Elsevier

Early Childhood Research Quarterly

Volume 53, 4th Quarter 2020, Pages 551-570
Early Childhood Research Quarterly

Exploring the role of quality in a population study of early education and care

https://doi.org/10.1016/j.ecresq.2020.06.005Get rights and content

Highlights

  • We explore micro-features of quality across early education and care in Massachusetts.

  • Children in middle-income families are the least likely to attend formal settings.

  • Public school prekindergarten show relative strengths on some indicators.

  • Quality measures did not explain differences in child skills across settings.

Abstract

This paper uses rich, representative data from the Early Learning Study at Harvard (ELS@H) to examine the landscape of early education and care among three- and four-year-olds in one state. We examine the distribution of children in setting types, taking socio-demographic characteristics of children and families into account, investigate how micro-features of quality differ across early education and care setting types (i.e., formal and informal settings), and explore whether variation in micro-features of quality is associated with children's language, literacy, math, executive functions, and social-emotional skills across the setting types. To address these aims, we employ measures of quality that capture detailed information about adult–child interactions and the nature of daily activities. Overall, we found a slight majority of three- and four-year-olds in the state were enrolled in formal settings, with fewer children enrolled in informal settings or parental care only. We observed different patterns of enrollment in formal and informal care settings based on child and family characteristics; younger children were less likely than older children to attend formal care settings, and middle-income families were less likely to use formal care relative to lower-income and higher-income families. Moderate differences in quality across setting types were observed. Children's skills also varied across setting types. However, controlling quality features did not change these patterns, and child and family characteristics accounted for much of the variation in child skills between setting types.

Introduction

To date, research on early education and care has tended to focus on a single early education or care type, most typically formal, center-based programs (e.g., Howes et al., 2008, Keys et al., 2013). Indeed, children today have relatively more opportunities than their predecessors to access formal early education and care programs. In 2016, the Current Population Survey (CPS) estimated that 42% of three-year-olds and 66% of four-year-olds in the United States were enrolled in early education and care programs (U.S. Department of Commerce, Census Bureau, 2017). Of these, 59% were enrolled in a public program and 41% were enrolled in some form of private program. This shifting landscape in part reflects a focus on expanding publicly funded programs. For example, from 2002 to 2017, the proportion of four-year-olds in state-funded public prekindergarten programs increased from 15% to 33% (Barnett, Hustedt, Robin, & Schulman, 2003; Friedman-Krauss et al., 2018). And in some cases, entire cities (e.g., New York) have rolled out public early education and care programs, seeking broader reach and scale than ever before (CityHealth & NIEER, 2019).

Yet, despite the increasing availability of formal, publicly funded early education and care programs, families with young children nevertheless continue to rely on a diverse constellation of education and care options. Many families rely on what is referred to as informal care (e.g., family child care or unlicensed relative care)—but these settings are rarely included in research. Studies that do examine informal programs typically rely on data from the Early Childhood Longitudinal Study, Birth Cohort (ECLS-B; Abner, Gordon, Kaestner, & Korenman, 2013; Bassok, Fitzpatrick, Greenberg, & Loeb, 2016), a dataset that is now over a decade old. There is therefore a need for new research that considers the realities of all of the settings in which young children spend their time, from formal public school prekindergarten to community-based centers to the federally funded Head Start program, to the informal settings where a small group of children may spend their days in the church up the street, or with a relative or family friend.

In broadening the landscape of research on early education and care settings, there is a simultaneous need to expand the way that quality is operationalized in these studies, to reflect the latest science of early learning and development. Quality features are generally thought to contribute to children's learning and are also thought to play a key role in explaining the effectiveness (or ineffectiveness) of early education and care programs (Burchinal, Kainz, & Cai, 2011; Lipsey, Farran, & Durkin, 2018). Despite some ongoing debate about what quality means and how to measure it, overall it is viewed as comprising key structural (e.g., appropriate ratios, health and safety regulations, and staff qualifications) and process features (e.g., content-rich curricula and instruction, as well as stimulating and supportive adult–child interactions; Hamre & Pianta, 2007; Layzer & Goodson, 2006; Markowitz, Bassok, & Hamre, 2017; Yoshikawa et al., 2013). However, to date, many of the measures used tend to emphasize structural features (e.g., the Early Childhood Environment Rating Scales [ECERS]; Harms & Clifford, 1980). Those that do include a focus on process quality (e.g., the Classroom Assessment and Scoring System™ [CLASS]; Pianta, La Paro, & Hamre, 2008) typically make a very broad and global assessment, without necessarily getting to the level of moment-to-moment interactions or activities (Phillips et al., 2017). We refer to such elements throughout this paper as micro-features of quality.

