School Performance in Third Grade After a Full-Day vs Part-Day Preschool Program

Abstract

This study assesses whether learning gains associated with full-day preschool were sustained through third grade by analyzing achievement scores and need for remediation.

Preschool improves educational outcomes and well-being, especially for economically disadvantaged children.¹ Most evidence is from part-day programs. Whether full-day preschool (FDP) promotes sustained learning gains has rarely been tested, despite its expansion in recent years.² The Child-Parent Centers (CPC) program began in 1967 and was designed to promote children’s school success. In a previous article, FDP compared with part-day preschool (PDP) in CPCs in Chicago, Illinois, was associated with greater readiness skills in 4 of 6 domains at the end of the year.³

To assess whether associated gains were sustained, we conducted a follow-up through third grade, analyzing achievement scores and need for remediation.

Methods

The Midwest CPC, an expansion of the CPCs to economically diverse communities in 4 Illinois and Minnesota districts, provides enriched services in 6 core components: effective learning experiences, family engagement, aligned curriculum, professional development, continuity and stability, and collaborative leadership (eTable and eAppendix in Supplement 1).⁴

This study includes the Chicago cohort of the Midwest CPC, which has tracked a nonrandomized group of 2630 economically disadvantaged children since 2012-2013.⁴ As in a previous study,³ we focused on 3- and 4-year-old children in 11 of the 16 Chicago CPCs offering FDP and PDP. Mean class sizes were 18.1 for FDPs and 16.0 for PDPs, with FDP lasting 7 hours per day vs 3 hours per day for PDPs.³ The study was approved by review boards of the University of Minnesota and districts. Informed consent from parents and guardians was written and oral.

Third-grade follow-up was conducted through 2017-2018, and 4 outcomes from school records included (1) English/language arts (ELA) scores and proficiency and (2) math scores and proficiency on the state Partnership for Assessment of Readiness for College and Careers test in third grade (range, 650-850; proficiency defined as level ≥4 [5 levels]),⁵ (3) grade retention through third grade, and (4) special education placement through third grade.

Data were analyzed using generalized estimating equations (SPSS version 29; IBM) with school-level clustering, adjusted for 13 covariates (eg, baseline skills, race and ethnicity; eAppendix in Supplement 1) and attrition, via inverse propensity score weighting. Statistical significance was defined as P < .05 (2-tailed). Standardized mean differences (SMDs) were reported for outcomes (dichotomies converted via probit transformation) to allow comparison across measures and studies. Differences of greater than 0.2 were defined as practically significant.^3,4 Sensitivity analyses were conducted for children at less than 130% of the poverty level. Interactions with implementation quality, measured by center ratings of collaborative leadership (high [≥4] vs low [<4]; 5-point scale) were tested for all outcomes.

Results

At follow-up, 88.3% (361 of 409) and 89.5% (513 of 573) of the FDP and PDP groups, respectively, were enrolled in the district with available outcomes. Demographic characteristics in the cohort (mean age, 8.8 years; Black, 91.9%; Hispanic, 7.3%) were similar to the original sample.³ FDP was associated with higher ELA scores compared with PDP (725.0 vs 716.1; difference, 8.9 [95% CI, 1.6 to 16.2]; P = .02; SMD, 0.24) and higher ELA proficiency (38.3% vs 26.8%; difference, 11.5% [95% CI, 0.9% to 22.2%]; P = .03; SMD, 0.32; Table). FDP, compared with PDP, was associated with higher math scores (725.5 vs 719.3; difference, 6.2 [95% CI, 0.2 to 12.2]; P = .04; SMD, 0.20). Math proficiency reached practical but not statistical significance (25% vs 17.4%; difference, 7.6% [95% CI, −0.1% to 15.4%]; P = .055; SMD, 0.27).

Table. Means, Rates, and Group Differences in Percentages for Third-Grade Outcomes in Midwest Child-Parent Center (CPC) Full-Day (FDP) vs Part-Day (PDP) Preschool Adjusted for Baseline Attributes and Inverse Propensity Score Weighting for Attrition^a.

