Time spent playing predicts early reading and math skills through associations with self-regulation

Abstract

Children’s play time has declined in recent decades, which could negatively impact early self-regulation—a vital component of school readiness. To date, studies have not fully explored how the time spent playing relates to children’s self-regulatory skills, and in turn, their early reading and math competencies. Using data from time diaries and direct assessments of self-regulation, prereading, and math skills, this study examined how minutes spent playing at home predict these skills in a sample of 128 children followed from age four to five. Additionally, it considered whether self-regulation explained links between play time and prereading and math. Results showed that the time spent playing positively related to children’s self-regulation. Moreover, through its association with self-regulation, play time had indirect effects on prereading and math skills measured one year later. Results suggest that fostering opportunities for play time during the preschool years may help to boost school readiness skills.

Introduction

Children’s play time is in jeopardy (Bodrova & Leong, 2003; Gray, 2011; Zigler & Bishop-Josef, 2009). A study documenting changes in children’s time use from 1981 to 1997 found that time for free play, which included things like playing cards, board games, and puzzles, playing social games like tag or hide-and-seek, pretend play, and playing with toys/objects, decreased by 25% for 6- to 8-year-old children (Hoefferth & Sandberg, 2001). This translates to eight fewer hours of play per week for today’s young children (Elkind, 2008; Hoefferth & Sandberg, 2001). This decline in play time may have troubling implications for children’s development, especially self-regulatory skills development (Ursache et al., 2012). Play helps children acquire skills that are necessary for success in adulthood, including decision making, problem solving, self-control, rule following, emotion regulation, and interpersonal skills (Gray, 2011). These self-regulatory skills are vital once children begin formal schooling; self-regulation underlies the cognitive and behavioral outcomes imperative for school success, including academic achievement (Blair & Razza 2007; McClelland & Tominey, 2016; Rimm-Kaufman et al., 2009).

Yet, while prior time diary studies have shown decreases in children’s play time (Hoefferth & Sandberg, 2001), research hasn’t addressed whether this reduction in children’s play time may be related to preschool-aged children’s self-regulatory skills. Moreover, no studies have examined whether play time has indirect links to children’s early prereading and math skills through its role in the development of self-regulation. Rather, two informative, but largely separate literatures have developed: one using time diaries documenting children’s time use, including, play time, but failing to relate variation in children’s play time with child outcomes and a second associating specific characteristics of play obtained from lab-based, adult-structured instances of play with children’s skill. Using time diary reports on children’s play time at home across a broad range of play activities and rich data on children’s early regulatory and preacademic skills, the current study aimed to bridge these literatures and fill a research gap by looking at associations between time spent engaging in play and self-regulation in a diverse group of 128 four-year-old children. It also examined whether play time is related to emerging reading and math skills measured one year later via its links to self-regulation. In addressing these questions, the results of this study provide useful information to researchers, educators, practitioners, and policy makers interested in fostering children’s school readiness skills and narrowing achievement gaps. Our focus on play time occurring in naturalistic everyday environments is significant for several reasons. First, it speaks to implications of reductions in children’s play time observed in national data. Second, it sheds light on whether unstructured, child-directed play opportunities may have benefits for children. This knowledge is informative for structuring the lives of children both at home and in early education and care settings.

Theoretical Framework

Although play is a complex concept that can be difficult to define (Eberle, 2014), it is important to articulate the definition of play that theoretically motivates the current study. Most scholars agree upon five general features: play is freely chosen and child-directed, intrinsically motivated and concerned more with means than ends, actively engaging and pleasurable, imaginative, and guided by mental rules but with room for creativity (Ashiabi, 2007; Gray, 2011; Zigler & Bishop-Josef, 2009). It can involve playing with objects, rough and tumble play, and pretend play, as well as fine motor and gross motor play (separately, or more often, in conjunction with each other; Hirsh-Pasek et al., 2009).

The importance of play in early childhood has been emphasized by decades of developmental psychological theory. The 20^th century developmental theorist, Lev Vygotsky, stressed the essential role of play time in cognitive development, and in particular the development of self-regulation (Vygotsky, 1978; Zigler et al., 2004). It is through play that children first learn to make decisions, gain a sense of agency and direct their own actions, solve problems, exert self-control, and follow rules in spite of their own desires (Gray, 2011). Furthermore, play provides an ongoing context for children to practice and refine these newly developed skills and capacities, to take on new social roles, attempt novel or challenging tasks, and solve complex problems (Hirsh-Pasek et al., 2009). More recent theories provide support for the importance of play in the development of early self-regulation (Becker et al., 2014a; Burdette & Whitaker, 2005). During play, children expend energy, encounter opportunities for decision making that require problem solving and creative thinking, and enhance visuospatial, gross, and fine motor skills, all of which help children regulate their emotions, attention, and behavior when presented with classroom demands (Burdette & Whitaker, 2005). Play helps children learn to inhibit impulsive behaviors and follow rules, develop internal representations, and act upon the meanings of objects. For example, when playing tag, children must establish and follow rules like who is “it” and how that changes, they define and act upon the meaning of objects like defining the porch as “jail”, taking captees there to “lock them up”, and generating internal representations like using a stick as the key to the “jail”. Moreover, aerobic activity associated with physically active/gross motor play may activate the prefrontal cortex (Becker et al., 2014a; Chaddock-Heyman et al., 2013; Davis et al., 2011), which has been related to better self-regulatory behaviors (Wagner & Heatherton, 2011), including inhibitory control (van Gaal et al., 2008; van Gaal et al., 2010) and working memory (Weerda et al., 2010).

Following these theories, preschool-aged children should be afforded ample time for play. Yet, research has documented declines in the time spent playing for young children, while time in school increased by about 20% and time spent doing academic work at home increased by a remarkable 145% during that same time (Hoefferth & Sandberg, 2001). At home, children’s opportunities for play time are being supplanted by hurried lifestyles, structured activities, and enrichment programs, as well as time spent using flashcards, educational toys, and media by parents interested in giving their child an advantage in kindergarten (Ginsburg, 2007). In many preschools too, the “trickle down” effect of state standards and high-stakes testing has pushed academic instruction and teacher-directed activities down to the preschool classroom, reducing the amount of time left for play compared to prior decades (Almon & Miller, 2011; Ashiabi, 2007; Guirguis, 2018). A recent study of changes in kindergarten classrooms from 1998 to 2010 found that the percentage of classrooms spending at least one hour in child-directed activities declined by up to 28% while those spending more than three hours doing teacher-directed activities more than doubled (Bassok et al., 2016).

The decrease in play time in recent decades and the accompanying focus on academic instruction and structured activities may be developmentally inappropriate at this age (Hirsh-Pasek et al., 2009). According to the aforementioned theoretical models, reductions in opportunities for play time may compromise children’s school readiness skills, particularly key self-regulatory skills like attention, executive functioning, and behavioral regulation (Burdette & Whitaker, 2005; Cameron et al., 2015). Indeed, early childhood is a sensitive period for the development of self-regulation due to rapid growth in the prefrontal cortex, which is the region of the brain associated with self-regulatory processes (Carlson et al., 2013). Accordingly, the preschool years may be a critical time to encourage and support children’s play time.

