Exploring the Features of the Self-Regulatory Environment in Kindergarten Classrooms

Abstract

Self-regulatory skills are increasingly recognized as critical early education goals, but few efforts have been made to identify all the features of the classroom that actually promote such skills. This study experiments with a new observational measure capturing three dimensions of the classroom environment hypothesized to influence self-regulation: classroom management, emotionally supportive interactions, and direct promotion of self-regulatory skills. These classroom dimensions were tested as predictors of change over the kindergarten year in both self-regulatory and academic skills in a sample of racially/ethnically-diverse low-income children in Tulsa, OK. Results showed that classroom management was associated with small gains in one of four measures of self-regulation, and four of six measures of academic skills. The other dimensions of the environment had weak or no associations with outcomes. These results indicate that further work is needed to refine both models and measures of the self-regulatory environment.

The kindergarten classroom environment is arguably one of the most important contexts in which children learn and grow. For many children, it is their first formal classroom experience and the launchpad for 13 years of K–12 education. As a result, decades of research have focused on understanding which features of the kindergarten classroom environment predict key child outcomes. Disproportionately more attention has been paid to features of the kindergarten instructional environment, and their ability to predict growth in academic skills, than to features of the classroom environment that may predict growth in other critical domains of development, such as self-regulation. Recently, however, there has been an explosion of interest in self-regulation and the environments that promote it.

Self-regulation encompasses a variety of skills that enable children to direct their own thoughts, emotions, and behaviors (Nigg, 2017). Among these are the traditional measures of executive function, namely attentional control, inhibitory control, and cognitive flexibility, as well as skills that display adaption to the demands of specific contexts (McClelland & Cameron, 2012; Ursache et al., 2012). In the classroom, these skills include, for example, waiting to take a turn, following directions, and staying seated. Self-regulatory skills accelerate in early childhood and are thought to be foundational because they predict high school and college success and better adult employment, health, and social outcomes (Breslau et al., 2009; McClelland et al., 2013; Mischel et al., 1989; Robson et al., 2020).

While self-regulatory skills are crucial in their own right, they are also thought to facilitate the acquisition of academic skills in early and middle childhood, making them an important priority for educators and scholars alike (Blair & Razza, 2007). It is commonly thought that the skills marshalled in the service of controlling thought, attention, emotion, and behavior enable children to attend to and remember lessons, switch attention across tasks, persist in challenging situations despite frustration, and think creatively to solve academic problems. Indeed, self-regulatory skills in preschool and kindergarten have been found to predict gains in language, literacy, and math achievement during the elementary school years (e.g., Blair & Razza, 2007; Brock et al., 2009; Duncan et al., 2007; Howse et al., 2003; McClelland & Wanless, 2012; McClelland et al., 2014; Welsh et al., 2010). Furthermore, it appears that growth in language (Bohlmann et al., 2015; Cadima et al., 2019), literacy (McClelland et al., 2014), and math (McClelland et al., 2014; Schmitt et al., 2017; Welsh et al., 2010) during preschool has a reciprocal beneficial effect on self-regulatory skills.

And yet, because features of the classroom that promote young children’s self-regulatory skills—referred to here as the “self-regulatory environment”—have received comparatively less attention than features that promote academic skills, there remains a paucity of research on which specific aspects of the kindergarten classroom can be harnessed to promote this goal. To be sure, instruction is certainly facilitated—and academic outcomes enhanced—in a classroom characterized by predictability, organization, and discipline—that is, a well-regulated classroom (Pianta, Hamre, & Allen, 2012). Yet there are other features of the classroom environment that may be particularly, or even exclusively, salient for fostering children’s burgeoning self-regulatory skills. These include the behaviors taught in interventions with teachers designed to promote student self-regulation, such as those surrounding the prevention of, and responses to, emotional distress and peer conflict. Scholars are now urgently calling for an instrument that adequately captures all aspects of the classroom thought to shape self-regulation (Phillips, 2023). This study attempts to answer this call by testing a new measure of the self-regulatory classroom environment and its association with children’s change in self-regulatory and academic skills over the kindergarten year.

The Classroom as a Self-Regulatory Environment

As the developmental importance of early self-regulation has become increasingly clear, education researchers have sought to identify the features of the classroom in early childhood that might be targeted to promote these skills. Barnes, Abenavoli, and Jones (2022) recently identified three pillars of the classroom as key ingredients of an environment that enables young children’s social and emotional learning. The first is classroom management, a function of the teacher’s organization and planning strategies for sustaining students’ attention and positive engagement in classroom activities. Such strategies include minimizing time spent on transitions between activities, the use of predictable daily routines, preparation before lessons, and the effective inclusion of children with differing abilities in classroom activities. The second pillar is emotionally supportive interactions between teachers and children. These are infused with teachers’ positive affect and support for students’ autonomy. The third pillar is the direct promotion of children’s social and emotional skills through validation of emotions, promotion of emotion awareness and expression, and assistance with strategies for emotional control and conflict resolution, particularly with peers.

All three of these pillars are directly relevant to the early emergence of self-regulatory skills in the classroom, and they may also support children’s academic skills. For example, by minimizing unpredictability and distraction, effective classroom management allows children to practice sustaining their attention on relevant stimuli and formulating and enacting goal-oriented behavior. Emotionally supportive interactions allow children to remain calm and focused enough to practice higher-order cognitive problem-solving skills instead of maintaining vigilance for potential threats and negative arousal. Direct promotion of social and emotional skills teaches children useful strategies to manage their emotions and behaviors, such as naming and expressing emotions, considering how their actions affect others, and identifying activities that allow them to cool off or feel comforted when in distress. Together, these facets of the classroom may be conceived of as contributing to an environment that fosters self-regulation—or what we refer to here as the self-regulatory environment. Such an environment may also promote academic skill growth; for instance, a well-regulated classroom that minimizes unpredictability and distraction should pave the way for more effective delivery and absorption of academic content (Pianta et al., 2012). A classroom that supports children’s emotion regulation should reduce the time and energy children might allocate away from academic learning to managing challenging emotions and peer interactions, and allow them to instead focus their attention on instruction. Yet to date no single study has evaluated these self-regulatory classroom environment features simultaneously as predictors of early self-regulatory and/or academic development.

