Prekindergarten and Kindergarten Center-based Child Care and Students’ Early Schooling Outcomes
by Michael A. Gottfried - 2015
Background/Context: In the United States, there has been an increased trend in parents’ utilization of center-based child care. Yet, though research has examined the effects of attending prekindergarten center-based care or the effects of attending center-based care during the kindergarten school year, little is known about the effects of having attended both.
Purpose/Objective: This study asks three questions: (a) Do children who attend both prekindergarten and kindergarten center-based care have different achievement outcomes, measured at the end of kindergarten? (b) Do children who attend both prekindergarten and kindergarten center-based care have different socioemotional outcomes, measured at the end of kindergarten? (c) Do these relationships differ by individual socio-demographic characteristics?
Population/Participants: This study utilizes data from the newly released Early Childhood Longitudinal Study – Kindergarten Class of 2011 (ECLS-K:2011). The ECLS-K: 2011 represents the most contemporary national-level data available to study the educational experiences of young students in the United States. Information was first collected from kindergartners (as well as parents, teachers, and school administrators) from U.S. kindergarten programs in the year 2010–2011.
Research Design: This study combines secondary data analyses and quasi-experimental methods. There are two achievement outcomes: reading and math. There are five socioemotional outcomes: externalizing behaviors, internalizing behaviors, self control, approaches to learning, and interpersonal skills. The study begins with a baseline, linear regression model. To address issues pertaining to omitted variable bias, the study employs various fixed-effects models.
Findings: The findings for the first research question indicated that academic outcomes do not differ for children in both years of center-based care compared to children who attended only one year of center-based care or none at all. As for the second research question, the findings show that multiple years of center-based care is related to increases in problem behaviors and decreases in prosocial behaviors—outcomes that are worsened by the number of years of center-based care attendance. The findings for the third research question suggest some minor differences between boys and girls in zero, one, or two years of center-based care.
Conclusions/Recommendations: This study has brought to surface new ways that center-based care attendance might influence children’s short-term outcomes. Therefore, researchers, policymakers, and practitioners must base future questions on empirical work concerning how to address children’s outcomes across multiple years of care, rather than simply focus exclusively on one year’s influence.
Across the board, the United States is witnessing an increasingly larger number of families who are placing their young children into out-of-home daycare (Yamauchi & Leigh, 2011). As one major trend, during the year prior to school entry, a larger proportion of prekindergarten-aged children in the United States are cared for outside of the home and by non-parental adults (Blau & Currie, 2004; Smith, Kleiner, Parsad, Farris, & Green, 2003). Another major trend is that a majority of kindergarten students (i.e., school-aged children) are also now receiving out-of-home care either before or after the school day (Brimhall, Reaney, & West, 1999; Claessens, 2012). Many researchers, policymakers, and practitioners hold that out-of-home care has the potential to boost a childs academic growth and social development (e.g., Cross, 2009; Howes et al., 2008). Therefore, given this proclaimed potential benefit to children combined with an increased trend in the utilization of out-of-home care in both prekindergarten and kindergarten years, research has understandably been focusing on assessing the effectiveness of out-of-home care alternatives.
Most research on the efficacy of these out-of-home care alternatives has focused on the role of center-based care, and this study will follow in this line of inquiry as well; the reason for this focus is that center-based care serves as a tangible policy lever when it comes to developing interventions (as opposed to, for instance, relative-based care). Within the scope of evaluating the effects of center-based care on academic and developmental outcomes, most research focuses on center-based care during the year immediately before school entry (i.e., prekindergarten center-based care). Limited research exists on the effects of center-based care on academic and developmental outcomes during the kindergarten school yearwhich this study is defining as care that kindergarten students receive before or after school in that year.
As for the researched effects on academic outcomes, the body of research in this area generally supports the claim that prekindergarten-aged children benefit academically from some form of center-based prekindergarten care, particularly as a way to prepare for the skills required at school entry (Burger, 2010; Loeb, Fuller, Kagan, & Carrol, 2004; Loeb, Bridges, Bassok, Fuller, & Rumberger, 2007; Magnuson, Meyers, Ruhm, & Waldfogel, 2004; National Institute of Child Health and Human Development [NICHD], 2006). The findings from preschool interventions, like Perry Preschool, have exhibited similar positive results (Campbell, Ramey, Pungello, Miller-Johnson, & Sparling, 2002; Loeb et al., 2004; Reynolds & Temple, 1998). One key reason why this area of research consistently supports the effect of prekindergarten care on academic outcomes is because center-based care may be providing a more structured formal learning environment than home-based care might (Claessens, 2012). Additionally, it may be the case that center-based instructors are more academically qualified than those guardians in other care alternatives, such as a relative (Barnett, Carolan, Fitzgerald, & Squires, 2011). In regards to the role of center-based kindergarten care on academic outcomes, for young school-aged children (i.e., those in kindergarten), the scant research indicates no effect of center-based care during kindergarten on academic outcomes (Claessens, 2012). It might be that kindergarten already serves as a structured learning environment. Therefore, there may not be any additional learning gains from spending more time in a second structured learning environment either before or after the formal kindergarten school day.
As for the researched effects on socioemotional outcomes, the research has generally upheld that center-based care during prekindergarten reduces positive socioemotional outcomes and increases negative behavioral outcomes compared to home-based/parental prekindergarten care (Baker, Gruber, & Milligan, 2008; Belsky et al., 2007; Herbst & Tekin, 2010; Loeb et al., 2007; NICHD, 2006; Magnuson, Christopher, & Jane, 2007; Yamauchi & Leigh, 2011; see Votruba-Dzral, Coley, & Chase-Lansdale, 2004, for one exception). Again, in the little work that has been conducted around center-based care for school-aged children before or after school during kindergarten, the findings suggest that these children tend to have more problem behaviors and lower social skills compared to those children in other care arrangements during that first year of schooling (Claessens, 2012). No single account fully explains this fairly consistent negative relationship between center-based care and socioemotional outcomes, though many possible reasons have been put forth, including: increased stress associated with shifting between multiple formal and structured environments; the demands of having to continually adapt to new situations; frequent changes disrupting adultchild relationships; and the long separation from parents at an early age (Claessens, 2012; Fiese et al., 2002; Loeb et al., 2007; Morrisey, 2009; Youngblade, 2003).
