One versus two years: Does length of exposure to an enhanced preschool program impact the academic functioning of disadvantaged children in kindergarten?☆
Introduction
One out of every five children in the United States currently lives in poverty (Children's Defense Fund, 2010). Children who grow up in poverty are significantly more likely to experience delays in early cognitive, language, and social-emotional development at school entry that undermine their subsequent academic achievement. Over time, the achievement gap between these students and their higher-functioning peers often widens (Campbell and von Stauffenberg, 2008, Janus and Duku, 2007). Children who begin kindergarten lagging behind their classmates are at greater risk for academic problems in elementary and high school, including school dropout (McClelland et al., 2006, Ryan et al., 2006). In an effort to reduce the achievement gap, a substantial amount of research has been devoted to developing models of early childhood education (ECE) and evaluating their benefits, particularly for disadvantaged children (Ramey and Ramey, 1998, Weikart and Schweinhart, 1997). This research began several decades ago with highly controlled randomized trials of several comprehensive ECE programs (Berrueta-Clement et al., 1984, Gray and Klaus, 1987). These relatively small “demonstration projects” became the basis for the federal Head Start program (Ludwig & Phillips, 2007) and similar public preschool programs (Campbell and Ramey, 1994, Reynolds, 1995). The findings from these trials and several subsequent intervention trials and evaluation studies suggest that comprehensive, high-quality ECE improves children's academic and social-emotional functioning (Camilli et al., 2010, Manning et al., 2010), and can positively impact long-term outcomes such as rates of high school graduation, criminal activity, and employment (Muennig et al., 2009, Reynolds et al., 2001).
As research has established that participation in ECE is a cost-effective strategy for improving school readiness and long-term academic achievement for disadvantaged children (Aos et al., 2004, Belfield et al., 2006), more states are allocating funds to create ECE programs in public school settings (Barnett, 2010). In the last six years, several large-scale evaluations of state-funded public preschool programs have been conducted. These studies document immediate positive effects on reading and math skills at school entry and reductions in grade retention in kindergarten (Huang, Invernizzi, & Drake, 2012, Magnuson, Ruhm, & Waldfogel, 2007). The purpose of the current study was to evaluate the relative impact on children's early academic outcomes of attending one or two years of a public preschool program that delivered an enhanced version of Head Start.
In the Head Start Impact Study, a recent randomized control trial, researchers found that children who attended Head Start experienced some positive benefits in cognitive domains (U.S. DHHS, 2010). Given the fact that Head Start enrolls children at ages 3 and 4, the effects of participation were examined separately for these two groups but not compared. At the end of preschool, small effect size impacts were found for pre-reading, vocabulary, and parent reports of literacy skills; these results were found for children who entered Head Start at age 3 and at age 4 (U.S. DHHS, 2010).
The modest findings from Head Start evaluations have encouraged researchers to evaluate the integration of evidence-based enhancements (i.e., practices and programs focused on literacy, math, social-emotional learning, and executive functioning) to existing curricula that aim to strengthen program quality and boost children's performance (Griffin, 2010). Other examples of interventions that have been integrated into Head Start and rigorously evaluated through randomized control trials include: Dialogic Reading (Wasik et al., 2006, Whitehurst et al., 1994), Head Start REDI (Bierman et al., 2008), Pre-K Mathematics (Starkey, Klein, & Wakeley, 2004), Tools of the Mind (Diamond, Barnett, Thomas, & Munro, 2007), and Incredible Years (Webster-Stratton, Reid, & Hammond, 2004).
One of the challenges of going to scale with public preschool has been achieving results comparable to the early demonstration projects that established the evidence base for ECE. Experts in the field note that none of the major federal or state programs have successfully replicated these findings (Barnett, 2010). This may be a function of the fact that preschool is now being compared to a wider range of early childhood services available to children in the comparison group. In some cases, the advantages are a function of quality, as higher-quality preschool programs achieve better student outcomes (Burchinal et al., 2008, Howes et al., 2008, Mashburn et al., 2008). Unfortunately, when classroom observations are conducted with standard measures (e.g., Early Childhood Environment Rating Scale-R [ECERS-R; Harms, Clifford, & Cryer, 1998]; Classroom Assessment Scoring System [CLASS; La Paro & Pianta, 2003]), the quality of many public preschools is rated as mediocre to poor, particularly in terms of the instructional practices of teachers (Barnett, 2010, Burchinal et al., 2008).