Counterintuitively, increases in academic time and structured activities employed to boost preschoolers’ school readiness and academic achievement at the expense of play time may ultimately hinder their academic performance, an idea to which the results of this study may be relevant. Declines in play time during the preschool years may negatively impact children’s self-regulation, which in turn could harm their academic skills. The self-regulatory skills that children acquire during early childhood are foundational to the development of their academic skills (e.g., Blair & Raver, 2012; Bull et al., 2011; Duncan et al., 2007; McClelland et al., 2014; Welsh et al., 2010). Self-regulation is a multidimensional construct that is often operationalized in varied ways (Cole et al., 2019). The present study focuses on aspects of cognitive and behavioral self-regulation that are particularly important when children enter formal school settings, including working memory, attentional flexibility (maintaining focus and adapting to changing goals), and inhibitory control (Cole et al., 2019; Fuhs et al., 2014; McClelland et al., 2010). These skills are important building blocks for the successful transition to formal schooling and future learning because they allow children to navigate structured learning environments, learn new material, ignore distractions, stay on task, and persist in the face of challenges (Cole et al., 2019; McClelland et al., 2014). And while a theoretical pathway from play time to self-regulation to achievement exists, research to date has never empirically examined whether self-regulation acts as a mediator between the time children spend playing and their early achievement. This study aims to address this gap in the literature.

The Importance of Play Time in Children’s Development

Both psychological and educational literature supports the importance of play time in children’s development, but most of it consists of theoretical arguments and there is less empirical data to support its importance (e.g., Savina, 2014). Empirical studies tend to focus on associations between the quality of play (specifically pretend play) and self-regulation as opposed to play time more generally (Matthews, 2008; Slot et al., 2017; Vieillevoye & Nader-Grosbois, 2008; Nader-Grosbois & Vieillevoye, 2012). Some studies have explored specific types of play, and their results also support the link between play and self-regulation, which informed our decision to focus broadly on time spent engaging in any type of play. These play types include block play (e.g., Schmitt et al., 2018), other object play (e.g., Whitebread & Jameson, 2005), gross motor play or physically active play (e.g., Becker et al., 2014a; Lundy & Trawick-Smith, 2020), music play (e.g., Williams et al., 2015; Zachariou & Whitebread, 2015), and more generally, play with peers (e.g., Ivrendi, 2016). Although these studies establish a link between specific characteristics of play and self-regulation, none measure time spent engaged in play naturalistically, in the context of children’s daily lives. Instead, they use experimentally induced play tasks and measure play quality, and thus the ecological validity of results is questionable when it comes to making generalizations to the play experiences that children engage in in their everyday lives (e.g., Schmitt et al., 2018; Slot et al., 2017; Vieillevoye & Nader-Grosbois, 2008; Whitebread et al., 2009). This study adds to the current literature by bridging the gap between time diary studies that measure children’s naturalistic, real-world play but do not contain rich measurement on their developmental outcomes and developmental studies focused on specific characteristics of play but reveal little about whether unstructured play opportunities for children have positive associations with their development. By examining how the time children spend playing in their daily lives is related to their early development, results shed light on potential ramifications of trends showing decreasing play time for children (Elkind, 2008; Hoefferth & Sandberg, 2001).

The Role of Self-Regulation in Early Reading and Math Skill Development

The importance of self-regulation in buttressing academic development in well known (Blair & Razza 2007; Duncan et al., 2007; Howse et al., 2003; McClelland & Tominey, 2016; Payton et al., 2008; Raver et al., 2007; Rimm-Kaufman et al., 2009). Self-regulation encapsulates children’s ability to regulate their behavior using attentional or cognitive flexibility, working memory, and inhibitory control (McClelland et al., 2014), and it is an important indicator of children’s readiness for school. It relates to their future academic achievement as well (Blair & Raver, 2015; Duncan et al., 2007; McClelland et al., 2007; McClelland et al., 2013). It helps children pay attention and remember instructions, shift attention from task to task, ignore distractions, and resist impulsive behavior (Blair & Diamond, 2008; Blair & Raver, 2015; McClelland et al., 2014). Children’s self-regulatory skills have been linked to an array of prereading and math skills, including some of the domains examined in the current study (Becker et al., 2014b; Cadima et al., 2015; Gestsdottir et al., 2014; McClelland et al., 2014; Van de Sande et al., 2013; von Suchodoletz & Gunzenhauser, 2013). These include prereading skills like letter identification, recognizing common sight words, reading words aloud, and receptive and expressive vocabulary (e.g., Becker et al., 2014b; Gestsdottir et al., 2014; von Suchodoletz & Gunzenhauser, 2013). Most of the studies linking self-regulation to early math skills have associated self-regulation with standardized math assessments (e.g., the Applied Problems subtest of Woodcock-Johnson Tests of Achievement) and other number skills, such as number line estimation, counting, cardinality, arithmetic calculations, and non-symbolic number processing (Becker et al., 2014b; Malone et al., 2019; McClelland et al., 2014; Purpura et al., 2017; Silver et al., 2020; Simanowski & Krajewski, 2019). Studies have also documented correlations between domains of self-regulation and spatial skills (Becker et al., 2014b; Hawes et al., 2019; Verdine et al., 2014), though associations between self-regulation and spatial skills remain understudied. Our utilization of a multi-faceted math skills measure that includes numeracy, non-symbolic number processing, and spatial skills is a strength of the present study.

Children with strong self-regulation in preschool score higher on measures of preacademic skills (McClelland et al., 2000, 2006; Ponitz et al., 2009; Robson et al., 2020). These associations between self-regulation and preacademic skills in young children have been demonstrated across samples varying in socioeconomic and demographic characteristics (Blair & Razza, 2007; Liew et al., 2008; McClelland et al., 2007), as well as with the self-regulation assessments used in the current study (Lipsey et al., 2014; Robson et al., 2020). In addition, self-regulatory skills observed in early childhood continue to predict long-term academic achievement, as well as high school and college completion (Breslau et al., 2009, McClelland et al., 2013; Robson et al., 2020; Valiente et al., 2008). Thus, the importance of self-regulation for academic success is clear and indicates that improving self-regulation in preschool may have long-term effects.

Despite the substantive body of literature linking self-regulation to early prereading and math, it is difficult to discern whether these associations are causal (Jacob & Parkinson, 2015). Random assignment intervention studies that have been shown to improve self-regulation, and prereading and math skills in turn, often include activities that improve both self-regulatory and preacademic skills, like Tools of the Mind (Barnett et al., 2008; Bodrova & Leong, 2003); thus, the links between self-regulation and preacademic skills are not necessarily causal. However, studies utilizing other rigorous analytical methods have shown these associations. For instance, McClelland et al. (2007) used change in self-regulation to predict change in early mathematics and prereading skills from fall to spring of the prekindergarten year, controlling for fall achievement levels, thereby eliminating omitted variable bias from time-invariant characteristics of children and families. Other studies have used autoregressive models, where an earlier measure of achievement is included as control variable, which also adjusts for time-invariant factors that affect achievement (e.g., Blair & Razza, 2007; Duncan et al., 2007; McClelland et al., 2007; Welsh et al., 2010). Taken together, the research suggests that associations between self-regulation and academic skills are, at least in part, causal. Thus, identifying early predictors of self-regulation is vital information for efforts to narrow achievement gaps.

