply factors and hours (mean = 28.5) attended Only detected when preschool slots per 3-5 year olds

interacted with type of residence Consider findings for provider measure within 20 miles

Suburban (as compared to urban) attends13.8 less hours Town/Rural (as compared to urban) attends18.3 less hours

These hours increase based upon preschool slots per 3-5 year old in Town/Rural area

If minimum of 0.017 then 0.63 more If mean of 0.177 then 6.58 more If maximum of 0.319 then 11.86 more

But never enough to overcome Town/Rural deficit

Conclusion

Public policy interventions designed to increase preschool attendance should not only target outreach/subsidies at the groups identified here, but also consider ways to increase providers in areas with low preschools slots per 3-5 year olds More public and Head Start providers Public subsidies to private providers Reducing zoning codes previously found to

negatively influence presence of providers (Queralt and Witte, 1998)

04/19/23

23

the relative importance of supply and demand factors in determining preschool attendance rob...

Documents