Measuring the Self-Regulatory Environment

Past research on classroom characteristics that foster self-regulation has had two significant limitations. First, much of this research is piecemeal, focusing on only one or two characteristics at a time. For example, emotional tone (Fuhs et al., 2013) and support for children’s autonomy (Cadima et al., 2019) have both been found to facilitate self-regulation in the classroom, but not in the same study, although they may well covary. Second, nearly all studies that have measured multiple classroom characteristics simultaneously have relied on one or two scales from the CLASS (Pianta et al., 2008), an observational measure designed to quantify global classroom quality. The CLASS Classroom Organization scale relates closely to Barnes et al.’s construct of classroom management, and assesses the teacher’s behavior management techniques, transitions between activities, preparedness, and routines. Several studies have found that students in pre-k and kindergarten classrooms scoring higher on this scale show greater improvement in self-regulation (Hamre et al., 2014; Nguyen et al., 2009), but others have had null findings (Guerrera-Rosada et al., 2021; McDoniel et al., 2022; Moffett et al., 2021). The evidence is similarly mixed with respect to the CLASS Emotional Support scale. This scale relates closely to Barnes et al.’s emotionally supportive interactions construct, and assesses the teacher’s sensitivity, support for autonomy, and positive and negative affect. One study finds that in classrooms scoring higher on Emotional Support children show greater improvement in self-regulation (Hatfield et al., 2022), but another does not (McDoniel et al., 2022). Thus to date there is only mixed evidence to support these two CLASS scales as valid measures of the classroom self-regulatory environment.

In addition, neither CLASS scale captures the third pillar of the self-regulatory environment: direct instruction in self-regulatory skills. Evaluations of classroom interventions to foster students’ self-regulation have shown that teacher training and specialized curricula can result in steeper self-regulatory growth in preschool and early elementary school. For example, the Head Start REDI intervention instructed preschool children in prosocial skills such as sharing and taking turns; taught children how to recognize and express emotions; and shared strategies for calming down when upset (Bierman et al., 2008). Games and projects allowed children to practice these skills, and teachers were instructed on how to coach children. The REDI intervention boosted growth over the school year in children’s cognitive flexibility and task orientation; task orientation, in turn, mediated effects of the intervention on selected language and literacy skills. The Tools of the Mind curriculum taught preschoolers how to tell themselves out loud what to do and trained them directly in memory and attention strategies (Diamond et al., 2007). Tools of the Mind resulted in higher scores on tasks tapping attentional and cognitive control. The Incredible Years Dinosaur curriculum, which trained preschool through first grade teachers to enhance children’s persistence, social problem-solving, anger control, and cooperation, succeeded in improving on-task behavior in the classroom among children who began the year above-average (Webster-Stratton et al., 2008). Yet the teacher behaviors targeted by these three interventions—emotional labeling and validation, instruction in self-regulatory strategies, and guided prosocial activities—are not explicitly measured by either CLASS scale.

The Adapted Teacher Style Rating Scale (ATSRS) was designed by Raver et al. (2012) precisely to capture teacher behaviors and practices explicitly aimed at fostering student self-regulation. In addition to a Classroom Structure and Management scale, it also includes a scale called Supporting Social and Emotional Skills, which has been used to measure the effectiveness of interventions with teachers designed to improve their students’ self-regulation (Mattera et al., 2013; Morris et al., 2014). For example, it captures teachers’ encouragement of student strategies for self-calming rather than scolding or giving a time-out to children in distress. To date, however, no one has tested the predictive validity of the ATSRS as a measure of the self-regulatory environment by attempting to demonstrate its associations with child self-regulation skills.

Additionally, while the ATSRS captures two of the three pillars proposed by Barnes et al. (2022)—classroom management and the direct promotion of self-regulatory skills—it neglects emotionally supportive interactions, the remaining pillar and the focus of much recent scholarship on young children’s self-regulation in the classroom (Murray et al., 2019; Rosanbalm & Murray, 2018). Emotionally supportive interactions are conceptualized by Barnes et al. (2022) as being grounded in attachment theory, and reflective of the “student-teacher interactions characterized by warmth, sensitivity, and support for children’s autonomy” (p. 205) that facilitate positive teacher-child relationships. They thus include a focus on dyadic teacher-child interactions and on interactions that are designed to praise, disapprove of, or redirect a child’s behavior. In this way, they are distinct from classroom management strategies that are more likely to be classroom-wide and emphasize broad rules and routines. As such, classroom management strategies may also be more affectively neutral for individual children, for whom a direct behavioral instruction (positive or negative) is likely to carry greater emotional salience, for better or for worse.

Indeed, accumulating research suggests the unique potential for negative teacher-child interactions that involve harsh punishment, use of sarcasm, and overcontrol of child behavior to undermine early self-regulation (Brendgen et al., 2006; Degol & Bachman, 2015; L’Ecuyer et al., 2021). The literature on parenting has similarly identified “frightening” behaviors (Dozier, 2019), such as yelling at or physically restraining children, that actually disrupt children’s ability to self-regulate by causing excessive stress and anxiety (Dozier et al., 2002; Gunnar et al., 2011; Schuengel et al., 1999). In an earlier study conducted with the sample used here, it was found that even the use of one “red flag” punitive behavior (e.g., yelling, eye rolling, making derogatory comments about a child, ignoring a child in emotional need) among preschool teachers predicted lower gains on self-regulation at year’s end as measured directly, reported by teachers, and reported by observers (Phillips et al., 2022). Notably, these results emerged in models controlling for the classroom’s scores on the CLASS Emotional Support and Classroom Organization scales.