The Current Study

The acquisition of self-regulatory skills is an important developmental milestone during early childhood because it acts to strengthen achievement as children enter and progress through formal schooling (Blair & Diamond, 2008; Gestsdottir et al., 2014; Valiente et al., 2010). Children’s access to and engagement in free play time may be a predictor of self-regulation in early childhood (Gray, 2011), and play time may in turn relate to children’s early prereading and math skills. This idea has not been fully studied despite recent trends showing decreases in children’s play time at home and at preschool. The current study bridges the gap between time diary research showing declines in play and play research that has largely failed to examine time spent in naturally occurring play at home or in broad categorizations of play.

Our first aim was to relate the time children engage in play, as reported by parents using time diaries, to their self-regulatory skills at age four. Next, it asked whether play time has links to children’s prereading and math skills at age five. Finally, it explored whether self-regulation acts as a pathway through which play time predicts later prereading and math. The use of time diary reports on minutes of play coupled with direct assessments of children’s skills is a major strength and contribution of the study. It answers questions regarding whether children’s naturalistic, self-directed play time at home relates to their early development, which is useful information when thinking on how to structure children’s time at home and in early care settings. In doing so, results provide important information for researchers, parents, and practitioners, as well as policymakers interested in fostering children’s early self-regulatory, prereading, and math skills and narrowing gaps in school readiness skills.

Method

Participants

Data were drawn from the Parents Promoting Early Learning Study (PPEL). PPEL is a community-based study of 128 four-year-old children and their parents. PPEL involved two waves of data collection: one when children were four and the second one year later, when they were about five years old. Families were recruited in preschools and childcare centers and through the distribution of fliers in the community in a large, mid-Atlantic metropolitan area in the United States. Recruitment efforts resulted in the enrollment of a socioeconomically diverse sample of families. At enrollment, parents reported annual incomes between $0 and $425,000, with a median of $90,000 (SD = $78,547). Approximately one-quarter of the sample reported incomes below 200% of the poverty line based on their household size. Highest levels of parental education (of either parent) varied, including parents who did not finish high school (2%), parents with high school diplomas/GEDs (7%), parents who attended some college or a two-year college (13%), parents with bachelor’s degrees (23%), and parents who were pursuing or completed a graduate degree (55%). Overall, children were on average four years, five months old at the first assessment (SD=.30, range=4–4.96) and five years, six months old at the second assessment (SD=.32, range =5.01–6.39). Most participating parents were biological or adoptive mothers (94%) and were on average 36 years old (range = 24–56 years). Parents also reported their employment (40% full-time, 27% part-time, 33% unemployed or out of the labor force), marital status (76% married), and race/ethnicity (76% White, 17% Black or African American, 7% Asian, Latino, multiracial, or other).

Of the 128 families that enrolled in the study, 15 participated in only wave 1 of the study, and others were missing data on at least one variable included in the present analyses. The amount of missing data per variable ranged from about 0%−10% for many of the demographic variables, like family SES, marital status, child sex assigned at birth, and race/ethnicity, to a high of 20%−25% for two of the direct assessments: the self-regulation and reading composites (these percentages include the 15 cases that did not complete the second wave). Children with complete data were more likely to have married parents and had, on average, higher non-verbal intelligence, expressive vocabulary, and reading scores. As these data were not missing completely at random, missing data were imputed in MPlus to create 50 imputed data sets (Royston, 2004). Multiple imputation imputes values for missing data by estimating these values as a function of several covariates and random error (Asparouhov & Muthen, 2022; Rubin, 1987; Shafer, 1997). Following imputation, regression models are estimated on each of the 50 data sets that include complete data. Parameters are averaged across the resulting regressions and standard errors are computed using methods described by Asparouhov and Muthen (2022). This approach to missing data imputation is superior to imputation using a single value, such as the mean or median, because it takes into account the underlying uncertainty in the distribution of the missing data. Importantly, this method of data imputation is recommended when data are not missing completely at random (Enders, 2012). Listwise deletion tends to inflate standard errors and may bias parameter estimates and even lead to model misspecification when data are not missing completely at random (Asendorpf et al., 2014; Enders, 2012; Newman, 2003).

Measures and Procedures

Data for this study were collected via three different procedures. At wave 1, direct assessments of children’s cognitive competencies, including self-regulatory skills and early prereading and math, were conducted during in-person visits. Children completed the direct assessments across two sessions. The average time in between these sessions was 14 days (SD=13), and 84% of children completed the second session within one month of the first session. The longest gap between sessions was 56 days. Testing occurred in a quiet location in the child’s home, childcare center, or our research lab. When tested at home or in the lab, parents were invited to observe the testing session but instructed not to provide help or feedback to the child. At the conclusion of each session, researchers completed a child behavior rating form. Second, demographic and other data, which we use for many of the covariates in the study, were collected from parents via electronic questionnaires. Lastly, time diary data were collected via phone calls with parents using a modified version of the American Time Use Survey (ATUS; U.S. Bureau of Labor Statistics, 2016). Importantly, these time diary methods have been used widely with families of diverse backgrounds (Nesteruk & Garrison, 2005; Lee et al., 2016). There were two time diary interviews, one in which parents reported all activities engaged in by parents and children during a workday, and one in which they reported this information for a non-workday. If the parent worked every day or was not employed, the time diaries were completed to reflect activities on a weekday and a weekend day. Parents reported all of their activities and their child’s activities on the prior day starting at 4 AM and ending at 4 AM the following day, including telling the interviewer what time activities started and what time they ended. They were given a paper time diary to assist in the recording of activities for the two days on which they were reporting. Parents reported on primary and secondary activities (when activities took place simultaneously), and where and with whom activities took place.

One year after wave 1 of data collection (age 4), families were invited to participate in wave 2 of the study (age 5). Due to the COVID-19 pandemic, follow up visits were conducted virtually using video conferencing software. Assessments were divided into three calls to keep testing sessions between 15 and 30 minutes each, and the order of testing sessions was counterbalanced between children, but the order of tasks within a testing session was fixed. The average time between the first and last session was 14 days (SD=13). At age 5, 93% of children completed all three sessions within one month. Assessment materials on PowerPoint slides were shown to participants through Zoom’s “Share Screen” function, and researchers recorded children’s responses during administration. Parents were with their children during the video sessions but were instructed to allow their children to answer all questions independently, and all tasks were designed so that children could complete the sessions without parent assistance once the call was begun. Additionally, at both waves, demographic information was collected from parents through visits, phone calls, and electronic questionnaires.

Child Outcomes

Self-regulation.