In these assessments of center-based care, research has focused separately on two issues: either (1) on the effects of prekindergarten center-based care, or (2) on the effects of center-based care during the kindergarten year. However, these issues have been examined in silos: No study has examined the effects of being a child who attended both. Though prior research has highlighted that attending either kind of center-based care may have academic or socioemotional effects, no study has evaluated implications for children who attend both center-based care in prekindergarten and during kindergarten. This study will be one of the first to do so, by asking three key research questions:
Do children who attend both prekindergarten and kindergarten center-based care have different achievement outcomes, measured at the end of kindergarten?
Do children who attend both prekindergarten and kindergarten center-based care have different socioemotional outcomes, measured at the end of kindergarten?
Do these relationships differ by individual socio-demographic characteristics?
As for the first research question, the lack of prior empirical research on the effectiveness of having attended center-based care in both years makes it challenging to hypothesize about the direction of the effects of two years of attendance. Research shows that prekindergarten center-based care attendance is associated with higher academic outcomes, and yet research also shows that there are null effects on achievement from attending kindergarten center-based care. Thus, the hypothesis based on attending both years here remains non-directional.
The hypothesis stemming from the second research question is more straightforward. Across the evaluations of either prekindergarten center-based care or of kindergarten center-based care (again, no research has considered the effect of attending both), research predominantly supports negative effects of one year of attendance on socioemotional skills, as mentioned above. Thus, although no prior research study has examined the role that attending both years of care has on socioemotional outcomes, this study speculates that socioemotional outcomes will be even lower for students attending both prekindergarten and kindergarten center-based care. The negative socioemotional effects that have been established for one year of center-based care attendance certainly could intensify over multiple years of attendance. With multiple years of shifting between formal and structured environments, having to adapt to an even greater number of new situations, establishing multiple new relationships with adults, and having an even longer separation from parents, it is probable that developmental issues arise from a longer duration in center-based care (Claessens, 2012; Loeb et al., 2007). This would be particularly true for children in their second year of center-based care during kindergarten, who must first adapt to the prekindergarten setting prior to school entry and then must adapt to a kindergarten setting as well as a center-based care setting in that same year (Claessens, 2012). It could also be speculated that the capacity of the center itself may be linked to socioemotional skills: It might be that the care center itself is not developmentally sensitive, and thus, multiple years of attendance may reduce skill formation that children might have received through other care alternatives.
The third research question is critical to address, as prior research has found mixed effects of a single year of center-based care on outcomes. Of those studies focusing on prekindergarten care, some research supports the claim that center-based care arrangements increase boys behavioral issues (Crockenberg, 2003; Crockenberg & Litman, 1991; Desai, Chase-Lansdale, & Michael, 1989; Youngblade, 2003). Yet other research implies increased behavioral issues for girls in these settings (Morrissey, 2009). Claessens (2012) found differential effects by gender of center-based care in kindergarten. Another key demographic trait evaluated in the care literature is socioeconomic status. Some prior research has demonstrated the possibility of closing the achievement and developmental gaps for high-poverty children with center-based prekindergarten care in the year before entering kindergarten (Barnett, 1995; Burchinal, Campbell, Bryant, Wasik, & Ramey, 1997; Loeb et al., 2004; Magnuson et al., 2004). That said, other research has found positive gains for middle-class children who also attended center-based care in the year before kindergarten (Bridges, Fuller, Rumberger, & Tran, 2004), and thus the evidence that one type of prekindergarten care differentially affects one socioeconomic group over another remains mixed. As for center-based care during kindergarten, Claessens (2012) found no effect by SES. Therefore, while these key traits (among others) will be explored, the hypotheses for individual characteristics remain non-directional.
In sum, little is known about those children who have attended both prekindergarten and kindergarten center-based careprior studies have only inquired into the influence of one years attendance. Hence, in order to determine the correct care fit for young students, it is important for policy decision-making to be based on research that addresses the effectiveness of children who are attending multiple years, especially considering this trend in the United States, as mentioned previously. Thus, to shed light on this issue, this study relies on a newly released, large-scale dataset of kindergarten students in the 20102011 school year. Through these data, it is possible to ascertain child care experiences for children who have recently been in prekindergarten and kindergarten. Given that this is the most recent cohort of students for which these data are available and given the increasing percentages of children in out-of-home care, issues pertaining to the role of center-based care during multiple years could not be more pertinent.
This study utilizes data from the newly released Early Childhood Longitudinal Study Kindergarten Class of 2011 (ECLS-K:2011). Given the recent release of these data, the ECLS-K: 2011 represents the most contemporary national-level data available to study the educational experiences of young students in the United States. This dataset, developed by the National Center for Education Statistics (NCES), involves a large-scale survey design and assessment data collection of kindergarten students and their families, teachers, classrooms, and school. Students were in kindergarten in the academic year 20102011. The ECLS-K:2011 used a three-stage stratified sampling strategy, in which geographic region represented the first sampling unit, public and private school represented the second sampling unit, and students stratified by race/ethnicity represented the third sampling unit. Hence, observations in the dataset are from a diversity of school types, socioeconomic levels, and racial and ethnic backgrounds. Currently, the fall and spring survey waves from kindergarten are available. Therefore, this study makes use of all available data from the ECLS-K:2011 dataset.