The ATSRS, the only existing instrument designed to measure the classroom as a self-regulatory environment, has only two items tapping these positive and negative emotionally-laden teacher-child interactions, and they are subsumed within the Classroom Structure and Management subscale. Fortunately, new ways to identify and measure features of the self-regulatory environment are currently under development (Phillips et al., 2023). Bardack and Obradovic (2019) created the Teachers’ Displays and Scaffolding of Executive Function (T-DASEF) Protocol, an observational measure of teachers’ promotion of executive function in third through fifth grade classrooms; they found that the promotion of executive functioning did not predict students’ spring executive function skills after controlling for their fall skills. Montoya et al. (2023) recently created observational scales of teachers’ scaffolding of self-regulatory skills in Chilean preschool classrooms, but they did not test for associations with children’s outcomes. Moffett et al (2021) examined pre-k teachers’ promotion of classroom organization drawing on the Individualizing Student Instruction (ISI) Coding System (Connor et al., 2009), an observational measure capturing how much time each individual student was exposed to the teacher’s organization strategies; results did not support an association between that measure and children’s change in academic achievement or executive function over the school year. Clearly, identifying dimensions of the self-regulatory classroom environment that predict self-regulatory skills is a challenge for the field. We take on that challenge, joining these efforts by adding a new measure for consideration to this burgeoning evidence base.

The Present Study

The present study adds to ongoing – and thus far, unsuccessful – efforts to identify features of the classroom environment that promote children’s early self-regulatory gains. We draw on a longitudinal study of children from low-income families in Tulsa, OK whose kindergarten classrooms were observed using three new observational tools capturing diverse features of the environment that may be expected to promote, or undermine, children’s self-regulatory skills. The first is the ATSRS (Barnes et al., 2022; Raver et al., 2012), which was originally designed to measure classroom-level impacts of social-emotional interventions such as the PATHS Curriculum on teacher behaviors. It includes many items describing classroom and behavior management, similar to the CLASS, but it also includes items describing the teacher’s explicit instruction in behaviors underlying effective self-regulation, such as emotion identification and labeling. The second observational tool used in kindergarten classrooms was the COPG/TOPG (Peabody Research Institute, 2017a, 2017b), which focuses the observer’s attention on both the teacher and individual children in the classroom, and includes particularly detailed measures of positive and negative teacher-child interactions aimed at managing a child’s behavior. The third is the Post Observation Rating Scale (Post; Farran, Meador, & Yun, 2014), which captures the observer’s overall impressions of teacher behavior, including their reliance on harsh, emotionally unsupportive discipline. We extracted selected items from the COPG/TOPG and Post and combined them with the ATSRS to yield a comprehensive measure of the self-regulatory environment, encompassing three subscales: Classroom Management, Emotionally Supportive Interactions, and Direct Promotion of Self-Regulatory Skills.

Our first research question is whether these three new subscales describing the self-regulatory environment, entered simultaneously in a single model, predict change in children’s self-regulatory skills between the fall and spring of the kindergarten year. To the extent that Classroom Management predicts child self-regulatory development, it will suggest that a well-regulated classroom—generally considered a key ingredient for academic growth (Pianta et al., 2008)—also promotes self-regulatory improvement. To the extent that Emotionally Supportive Interactions predicts self-regulation, it will suggest that the contribution of teachers’ interactions with individual children aimed at managing behavior may have been underestimated by previous tools. To the extent that Direct Promotion of Self-Regulatory Skills predicts self-regulatory development, it will suggest the value of specific guidance in self-regulation.

Our second research question goes on to ask whether the three subscales capturing the self-regulatory environment predict change in children’s academic skills between the fall and spring of kindergarten. Specifically, we test whether a higher-quality self-regulatory classroom environment boosts change in three measures of language/literacy (letter-word identification, expressive vocabulary, and sentence structure) and three measures of math knowledge (applied problem solving, numerical fluency, and conceptual counting) over the kindergarten year. If the self-regulatory environment is found to foster change in academic skills, its importance for young children’s developmental success will be even greater than previously understood.

The study design has several strengths. First, the sample is exclusively low-income. Because their families experience greater stress due to economic hardship, limited access to resources, and often discrimination, frequently across multiple generations, children who come from low-income backgrounds tend to enter school with fewer of the self- regulatory skills teachers value (Noble et al., 2005; Raver, 2012; Rimm-Kaufman et al., 2000), and lower academic scores (Duncan & Brooks-Gunn, 1997; Reardon & Portilla, 2016), than their higher-income peers. Thus there is an especially urgent need to identify the characteristics of classrooms that may be leveraged to improve outcomes in this population. Second, multiple modes of data collection were used to measure children’s self-regulatory skills. We directly assessed inhibitory control and cognitive flexibility in-person, asked teachers to report on the child’s self-regulatory behaviors in the classroom, and drew on reports from data collectors to describe the child’s self-regulatory behaviors during the assessment session. This approach accounts for the relatively weak overlap generally found in measures of self-regulation across sources due to variation in tasks’ demands, context, and emotional salience (Eisenberg et al., 2019). Our use of a broad array of classroom characteristics, self-regulatory assessments, and academic achievement should allow greater specificity in our understanding of how classroom processes translate into student success.

Methods

Participants

Data are drawn from the Tulsa School Experiences and Early Development (SEED) Study, an ongoing longitudinal study of the preschool and early elementary school experiences of children from low-income households in Tulsa, OK. It recruited children in the Tulsa Public Schools (TPS) district who were low-income (family had income below 185% of the federal poverty line or received public benefits in the last year) and were in one of three preschool arrangements at ages 4–5 (Head Start, TPS public pre-k, or parental/relative care). Children were enrolled at age 3 (n = 611; academic year 2016–2017), age 4 (n = 687; academic year 2017–2018) or kindergarten (n = 130; academic year 2018–2019). Recruitment strategies differed across years in order to target children from all three preschool arrangements (see Johnson et al., 2022, 2023). The University of Oklahoma-Tulsa IRB approved all study protocols.