Finally, studies linking aspects of quality to child outcomes largely focus on a narrow set of outcomes, concentrating most frequently on traditional early academic skills (e.g., reading and math) that are viewed as important to understanding the effects of early learning opportunities on children's later schooling outcomes. However, decades of research have shown that children's early language and their social-emotional skills are both linked to their experiences in early education and care (Hanno, Jones, & McCoy, 2019; Jones, Bub, & Raver, 2013; Rowe, Silverman, & Mullan, 2013) and predictive of their outcomes in school and beyond (Moffitt et al., 2011). With a broader set of developmental domains represented, including social-emotional skills and competencies, there is reason to examine whether and how micro-features of quality are differentially associated with these multiple dimensions of learning and development (e.g., language, literacy, executive functions, social and emotional skills; Downer, Sabol, & Hamre, 2010; Hamre, Hatfield, Pianta, & Jamil, 2014).

In sum, there is a need for research that considers the full array of settings in which young children spend their time, broadens our understanding of quality, and considers a more comprehensive set of developmental domains. In addition, as the field of early education and care continues to expand and faces ongoing pressure to improve its quality at scale, there is a need for a research strategy that moves beyond the broad question of whether specific programs “work;” one that is designed to inform the more complex questions of how to build and sustain a high-quality early education and care system that meets today's needs. That is, there is not yet a robust science that matches today's population and contexts to inform the design and scaling of high-quality early learning environments for all children, and across different setting types.

In this paper, we begin to address the need for a next generation of research in this area, using data drawn from the first wave of the Early Learning Study at Harvard (ELS@H), a longitudinal, population study of early education and care in one state. The first wave includes comprehensive information (derived from direct assessments, parent and provider surveys, and setting observations) on over 3000 three- and four-year-old children, their families, and their primary early education and care settings including those that are formal and informal. Critical to the questions we address, and those of the field more generally, ELS@H is the first study designed to include a representative sample of early education and care settings in conjunction with a population representative sample at the child-level.

We use these data to first document the broad landscape of early education and care settings in the state, shedding light on the settings in which young children spend their time before elementary school. We then consider how micro-features of quality vary within and across different early education and care setting types including both formal and informal settings. Finally, we examine how children's skills vary across these setting types.

Public and private programs today range in type from formal public school prekindergarten to community-based preschool, to Head Start, to family-based child care, to relative or other family care. Families likely choose options based on a host of interrelated factors including cost, location, hours, availability of different choices, etc. (Bassok, Magouirk, Markowitz, & Player, 2018). To date the most comprehensive view into children's participation in these various early education and care settings comes from the ECLS-B, a nationally representative study that tracked 14,000 children from their year of birth in 2001 until kindergarten entry (National Center for Education Statistics, n.d.). Studies using data from the ECLS-B have distinguished between formal and informal settings, with child care centers, Head Start, and prekindergarten programs comprising the formal category and family child care and other care settings (e.g., non-relative care and non-parental relative care) comprising the informal category (Bassok et al., 2016). In 2005, when the sample was four years old, 44.8% of children were enrolled in child care centers; 13.1% in home-based non-parental, relative care; 12.7% in Head Start; and 7.6% in non-relative care (Chernoff, Flanagan, McPhee, & Park, 2007). A sizeable proportion (20%) of participating children were not enrolled in any early education or care, representing those we would classify as being in “parental care,” which we also refer to as “parent only care” and “parent care only.” Despite the richness of the ECLS-B data, these patterns are likely to have shifted given the changes in the early education and care sector since 2005 that we noted above (e.g., increases in state funding). More recent enrollment estimates are available for specific education and care types, like Head Start and state-funded prekindergarten, due to tracking by the federal government and the National Institute of Early Education Research (NIEER). However, comprehensive enrollment estimates from across the mixed-delivery system, like those yielded from the ECLS-B, are largely unavailable; few states track enrollment across all formal and informal education and care types, and those that do track enrollment for some types are often not readily linkable to other states’ systems.