Head Start provides a comprehensive model that includes social and health services to families in addition to the educational experiences that were part of the seminal ECE programs, but public schools usually provide only an educational component and rarely include the home visits required by Head Start. On the other hand, public preschool teachers receive higher salaries and employ certified teachers with at least bachelor's degrees that are comparable to those of the teachers in the early demonstration projects, based on the assumption that higher levels of education will result in higher levels of program quality. Recent studies of public preschool initiatives have focused on identifying the characteristics of programs associated with improved child outcomes (Curby et al., 2009, Pianta et al., 2005). These studies suggest that the quality of teacher–student interactions is a key factor associated with positive child outcomes (Early et al., 2006, Mashburn et al., 2008). This type of learning environment can be promoted with curricular enhancements and high-quality, embedded professional development of teachers (Domitrovich et al., 2009, Pianta et al., 2008).
Despite the potential policy implications, one structural feature of preschools that has not been researched extensively is program duration. From a theoretical standpoint, entry into preschool at age 3 should contribute to improved school readiness for several reasons. Younger children have greater neural plasticity. Given the age of entry into kindergarten, children who enter preschool earlier have the opportunity to remain in an educational program for twice as long as their older peers. The critical period for developing a range of cognitive (e.g., inhibitory control) and language (e.g., vocabulary) skills, which are the building blocks of later social and academic capacities, begins even earlier than the age of 3 and continue throughout childhood and adolescence (Shonkoff, 2011, Whitehurst and Lonigan, 1998). These skills develop and are refined through repeated experiences with others, so it is possible that children who are exposed to a stimulating preschool classroom environment earlier and for longer would make greater progress compared to those who entered later or participated for only one year.
Head Start and other public ECE programs enroll children younger than four years old but this is less common in public preschool programs. Advocates of the two-year preschool model argue that the longer children spend in a cognitively stimulating environment, the more they are likely to benefit and that two years are necessary to deliver a comprehensive preschool program effectively (Zigler & Styfco, 1994). However, findings from the National Household Education Survey indicate that only 65% of children in the lowest two income quintiles attended preschool at age 4 (Iruka & Carver, 2006), suggesting that children from the most disadvantaged backgrounds may not even be receiving one year of preschool. Those opposed to the two-year model argue that given the limited funds available for preschool, it is more important to serve more children for one year beginning at age 4 (Reynolds, 1995).
Only a handful of studies have compared outcomes for children receiving one versus two years of preschool and evidence from these studies is mixed. Some of the early work in this area has not shown a benefit of extended participation on child outcomes (Ritblatt et al., 2001, Schweinhart and Weikart, 1989, Sprigle and Schaefer, 1985), while more recent studies have demonstrated positive effects on children's language and early literacy skills (Lee, 2011, Reynolds, 1995, Skibbe et al., 2011, Wen et al., 2012). In a meta-analysis of longitudinal studies of ECE programs, the average weighted effect size for longer and more intense programs was higher (d = .37–.44) compared to the effect size of programs that were shorter and less intense (d = .11–.18) (Nelson, Westhues, & MacLeod, 2003). Reynolds (1995) examined the effect of dosage within the Chicago Parent Center (CPC) program and found that those who attended two years of the classroom program demonstrated higher cognitive readiness scores made up of several scales of the Iowa Tests of Basic Skills (word analysis, mathematics) compared to those who attended for one year but the advantage was not maintained across the elementary grades.
Several recent large-scale studies of Head Start, including the impact study (Puma et al., 2010, Puma et al., 2006) and the Family and Child Experience Survey (FACES) study, were designed in a way that enabled researchers to explore dosage effects. Puma et al., 2006, Puma et al., 2010 and Lee (2011) found that outcomes were better for children enrolled at age 3 compared to those who enrolled at age 4. Randomized control trials are generally considered to be the gold standard design for determining the effect of an early intervention such as preschool or a program feature such as duration. This approach is often not feasible in the context of community programming (McCall & Green, 2004) and was not the case in these evaluations of Head Start. The advantage of randomization is that it eliminates sources of bias that undermine causal inference. This is important in the case of preschool dosage because factors that affect caregivers’ decisions to enroll their child at either age 3 or 4 may also account for differences in the pattern of child outcomes (McGowan, Nix, Murphy, Bierman, & Conduct Problems Prevention Research Group, 2010). Examples of confounding variables include parental attitudes (e.g., educational expectations) and demographic variables (i.e., income, work status) that may directly or indirectly influence parents’ decisions about both age of enrollment and children's functioning.