An agreement with the TPS district allowed the study access to most schools in order to conduct observations of kindergarten classrooms in the spring of 2019, and to draw children out of class for individualized assessment sessions in the fall of 2018 and spring of 2019.

The current analysis draws on the 1,093 children who attended a kindergarten program in the TPS district. Of these, 244 children were dropped because their classrooms were not observed. The remaining 849 children were racially/ethnically diverse (Table 1). Nearly half (48%) the children were Hispanic/Latinx; 21% were Black (non-Hispanic); 14% were White (non-Hispanic); 9% were multiracial; and 7% were of another group. The sample was in general socioeconomically disadvantaged. About one-third (36%) of the children’s mothers had more than a high school education, and the average monthly household income was $2,013 (SD = 1,375). Eight percent of the children had an Individualized Education Plan (IEP).

Table 1.

Sample characteristics

	M (SD)	%	n
Covariates
Child’s age	5.57 (0.30)		849
Child is male		49%	849
Child race/ethnicity			849
Hispanic/Latinx		48%	849
Black, non-Hispanic		21%	849
White, non-Hispanic		14%	849
Multiracial		9%	849
Other		7%	849
Parent has more than HS degree		36%	700
Monthly household income ($)	2,013 (1,375)		551
Child is dual language learner		46%	849
Child has IEP		8%	843
Self-Regulation
Fall inhibitory control	61.92 (19.83)		801
Spring inhibitory control	68.86 (18.65)		826
Fall cognitive flexibility	63.32 (19.24)		803
Spring cognitive flexibility	69.90 (19.82)		825
Fall teacher-rated BRIEF	71.67 (16.75)		737
Spring teacher-rated BRIEF	72.21 (16.53)		764
Fall assessor rating	45.01 (7.88)		812
Spring assessor rating	46.33 (7.29)		831
Academics
Fall letter-word identification	11.63 (5.69)		792
Spring letter-word identification	18.39 (7.38)		833
Fall expressive vocabulary	12.58 (8.33)		807
Spring expressive vocabulary	16.70 (9.42)		837
Fall sentence structure	15.55 (4.98)		807
Spring sentence structure	18.34 (4.76)		837
Fall applied problems	14.57 (4.69)		791
Spring applied problems	18.02 (4.47)		832
Fall numerical fluency	5.33 (8.45)		765
Spring numerical fluency	13.13 (9.42)		817
Fall conceptual counting	0.41 (0.28)		801
Spring conceptual counting	0.67 (0.26)		836

Note. HS = high school. IEP = Individualized Education Plan; BRIEF = Behavior Rating Inventory of Executive Function.

The study sample differed demographically from the overall population of public school attendees in TPS that year insofar as it overrepresented Hispanic/Latinx students (48% vs. 35% in TPS; Oklahoma Office of Educational Quality and Accountability; edprofiles.info/report-card) and underrepresented White students (14% vs. 24% in TPS). Our sample also included a higher proportion of dual language learners (46% vs. 22% in TPS).

Procedures

Between January and March of the kindergarten year, a trained observer was sent to all study children’s classrooms to complete the COPG/TOPG, the ATSRS, and the Post. The observation period occurred between approximately 8:30 am and 1:00 pm.

To measure children’s self-regulation and academic achievement, trained assessors visited schools in the fall and spring of the kindergarten year to conduct individualized assessment sessions. The child was taken out of the classroom to a designated location in the school to meet with an assessor. If a child was a Spanish-speaking dual language learner (DLL), the assessor was bilingual in both English and Spanish. Sessions, each lasting about 35 minutes, were separated into two days to reduce student fatigue. Immediately after the child returned to their classroom, the assessor rated that child’s self-regulatory behaviors during the testing session. Finally, each fall and spring, teachers of study children were asked to rate each child’s self-regulatory abilities in the classroom, receiving $20–$60 in gift cards depending on the number of study children in their classroom.

Parents (specifically, children’s primary caregiver, which was typically the mother) were sent a survey in the spring of the kindergarten year in which they reported on family information included here as covariates. Surveys were distributed via email, text, and paper, and were available in English and Spanish. Parents received a $30 gift card for its completion.

Measures

Classroom Self-Regulatory Environment

Three observational tools were administered to measure the classroom self-regulatory environment. The first, the ATSRS, includes 13 teacher practices rated for frequency on a 5-point scale with item-specific anchors for each point. Some items capture behaviors that promote classroom order and predictability, such as preparation of materials before lessons, consistent use of rules and routines, and use of gestures and strategies to maintain student attention and engagement. Other items capture the strategies teachers use to promote children’s use of self-regulatory skills such as emotion awareness and management. For example, the item capturing social awareness measures how often the teacher draws students’ attention to the thoughts and feelings of others or the interpersonal consequences of their behavior (e.g., by asking a child to think about how they would feel if they were in a peer’s position), where 1 signifies not at all, 3 signifies occasionally, and 5 signifies consistently. An additional two items capture teachers’ display of positive and negative interactions to manage children’s behavior. Recent incarnations of the ATSRS have varied according to their construction of either two or three scales (Phillips et al., 2022; Rojas et al., 2022; Wright et al., 2023), following encouragement of the scale’s creators to adapt the tool to local conditions (Barnes et al., 2022).

Observers received one day of remote training by one of the tool’s developers (Dr. Stephanie Jones), and then conducted self-study and coded three practice videos. They were certified as reliable if they reached within 1-point of agreement on 80% of items with a gold-standard coder in two practice classrooms. Ongoing quality was assured by having gold-standard coders review all observations for consistency between notes and scores. In cases of disagreement, observers and gold-standard coders reached agreement through consensus.