In Massachusetts, the setting of the ELS@H, children participate in the complete range of formal and informal setting types. Prior estimates indicate that nearly 70% of children from birth to age five in Massachusetts spend at least 10 h per week in non-parental care (Committee for Economic Development of The Conference Board, 2019). With only a small-scale state-funded prekindergarten program in the state (Friedman-Krauss et al., 2018), it is assumed that families of three- and four-year-olds rely on a number of formal and informal alternatives. Although the state collects data on the number and types of early education and care programs licensed by the state (i.e., public school prekindergarten [PSP], Head Start [HS], community center-based care [CCC], and family child care [FCC]), little is known about the exact number of children enrolled in or the characteristics of these various settings. Moreover, nothing is known about the proportion of children in unlicensed settings (e.g., unlicensed non-relative care [UNC], unlicensed relative care [URC], and parental care only [PC]). Although this study does not necessarily represent the landscape of early education and care across the United States, the patterns in Massachusetts echo the nation's broader reliance on a mixed-delivery system.

As noted above, high-quality education and care is generally thought to comprise a blend of key structural and process features (Hamre and Pianta, 2007, Markowitz et al., 2017, Yoshikawa et al., 2013). Consistent with foundational theories of human development and learning, a cornerstone of quality is the interactions between adults and children (e.g., Bowlby, 1999, Bronfenbrenner and Morris, 2006, Vygotsky, 1978). Adults play a key role in scaffolding and supporting learning, providing emotional warmth and a sense of security, and in turn, guide and shape children's interactions with others (Pianta, 1999, Sabol and Pianta, 2012), ultimately fueling children's development in positive ways.

A number of studies address variation in structural and process quality features across education and care settings (Abner et al., 2013; Bassok et al., 2016; Dowsett, Huston, Imes, & Gennetian, 2008; Hatfield, Lower, Cassidy, & Faldowski, 2015; Kontos, Howes, Shinn, & Galinsky, 1997; Rigby, Ryan, & Brooks-Gunn, 2007; Zellman & Karoly, 2015). However, because they vary in the number of education and care types they consider and in the measures of quality they use, our ability to compare findings across these studies is somewhat limited. Bassok and colleagues (2016) present the most complete comparison of quality across the full range of early education and care types to date. Using data from ECLS-B, the authors compared quality at the national-level across formal and informal settings, including child care centers, Head Start, public school prekindergarten, family child care centers, and other care arrangements. They used a robust set of structural quality indicators (i.e., caregiver characteristics, safety features, and activities/curriculum) and two observational tools (i.e., ECERS and the Arnett Caregiver Observation Scale) that captured both structural and process features (e.g., adult–child interactions). Results indicated that adult–child ratios were lower in informal than formal settings, but that children in informal settings were less likely to engage in learning activities (e.g., read alouds by caregivers, math activities). Informal settings also scored significantly lower on both observational measures than did formal settings.

These estimates suggest that children in formal and informal settings have different experiences, yet the quality measures used offer little insight into the specific nature of children's activities and interactions in these settings. In other words, they lack sufficient detail about the micro-features of quality that might ultimately inform a concrete improvement process, as the system continues to expand and aims to increase impacts on children's outcomes. First, it appears that some structural features do not necessarily lead to higher quality experiences for children (Cryer, Tietze, Burchinal, Leal, & Palacios, 1999; Early et al., 2006, 2007; Phillipsen, Burchinal, Howes, & Cryer, 1997; Slot, Leseman, Verhagen, & Mulder, 2015). For example, as illustrated by Bassok and colleagues (2016), informal settings tended to have lower teacher-child ratios, but, counter to traditional logic, also tended to have lower scores on observational measures of quality. Second, the observational tools employed to capture features of process quality tended to document only global dimensions. For example, the ECERS tool (Harms & Clifford, 1980), used by Bassok and colleagues (2016), includes a single item to capture the nature of adult–child interactions. Scores on this single item represent tone, emotionality, and respect evident in adult–child interactions across the observational period. The bundling of distinct features into individual items and subscales suggests that two settings could have identical quality scores but children's experiences within them might actually be quite different (e.g., one setting could be marked by positive tone but mixed levels of respect and another by mixed tone but high levels of respect). Relatedly, global measures of quality can obscure the direct experiences of individual children (Farran & Hofer, 2013). For example, an adult might be observed to have, on average, positive interactions with children in the setting, but that average might obscure the fact that she consistently has quite negative interactions with a small group of individual children. These limitations motivated us to employ an approach to measuring quality that builds on and deepens existing measures, and therefore the existing knowledge-base on quality, by capturing the micro-level interactions and activities of adults and individual children in early education and care settings (Downer, Booren, Lima, Luckner, & Pianta, 2010; Farran & Hofer, 2013).