One statistical approach to managing selection bias that improves the ability to draw causal inference from non-randomized data is propensity score analysis (Rosenbaum & Rubin, 1983). A propensity score is a simple statistical tool that allows researchers to balance non-equivalent groups and adjust for confounding. In the case of preschool duration, the propensity score is each child's probability of having attended two years of preschool instead of one, conditional on a set of potential confounding variables.
There are a variety of methods for using propensity scores (Harder et al., 2010, Lanza et al., 2012Rosenbaum, 2002, Stuart, 2010). Regardless of the particular method, Lanza et al. (2012) described the process as involving four main analytic steps: (1) estimate propensity scores in a separate analysis, (2) use propensity scores to adjust for confounding variables by using one of the several techniques, i.e., inverse probability of Treatment Weighting (IPTW; Hirano & Imbens, 2001), matching (Stuart, 2010), and subclassification (Rosenbaum & Rubin, 1984), (3) assess balance to determine if there are mean differences between treatment and control among the covariates, and (4) estimate the treatment effect. In the case of matching, this involves comparing the new groups that are constructed. For other methods, it involves estimating the effect of interest and including the propensity score in the model to adjust for confounding.
There is consensus among experts that when estimating the propensity score (Step 1), all measured confounders that are potentially predictive of selection into the treatment groups and the outcome should be included in the propensity model as this improves the causal inference that can be applied to the findings (Rosenbaum, 2002). This means that measures of constructs theoretically related to selection and the outcome are also recommended for inclusion, regardless of statistically significant differences between the groups. There are several comparable techniques (e.g., IPTW, matching, subclassification) for adjusting for confounds when using propensity scores (Rosenbaum and Rubin, 1984, Rosenbaum and Rubin, 1985). Typically, once the adjustment is made to the data only the treatment indicator is included as a predictor in the outcome analyses but in the doubly robust method the propensity score is also included in the model (Schafer & Kang, 2008).
Several recent studies have applied propensity score analysis methods to examine the impact of preschool dosage on child functioning. In a small sample (N = 76) of children from predominantly middle- to upper-SES background, Skibbe et al. (2011), used HLM to determine whether program duration had a differential impact on kindergarten children's academic skills. A propensity score constructed from a set of demographic variables was included in the models. This study found that children who attended preschool for two years had significantly higher scores on measures of letter knowledge and decoding skills. Preschool dosage was not related to vocabulary levels or self-regulation skills. In a large sample (N = 1778) of low-income children drawn from the nationally representative Head Start Family and Children Experiences (FACES) Study (U.S. DHHS, 2008), Wen et al. (2012) recently used subclassification propensity score analysis to compare demographically matched samples of children who attended Head Start for one year or two years on a range of school readiness outcomes. In this study, the propensity score included a range of variables reflecting child characteristics, family demographics, parenting attitudes and behaviors, baseline child receptive language skill, child early intervention experience, and degree of Head Start services (Wen et al., 2012). The findings were that children who participated in Head Start for two years had better receptive vocabulary, emergent literacy, and math skills compared to the demographically matched sample of children with only one year of exposure.
The current study is similar to the work of Wen et al. (2012) in that it examines the effects of preschool dosage on the early academic functioning of children in kindergarten drawn from a low-income sample. Rather than using a subclassification technique to adjust for confounding, it applies the comparable technique of one-to-one matching to create a quasi-experimental evaluation of preschool dosage in a Head Start preschool program based within a public school system. The propensity score in the current study is similar to that used by Skibbe et al. (2011) in that the content of the items was primarily demographic but the number of variables included was comparable to the score created by Wen and colleagues. The current study extends all previous research on preschool dosage by examining how length of program exposure to an enhanced Head Start model impacts the functioning of participants. The main enhancements were the integration of evidence-based curricular components into standard Head Start practice. This approach has been identified as an important strategy for Head Start to use in order to strengthen child outcomes and one that is increasingly more common as the evidence-base for these types of intervention components expands (Griffin, 2010). The current study also assessed quality of programming in the participating classrooms, which has not been done in previous work, and holds constant an important factor that is predictive of child outcomes.