The other tools administered were the COPG/TOPG, a revised version of the COP/TOP (Peabody Research Institute, 2017a, 2017b), and the Post (Farran, Meador, & Yun, 2014). Unlike the ATSRS, which observes the teacher in the context of the classroom as a whole, the COPG/TOPG rotates observations (called sweeps) across all individual children and the lead teacher to capture teacher-child interactions. Each sweep lasts three seconds, after which the observer records information about the person being observed. All scores are averaged across sweeps. Only the TOPG, which records data on the teacher, was used in this study. After the conclusion of the observation period, the observer completed the Post, a series of items describing both teacher (e.g., quality of individual instruction and scaffolding) and child (e.g., aggression, signs of being upset) behavior in the classroom seen over the course of the observation period.

Observers were trained on the COPG/TOPG and the Post during a 2-day in-person training by the tool’s developers at Vanderbilt University. They were certified as reliable if they reached an average of 80% agreement with a gold standard coder across all dimensions and counts, with no single dimension or count falling below 70% agreement. After each observation, gold-standard coders reviewed all codes and met with the observer in cases of disagreement. Weekly meetings were also held to reach consensus on how to code new situations as they arose.

For this study, a very few select items were drawn from the COPG/TOPG and the Post for their relevance to the self-regulatory environment based on theory and past research demonstrating links with child self-regulation scores. We were particularly interested in capturing teacher-child interactions that are pertinent to Barnes et al.’s (2022) conceptualization of emotionally supportive interactions. We thus selected Behavior Approvals, Behavior Disapprovals, and Emotional Tone from the TOPG following Fuhs, Farran, and Nesbitt (2013), and Red Flags from the Post following Phillips et al. (2022). Behavior Approvals measures the proportion of sweeps in which the teacher praised a specific child for their behavior. Behavior Disapprovals–Negative measures the proportion of sweeps in which the teacher directed a disapproval to a child in the form of a criticism. Behavior Disapprovals–Redirecting measures the proportion of sweeps in which the teacher expressed disapproval to a child by redirecting their behavior from an undesired behavior to a desired one. Emotional Tone is an average of the teacher’s affective tone across sweeps for each child in the classroom on a 5-point scale where higher scores are increasingly negative (1 = vibrant, 2 = pleasant, 3 = flat, 4 = negative, 5 = extreme negative). Red Flags comprises a checklist of 10 punitive behaviors such as eye-rolling, sarcasm, yelling or cursing at an individual child, or putting a child in isolation for more than 4 minutes. A count of the demonstration of such behaviors is used here.

Three scales were created to capture the self-regulatory environment, consistent with the three pillars proposed by Barnes et al. (2022). The first, called Classroom Management (α = .93), includes the ATSRS items tapping Consistency/Routine (rules and routines provide clear expectations and create a sense of community), Preparedness (clear, detailed instructional plans keep children engaged in learning), Classroom Awareness (awareness of the overall classroom environment and reliance on proactive classroom-level behavior management minimizes disruption), and Attention Support (different gestures, cues, and signals maintain students’ attention).

The second scale, called Emotionally Supportive Interactions (α = .78), includes the Positive Behavior Management (rewards for good behavior with specific praise, and clear statement of rules with logical consequences) and Negative Behavior Management (yells, makes harsh threats, uses sarcasm, or is overly restrictive) items from the ATSRS. It also includes the four items from the TOPG—Behavior Approvals, Behavior Disapprovals–Redirecting, Behavior Disapprovals–Negative, and Emotional Tone—and Red Flags from the Post. All items except Positive Behavior Management and Behavior Approvals were reverse-coded so that higher scale scores represented more positive and less negative interactions. (It should be noted that although the Classroom Awareness item involves proactive behavior management through the prevention of disciplinary problems, it was placed in the Classroom Management scale instead of the Emotionally Supportive Interactions scale because it relates to the classroom as a whole, not individual teacher-child interactions.)

The third scale, called Direct Promotion of Self-Regulatory Skills (α = .95), comprises the following items from the ATSRS: Emotion Modeling (identifies children’s or own emotions), Emotion Expression (validates children’s emotional expression), Emotion Regulation (acknowledges when children are upset and encourages effective strategies to prevent and help self-modulate emotion displays), Social Awareness (draws attention to consequences of actions and peers’ emotional experiences), Social Problem-Solving (approaches interactions between children as opportunities for learning), Scaffolding Peer Interactions (helps children expand collaborative play), and Provision of Interpersonal Support (uses verbally or physically supportive strategies to help children regain control).

To enable the combination of items measured on different scales (e.g., the ATSRS items on a 5-point scale, Behavioral Approvals as a proportion, and Red Flags as a count), all items were standardized and then averaged to form their respective scales.

Child Self-Regulation

Four measures were used to capture children’s self-regulatory skills. The first two measures were collected via direct assessment. The tablet-based National Institutes of Health Toolbox (NIHT; Zelazo et al., 2013) Flanker and Dimensional Change Card Sort tasks were used to assess inhibitory control and cognitive flexibility, respectively. The Flanker task shows children an image of a central stimulus (namely, a picture of a fish) that is flanked by other stimuli (in this case, fish) on either side. Children are asked to identify the direction the fish is facing across multiple trials which vary according to whether this fish faces the same direction as the flanker fish. The Dimensional Change Card Sort task asks children to match pictures on a feature such as color, and then asks them to match the same pictures on another feature such as shape. Children are scored on the Flanker and the Dimensional Change Card Sort tasks according to both accuracy and reaction time. Uncorrected standard scores (enabling comparison to the U.S. population) were used.

The third measure of self-regulation was teacher-reported using the 24-item Behavior Rating Inventory of Executive Function (BRIEF) adapted by LeJeune et al. (2010). For this tool, teachers are asked to rate how well each item describes the study child on a scale from 1 (strongly disagree) to 4 (strongly agree). Items fall generally into two categories: behavioral regulation (e.g., has trouble putting the brakes on his/her actions, talks at the wrong time) and metacognition (e.g., needs help from an adult to stay on task, does not plan ahead for school assignments, written work is poorly organized). Negatively valenced items were reverse-coded and all items were summed (α = .96).