The empirical evidence on the relationship between measures of quality and child skills has been mixed (Zaslow, Burchinal, Tarullo, & Martinez-Beck, 2016). In their study of a nationally representative sample of young children, Bassok and colleagues (2016) reported that variation in process and structural quality was linked to differences between sectors (i.e., informal and formal early education and care settings) in children's reading and math skills at age five. Yet, Abner et al. (2013) found that a combination of structural and process quality (measured using the ECERS) did not account for associations between setting type and children's academic skills. Other correlational studies linking quality measures to child skills have been similarly equivocal (Burchinal, Vandergrift, Pianta, & Mashburn, 2010; Burchinal et al., 2011; Early et al., 2007; Keys et al., 2013; Love et al., 2003; Mashburn et al., 2008; Sabol, Hong, Pianta, & Burchinal, 2013; Vandenbroucke, Spilt, Verschueren, Piccinin, & Baeyens, 2018). For example, Keys et al. (2013) found that across a number of studies process quality was positively, but weakly, associated with children's language and math skills but not with their social skills. In another meta-analysis, Vandenbroucke and colleagues (2018) found that process quality, specifically the quality of teacher-child interactions, was positively linked to only some components of executive functioning (i.e., working memory and inhibition, but not cognitive flexibility). More rigorous studies isolating the contribution of changes in quality to child outcomes have found small or no associations between quality and child skills (Auger, Farkas, Burchinal, Duncan, & Vandell, 2014; Gonzalez, McCoy, & Sabol, 2018).

In our view, many of the same factors that limit a deeper understanding of quality across education and care types, also affect our ability to link features of quality to child outcomes. Measures that capture global characteristics of quality are likely to have inconsistent associations with children's outcomes as they bundle multiple dimensions together and fail to account for individual children's experiences in settings (Farran & Hofer, 2013). Relatedly, existing measures of quality fail to isolate individual features of quality, or specific strategies, that promote children's learning (Weiland, 2018). More detailed knowledge of micro-features and dimensions, because they capture more directly children's actual everyday experiences, may show stronger associations with children's skills. Some relatively newer tools such as the Child Observation in Preschool and Teacher Observation in Preschool (COP-TOP; Bilbrey, Vorhaus, & Farran, 2007; Farran & Anthony, 2014) respond to these limitations, because they are designed to capture micro-level processes and the experiences of individual children. Although evidence linking these newly-developed tools to children's skills so far is limited, recent research is encouraging. For example, using the COP-TOP, Farran, Meador, Christopher, Nesbitt, and Bilbrey (2017) observed 26 public school prekindergarten classrooms, noting that gains in children's early language and literacy, mathematics, and self-regulation were related to eight specific practices (i.e., the “Magic 8™”) measured by the tool.

Recognizing that the relationship between features of quality and children's skills may be bidirectional, more work is needed to describe whether and how specific micro-features of quality are differentially linked to children's skills (e.g., language, literacy, executive functions, social and emotional skills). For example, math opportunities (i.e., the amount of time children spend engaged in math-related activities), one of the Magic 8 practices, are likely to be associated with children's math skills, but others, like the general rigor of instruction in the setting, may be tied to child skills across domains. Some prior work has explored relationships between specific features of quality and learning domains, although most again rely on global measures of quality and focus on a limited range of learning domains (Downer, Sabol, et al., 2010b; Hamre et al., 2014). For example, Hamre et al. (2014), using a bi-factor model of the CLASS, showed that a global factor (responsive teaching) was weakly associated with children's skills across a number of domains, including language, literacy, and self-regulation, whereas two domain-specific factors were differentially linked to specific skills. Cognitive facilitation was associated with language and literacy skills and positive management and routines was associated with children's self-regulation.