There are approximately n = 11,250 kindergarten student observations available in this study. This analytical sample size is consistent with prior research using ELCS-K data (Magnuson, Ruhm, & Waldfogel, 2007). Note that sample sizes are rounded to the nearest tens digit due to restrictions on using these data. The analytic sample is limited to children who had non-missing information on outcomes, as consistent with prior research conducted using the earlier version of ECLS-K. That said, comparing children in the analytic sample versus those excluded did not generate any meaningfully significant differences, which is also consistent with prior work using ECLS-K (e.g., Gottfried, 2013; Claessens, 2012).
The means and standard deviations of all outcomes are presented at the top of Table 1. The outcomes are measured during the spring survey wave, with a baseline measure serving as an independent variable in the model. Outcomes can be grouped into two subsets. The first subset includes students scale scores on math and reading assessments. The math assessment consisted of questions on number sense, properties and operations, measurement, geometry and spatial sense, data analysis, statistics and probability, patterns, algebra, and functions. The reading assessment assessed print familiarity, letter recognition, and recognition of common words. The alpha reliability coefficient ranged from 0.92 to 0.99 (Tourangeau et al., 2012).
Table 1. Descriptive Statistics (n = 11,250)
The second subset of outcomes includes teachers ratings of childrens problem behaviors and social skills. The problem behaviors construct consists of the externalizing and internalizing scales. The five-item externalizing problem behaviors scale assesses the frequency with which a child argues, fights, gets angry, acts impulsively, and disturbs ongoing activities. The four-item internalizing problem behaviors scale measures the extent to which the child exhibits anxiety, loneliness, low self-esteem, and sadness.
The social skills construct is comprised of four scales: self-control, approaches to learning, interpersonal skills, and eagerness to learn. The four-item self-control scale measures the extent to which the child is able to control his or her temper, respect others property, accept his or her peers ideas, and handle peer pressure. The seven-item approaches-to-learning scale rates the frequency with which the child keeps his or her belongings organized, shows eagerness to learn new things, adapts to change, persists in completing tasks, pays attention, and follows classroom rules. The five-item interpersonal skills scale measures the frequency with which a child gets along with others, forms and maintains friendships, helps other children, shows sensitivity to the feelings of others, and expresses feelings, ideas, and opinions in positive ways.
The scales are on a four-point Likert-metric, with higher scores indicating that the child engaged in those behaviors more frequently. Thus, a higher score on the problem behavior scales indicates an unfavorable outcome, whereas a higher score on the social skills scales reflects a favorable outcome. All of these scales have high internal consistency, with the alpha reliability coefficients ranging from 0.79 to 0.89 (Tourangeau et al., 2012). Note that no item-level information is available in the dataset, even at the restricted-users level.
CENTER-BASED CARE ATTENDANCE
The key measures in this study designate whether a student had attended center-based care in the year before kindergarten and/or during kindergarten (i.e., outside of kindergarten time). In the fall survey administered to parents, they responded to questions pertaining to child care in the prior-to-kindergarten and kindergarten years. In total, there are several categories of child care on the survey: center-based care, Head Start care (prekindergarten only), relative care, non-relative care, and parental care.
This study focuses on the influence of center-based care attendance. Based on the parent survey questions, a student was identified as having attended center-based care if his or her parents indicated that their child went to any type of center care during the prekindergarten and/or kindergarten years, with the exception of Head Start. Designating Head Start as a separate prekindergarten category of care is consistent with prior research (Loeb et al., 2007).
Three binary indicators were created for the purposes of this evaluation. The first was whether a student attended center-based care in both prekindergarten and kindergarten years. This serves as the reference category in the analyses to follow. The second category indicated all students who attended only one year of center-based care. Of all the students included in this category, only 3% attended that one year of center-based care in kindergarten and not in the prekindergarten years. Therefore, having two separate measures indicating the precise designation of when this single year of center-based care attendance had actually occurred would be redundant with the only one year indicator. Finally, a binary variable was created to indicate if a student did not attend center-based care in prekindergarten nor kindergarten years.
This analysis employs a widely used set of socio-demographic student variables. Student measures include gender, race, English language learner (ELL) status, an indicator for attending full-day versus part-day kindergarten, age at kindergarten entry, and a continuous measure for body mass index. These student-level covariates have been found to be associated with achievement and social outcomes (Gottfried, 2013; Datar, 2006; Elder & Lubotsky, 2009; Wingfield, McNamara, Janicke, & Graziano, 2011).
The other care designations are utilized as control variables in the analysis, which is consistent with prior research (e.g., Claessens, 2012; Votruba-Dzral, Li-Grining, & Maldonado-Carreno, 2008). Parental care serves as the reference group. Additionally, the number of hours of non-parental care in prekindergarten and in kindergarten years were utilized as control variables in this study, as seen in the latter portion of Table 1.
Family measures include an NCES-constructed measure of SES, which is based on parental educational attainment, occupational status, and family income as reported by parents. Additional measures include an indicator for whether a childs parents are married, a continuous measure for the number of siblings, and age of mother when she first gave birth to any child.
Additionally, two other family measures are included, which were replicated from the Votruba-Drzal et al. (2008) study. Both scales assessed the childs home environment, activities, and cognitive stimulation. The first scale, relating to parental involvement, was measured on a four-point Likert scale. The 10-item parental involvement scale assessed the frequency with which parents engaged the child in various activities, such as playing games, singing songs, reading books, and doing arts and crafts. The second scale, which was comprised of 15 dichotomously scored items, measured the number of learning activities in which children participated. This scale assessed whether in the past month, the child engaged in activities such as visiting a bookstore, taking music lessons, or attending tutoring lessons.