At age four, self-regulation was directly assessed using the Head-Toes-Knees-Shoulders task (HTKS; McClelland et al., 2007). HTKS is a measure of behavioral self-regulation that integrates various aspects of working memory, attentional flexibility, and inhibitory control into a game appropriate for children aged 4–8 years of age (McClelland & Cameron, 2012; Wanless et al., 2011). The HTKS has three sections with four paired behavioral rules: 1) “touch your head” and “touch your toes;” and 2) “touch your shoulders” and “touch your knees.” First, children are asked to play naturally. In the second and third parts, they are instructed to switch rules by responding in the “opposite” way (e.g., touch their head when told to touch their toes), and then pairings are switched (i.e., head goes with knees and shoulders go with toes). In all sections, incorrect responses were scored as 0, self-corrections were scored as 1, and correct responses were scored as 2. The HTKS assessment been examined and cross-validated for construct validity, developmental change, test-retest reliability, convergent validity with teacher ratings of children’s self-regulation, and predictive validity for subsequent preacademic achievement and achievement gain in children aged four (Lipsey et al., 2014).

Additionally, after each home visit, research assistants completed the Preschool Self Regulation Assessment (PSRA) assessor report, which rates children’s behavior during the assessment battery (Smith-Donald et al., 2007). The PSRA assessor report includes 28 items, scored on a 0–3 scale, that ask the assessor to report on children’s attention, inhibition/impulse control, and social behaviors (Smith-Donald et al., 2007). The PSRA assessor report has been found to be reliable for the evaluation of preschool-aged children’s self-regulation skills (Smith-Donald et al., 2007; Tanribuyurdu & Yildiz, 2014). We collapsed two subscales tapping children’s attention and impulse/inhibitory control based on prior studies using this measure (Daneri et al., 2018). For this study, scores on HTKS and PSRA measures were standardized and averaged (r=.52). The decision to create a self-regulation composite was based on research showing that composite scores created from using these measures performed better than any single measure (Lipsey et al., 2014).

Early prereading skills.

At age five (wave 2), children’s emergent prereading skills were assessed directly via the Sound Matching subtest of the Comprehensive Test of Phonological Processing, second edition (CTOPP-2; Wagner et al., 2013). The CTOPP-2 is a comprehensive, norm-referenced instrument designed to assess phonological processing abilities as prerequisites to reading fluency. The Sound Matching subtest measures children’s ability to identify common words with matching initial or final sounds (e.g., “Which word starts with the /n/ sound like neck? Nut, bed, or cake?”). Children also completed the Letter-Word Identification subtest of the Woodcock-Johnson Tests of Achievement IV (Schrank et al., 2014). This subtest required children to recognize (i.e., by pointing) or recall (i.e., by labeling) letters of the alphabet, followed by recognizing common sight word and finally reading words aloud. Standardized scores were calculated based on children’s ages at the time of assessment. Reliability of this subtest among 5-year-olds is estimated to be 0.98 (McGrew et al., 2014).

Early math skills.

At age 5 (wave 2), children’s early math skills were measured using several direct assessments covering a broad range of skills relevant for early math competence including numeracy, spatial, and non-symbolic number processing skills (Clements, 2004; Pritulsky et al., 2020; Geary & vanMarle, 2016; Rittle-Johnson et al., 2017). First, they completed the Applied Problems subtest of the Woodcock-Johnson Tests of Achievement, 3^rd edition (WJ-AP III; Woodcock et al., 2007), which measured their ability to analyze and solve math problems. The problems became progressively more difficult, with initial items requiring the application of basic number concepts, such as counting, to items requiring arithmetic and knowledge of units, such as currency and temperature. Children’s spatial skills were assessed with a shortened version of the Children’s Mental Transformation Task (CMTT; Levine et al., 1999; Ehrlich et al., 2006). Children were presented with shapes and asked to identify the shape the pieces would create if they were put together from a set of four response options. Sixteen of the original set of 32 CMTT items were administered, requiring children to perform two-dimensional mental transformations, including horizontal and diagonal translations and rotations. The Geometric Sensitivity task assessed children’s geometric skills (GS; Dehaene et al., 2006). Children were presented with six images—five out of the six images represented a geometric property (e.g., of distance or angles) that was not present in the sixth image—and were asked to identify the one image that didn’t belong. Children also completed a math in the context of money task, which included two sections. First, children were asked to identify the values of different coins and bills. Then they completed six items requiring magnitude comparison or arithmetic in applied word problems relating to money. Children’s knowledge of Arabic numerals was assessed using a brief assessment requiring children to label numerals shown on the screen. Specifically, children were shown 12 trials with one-, two-, and three-digit numbers. In each trial, a single number was shown on the screen and children were asked to identify it. Lastly, children completed a non-symbolic number comparison task designed to assess the precision of their approximate number system (ANS; Halberda et al., 2008). Participants were presented with arrays of yellow and blue dots on a tablet and were asked to indicate which of the two sets contained the larger number of dots. Numbers ranged between 5 to 21 and pair of numbers varied in ratio (larger number divided by smaller number: 3, 2, 1.5, or 1.3). Trials included three different conditions to control for non-numerical visual confounds of the displays. In correlated trials, cumulative surface area positively correlated with the number of dots and thus, the array with more dots had a larger surface area, as average dot size was the same across arrays. In neutral trials, the cumulative surface areas of each pair of arrays were equated and thus, the arrays with more dots necessarily had smaller-sized dots. In anti-correlated trials, cumulative surface area was negatively correlated with dot numerosity while cumulative perimeter was equated, and therefore, the set with more dots had a smaller surface area and smaller-sized dots.

Our prior research with these data has shown that all of these math tasks are significantly related, with correlations ranging from .4-.6 (Author, 2022). Accordingly, scores in these six math tasks were standardized and then averaged to create an early math skills composite (α=.84).

Minutes of Play

At the first wave, parents completed time diaries. Parent-reported time diaries provide a reliable measure of the duration of children’s activities, like time spent playing (Bachman et al., 2020). Ecological validity of time diary data on play is likely stronger than structured observations or experimental tasks because it captures naturalistic daily family rhythms and activities. Time diaries can also minimize recall bias because participants report on a full day of occurrences instead of being prompted to report on one prioritized activity. Also, time diaries were collected on the next day, so participants’ memory errors decrease in comparison to retrospective surveys that inquire about the time spent in activities over the prior week or month.

To measure play time, we used modified ATUS codes to more precisely capture the activities in which four-year-old children may engage, including play. Our play codes included broad categories of play at home, including minutes children spent engaging in arts and crafts play, music play, block play, building, puzzles and board games, dramatic/pretend play, gross motor play, fine motor play, and outdoor/nature/science play. We chose to collapse these play types for multiple reasons. First, it increased variability in the play time measure, and some categories had too few endorsements to estimate reliable associations. Second, in many instances, children were engaged in several types of play at one time (e.g., playing “sailors and pirates” outside involves pretend play, gross motor play, and outdoor play). Third, our reliance on parent-reported time diaries means our time diary variables were dependent on parents’ indication that children were playing and a description of the activities. Prior studies examining specific play types were based on researchers’ observations; we had limited information to categorize specific play. Lastly, initial analyses showed no systematic differences in correlations between types of play and child outcomes. Accordingly, we decided to examine play broadly as opposed to relying on specific categorizations. Time spent playing video games or other electronic media (e.g., iPads, etc.) was not included in our measure of play because prior research has shown negative associations between electronic media and children’s self-regulatory and cognitive skills (Kerai et al., 2022; MacGowan & Schmidt, 2021; Munzer et al., 2018; Radesky et al., 2014), and our initial analyses echoed these studies—time spent doing these activities had negative correlations to the child outcomes studied here.