Children in the current study who entered the preschool program at age 3 and participated for two years were expected to perform better on measures of literacy and numeracy at kindergarten compared to those who entered the program at age 4 and participated for only one year. The fact that the preschool program was conducted in only one school district is considered a strength and a unique feature compared to similar studies because it reduces the random error associated with variability in program quality.
Section snippets
Participants
The study took place in the context of an evaluation of a public preschool program that was implemented by the Harrisburg City School District in Harrisburg, PA. In this district, approximately 90% of students are eligible for free or reduced lunch each year. The program served children aged 3 and 4 and was delivered in four of the district's 12 buildings (12 classrooms) during the 2002–2003 academic year and six buildings (26 classrooms) in the 2003–2004 academic year. In the first year of the
Results
MLM estimates, standard errors, and beta coefficients for each of the outcomes are presented in Table 2. The intercept represents the level of each outcome variable in kindergarten. A significant coefficient for years in program denotes differences between the one-year and two-year groups on the outcome variable at kindergarten. The one-year preschool group was the reference group against which the two-year preschool group was compared. Adjusted means for the one- and two-year groups and the
Discussion
This study examined the impact of duration of preschool experience on academic functioning at kindergarten by comparing the outcomes of children who participated for either one (entered at age 3) or two years (entered at age 4) in a public program that was enhanced in two ways. First, the program was an integrated model that resulted from a partnership between the public preschool and Head Start. As such, it met both the Head Start and the public school system's requirements. Second, the
Conclusions and implications
Despite these limitations, the study had several methodological strengths. Multilevel modeling was used to account for the nesting of children within classrooms. The study used age-standardized scores which controlled for variation in children's age at the time of testing and provided a benchmark against which to compare participants’ scores. Baseline PPVT-III scores were included for analyses examining the subscales of WJ-R as a way to control for baseline differences since the WJ-R cannot be
References (82)
- et al.
Are teachers’ education, major, and credentials related to classroom quality and children's academic gains in pre-kindergarten
Early Childhood Research Quarterly
(2006) Research on the implementation of preschool intervention programs: Learning by doing
Early Childhood Research Quarterly
(2010)- et al.
Ready to learn? Children's pre-academic achievement in pre-kindergarten programs
Early Childhood Research Quarterly
(2008) - et al.
The differential effects of preschool: Evidence from Virginia
Early Childhood Research Quarterly
(2012) - et al.
Does prekindergarten improve school preparation and performance?
Economics of Education Review
(2007) - et al.
A meta-analysis of the effects of early developmental prevention programs in at-risk populations on non-health outcomes in adolescence
Children and Youth Services Review
(2010) - et al.
The impact of kindergarten learning-related skills on academic trajectories at the end of elementary school
Early Childhood Research Quarterly
(2006) - et al.
Effects of web-mediated professional development resources on teacher–child interactions in pre-kindergarten classrooms
Early Childhood Research Quarterly
(2008) One year of preschool intervention or two: Does it matter?
Early Childhood Research Quarterly
(1995)- et al.