Additional Child-Care Measures
Aside from center-based care, the other categories of child-care measures are included as control variables. These are denoted by binary indicators for the type of care received in the year prior to kindergarten entry, and the type of care received while simultaneously enrolled in kindergarten. The models account for the fact that students can be enrolled in multiple types of care. Additionally, the models include a measure for the number of hours of care received.
The main model utilized to assess the relationship between center-based care attendance in prekindergarten and kindergarten years and achievement/socioemotional outcomes is presented as follows:
YiSK = β0 + β1OiK + β2NiK + β2SiK + β3FiK + β4CiK + εiK
In the model, Y represents an achievement or socioemotional outcome for student i, where the outcome was assessed during the spring kindergarten wave of data collection, as described previously. The center-based care variables are represented as: O, indicating having only attended one year of center-based care, as denoted on the fall kindergarten parent survey; and N, indicating that on the fall kindergarten parent survey, it was denoted that the child never attended any center-based care, not in the prekindergarten year nor in the kindergarten year. SiK represents individual student characteristics measured in kindergarten, FiK represents family characteristics measured in kindergarten, and CiK represents parents indication on the kindergarten survey wave of their childs additional prekindergarten and kindergarten care arrangements and the associated number of hours per week that a child was in care during those years.
It is in the error term where the nested structure of the data is taken into account, which is consistent with prior work in center-based care on large-scale datasets (e.g., Claessens, 2012). Empirically, εiK is estimated with robust standard errors, adjusted for classroom clustering (see, e.g., Woolridge, 2002). Adjusting for classroom clustering is key when using the ECLS-K:2011 dataset notably because NCES collected data for more than one student per kindergarten classroom. Therefore, since students are nested in schools by classroom and hence share common but unobservable characteristics and experiences, clustering student data at the classroom level provides for a corrected error term given this non-independence of individual-level observations when assessing outcomes in the kindergarten year.
Tests of Validity
The main model in this analysis has certainly taken into account a rich set of control variables to improve the precision of the estimates of center-based care over multiple years. That said, while cluster-adjusted robust standard errors may account for the nested structure of the data and therefore the unobserved experiences of students in the dataset (Rogers, 1993; White, 1980; Woolridge, 2002), it is possible that additional unobserved factors are biasing the estimates. It might be the case that school-level factors are influencing the estimates of center-based care. For example, certain types of schools might contain highly involved parents. Highly involved parents may have chosen in-home pre-kindergarten care and in-home (afterschool kindergarten) care during the kindergarten year. At the same time, highly involved parents might also be making additional investments that would boost their childrens achievement and socioemotional outcomes. If it were common for schools to contain similar types of parents, then the estimate of β1 and β2 would be systematically biased. Or, it might be the case that certain school administrators are going out into the community urging parents to send their children to center-based care in both prekindergarten and kindergarten years in order to prepare them academically for school entry. This same principal might be making investments in the school itself to ensure students have high achievement levels.
To address these issues, the main model in this study is augmented to employ school fixed effects:
YiSK = β0 + β1OiK + β2NiK + β2SiK + β3FiK + β4CiK + δj + εiK
where δj are school fixed effects for school j. Technically, the term δj is a set of binary variables that indicates whether a student had attended a particular school (for each school variable in the dataset, 1 indicates yes, and 0 indicates no). This set of school indicator variables leaves out one school as the reference group (this process is analogous to creating indicator variables for race, where one racial category is left out as the reference group).
Gottfried (2012) demonstrates that school fixed effects address an important issue related to this study. Because similar types of families often attend the same school (Gottfried, 2012)as delineated by SES, unmeasured tastes, involvement, etc.school fixed effects account for biases that might arise due to these common but unobserved family factors. If students are in a school that attracts highly involved parents, then a school fixed effects model controls factors like high parental involvement for each student within that school. That is, school fixed effects control for unobserved aspects that may be shared among all students in that school.
Two variants on this model are also tested. First, it might be that there are unobserved county-level factors influencing attendance of center-based care in either prekindergarten or kindergarten years. These factors might include, for instance, the availability of services (Cannon, Jacknowitz & Painter, 2011). Hence, county fixed effects are employed in a model analogous to the one above. Consequently, this model holds constant all county-by-county variation, and the identification of variance in center-based attendance in prekindergarten and kindergarten years occurs within each specific county. Indeed, using county fixed effects has been supported in the literature as effective in accounting for the care and educational options that parents have available to them (Cannon et al., 2011).
Second, a state fixed effects model is employed, which represents a series of indicator variables for states. In this case, the model is accounting for all state-by-state variation, thereby controlling for unobserved state-level policies and practices (e.g., requiring the offering of center-based care in prekindergarten and kindergarten years).
Table 2 presents coefficient and standard error estimates from the main regression model, as described in the previous section. Across achievement and socioemotional outcomes, coefficients are presented for each independent variable along with the cluster-adjusted standard errors in parentheses. Notably, the key measures in this study are found in the first two rows of resultsbinary indicators for whether students had attended only one year of center-based care or neither years of center-based care. The omitted reference category includes students who attended center-based care during both prekindergarten and kindergarten years.
Table 2. The Effects of Years of Center-Based Care on Achievement and Socioemotional Skills
Note. *** p < 0.01, ** p < 0.05, * p < 0.10. Robust standard errors adjusted for clustering in parentheses.
Math and reading outcomes are presented in the first two columns of the table. The evidence suggests that the number of years in center-based care does not predict differences in achievement for students in the sample. This is fairly consistent across both center-based measures and both math and reading outcomes. Overall, then, the results for achievement suggest that there is no difference in math or reading outcomes based on the number of years a child has spent in center-based care. That is, there is no added benefit or loss to academic achievement that is differentiated by whether a student spent zero, one, or two years in center-based care during the prekindergarten and kindergarten years.