The fourth measure of self-regulation was the assessor’s rating of the child’s behavior during the assessment session using the Long Form Attentive Impulse Control subscale from the Preschool Self-Regulation Assessment (PSRA) Assessor Rating Scale (Smith-Donald et al., 2007). This scale consists of 19 items capturing a variety of self-regulatory behaviors children could have displayed during assessment (e.g., child remains in seat appropriately during test). A binary variable was coded positively if either of two items indicated that the child expressed aggression. These variables and the remaining 17 items were added together for each child to yield their total score (α = .94).

Child Academic Achievement

English Literacy/Language.

Three measures of English literacy and language were collected. The Woodcock-Johnson-III Letter-Word Identification subtest (Woodcock et al., 2001) evaluated children’s ability to identify letters and pronounce words read on sight. The Clinical Evaluation of Language Fundamentals (CELF) Expressive Vocabulary subtest (Semel et al., 2013) measured children’s expressive vocabulary. This task asks children to label pictures orally and allows a range of acceptable answers (e.g., a tree branch may be labeled a branch or a limb). The CELF Sentence Structure subtest measured children’s understanding of both syntax (grammar) and semantics (meaning). In this task, children are asked to point to the image that corresponds to a sentence read out loud (e.g., “Choose the picture where the girl who is standing in the front of the line is wearing a backpack”). All three subtests accepted only English answers, and children who answered in Spanish were instructed to answer again in English. Raw scores on all three measures are used here. Children who failed the trial items on the CELF subtests were assigned a score of 2 SD below the mean.

Math.

Three instruments were used to assess children’s math skills. The Woodcock-Johnson III Applied Problems subtest measured children’s ability to solve math word problems by counting and performing simple addition and subtraction. Raw scores were used here. The Numeracy Screener Symbolic Numeral Comparison subtest (Lyons et al., 2018; Nosworthy et al., 2013) assessed children’s numerical fluency, or the efficiency with which they comprehend numerals’ underlying meaning. This task presents a series of two numerals (1–9) side-by-side and, for each pair, asks children to “mark the box with the number that means the most thing.” Scores are calculated as the number of correct responses returned in one minute minus the number of incorrect responses to adjust for guessing.

The modified short-form Research-Based Early Mathematics Assessments (SF-REMA; Johnson et al., 2019), an adaptation of the REMA-SF (Weiland et al., 2012), was used to assess conceptual counting skills. Children were given a pile of 31 pennies and asked to count them. They were scored on knowledge of 1:1 correspondence (the principle that each item may only be counted once) across four sets of pennies (1–5, 6–10, 11–20, and 21–31). The total possible score ranged from 0% (never followed 1:1 correspondence) to 100% (followed 1:1 correspondence across all sets). Children were also scored on their knowledge of cardinality (the principle that the last number counted represents the total number in the set) if the child answered the question “How many?” after they finished counting. Children were also scored on their knowledge of stable order (the principle that numbers always appear in the same order). They were assigned the highest number they counted to without error expressed as a percentage of the total possible score. These three scores were averaged to form an overall score ranging from 0–100%. All math instruments were administered in English but instructions were given in Spanish as needed, and Spanish-language answers were accepted.

Controls

Child and household demographic controls were drawn from parent surveys and school district administrative data. These variables include child age at the fall of kindergarten, gender, race/ethnicity (Hispanic/Latinx, Black non-Hispanic, White non-Hispanic, Multiracial, Other), whether the child was a DLL, whether the child had an IEP, whether the mother had more than a high school education, and monthly household income (log-transformed).

Analytic Plan

As a first step, a confirmatory factor analysis was conducted to assess how well the empirical data supported the scales capturing the self-regulatory environment that were constructed to mirror those proposed by Barnes et al. (2022): Classroom Management, Emotionally Supportive Interactions, and Direct Promotion of Self-Regulatory Skills. We also examined pairwise correlations between the scales measuring the self-regulatory environment and child outcomes in both the fall and spring of kindergarten. We then addressed our first research question, which asked whether the self-regulatory environment predicted change in children’s self-regulatory scores over the kindergarten year. To answer this, we separately regressed each of the four self-regulatory outcomes, measured in the spring of kindergarten, on the three self-regulatory environment scales simultaneously in a model including the fall measurement of that outcome and all child and household controls described above. To address our second research question, which asked whether the self-regulatory environment predicted change in children’s academic scores over the kindergarten year, we separately regressed each of the six academic outcomes, measured in the spring of kindergarten, on the three self-regulatory environment scales simultaneously in a model including the fall measurement of that outcome and all controls.

All predictors in the model that were measured continuously were standardized to facilitate interpretation of results. Mixed models with random intercepts by classroom were estimated to account for the clustering of children within classrooms. The rate of missing data on covariates ranged from 0–35%. Missing values were imputed using imputation by chained equations via the ice command in Stata; this method fills in missing values by iteratively predicting them. In this analysis, 25 data sets were imputed and Rubin’s Rules were used to combine estimates across imputations in Stata. Dependent variables were not imputed and because not all children had valid data on all outcomes, sample size varies by outcome and is denoted accordingly.

Results

Results of a confirmatory factor analysis indicated that the theoretical model with three factors capturing the self-regulatory environment did not fit the data well, χ² (132) = 282.87, p < .000, CFI = 0.91, RMSEA=0.10, 90% CI [0.09, 0.12]. Standardized factor loadings on the Classroom Management and Direct Promotion of Self-Regulatory Skills factors were high (.82–.94 and .72–.91, respectively; Table 2). However, factor loadings on Emotionally Supportive Interactions ranged from low to high (.14–.92). This result may be attributable to poor conceptual coherence among the items or the fact that this scale alone combined measures across all three instruments (ATSRS, TOPG, and Post) using different response metrics (e.g., a scale, a proportion, and a count). Despite the weakness of the third factor, the model was retained on theoretical grounds.

