Does Preschool have long term effects

1
One Year of Preschool or Two Does it Matter for
Long-term Outcomes ?
Irma Arteaga, Sarah Humpage, Arthur Reynolds
Judy Temple University of Minnesota
April 27, 2009 HCRC Mini-Conference Promoting
Child Health and Wellbeing
2
State Public Preschool Percent of National
Population Enrolled
Source National Institute for Early Education
Research (NIEER). The state of preschool 2007.
Yearbook.
3
MN Preschool Enrollment, 2008
Source National Institute for Early Education
Research (NIEER) and American Community Survey.
4
Policy Issues

• Policies on 1 versus 2 years of preschool vary
  by state.
• There is evidence that some preschool is better
  than none.
• Studies have found that a second year of
  preschool has short-term effects beyond one year

• Head Start (Ritblatt et al., 2001) ? children
  scored higher on the intellectual-cultural and
  active-recreation outcomes, parents read
  significantly more often to their children.

• Abbott (Barnett and Lamy, 2006) ? increase in
  childrens scores on vocabulary at the beginning
  of kindergarten.

• ECLS-K (Loeb et al., 2007 ) ? increase in
  childrens scores on reading and math at
  kindergarten.

5
Research Questions

• What is the effect of a 2nd year of preschool on
  early school adjustment over and above the first
  year?
• Does the effect of a 2nd year of preschool
  persist into adolescence and adulthood?
• 3. Does the impact of a 2nd year vary by child,
  family, and program characteristics?

6
Chicago Longitudinal Study

• Follows a single cohort of 1,539 children.
• The treatment group was composed of 989 children
  who graduated from the Chicago Child-Parent
  Centers. Centers are located in the highest
  poverty areas of Chicago.
• A comparison group of 550 children also received
  Title 1 services, which consisted of all-day
  kindergarten rather than the CPC preschool.

7
Child-Parent Centers

• Title I funded.
• Serves children at risk of school failure
• Educational and family-support services.
• Teachers have bachelors degrees and are
  certified in early childhood education.
• Reduced class sizes. Ratio is 17 to 2.
• Intensive parent program.

8
MethodologyQuasi-experimental design

• Selection into program

• Dosage

9
Propensity Score Weighting (PSW) to address
selection bias

• Controls for confounding characteristics that
  determine treatment and outcomes

• The PSW estimator creates a reweighted data set
  that better resembles a randomized experiment.

Implementation
Step 1 Estimate the probability of being
assigned to the 0, 1 or 2 group.
Step 2 Use these estimated probabilities to
construct the weight
Step 3 Estimate the outcome equation weighting
by PSW.
10
With and Without PSW Correction
Note For all results, plt0.001, plt0.01,
plt0.05 plt.10.
11
CLS Readiness Reading (at/above national
average)

Note plt0.001, plt0.01, plt0.05 plt.10.
12
Parent involvement (5-point scale)

Note For all results, plt0.001, plt0.01,
plt0.05 plt.10.
13
Grade Retention(cumulative)

14
Special Education(cumulative)

15
Juvenile Outcomes

16
Adult Crime

17
Education Attainment

18
Occupation and Socioeconomic Status

19
By Mothers Education
20
By Mothers Education

21
Conclusions

• A second year of preschool increases levels of
  school readiness and early achievement
• A second year of preschool has significant
  effects on grade retention and special education
  these can be of greater magnitude than the first
  years.
• A second year of preschool has lasting effects on
  behavioral outcomes (juvenile and adult crime)
  but not on educational attainment.
• The positive effects of a second year are greater
  and more significant for more disadvantaged
  children.
• The use of PSW on dose-response interventions is
  a versatile approach to bias reduction and can be
  widely used in other situations.

22
Policy Implications

• If the goal of preschool is school readiness and
  early achievement, two years are better than one.
• If the goal is enhancing long-term outcomes, the
  effects of two years are more mixed, but
  preventive effects were found for child
  maltreatment and juvenile arrest.
• Facing budget constraints, results suggest that
  resources should be targeted to disadvantaged
  children.
• For optimal effectiveness, interventions that
  extend beyond preschool are important to consider
  in policy decisions.