We included minutes spent playing whether play was reported as the primary or secondary activity. To ensure inter-coder reliability, 20 percent of time diaries were double-coded (Chorney et al., 2015). The interclass correlations (ICC) across coders for the play measures showed excellent reliability, averaging 0.95 (Koo & Li, 2016). Our final measure of play time consisted of the sum of minutes of play time across both reported days and all play categories.¹ We transformed the measure of play time to reflect hours of play per day.

Child and Family Covariates

We included child and family covariates to control for their associations with self-regulation and early prereading and math skills. We controlled for child age at the time the outcome was assessed and sex assigned at birth (male reference). We also included several demographic characteristics that have been linked to children’s academic and behavioral skills (Brooks-Gunn & Duncan, 1997). These include child race/ethnicity (codified as a dichotomous variable indicating whether the child was a member of a racially/ethnically minoritized group) and marital status (dichotomous indicator representing that the respondent was married). We included family socioeconomic status (SES) as a control variable. Parents reported their total household income from various sources including wages/salary, Social Security, retirement accounts, and government assistance programs. They also reported the highest educational degree held by either of the child’s parent. To create the family SES variable, we standardized income and education and averaged them (Dickinson & Adelson, 2014). All of these variables were measured in the age-four parent questionnaire, with the exception of child age at the wave 2 assessment, which was drawn from the age-five questionnaire.

Lastly, and importantly, reduce omitted variable bias concerns, we controlled for children’s scores on the Primary Test of Nonverbal Intelligence (PTONI; Ehrler & McGhee, 2008) and the Developmental Vocabulary Assessment for Parents (DVAP; Libertus et al., 2013), both measured at age 4. The PTONI is a measure of children’s nonverbal intellectual ability that is appropriate from children as young as age three. At the wave 1 in-person visit, children were asked to examine pictures on a page and select the picture that did not belong. Items range in difficulty from lower order reasoning skills like visuospatial perception to advanced reasoning skills like sequential reasoning and categorical formulation. The standardized score was used as a control in models. Importantly, while some PTONI items require children to use visuospatial perception, they are not required to perform spatial transformations or use geometric knowledge (e.g., knowledge about angles, distance, etc.), which distinguishes this measure from our spatial math skills assessments. To measure language abilities, parents completed the Developmental Vocabulary Assessment for Parents (DVAP; Libertus, Odic, Feigenson, & Halberda, 2013) in the wave 1 questionnaire, which provides parents with a list of 212 words taken from the Peabody Picture Vocabulary Test (PPVT-4, Dunn & Dunn, 2007) and asks them to mark which words they have heard their child say. Concurrent and predictive validity of the DVAP has been established in samples of 2- to 7-year-old children (Libertus et al., 2013). The DVAP is scored by totaling the number of words that the parent indicates the child has said before. We included scores on the PTONI and DVAP as control variables to all models to ensure that any observed associations between play, self-regulation, and prereading and math skills were not driven by children’s underlying cognitive competencies or language skills, which are related to these characteristics (Becker et al., 2014b; Bull et al., 2011; McClelland et al., 2007).

Data Analysis

Data analyses, conducted in Mplus 8, aimed at testing the mediation model in Figure 1 using structural equation modelling methods. First, we tested whether four-year-old children’s play time predicted their self-regulation skills (“a” path). We predicted self-regulation with time spent engaged in play, controlling for children’s age, sex assigned at birth, non-verbal intelligence and vocabulary, and parental marital status, SES, and race/ethnicity. Next, we tested whether play time had direct links to children’s prereading and numeracy skills at age 5 by estimating similar regression models, but with play predicting prereading and math skills in separate models. All child and family covariates were included in these models. Our final aim was to test whether self-regulation functions as a pathway through which children’s play time relates to their prereading and math skills. To address this aim, we added our measure of self-regulation to both models predicting prereading and math scores with time spent playing to estimate associations between self-regulation and prereading and math (“b” path). Formal tests of mediation using the product of coefficients method with bootstrap confidence intervals were conducted to estimate the statistical significance of the indirect effects of play time on prereading and math skills operating through time spent in play (c’) (Hayes, 2021; Preacher & Hayes, 2004).

Figure 1.

Mediation Model

Our sample size of 128 required us to be parsimonious with the number of covariates included in our models. However, we tested some alternative model specifications with additional covariates to test the robustness of our results and strengthen the findings. Specifically, we re-estimated the mediational paths controlling for other factors that could be correlated with children’s play time, self-regulation, and prereading and math skills. First, to rule out concerns that any observed links between play, self-regulation, and achievement were an artifact of children’s time spent in preschool, we ran our models with a measure of the number of hours per week the child spent in center-based childcare as reported by parents in the questionnaire as an additional covariate. Second, drawing data from the time diaries, we also controlled for the total time children spent with the parent during the two days captured to ensure that observed links between play time and outcome variables were not driven by correlations between increased play time and increased time spent with parents engaging in enriching interactions. Lastly, we added time spent doing academic enrichment activities over the two days sampled, also drawn from time diary data, as a control variable to determine whether associations between time spent playing, self-regulation, and achievement are spurious and explained by the fact that children who spend more time playing also engage in more academically enriching activities or that children with more self-regulation do more academic activities, thus exploring the concern that the time spent in academic enrichment rather than play drives increased self-regulation or academic enrichment rather than self-regulation that drives higher prereading and math scores.

Results

Descriptive statistics for the full sample are presented in Table 1. These statistics show that our analytic sample was diverse in terms of sociodemographic characteristics. Almost 1/5 of the children in this study were of racially/ethnically minoritized backgrounds and the sample was split almost evenly by sex assigned at birth (49.6% male). There was also considerable variability in the composite SES measure. With respect to play time, on average, children played for approximately 1.05 hours per day, though play time ranged from 0 hours to 3.55 hours per day. Additionally, Table 2 shows pairwise correlations between our variables.

Table 1.

Descriptive Statistics (N = 128)

Variable	Mean or %	SD	Max-Min
Child Outcomes
Self-regulation	−0.03	0.82	−1.95–1.26
Early reading skills	−0.12	0.94	−1.81–2.57
Early math skills	−0.10	0.75	−1.77–2.00
Play Time (hrs. per day)	1.05	0.77	0–3.55
Child/Family Covariates
Non-verbal cognitive skills	108.87	21.83	23–149
Vocabulary	104.59	29.01	11–173
Child age at wave 1 (yrs.)	4.39	0.30	4–4.96
Child age at wave 2 (yrs.)	5.47	0.32	5.01–6.39
Child sex assigned at birth: male	49.6%
Racially/ethnically minoritized child	18.1%
Responding parent is married	78.1%
Family SES	0.04	0.81	−2.39–1.39

Table 2.