Table 2.

Standardized factor loadings from confirmatory factor analysis (n = 110)

	Classroom Management	Emotionally Supportive Interactions	Direct Promotion of Self-Regulatory Skills
ATSRS Consistency/routine	0.94
ATSRS Preparedness	0.85
ATSRS Classroom awareness	0.90
ATSRS Attention support	0.82
ATSRS Positive behavior management		0.92
ATSRS Negative behavior management		−0.90
TOPG Behavior approvals		0.14
TOPG Behavior disapprovals - redirecting		−0.31
TOPG Behavior disapprovals - negative		−0.37
TOPG Emotional tone		−0.61
Post Red flags		−0.73
ATSRS Emotion modeling			0.84
ATSRS Emotion expression			0.89
ATSRS Emotion regulation			0.91
ATSRS Social awareness			0.85
ATSRS Social problem-solving			0.88
ATSRS Scaffolding peer interactions			0.72
ATSRS Provision of interpersonal support			0.87

Note. ATSRS = Adapted Teacher Style Rating Scale; Post = Post Observation Rating Scale. Items negatively loading on Emotionally Supportive Interactions were reverse-coded to create the scale.

Pairwise correlations among the three scales ranged from 0.61 to 0.72 (Table 3). Pairwise correlations between the self-regulatory environment scales and child self-regulatory and academic skills in the fall of kindergarten revealed few associations (Table 3). Both Classroom Management and Emotionally Supportive Interactions had small but positive significant associations with cognitive flexibility, assessor-rated self-regulation, and letter-word identification (r = .07–.12). Pairwise correlations between the self-regulatory environment scales and child skills in the spring of kindergarten also revealed few associations (Table 4). Classroom Management had small but positive significant associations with letter-word identification, sentence structure, and numerical fluency (r = .07–.11).

Table 3.

Correlations among the kindergarten classroom self-regulatory environment and children’s self-regulatory and academic skills at the fall of kindergarten

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
(1) Classroom Management
(2) Emotionally Supportive Interactions	0.72***
(3) Direct Promotion of Self-Regulatory Skills	0.64***	0.61***
(4) Inhibitory control	−0.01	0.02	−0.02
(5) Cognitive flexibility	0.09*	0.12***	0.04	0.48***
(6) Teacher-rated BRIEF	−0.01	−0.02	−0.03	0.34***	0.29***
(7) Assessor rating	0.07*	0.08*	0.01	0.41***	0.31***	0.38***
(8) Letter-word identification	0.09**	0.07*	0.02	0.23***	0.18***	0.24***	0.26***
(9) Expressive vocabulary	0.03	0.00	−0.01	0.08*	0.12**	0.10**	0.13***	0.39***
(10) Sentence structure	0.03	−0.01	−0.01	0.19***	0.23***	0.22***	0.17***	0.30***	0.61***
(11) Applied problems	0.04	0.02	−0.03	0.32***	0.27***	0.31***	0.34***	0.54***	0.53***	0.61***
(12) Numerical fluency	0.02	−0.01	−0.04	0.33***	0.24***	0.25***	0.22***	0.39***	0.35***	0.37***	0.48***
(13) Conceptual counting	0.03	0.03	−0.03	0.22***	0.14***	0.18***	0.20***	0.56***	0.35***	0.32***	0.50***	0.44***

Note. BRIEF = Behavior Rating Inventory of Executive Function

^*p < .05;

^**p < 0.01;

^***p < 0.001.

Table 4.

Correlations among the kindergarten classroom self-regulatory environment and children’s self-regulatory and academic skills at the spring of kindergarten

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
(1) Classroom Management
(2) Emotionally Supportive Interactions	0.72***
(3) Direct Promotion of Self-Regulatory Skills	0.64***	0.61***
(4) Inhibitory control	0.06	0.03	0.02
(5) Cognitive flexibility	0.03	0.01	−0.02	0.49***
(6) Teacher-rated BRIEF	0.02	0.01	−0.00	0.34***	0.35***
(7) Assessor rating	−0.01	0.01	−0.05	0.39***	0.32***	0.33***
(8) Letter-word identification	0.11**	0.04	−0.01	0.27***	0.30***	0.31***	0.22***
(9) Expressive vocabulary	0.04	−0.02	−0.03	0.09**	0.14***	0.02	0.08*	0.41***
(10) Sentence structure	0.07*	0.01	0.01	0.24***	0.29***	0.16***	0.14***	0.38***	0.58***
(11) Applied problems	0.05	0.02	−0.01	0.40***	0.36***	0.29***	0.57***	0.49***	0.62***
(12) Numerical fluency	0.08*	0.04	−0.03	0.36***	0.33***	0.24***	0.19***	0.44***	0.35***	0.47***	0.58***
(13) Conceptual counting	0.05	0.02	−0.04	0.30***	0.29***	0.31***	0.26***	0.56***	0.34***	0.37***	0.57***	0.46***

Note. BRIEF = Behavior Rating Inventory of Executive Function

^*p < .05;

^**p < 0.01;

^***p < 0.001.

In multivariate models, only one of the four self-regulatory outcomes in the spring of kindergarten was predicted by self-regulatory environment scales (Table 5). Inhibitory control was positively, albeit weakly, associated with higher scores on the Classroom Management scale (β = 0.12, se = 0.05, p = .011). Neither the Emotionally Supportive Interactions scale nor the Direct Promotion of Self-Regulatory Skills scale was predictive of any self-regulatory outcomes.

Table 5.