The findings, however, are more nuanced when considering the set of socioemotional outcomes. The two columns after achievement examine problem behaviors (where negative coefficients reflect more favorable outcomes). Children attending center-based care during both prekindergarten and kindergarten years (the omitted reference category) have higher externalizing problems compared to all other children in the samplei.e., those in one or zero years of center-based care. Importantly, these results remain statistically significant even after controlling for a wide range of independent measures, including child and family background, as well as all other types of care.
To put these results into perspective, the standardized beta coefficient is utilized, where both one year of attendance and neither years of attendance coefficients have a beta of -0.08σ and -0.09σ, respectively. The largest difference in outcomes occurs between those in both years of center-based care versus those who did not attend center-based care in either year; this is followed by a similar, but slightly smaller difference between those in two versus one year of center-based care. Therefore, students who attended zero years of center-based care have the lowest frequencies of externalizing problem behaviors, followed by students in only one year of center-based care. Those in the omitted center-based care category of both years of attendance had the highest frequencies of externalizing problem behaviors in the sample. Note that the effect sizes in this study are consistent with the sizes reported in previous studies (e.g., Claessens, 2012; Lee, Zhai, Brooks-Gunn, Han, & Waldfogel, 2013; Loeb et al., 2007), thereby suggesting that the empirical specification in the present study is consistent with the literature. Also consistent with prior research (e.g., Claessens, 2012) is the negative association between center-based care and externalizing problems, though not with internalizing problems.
The final three columns constitute social skills (where positive coefficients reflect more favorable outcomes). Consistent with externalizing problems, children who attended center-based care during both prekindergarten and kindergarten years have lower frequencies of self-control compared to all other students. Moreover, based on the effect sizes, children who never attended center-based care (effect size: 0.12σ) have the highest frequencies of self-control out of all students in the sample. Displaying the second-highest frequencies of self-control are those who only attended one year of center-based care (effect size: 0.07σ), with again the poorest outcomes occurring among those who attended center-based care in both prekindergarten and kindergarten years (the omitted reference group).
A similar interpretation exists for the approaches-to-learning scale. Children who attended center-based care during both prekindergarten and kindergarten years have lower frequencies of approaches to learning compared to those students who only attended center-based care in one year, and compared to those students who never attended center-based care at all. As with self-control, the effect sizes are largest for students who never attended center-based care (0.10σ), whereas the effect size is 0.06σ for one year of attendance.
Finally, there is evidence that the interpersonal skills are lower for those students who attended center-based care during both prekindergarten and kindergarten years. The effect sizes are fairly similar, though slightly larger for those students who never attended center-based care. Therefore, again, the greatest difference in outcomes occurs for students who never attended center-based care versus those who attended in both years, suggesting that the largest difference exists between no attendance and two years of attendance. Regardless, however, the evidence suggests that students who did attend center-based care in both years have lower frequencies in these social outcomes compared to any other student in the sample. These results are consistent with prior work that has examined center-based cares effect on socioemotional functioning (e.g., Claessens, 2012; Loeb et al., 2007).
Briefly turning to the wide span of control variables implemented in this study, the findings are consistent with prior research. First, higher baseline scores predict higher spring scores. The results are also differentiated by gender and race in commonly found patterns, particularly for the achievement outcomes. Kindergarten entry age has a small but significant effect, which is consistent with Elder and Lubotsky (2008). While having attended full-day kindergarten might be related to several of the outcomes in patterns that are consistent with prior research (DeCicca, 2007; Votruba-Dzral et al., 2008), the interaction (not shown) between full-day kindergarten attendance and the center-based care indicators was not statistically significant. Therefore, there is no evidence that the main findings are differentiated by type of kindergarten attended.
Generally speaking, higher family socioeconomic status, married parents, and higher parental involvement predict better student outcomes. Other, non-center-based child care characteristics in prekindergarten and kindergarten are statistically significant in some models, but not in a consistent way. Finally, the coefficients on weekly hours of care might be statistically significant; the values approximate 0.0 in almost all models for both prekindergarten and kindergarten hours. Thus, even though there may be statistical significance, the effect sizes of approximately 0.0 indicate that practical significance is quite limited. Note that, although not presented, the interactions between weekly hours of care and the center-based care indicators do not provide significant findings. This is consistent with prior research examining similar interactions with hours of care (e.g., Claessens, 2012).
TESTS OF VALIDITY
As described above, in order to account for unobserved influences that may be affecting both the key center-based indicators and the set of outcome variables, a series of fixed-effects models are examined. The first panel of Table 3 presents the coefficients and robust clustered standard errors from the main models in Table 2. The second panel presents coefficients and standard errors from school fixed effects models. The third panel presents county fixed effects and the fourth panel presents state fixed effects. Though not presented for the sake of clarity, all control variables are included.
Table 3. Tests of Validity
Note. *** p < 0.01, ** p < 0.05, * p < 0.10. Robust standard errors adjusted for clustering in parentheses.
Examining all of the panels in the table, the estimates suggest a great deal of consistency between all estimation approaches. There remains a lack of statistical significance on the center-based care coefficients in achievement columns, thereby suggesting that any differences in the number of years in center-based care does not significantly relate to differences in achievement. The results pertaining to externalizing behaviors, self-control, and approaches to learning also remain statistically significant in similar patterns. As before, the results generally present the largest effect size for having never attended center-based care in either year, though having attending one year of center-based care is nonetheless related to displaying a greater frequency of positive outcomes compared to having attended both years of center-based care. There is no longer consistent evidence pertaining to the model predicting interpersonal skills. That said, given that the center-based care estimates from Table 2 were significantly only at p < 0.10, this result is not surprising once having restricted the variance to within-school, within-county, or within-state variation.
In the end, the interpretation is similar to the findings from Table 2. It is still the case that for students in the dataset, being in center-based care for both years is related to reduced socioemotional outcomes and does not provide any improvement in achievement.