Pairwise Correlations between Variables

	Self-reg.	Reading	Math	Play time	Cog. skills	Vocab.	Wave 1 age	Wave 2 age	Male	Minoritized	Married
Self-reg.
Reading	0.58*
Math	0.68*	0.76*
Play time	0.22*	0.15	0.26*
Cog. skills	0.59*	0.53*	0.52*	0.002
Vocab.	0.34*	0.44*	0.49*	0.11	0.35*
Wave 1 age	0.26*	0.07	0.28*	−0.18*	0.19*	0.16*
Wave 2 age	0.21*	0.04	0.19	−0.06	0.19	0.08	0.91*
Male	−0.24*	−0.11	−0.06	0.06	−0.27*	−0.02	−0.06	−0.04
Minoritized	−0.29*	−0.12	−0.24*	−0.26*	−0.19*	−0.24*	0.08	0.08	−0.01
Married	0.26*	0.24*	0.35*	0.26*	0.27*	0.38*	−0.13	−0.13	0.14*	−0.29*
SES	0.37*	0.32*	0.36*	0.16*	0.34*	0.32*	−0.13	−0.23*	0.01	−0.29*	0.71*

Degrees of freedom range from 76–125

Association Between Play Time and Children’s Self-Regulation

Results from our analyses linking time spent in play with children’s self-regulatory skills (“a” path in Fig. 1) show that, controlling for children’s non-verbal intelligence and vocabulary, age, sex assigned at birth, and several family covariates, play is positively related to children’s self-regulation (Table 3). For every additional hour spent playing, self-regulatory skills increased by .20 points (or .24 of a standard deviation (SD)). Children’s non-verbal intelligence and age were also positively linked to self-regulation.

Table 3.

Results from Regression Predicting Self-Regulation with Play Time and Covariates

	Self-Regulation
	Coeff.
	(SE)
	β
Play time (hrs. per day)	0.21**
	(0.08)
	0.20
Non-verbal cognitive skills	0.02***
	(0.004)
	0.41
Vocabulary	0.001
	(0.002)
	0.03
Child age (wave 1)	0.51*
	(0.21)
	0.19
Child sex assigned at birth: male	−0.18
	(0.14)
	−0.11
Racially/ethnically minoritized child	−0.30
	(0.19)
	−0.14
Responding parent is married	−0.11
	(0.21)
	−0.06
Family SES	0.24*
	(0.11)
	0.24

N=128.

^***p < .001

^**p < .01

^*p < .05.

Associations Between Play Time and Children’s Early Prereading and Math Skills

Next, we examined direct associations between play time and prereading and math (“c” path in Fig. 1). Play time directly related to children’s math skills, but not prereading skills (Table 4). Controlling for our host of covariates, including non-verbal intelligence and vocabulary, increases in time spent in play were related to higher math skills. Specifically, for every one hour increase in play time, there were associated increases of .21 points on the math composite. This constitutes an increase in early math skills of .28 SD per hour of play. While the direct link between play time and prereading was not significant, play time could still have indirect effects on prereading skills via self-regulation, which we addressed in the next aim.

Table 4.

Results from Regression Predicting Reading and Math at Age Five (Wave 2) with Play Time and Covariates

	Reading Skills	Math Skills
	Coeff.	Coeff.
	(SE)	(SE)
	B	B
Play time (hrs. per day)	0.17	0.21**
	(0.11)	(0.08)
	0.14	0.21
Non-verbal cognitive skills	0.01**	0.01**
	(0.005)	(0.004)
	0.34	0.30
Vocabulary	0.01*	0.01*
	(0.004)	(0.002)
	0.24	0.21
Child age (wave 2)	0.02	0.44*
	(0.26)	(0.20)
	0.01	0.18
Child sex assigned at birth: male	0.03	0.08
	(0.18)	(0.13)
	0.02	0.05
Racially/ethnically minoritized child	0.12	−0.18
	(0.23)	(0.17)
	0.05	−0.09
Responding parent is married	−0.09	0.08
	(0.34)	(0.24)
	−0.04	0.04
Family SES	0.24	0.20
	(0.17)	(0.12)
	0.21	0.20

N=128.

^***p < .001

^**p < .01

^*p < .05.

Indirect Effects of Play Time on Prereading and Math Through Self-Regulation

Our final aim was to test whether play time had indirect effects on children’s prereading and math skills at age 5 that operated through its association with children’s self-regulatory skills. Table 5 illustrates how time spent in play and self-regulation predict prereading and math skills controlling for all covariates (“b” path in Fig. 1). Self-regulation was positively related to both prereading and math skills, with effect sizes ranging from moderate for prereading to large for math. Each SD increase in children’s self-regulatory skills predicted increases in prereading skills of .34 SD and math skills of .83 SD. Links between play time and prereading and math skills were reduced when self-regulation was added to models. Play time was no longer significantly related to prereading and math once we added self-regulation to the models. Formal tests of mediation showed that self-regulation at age four significantly mediated links between time engaged in play and both prereading (coeff.=.08, SE=.04, p = .03) and math skills (coeff.=.10, SE=.04, p = .02) measured one year later (c’ path in Fig. 1).

Table 5.

Indirect Effects of Play Time on Early Reading and Math Via Self-Regulation

	Reading Skills	Math Skills
	Coeff.	Coeff.
	(SE)	(SE)
Play time (hrs. per day)	0.08	0.14
	(0.11)	(0.08)
Self-regulation	0.45***	0.46***
	(0.13)	(0.09)
Non-verbal cognitive skills	0.01	0.004
	(0.01)	(0.003)
Vocabulary	0.01*	0.01*
	(0.003)	(0.02)
Child age (wave 2)	−0.18	0.32
	(0.25)	(0.20)
Child sex assigned at birth: male	0.12	0.16
	(0.17)	(0.12)
Racially/ethnically minoritized child	0.25	0.003
	(0.23)	(0.16)
Responding parent is married	−0.04	0.04
	(0.33)	(0.20)
Family SES	0.13	0.12
	(0.17)	(0.11)
Indirect effect of play time via self-regulation (c’)	0.08*	0.10*
	(0.04)	(0.04)
	p = .03	p = .02

N=128.

^***p < .001

^**p < .01

^*p < .05.

Sensitivity Analyses

To test the robustness of our findings, we estimated some alternative models that included other predictors (results presented in supplementary materials). First, we tested whether the observed associations were driven by children’s attendance in preschool (model 1). There was no evidence that this was the case. Links between play time and self-regulation were unchanged by the addition of hours in preschool. Nor did preschool attendance weaken links between self-regulation and prereading and math skills. Next, we tested whether time spent playing was still associated with self-regulation after controlling for the amount of time the child spent with a parent (model 2). The inclusion of this variable in the models did not weaken the associations between play and self-regulation or self-regulation and prereading and math. Lastly, there was no evidence that time spent doing academic activities was driving the observed mediation (model 3). Play time was still significantly related to self-regulation with academic enrichment time included in models, and academic enrichment time was not related to age 5 prereading and math. Indeed, none of the additional variables tested were significantly related to either self-regulation or prereading and math skills.