Outcome/sample group	FDP (n = 361)	PDP (n = 513)	Difference (95% CI)	P value	Standardized mean difference
Total sample (n = 874)
English PARCC achievement	725.0	716.1	8.9 (1.6 to 16.2)b	.02	0.24
Proficient, %	38.3	26.8	11.5 (0.9 to 22.2)b	.03	0.32
Math PARCC achievement	725.5	719.3	6.2 (0.2 to 12.2)	.04	0.20
Proficient, %	25.0	17.4	7.6 (−0.1 to 15.4)c	.055	0.27
Special education placement, %	11.4	9.8	1.6 (−5.5 to 8.7)	.66	0.06
Grade retention, %	3.0	9.1	−6.1 (−11.1 to −1.1)b	.02	−0.54
Family income <130% poverty level (n = 798)
English PARCC achievement	726.1	716.1	10.0 (3.4 to 16.5)b	<.01	0.29
Proficient, %	37.8	24.8	13.0 (3.1 to 22.9)b	.01	0.37
Math PARCC achievement	728.5	724.7	7.2 (1.7 to 12.5)b	.01	0.25
Proficient, %	24.2	15.6	8.6 (1.7 to 15.5)b	.02	0.31
Special education placement, %	12.4	10.1	2.3 (−5.3 to 9.9)	.56	0.12
Grade retention, %	3.2	9.5	−6.3 (−11.9 to −0.6)b	.03	−0.56

Abbreviation: PARCC, Partnership for Assessment of Readiness for College and Careers.

^a Values are adjusted for 13 baseline covariates (eAppendix in Supplement 1) and attrition. Standardized differences for the continuous achievement outcomes are based on the within-group standard deviation for the total sample. For the dichotomous outcomes, coefficients from the generalized estimating equations were converted to standardized differences via the probit transformation of percentages. Proficiency is defined by performance in the top 2 categories (4 or 5 of 5 levels; score ≥750) on the PARCC. Estimates are based on the total cohort sample with dichotomous contrast codes but excluding the 5 sites not offering FDP (n = 742). A total of 874 children (FDP and PDP) had valid values for ≥1 outcomes (range, 786-874). Original sample sizes were 409 and 573, respectively. Total sample sizes for low-income family status ranged from 722 to 798. The maximum FPD and PDP sample sizes were 323 and 475, respectively. Implementation quality was assessed for subgroup differences.

^b Confidence interval does not include 0 (P < .05; 2-tailed).

^c P < .10.

FDP was associated with lower percentages of grade retention vs PDP (3.0% vs 9.1%; difference, −6.1% [95% CI, −11.1% to −1.1%]; P = .02; SMD = −0.54), but not special education placement. Group differences among low-income children were similar to the overall sample (Table). The implementation quality by FDP status interaction P value was only significant for ELA proficiency (P = .044). FDP vs PDP was more strongly associated with ELA proficiency in centers with high implementation quality (39.6% vs 20.3%; difference, 19.3% [95% CI, 14.5% to 24.1%]; SMD, 0.59) vs low implementation quality (27.2% vs 21.0%; difference, 6.2% [95% CI, −0.4% to −12.8%]; SMD, 0.19).

Discussion

This study found that an evidence-based preschool program implemented full day vs part day was associated with sustained achievement gains and reduced grade retention. FDP may be preferable to PDP, especially among children in economically disadvantaged communities.^1,6 Implementation quality appears to bolster gains.³ The individual components of CPCs warrant broader investigation in a variety of contexts. Study limitations include unknown generalizability beyond Chicago or to other programs.

Section Editors: Jody W. Zylke, MD, Deputy Editor; Karen Lasser, MD, and Kristin Walter, MD, Senior Editors.

Supplement 1.

eAppendix. Study Sample and Background for the Midwest Child-Parent Centers

eReferences

eTable. Midwest Child-Parent Center (CPC) Program Core Elements and Implementation

Supplement 2.

Data Sharing Statement