Schooling effects on preschoolers’ self-regulation, early literacy, and language growth
Early Childhood Research Quarterly
(2011)
Enhancing young children's mathematical knowledge through a pre-kindergarten mathematics intervention
Early Childhood Research Quarterly
Are two years better than one? A propensity score analysis of the impact of Head Start program duration on children's school performance in kindergarten
Early Childhood Research Quarterly
Benefits and costs of prevention and early intervention programs for youth
Universal and targeted approaches to preschool education in the United States
International Journal of Child Care and Education Policy
The High/Scope Perry Preschool Program: Cost–benefit analysis using data from the age—40 follow-up
Journal of Human Resources
Promoting academic and social-emotional school readiness: The Head Start REDI program
Child Development
Predicting child outcomes at the end of kindergarten from the quality of pre-kindergarten teacher–child interactions and instruction
Applied Developmental Science
Meta-analysis of the effects of early education interventions on cognitive and social development
Teachers College Record
Effects of early intervention on intellectual and academic achievement: A follow-up study of children from low-income families
Child Development
Child characteristics and family processes that predict behavioral readiness for school
The state of America's children 2010
Statistical power analysis for the behavioral sciences
Initial impact of the Fast Track Prevention Trial for conduct problems: II Classroom effects
Journal of Consulting and Clinical Psychology
Teacher–child interactions and children's achievement trajectories across kindergarten and first grade
Journal of Educational Psychology
Tutorial in biostatistics: Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group
Statistics in Medicine
Preschool program improves cognitive control
Science
Prefrontal cortex cognitive deficits in children treated early and continuously for PKU
Monographs of the Society for Research in Child Development
Fostering high-quality teaching in Head Start classrooms: Experimental evaluation of an integrated curriculum
American Education Research Journal
The preschool PATHS curriculum
Peabody Picture Vocabulary Test III
An experimental preschool program for culturally deprived children
Peabody Journal of Education
Propensity score techniques and the assessment of measured covariate balance to test causal associations in psychological research
Psychological Methods
Early Childhood Environment Rating Scale—Revised
Estimation of causal effects using propensity score weighting: An application to data on right heart catheterization
Health Services & Outcomes Research Methodology
Learning to read and write: Developmentally appropriate practices for young children
Young Children
Initial results from the 2005 National Household Education Survey (NHES) Early Childhood Program Participation Survey (National Center for Education Statistics 2006-075). U.S. Department of Education. Institute for Education Sciences
The school entry gap: Socioeconomic, family, and health factors associated with children's school readiness to learn
Early Education and Development
Demystifying double robustness: A comparison of alternative strategies for estimating a population mean from incomplete data
Statistical Science
Cited by (33)
The externalities of preschool attendees in middle school classes
2023, China Economic ReviewAn early numeracy intervention for first-graders at risk for mathematical learning difficulties
2021, Early Childhood Research QuarterlyOne or two years of participation: Is dosage of an enhanced publicly funded preschool program associated with the academic and executive function skills of low-income children in early elementary school?
2017, Early Childhood Research QuarterlyCitation Excerpt :In a different study that used data from the Chicago Longitudinal Study, Arteaga, Humpage, Reynolds, and Temple (2014) did account for selection using propensity score methods and found that although two years of participation in the Child-Parent Centers preschool did not have significant benefits on achievement test scores in elementary school, it did have significant effects on grade retention and special education placement as well as juvenile and adult crime. The current study investigates whether the academic benefits of two years versus one year of preschool participation that were evident for this sample at kindergarten (Domitrovich et al., 2013) are maintained in early elementary school and whether the positive effects of longer participation are also reflected on measures of children’s EF skills. Specifically, this study assesses whether children who received two years of preschool (two-year group) have better (1) academic achievement (vocabulary, reading, and math); (2) EF skills (set-shifting and working memory); and (3) school adjustment (i.e., reduced likelihood of being retained or receiving special education services) by second grade than those who attended preschool for only one year (one-year group).
The effects of universal state pre-kindergarten on the child care sector: The case of Florida's voluntary pre-kindergarten program
2016, Economics of Education ReviewEarly childhood education and child development in four countries in East Asia and the Pacific
2019, Early Childhood Research QuarterlyCitation Excerpt :However, other studies suggest that, in addition to improved academic outcomes, ECE intensity has also been linked with increased behavioral difficulties (e.g., Gibbs, 2014; Loeb et al., 2007; Votruba-Drzal et al., 2008; Walters, 2014). Consistent with prior research (Biedinger et al., 2008; Domitrovich et al., 2013; Skibbe et al., 2011), we found that ECE duration was positively associated with children’s Cognitive Development after controlling for potential confounders at the regional level (data from all four countries combined) and for China (when data from four countries were analyzed separately). There were also non-linear associations between ECE duration and child performance in Cognitive Development and Language and Emergent Literacy.
- ☆
This project was supported with funding from the W.K. Kellogg Foundation. The authors would like to thank the administration of the Harrisburg School District in place at the time of this project's initiation, Capital Area Head Start, and the teachers of the Harrisburg Preschool Program, who are devoted and skilled educators of young children. Dr. Domitrovich and Dr. Greenberg are authors of the PATHS Curriculum and have a royalty agreement with Channing-Bete, Inc. Dr. Domitrovich receives income from PATHS Training, LLC. This has been reviewed and managed by Penn State's Individual Conflict of Interest Committee.