Discussion

Play, a fundamental part of childhood, has been touted as central to children’s learning and development (Gray, 2011; Hirsh-Pasek et al., 2003; Singer et al., 2006). Yet few studies explore how children’s play time, in naturalistic settings, relates to their development. In a diverse sample of 128 preschool-aged children, this study uses parental time-diary reports on children’s time spent playing and direct assessments of children’s self-regulatory and early reading and math skills to provide evidence of associations between play time, self-regulation, and prereading and math skills. Results showed that self-regulation is a pathway through which increased play time relates to more well developed prereading and math skills at age 5. This is an important first step to guide future research aimed at exploring the effects of play time and its usefulness as a lever to promote school readiness skills.

We found that the amount of time children spent engaged in play was related to their self-regulation as well as early math skills. Specifically, for every additional hour that children spent playing per day, self-regulatory skills were .20 SD higher. Given that children’s play time decreased from the early 1980’s to the late 1990’s by slightly more than one hour per day (Elkind, 2008; Hoefferth & Sandberg, 2001), these findings have real implications for the self-regulatory skills of young children today, especially if this trend in decreased play time continued through present day. The association observed between play time at age four and math skills at age five was explained by play time’s relation to self-regulation at age four. Moreover, play time had indirect effects on both prereading and math skills at age five via self-regulation. Indeed, results showed strong links between self-regulation measured at age four and preacademic skills assessed one year later, especially with respect to early math skills. These associations between play time, self-regulation, and prereading and math were robust to the inclusion of several child and family covariates, including a measure of children’s non-verbal intellectual capabilities and their vocabulary, which strengthens our confidence that observed associations between play time and development are real and not driven by some unmeasured characteristics of children or parents (e.g., Cameron & Heckman, 2001; Jaffe et al., 2011; Sattler, 1974).

Findings that the minutes engaged in play relates to important self-regulatory skills supports long-standing developmental psychological theory that has argued for the importance of play time in cognitive development, and in particular the development of self-regulation (Gray, 2011; Vygotsky, 1978; Zigler et al., 2004). Results also help bridge the literature gap between time diary studies illustrating declines in play time for American children and lab-based studies that link specific types of play, such as guided play or dramatic play, with children’s self-regulation. We found robust links between children’s play time and their early self-regulation, and through this link play time was associated with preacademic skills. These findings shed light on the developmental processes driving relations between naturalistic, child-driven, open-ended play and child development, as well as potential consequences of diminishing playtime for children.

The findings also have implications for theoretical arguments of the importance of play. While Vygotskian and other theories of play often rely on adult guidance as a key driver of the acquisition of skills through play (Hakkarainen & Bredikyte, 2008; Slot et al., 2017; Weisberg et al., 2016), more recent theories of play and development focus less on specific characteristics of play, like adult scaffolding or amount of symbolization or metacommunication in play, and more on a broad measure of unstructured, natural play (Burdette & Whitaker, 2005; Gray, 2011). These theories argue that play of any type, that it is freely chosen and child-directed, intrinsically motivated, imaginative, and pleasurable, may have benefits for children’s self-regulatory skill development (Burdette & Whitaker, 2005; Gray, 2011). Unstructured, child-directed play allows children to make decisions, gain a sense of agency and direct their own actions, solve problems, exert self-control, and follow rules in spite of their own desires (Gray, 2011). Studies like the present that explore the time children spend playing, defined broadly, in naturalistic settings are useful to support theories of the benefits of children’s free play. Our findings that a broad measure of time spent playing at home, regardless of specific characteristics, was positively related to children’s self-regulatory skills, and indirectly, their preacademic skills suggest benefits for a broad range of play. Further, results suggest that underlying developmental processes explaining benefits of play are shared across many forms of self-directed play. To see positive impacts of play, it may not be necessary for parents and educators scaffold or manage all of children’s play. Moreso, it may be important to provide children with space, materials, and time for child-directed play.

Strengths and Limitations of the Present Study

Before discussing potential implications of and future directions for this work, it is important to state its strengths as well as its limitations. First, this work improves upon the prior studies that made little attempt to control for omitted variables. By controlling for children’s non-verbal intelligence and vocabulary, we statistically account for the role of children’s general cognitive and verbal abilities on their self-regulatory skills and early prereading and math. These are robust covariates since their inclusion reduces the likelihood that observed associations between time spent engaging in play, children’s self-regulation, and their preacademic skills are driven by children with higher abilities, which may be related to their sustained engagement in play as well as support their self-regulatory skills.

Second, we measured play time using time diary data—the use of which is both novel to the literature on play and a particularly strong method of measuring children’s and families’ activities. A self-completed time diary is considered to be the most reliable and accurate method for obtaining information on time-use of study participants (Chatzitheochari et al., 2018; Michelson, 2005; Robinson & Godbey, 2010). The time diary covers a full 24-hour day, and respondents provide a sequential account of all daily activities, which corresponds to the way daily events are stored in memory, thus increasing the validity of the obtained data (Robinson & Godbey, 2010). Using parent-reports of play time gleaned from the time diaries provided us with a highly reliable and valid measure of children’s natural play in the context of their daily lives. Engagement in play was not observed in a lab or in another unnatural setting nor was it retroactively reported over longer periods of time (e.g., play time during the last month). This provides a better picture of the potential consequences of decreases in children’s play time that have been documented in recent decades.

There were also limitations of our measure of play time. First, this study included play occurring when children were at home only. Critically, we do not know anything about children’s play time during preschool or times that children spent in care without the target parent. Over 90% of our sample experienced some non-parental care, and more than half experienced regular non-parental care for 20 or more hours per week, which means we are missing part of the picture of children’s daily lives. We also do not have information regarding whether children played alone or with an adult or other child. Studies document the benefits of guided play, that is play that combines the child-directed nature of free play with adult participation and scaffolding toward a learning objective (Weisberg et al., 2016). Similarly, there is evidence that play time may be more beneficial to children when engaging with another person, including adults, siblings, or peers (Jent et al., 2011). It may be important to separate guided play or play with adults from play alone, with peers, or other children, and our inability to parse play partners prevents us from directly testing this in the current study. This is a good direction for future work.

Lastly, it must be acknowledged that this study is correlational, and thus, we cannot conclude from the results that children’s play time has a causal effect on either self-regulation or early reading or math skills. Testing the effects of play time in an experimental framework can be difficult since the measurement of children’s naturalistic play is not conducive to a laboratory setting. Evaluation studies of play-based preschool curricula may provide one way to test the effects of children’s play in an experimental framework. The use of highly rigorous quasi-experimental data analytic techniques is another avenue to strengthen causal links between play time and children’s development (Miller et al., 2016). Second, it remains possible that the links between play and self-regulation and achievement represent a substitution effect. In other words, it may not be time spent playing that relates to improved self-regulation and in turn preacademic skills, but what children avoid doing when they are playing instead. For instance, if children are playing instead of watching television, and television watching is negatively linked to self-regulation or preacademic skills (Shin, 2004), it may not be more play time that relates to better self-regulation or achievement but reduced television watching instead. While this was not an aim of this paper, exploratory analyses of these data suggest this is likely not the case here. Of the other categories of time use, time spent watching television and running errands were both negatively correlated with child outcomes, but watching television is positively correlated with play time and running errands is unrelated.