HETEROGENEITY IN RESULTS
Prior research has established that gender and/or socioeconomic status might moderate the effects of child care (Claessens, 2012; Crockenberg & Litman, 1991; Desai et al., 1989; Youngblade, 2003). To test the moderating effects of these characteristics, Table 4 presents the interactions between gender and family socioeconomic status. Analogous to Table 3, the first panel of results presents the child-care measures from Table 2. The interactions between these measures and gender are presented next, followed by the interaction between these measures and family socioeconomic status. The models for the interactions are based on the main model specification, though the results are consistent to more complex, fixed-effects specifications, and are available upon request.
Table 4. Heterogeneity in Effects
Note. *** p < 0.01, ** p < 0.05, * p < 0.10. Robust standard errors adjusted for clustering in parentheses.
Turning first to the coefficients for the interaction between gender and the center-based care measures, there are no moderating effects for achievement, regardless of gender. Consistent with prior findings in this study, no student in the sample has differential achievement outcomes based on having attended zero, one, or two years of center-based child care. The models predicting socioemotional outcomes suggest some differences between boys and girls in zero, one, or two years of center-based care, though the evidence is not consistent in the table. Externalizing behaviors might be one exception: Boys in two years of center-based care tend to have higher externalizing problem behaviors than do girls in two years of center-based care.
The final panel examines the interaction between the two center-based care measures and family socioeconomic status. There does not appear to be any systematic pattern across socioeconomic status. Note that other socio-demographic measures were also tested, including race and ELL status. That said, the interaction terms were not statistically significant in any model.
When it comes to evaluating the role of center-based care, prior research has examined one of the two following issues: first, whether attending prekindergarten center-based care influences a students academic or socioemotional outcomes; or second, whether attending kindergarten center-based care (before or after school) has any effect on these outcomes. This study contributed new evidence by examining the outcomes of children who attend both. That is, while previous studies inquired into an isolated effect of going to center-based care in a single year, the present research provided new insight into the potential for an effect based on a childs multiple years of center-based care attendance. This new perspective is critical, given the growing national trend of children attending both prekindergarten and kindergarten center-based care, as mentioned in the introduction.
To examine this multiple-year effect, this study relied on the most recent national dataset of kindergarten childrenfrom the 20102011 school year, which thus reflects the most recent trends in center-based care attendance in a national dataset. Relevant to this study, this dataset was unique in that for each student observation, it was possible to link the years of center-based care attendance to a range of academic and socioemotional outcomes, as well as to other key contextual information pertaining to individual- and family-level sociodemographics.
This study addressed three research questions. The findings for the first research question indicated that academic outcomes do not differ for children who attended both years of center-based care compared to children who attended only one year of center-based care or none at all. These null findings on school-aged academic outcomes are consistent with prior research (e.g., Claessens, 2012), which in general has not found that center-based care in kindergarten affects end-of-kindergarten achievement. While it may be the case, then, that prekindergarten care may yield academic benefits at school entry (Burger, 2010; Loeb et al., 2004; Loeb et al., 2007; Magnuson et al., 2004; National Institute of Child Health and Human Development [NICHD], 2006), there does not appear to be any differential benefit to the outcomes measured at the end of kindergarten (a) in having attended any number of years of center-based care, or (b) for those school-aged children who are already in kindergarten and are thereby being exposed to daily formal learning activities in kindergarten itself.
As for the second research question, the findings in this study were in the direction hypothesized and consistent with previous literature, which has shown that center-based care is related to increases in problem behaviors and decreases in prosocial behaviors (Baker et al., 2008; Belsky et al., 2007; Herbst & Tekin, 2010; Loeb et al., 2007; NICHD, 2006; Magnuson et al., 2007; Yamauchi & Leigh, 2011). Interestingly, for the socioemotional outcomes examined in this study, the findings are differentiated by the number of years in center-based care (as opposed to a binary relationship: having attended vs. not having attended). The results indicated that that there is a differential in this relationship for those students who attended both years of center-based care versus those who only attended one year. The students with the largest frequencies of positive socioemotional behaviors were those in zero years of center-based care, followed by the second-largest frequencies for those in only one year of center-based care; those in both years of center-based care had the lowest frequencies of positive behaviors. As explained in the introduction of this article, there is no one set explanation guiding the relationship between center-based care and behavioral outcomes. That said, center-based care involves changes in settings and new interactions, and potentially greater stress as a result; therefore, the negative behavioral outcomes found in previous works examination of a single year of center-based care may be exacerbated over two years of contextual changes and new interactions.
The final research question inquired into the heterogeneity in results. Prior research in the evaluation of prekindergarten or kindergarten center-based care has indicated potential differentials by gender or socioeconomic status (Claessens, 2012; Crockenberg & Litman, 1991; Desai et al., 1989; Youngblade, 2003). The results for childrens socioemotional outcomes suggest some minor differences between boys and girls in zero, one, or two years of center-based care. For example, boys in two years of center-based care tend to have higher externalizing problem behaviors compared to girls in two years of center-based care.
Like Claessens (2012), the results were not differentiated by family socioeconomic status. This null finding, however, is a potentially important result. One limitation of the dataset is that the quality of centers is not identified. Therefore, there might be a wide variation in the types of programs offered in both prekindergarten and kindergarten center-based care options. That said, one speculation is that higher SES families might be sending their children to higher quality center-based care options based on a number of distinguishing factors. One might also speculate that higher-quality center-based care programs promote academic and socioemotional skills and can employ people with the training and experience to do so. Thus, exploring the differentiation in results by SES is a step in examining how the effects might differ based on subsamples of students attending care centers of potentially differing qualitythat is, any differentiation in quality might be reflected in differential results based on students SES. The results, however, were not distinguishable by SES (hence the null findings), thereby implicating that at least this proxy for quality may not be accounting for the main set of results. As mentioned below, however, further research with different data might delve into measures of quality.