In the present study we use rich, population data from the Early Learning Study at Harvard to both explore the landscape of early education and care across the state of Massachusetts and describe the nature of the settings in which children receive early education and care. We then examine the features of different early education and care settings, and explore whether children's language, literacy, math, executive functions, and social-emotional skills vary between setting types when child and family characteristics are controlled. Specifically, we address the following research questions:

  • 1.

    What is the landscape of early education and care among three- and four-year-olds in Massachusetts?

  • 2.

    How do micro-features of quality vary within and across different types of early education and care settings, and specifically, between settings classified as formal and those as informal?

  • 3.

    How do child mathematics, literacy, language, executive functions, and social-emotional skills vary within and across different types of formal and informal early education and care settings?

The sampling and measurement approach of ELS@H offers an opportunity in the present study to build on existing work on quality in early education and care in three primary ways. First, we study the full range of early education and care settings using a sample representative of children and settings in the state, in contrast to the majority of prior studies that rely on non-representative samples or focus on only one or two specific education and care types. Second, we use an observational measure of quality (i.e., the COP-TOP) that is designed to document micro-features of quality and illuminate the more nuanced experiences of teachers and individual children in early education and care settings. Third, we include information on children's skills across a broad range of developmental domains (i.e., language, literacy, math, executive functions, and social-emotional).

Section snippets

Sample

Data were drawn from the first wave of the Early Learning Study at Harvard (or ELS@H), a longitudinal, population study of young children's early education and care settings. The ELS@H was designed to produce estimates representative of the state's population of three- and four-year-old children. The final sample from the first wave includes 3222 children; we additionally used sample weights (described in greater detail below) such that results from analyses were representative of all three-

Results

Question 1: What is the landscape of early education and care among three- and four-year-olds in Massachusetts? Table 1 presents the characteristics of three- and four-year-olds in Massachusetts. The percentages of children in each category incorporate sample weights such that results are representative of three- and four-year olds in Massachusetts. Slightly over half of children were four years old. The majority of households had family incomes above $75k, and approximately one-fifth of

Discussion

This paper presents initial findings from a population study of early education and care designed to describe the landscape of early education and care in one state. Specifically, we used this unique sample to examine variation in micro-features of quality across a subset of setting types, and to explore whether or not these indicators of quality were associated with any observed differences across setting types in child skills. These questions are critical as it is well-documented that the

Conclusion

The present study offers a comprehensive look at the early education and care system of one state, and all the settings within that system. In the context of a mixed-delivery system for early education and care, families use a variety of arrangements. There is some evidence that children and families who use different types of programs vary in their socio-demographic profiles, and that some micro-features of quality also vary across the different types of education and care. Child skills also

Conflict of interest

None declared.

Funding

Research activities were funded by the Saul Zaentz Charitable Foundation through its generous support of the Saul Zaentz Early Education Initiative at the Harvard Graduate School of Education.

Acknowledgements

We thank Abt Associates, in particular Amy Checkoway, Barbara Goodson, and Kerry Hofer, for their expertise, ongoing collaboration, and support of this work. We are also deeply grateful to all the children, families, educators, and caregivers who have so generously shared their time and perspectives with us.

References (79)

  • L.A. Karoly et al.

    Understanding variation in classroom quality within early childhood centers: Evidence from Colorado's quality rating and improvement system

    Early Childhood Research Quarterly

    (2013)
  • M.W. Lipsey et al.

    Effects of the Tennessee Prekindergarten Program on children's achievement and behavior through third grade

    Early Childhood Research Quarterly

    (2018)
  • L.C. Phillipsen et al.

    The prediction of process quality from structural features of child care

    Early Childhood Research Quarterly

    (1997)
  • M.L. Rowe et al.

    The role of pictures and gestures as nonverbal aids in preschoolers’ word learning in a novel language

    Contemporary Educational Psychology

    (2013)
  • C.M. Setodji et al.

    Using the early childhood environment rating scale-Revised in high stakes contexts: Does evidence warrant the practice?

    Early Childhood Research Quarterly

    (2018)
  • P.L. Slot et al.