Also, the correlational nature of these data raises the possibility of simultaneity bias (Miller et al., 2016). This study cannot determine whether play time increases children’s self-regulation or whether children with higher regulatory skills engage in play more frequently. To support our theory that play affects self-regulation, however, we draw from intervention studies that included play as an element to increase children’s self-regulation or executive functioning skills (Diamond et al., 2007; Hillman et al., 2014; Mulvey et al., 2018; Schmitt et al., 2018). These studies, which include Tools of the Mind involving dramatic play and interventions like Successful Kinesthetic Instruction for Preschoolers (SKIP), which involve physically active, gross motor play, have shown positive effects of interventions featuring play (Diamond et al., 2007; Hillman et al., 2014; Mulvey et al., 2018). To add to this evidence of causal connections, future observational studies should use longitudinal information on play time and self-regulation to test whether time spent playing at the earlier time point predicts changes in self-regulation from the earlier to the later time point (Miller et al., 2016). Unfortunately, the direct self-regulation assessment was not given at the second wave, so we cannot explore this particular issue here.

Implications of this Work and Future Directions

Acknowledging the limitations of this study, the results may have implications for early childhood researchers, policymakers, and practitioners concerned with promoting children’s school readiness skills. If indeed children’s play time is important for enhancing children’s self-regulation then policies, procedures, and routines (both within family and preschool settings) that encourage play may be warranted. Self-regulation is a vital skill that undergirds learning and social development (McClelland & Cameron, 2011). The benefits of self-regulation extend far beyond childhood; self-regulatory skills developed in early childhood appear to increase the capacity of individuals to take advantage of good opportunities or avoid harmful opportunities presented later in life (e.g., Blair, 2010; Heckman, 2006, 2007). Indeed, stronger self-regulatory skills promote adult socioeconomic attainments, health, and behavioral adjustment, including postsecondary educational success, employment, physical and mental health. Additionally, more advanced self-regulatory skills reduce the risk of substance dependence and criminality (McClelland et al., 2013; Moffit et al., 2011; Robson et al., 2020). Early childhood is an ideal moment to intervene to boost self-regulation because self-regulatory skills undergo rapid change from 2 to 6 years of age (Montroy et al., 2016). More research focused on the design, implementation, and evaluation of developmentally appropriate play interventions deserves our attention.

It is also worth considering whether the short-term benefits of increasing academic instruction at home and/or in preschool if at the expense of play time, which seems to be happening in early care and education settings in the U.S., may have long-term negative impacts on children’s school performance. Studies have shown that the academic benefits of center-based preschool programs fade out by first or second grade (Barnett, 1995; Bassok et al., 2015; Currie, 2001; Puma et al, 2010). Analyses of existing preschool studies show that end-of-treatment preschool impacts on achievement decline by about 50% the year following preschool, and then by another 50% over the next two years (Aos et al., 2004; Li et al., 2020). One line of research into the cause of this fade-out suggests that elementary school contexts do not adequately support students’ prior academic gains or meet students’ instructional needs (Jenkins et al., 2018), but the decline of opportunities for play time in preschool may be another promising explanation for diminishing returns from early education over time. Preschool programs that focus on didactic instruction unquestionably produce gains in discrete skills, like letter and number identification, but if the programs do not have a component aimed at improving children’s self-regulation (or alternatively, if they have supplanted play time with didactic instruction, which may have negative implications for children’s self-regulation according to our results) these gains may not be sustained when children are faced with more difficult academic work that requires increased regulatory demands (Almon & Miller, 2011). This is not to say preschool or pre-k programs are harmful to children; abundant research shows that high quality preschool programs boost early cognitive ability and achievement, especially for at-risk children (Duncan & Magnuson, 2013; Yoshikawa et al., 2013). More research is necessary, however, to determine whether and how preschool curricula should be infused with ample opportunity for children to engage in child-directed play that promotes both early preacademic skills and self-regulation, which will support not only children’s long-term achievement but also their wellbeing in other domains (McClelland et al., 2010; Moffit et al., 2011; Robson et al., 2020).

In a notable result, time spent doing academic activities was unrelated to prereading and math skills. There are a couple of possible explanations for this. First, our measure of academic activities did not distinguish between different types of activities or the quality of these activities. Previous research has shown that specific types of math activities are linked with children’s skills in specific subdomains of math such that e.g., greater engagement in patterning activities are associated with better patterning skills (e.g., Leyva et al., 2021). Moreover, other research has shown that formal and informal academic activities relate to children’s learning in different ways (e.g., Ramani & Siegler, 2014; Skwarchuk et al., 2014). Thus, the broad category of academic time captured by the time diaries may be obscuring selective associations between certain types of activities or outcomes. Second, it is possible that many parents whose children have below average prereading and/or math levels are engaging their children in more academic activities to improve their skills. In this case, academic activities and prereading and math skills would be negatively correlated solely due to the low baseline skills of the children engaging in above-average academic activity time. Thus, associations between academic activities and early reading and math skills may be non-linear or only be significant in subgroups of children (Elliott et al., 2022).

Putting achievement aside, reductions in play time to increase formal instruction occurring in preschool programs may be developmentally inappropriate, and furthermore may have negative implications for children’s self-regulatory and behavioral competencies (Almon & Miller, 2011). In a study with nearly 4,000 preschool teachers, researchers found that preschoolers were being expelled at three-times the rate observed in K-12 classrooms (Gilliam, 2005). These data further showed negative correlations between play time in preschools and expulsions; less play time was related to more expulsions (Gilliam, 2005). An evaluation of the effects of a preschool intervention found that the negative implications of supplanting play with direct instruction follow children through adulthood (Schweinhart & Weikart, 1997). The High Scope Preschool Curriculum Comparison Study randomly assigned low-income children to different classrooms; 1) a direct instruction classroom; 2) a play-based classroom; and 3) a classroom employing the “High Scope” curriculum where children learned through guided play in which teachers observe, support, and extend children’s self-directed play. At age 23, compared to students in the play-based and High Scope classrooms, participants that had experienced the direct instruction had significantly higher arrest rates and were more likely to be fired from a job and had been in special education classes during school (Schweinhart & Weikart, 1997). These studies provide additional evidence supporting the continued research attention on the role of play in fostering behavioral and regulatory skills.

Conclusion

Play is so important that it has been recognized as a basic human right of every child (Office of the United Nations High Commission for Human Rights, 1989). Yet play time has been decreasing in the U.S. and we need additional empirical research illuminating potential effects of declines in opportunities for children to engage in play. Using parental reports of children’s daily activities and direct assessments of children’s skills, this study found positive links between time spent playing at home to self-regulation and, indirectly, their early prereading and math skills a year later. Links between play time and preacademic skills were explained by the association between play time and self-regulation. These results suggest that play may be an important tool in efforts to boost children’s school readiness skills in the short term and their healthy development in the long run.