While other moderating effects were tested, such as whether the relationships differ for students in full- versus part-day kindergarten classes, by hours of care, and by other sociodemographic measures, the results were fairly consistent across all children in the sample.
Given the findings in this study, there are several concluding implications. First, the fact that students in multiple years of center-based care had the lowest socioemotional outcomes implies that perhaps additional attention needs to be paid to the social development of children who are in these care settings for multiple years (as opposed to just one single year). While this is certainly a point to raise for children who attended just one year of center-based care (see, e.g., Love et al., 2003; McCartney et al., 2010), the findings in this study suggest that there is an effect of multiple years of attendance, specifically on socioemotional outcomes. Second, and related, given these significant findings of the joint influence of having attended both prekindergarten and kindergarten center-based care, researchers, policymakers, and practitioners must base future questions on empirical work that addresses how to improve schooling outcomes across multiple years, rather than simply focus exclusively on one years influence. Finally, the fact that some results were differentiated by gender indicates that additional attention might be paid to the adjustment of boys and their relationship to the interplay between multiple care settings, including school. Boys in center-based care in prekindergarten and kindergarten years may require additional supports and resources to ensure success.
There are some limitations to the present study. First, while the study relied on a recently released large-scale national dataset of young children, the caveat with any survey dataset is that many of the responses were self-reported, such as parent-reported child-care arrangements. Additionally, the socioemotional outcome scales are based on teachers survey responses, and as such, there may be some degree of subjectivity in the ratings (see, e.g., DiPerna, Lei, & Reid, 2007; Galindo & Fuller, 2010). Therefore, one must take this into account when interpreting the results. Second, while precise characteristics of quality are important to consider when evaluating child-care arrangements (Anders et al., 2011; Camilli, Vargas, Ryan, & Barnett, 2010; Vandell et al., 2010), measures of quality are not contained in the ECLS-K:2011 dataset. Therefore, while this present study has shed new light on the aggregate pattern between center-based care and childrens outcome, it also calls for additional research (using alternative sources of data) to delve into the role of quality in moderating the effects in this research. Finally, though not necessarily a limitation in its own right, the methods presented in this study are quantitative. Therefore, a follow-on study may also employ a qualitative approach as a way to derive more detail on how the findings described in this study are actualized. Assessing qualitative data from both families and teachers in conjunction with the quantitative findings from the current study will lead to an even greater understanding of the mechanisms driving the influence of center-based care.
Nonetheless, this study has contributed new research into the relationship between out-of-home care and child attainment. It has done so by examining not simply the individual role of prekindergarten center-based care or of kindergarten center-based care, but rather how childrens academic and socioemotional outcomes differ by having attended both. Using a newly released national dataset of children is critical, given that families are increasingly utilizing center-based care both before school and during the first year of schooling. Thus, the results of this study reflect current trends and patterns. The findings have indicated differential relationships to child outcomes depending on the number of years in child-based care. Given these findings, this study serves to spur additional research and dialog as to how to develop supportive policies and practices in these early educational years.
Anders, Y., Sammons, P., Taggart, B., Sylva, K., Melhuish, E., & Siraj-Blatchford, I. (2011). The influence of child, family, home factors and pre-school education on the identification of special needs at age 10. British Educational Research Journal, 37, 421441.
Baker, M., Gruber, J., & Milligan, K. (2008). Universal childcare, maternal labor supply and family well-being. Journal of Political Economy, 116, 709745.
Barnett, S. (1995). Long-term effects of early childhood programs on cognitive and school outcomes. The Future of Children, 5, 2550.
Barnett, W. S., Carolan, M. E., Fitzgerald, J., & Squires, J. H. (2011). The state of preschool 2011. Brunswick, NY: National Institute for Early Education Research (NIEER).
Belsky, J., Vandell, D. L., Burchinal, M., Clarke-Stewart, K. A., McCartney, K., Owen, M. T., & NICHD Early Child Care Research Network. (2007). Are there long-term effects of early child care? Child Development, 78, 681701.
Blau, D., & Currie, J. (2004). Preschool, day care and afterschool care: Whos minding the kids? (NBER Working Paper #10670).
Bridges, M., Fuller, B. C., Rumberger, R., & Tran, L. (2004). Preschool for Californias children: Promising benefits. PACE Policy Brief, 04-3. Berkeley, CA: Policy Analysis for California Education (PACE).
Brimhall, D. W., Reaney, L. M., & West, J. (1999). Participation of kindergartners through third-graders in before- and after-school care. Washington, DC: National Center for Education Statistics.
Burchinal, M. R., Campbell, F. A., Bryant, D. B., Wasik, B. H., & Ramey, C. T. (1997). Early intervention and mediating processes in cognitive performance of children of low-income African-American families. Child Development, 68, 935954.
Burger, K. (2010). How does early childhood care and education affect cognitive development? An international review of the effects of early interventions for children from different social backgrounds. Early Childhood Research Quarterly, 25, 140165.
Camilli, G., Vargas, S., Ryan, S., & Barnett, W. S. (2010). Meta-analysis of the effects of early education interventions on cognitive and social development. Teachers College Record, 112(3), 579620.
Campbell, F. A., Ramey, C. T., Pungello, E. P., Miller-Johnson, S., & Sparling, J. J. (2002). Early childhood education: Young adult outcomes from the Abecedarian Project. Applied Developmental Science, 6, 4257.
Cannon, J. S., Jacknowitz, A., & Painter, G. (2011). The effect of attending full-day kindergarten on English learner students. Journal of Policy Analysis and Management, 30(2), 287309.
Claessens, A. (2012). Kindergarten child care experiences and child achievement and socioemotional skills. Early Childhood Research Quarterly, 27, 265376.