    Associations between structural quality aspects and process quality in Dutch early childhood education and care settings

    Early Childhood Research Quarterly

    (2015)
  • R. Smith-Donald et al.

    Preliminary construct and concurrent validity of the Preschool Self-regulation Assessment (PSRA) for field-based research

    Early Childhood Research Quarterly

    (2007)
  • C. Weiland

    Commentary: Pivoting to the “how”: Moving preschool policy, practice, and research forward

    Early Childhood Research Quarterly

    (2018)
  • G.L. Zellman et al.

    Improving QRISs through the use of existing data: A virtual pilot of the California QRIS

    Early Childhood Research Quarterly

    (2015)
  • K.S. Abner et al.

    Does child care quality mediate associations between type of care and development?

    Journal of Marriage and the Family

    (2013)
  • A. Auger et al.

    Preschool center care quality effects on academic achievement: An instrumental variables analysis

    Developmental Psychology

    (2014)
  • W.S. Barnett et al.

    The state of preschool 2002

    (2003)
  • D. Bassok et al.

    Within- and between-sector quality differences in early childhood education and care

    Child Development

    (2016)
  • J.R. Berrueta-Clement et al.

    Changed lives: The effects of the Perry Preschool Program on youths through age 19 (No. 8; Mongraphs of the High/Scope Educational Research Foundation).

    (1984)
  • C. Bilbrey et al.

    Teacher observation in preschool

    (2007)
  • J. Bowlby

    Attachment and loss, Vol 1: Attachment

    (1999)
  • U. Bronfenbrenner et al.

    The bioecological model of human development

    Handbook of Child Psychology

    (2006)
  • M. Burchinal et al.

    How well do our measures of quality predict child outcomes?

    A meta-analysis and coordinated analysis of data from large-scale studies of early childhood settings

    (2011)
  • F.A. Campbell et al.

    Early childhood education: Young adult outcomes from the Abecedarian Project

    Applied Developmental Science

    (2002)
  • S. Carlson et al.

    Minnesota Executive Function Scale: Test manual

    (2014)
  • A. Chaudry et al.

    Child care choices of low-income working families

    (2011)
  • J.J. Chernoff et al.

    Preschool: First findings from the preschool follow-up of the early childhood longitudinal study. Birth cohort (ECLS-B). First look. NCES 2008-025

    (2007)
  • CityHealth et al.

    Pre-k in American cities. CityHealth and National Institute for Early Education Research

    (2019)
  • Committee for Economic Development of The Conference Board

    Child care in state economies: 2019 update

    (2019)
  • J.T. Downer et al.

    Teacher–child interactions in the classroom: Toward a theory of within- and cross-domain links to children's developmental outcomes

    Early Education & Development

    (2010)
  • S.E. Duncan et al.

    PreLAS 2000

    (1998)
  • D.M. Early et al.

    Teachers’ education, classroom quality, and young children's academic skills: Results from seven studies of preschool programs

    Child Development

    (2007)
  • D.C. Farran et al.

    Child Observation in Preschools (COP): MNPS edition

    (2014)
  • D.C. Farran et al.

    Evaluating the quality of early childhood education programs

    Handbook of research on the education of young children

    (2013)
  • Cited by (16)

    • Residential and preschool neighborhoods: Exploring patterns of socioeconomic match and its association with child skills across Massachusetts

      2023, Early Childhood Research Quarterly
      Citation Excerpt :

      The use of different EEC settings likewise appears associated with household characteristics. Among Massachusetts families, middle-income households are less likely than lower- or higher-income households to send their children to center-based care (Jones et al., 2020). One survey from a Southern state found that non-White mothers are less likely than White mothers to prefer nonparental care (Rose & Elicker, 2010).

    • An ecological perspective on early educator well-being at the start of the COVID-19 pandemic

      2022, Early Childhood Research Quarterly
      Citation Excerpt :

      Non-emergency programs remained closed through at least the end of June 2020. Educators were invited to participate in our survey because they or someone else in their setting had previously participated in the Early Learning Study at Harvard (ELS@H), a longitudinal, population-level study of young children and their early education and care settings in Massachusetts (Jones et al., 2020). At the end of March 2020, we attempted to contact 925 early education and care programs that enrolled a child participating in the first two years of ELS@H (2017-2018 and 2018-2019).

    View all citing articles on Scopus
    View full text