Crockenberg, S. C. (2003). Rescuing the baby from the bathwater: How gender and temperament (may) influence how child care affects child development. Child Development, 74, 10341038.
Crockenberg, S., & Litman, C. (1991). Effects of maternal employment on two-year-old child behavior. Child Development, 62, 930953.
Datar, A. (2006). Does delaying kindergarten entrance given children a head start? Economics of Education Review, 25, 4362.
DeCicca, P. (2007). Does full-day kindergarten matter? Evidence from the first two years of schooling. Economics of Education Review, 26(1), 6782.
Desai, S., Chase-Lansdale, L., & Michael, R. (1989). Mother or market? Effects of maternal employment on intellectual development of four-year-old children. Demography, 26, 545562.
DiPerna, J. C., Lei, P. W., & Reid, E. E. (2007). Kindergarten predictors of mathematical growth in the primary grades: An investigation using the Early Childhood Longitudinal Study Kindergarten Cohort. Journal of Educational Psychology, 99, 369379.
Elder, T. E., & Lubotsky, D. H. (2009). Kindergarten entrance age and childrens achievement: Impacts of state policies, family background, and peers. Journal of Human Resources, 44(3), 641683.
Fiese, B., Tomoko, T., Douglas, M., Josephs, K., Poltrock, S., & Baker, T. (2002). A review of 50 years of research on family rituals: A cause for celebration? Journal of Family Psychology, 16, 381390.
Galindo, C., & Fuller, B. (2010). The social competence of Latino kindergartners and growth in mathematical understanding. Developmental Psychology, 46, 579592.
Herbst, C. M., & Tekin, E. (2010). Child care subsidies and child development. Economics of Education Review, 29, 618638.
Lee, R., Zhai, F., Brooks-Gunn, J., Han, W., & Waldfogel, J. (2013). Head Start participation and school readiness: Evidence from the Early Childhood Longitudinal Study Birth Cohort. Developmental Psychology. doi:10.1036/a0032280.
Loeb, S., Bridges, M., Bassok, D., Fuller, B., Rumberger, R. W. (2007). How much is too much? The influence of preschool centers on childrens social and cognitive development. Economics of Education Review, 26, 5266.
Loeb, S., Fuller, B., Kagan, S. L., & Carrol, B. (2004). Child care in poor communities: Early learning effects of type, quality, and stability. Child Development, 75, 4765.
Love, J. M., Harrison, L., Sagi-Schwartz, A., van IJzendoorn, M. S., Ross, C., Ungerer, J. A., . . . Chazan-Cohen, R. (2003). Child care quality matters: How conclusions may vary with context. Child Development, 74, 10211033.
Magnuson, K. A., Christopher, R., & Jane, W. (2007). Does prekindergarten improve school preparation and performance? Economics of Education Review, 26, 3351.
Magnuson, K. A., Meyers, M., Ruhm, C. J., & Waldfogel, J. (2004). Inequality in preschool education and school readiness. American Educational Research Journal, 41, 115157.
McCartney, K., Burchinal, M., Clarke-Stewart, A., Bub, K., Owen, M., Belsky, J., & the NICHD Early Child Care Research Network. (2010). Testing a series of causal propositions relating time in child care to externalizing behavior. Developmental Psychology, 46, 117.
Morrissey, T. (2009). Multiple child-care arrangements and young childrens behavioral outcomes. Child Development, 80, 5976.
National Institute of Child Health and Human Development Early Child Care Research Network. (2006). Child-care effect sizes for the NICHD study of early child care and youth development. American Psychologist, 61, 99116.
Reynolds, A. J., & Temple, J. A. (1998). Extended early childhood intervention and school achievement: Age 13 findings from the Chicago Longitudinal Study. Child Development, 69, 231246.
Rogers, W. H. (1993). Regression standard errors in clustered samples. Stata Technical Bulletin, 13, 1923.
Smith, T., Kleiner, A., Parsad, B., Farris, E., & Green, B. (2003). Prekindergarten in the U.S. public schools. Washington, DC: National Academy Press.
Tourangeau, K., Nord, C., Le, T., Sorongon, A. G., Hagedorn, M. C., Daly, P., & Najarian, M. (2012). Early childhood longitudinal study, kindergarten class of 201011 (ECLS- K:2011), Users manual for the ECLS-K:2011 kindergarten data file and electronic codebook (NCES 2013-061). U.S. Department of Education. Washington, DC: National Center for Education Statistics.
Vandell, D., Belsky, J., Burchinal, M., Steinberg, L., Vandergrift, N., & NICHD Early Child Care Research Network. (2010). Do effects of early child care extend to age 15 years? Results from the NICH study of early child care and youth development. Child Development, 81, 737756.
Votruba-Drzal, E., Coley, R. L., & Chase-Lansdale, P. L. (2004). Child care and low-income childrens development: Direct and moderated effects. Child Development, 75, 296312.
Votruba-Drzal, E., Li-Grining, C. P., & Maldonado-Carreno, C. (2008). A developmental perspective on full- versus part-day kindergarten and childrens academic trajectories through fifth grade. Child Development, 79(4), 957978.
Wingfield, R. J., McNamara, J. P. H., Janicke, D. M., & Graziano, P. A. (2011). Is there a relationship between body mass index, fitness, and academic performance? Mixed results from students in a southeastern United States elementary school. Current Issues in Education, 14(2), 112.
White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica, 48, 817830.
Woolridge, J. M. (2002). Economic analysis of cross section and panel data. Cambridge, MA: MIT Press.
Yamauchi, C., & Leigh, A. (2011). Which children benefit from non-parental care? Economics of Education Review, 30, 14681490.
Youngblade, L. (2003). Peer and teacher ratings of third- and fourth-grade childrens social behavior as a function of early maternal employment. Journal of Child Psychology and Psychiatry, 44, 477488.