Prediction of self-regulatory outcomes by the self-regulatory environment in kindergarten

	Inhibitory control		Cognitive flexibility		Teacher-rated BRIEF		Assessor rating
	β	se	β	se	β	se	β	se
Self-regulatory
environment
Classroom Management	0.12*	0.05	0.08	0.05	0.04	0.05	−0.02	0.05
Emotionally Supportive Interactions	−0.07	0.05	−0.08	0.05	−0.01	0.05	0.00	0.05
Direct Promotion of Self-Regulatory Skills	−0.02	0.04	−0.05	0.04	−0.01	0.04	−0.04	0.04
Controls
Fall inhibitory control	0.42***	0.03
Fall cognitive flexibility			0 39***	0.03
Fall teacher-rated BRIEF					0.74***	0.02
Fall assessor rating							0.50***	0.03
Age	0.09**	0.03	0.09**	0.03	0.02	0.02	0.03	0.03
Male	−0.12	0.06	−0.18**	0.06	−0.10*	0.04	−0.18**	0.06
Black, non-Hispanic	−0.26*	0.13	−0.32*	0.13	−0.12	0.09	−0.24	0.12
White, non-Hispanic	−0.21	0.14	−0.19	0.14	−0.11	0.09	−0.11	0.13
Multiracial	0.0	0.14	0.03	0.15	−0.10	0.10	−0.06	0.14
Other race/ethnicity	−0.07	0.14	−0.26	0.15	−0.01	0.10	0.09	0.14
Dual language learner	0.14	0.11	−0.02	0.12	0.15	0.08	0.13	0.11
Has IEP	−0.42***	0.12	−0.36**	0.12	−0.27***	0.08	−0.09	0.11
Mother has more than HS education	0.12	0.08	0.09	0.08	−0.03	0.05	0.09	0.07
Household income	0.03	0.04	−0.05	0.04	0.05*	0.02	0.03	0.04
n	826		825		764		831

Note. BRIEF = Behavior Rating Inventory of Executive Function. IEP = Individualized Education Plan. HS = high school.

^*p < .05;

^**p < .01;

^***p < .001.

Higher scores on Classroom Management were associated with higher scores on four of the six academic outcomes in the spring of kindergarten, though effect sizes were small (Table 6). Classroom Management was positively associated with all three literacy/language outcomes. Specifically, it predicted better letter-word identification (β = 0.15, se = 0.05, p = .005), expressive vocabulary (β = 0.09, se = 0.03, p = .004), and sentence structure (β = 0.10, se = 0.04, p = .013). Among math problems, Classroom Management predicted better conceptual counting (β = 0.11, se = 0.05, p = .036).

Table 6.

Prediction of academic outcomes by the self-regulatory environment in kindergarten

	Literacy/Language						Math
	Letter-word identification		Expressive vocabulary		Sentence structure		Applied problems		Numerical fluency		Conceptual counting
	β	se	β	se	β	se	β	se	β	se	β	se
Self-regulatory environment
Classroom Management	0.15**	0.05	0.09**	0.03	0.10*	0.04	0.04	0.05	0.10	0.05	0.11*	0.05
Emotionally Supportive Interactions	−0.09	0.05	−0.06*	0.03	−0.05	0.04	−0.05	0.05	0.01	0.05	−0.04	0.05
Direct Promotion of Self-Regulatory Skills	−0.08	0.05	−0.04	0.03	−0.02	0.04	0.01	0.05	−0.08	0.05	−0.08	0.05
Controls
Fall letter-word identification	0.68***	0.03
Fall expressive vocabulary			0.80***	0.02
Fall sentence structure					0.64***	0.03
Fall applied problems							0.64***	0.03
Fall numerical fluency									0.46***	0.03
Fall conceptual counting											0.42***	0.03
Age	0.03	0.02	0.01	0.02	0.10***	0.03	0.05*	0.03	0.08*	0.03	0.09*	0.03
Male	−0.10*	0.05	0.02	0.04	−0.12*	0.05	0.02	0.05	0.04	0.06	−0.01	0.06
Black, non-Hispanic	−0.13	0.10	0.02	0.08	−0.13	0.10	−0.16	0.11	−0.18	0.13	−0.20	0.12
White, non-Hispanic	−0.01	0.10	0.05	0.08	−0.02	0.11	0.00	0.11	−0.07	0.13	−0.16	0.13
Multiracial	−0.03	0.11	0.12	0.09	0.01	0.12	−0.04	0.12	0.02	0.14	−0.03	0.14
Other race/ethnicity	0.15	0.11	−0.09	0.09	−0.20	0.12	−0.24*	0.12	−0.16	0.14	0.02	0.14
Dual language learner	−0.02	0.09	−0.10	0.07	−0.15	0.10	−0.06	0.10	−0.09	0.11	−0.11	0.11
Has IEP	−0.37***	0.09	−0.06	0.07	−0.18*	0.09	−0.40***	0.10	−0.36**	0.11	−0.78***	0.11
Mother has more than HS education	0.04	0.06	0.13**	0.05	0.09	0.07	0.09	0.07	0.14	0.08	0.03	0.08
Household income	0.01	0.03	0.00	0.02	0.00	0.03	0.07*	0.03	0.04	0.04	0.08*	0.04
n	833		837		837		832		817		836

Note. IEP = Individualized Education Plan. HS = high school.

^*p < .05;

^**p < .01;

^***p < .001.

Higher scores on Emotionally Supportive Interactions had a negative but small association with one academic outcome: expressive vocabulary (β = −0.06, se = 0.03, p = .035). Scores on Direct Promotion of Self-Regulatory Skills were not associated with any of the academic outcomes.

A series of sensitivity tests was run to determine the robustness of results to differing model specifications. First, models of each outcome were re-run entering each self-regulatory environment scale individually rather than simultaneously (Supplementary Tables 1–10). Unexpectedly, any associations found between self-regulatory environment scales and self-regulatory and academic outcomes emerged only when all scales were entered simultaneously. Second, findings did not substantively change when models were re-estimated using only unimputed data (Supplementary Tables 11–12), although a positive association between Classroom Management and numerical fluency became significant. Finally, models using alternate metrics to measure outcomes where available (e.g., Woodcock-Johnson III standard scores) also yielded similar findings (Supplementary Table 13).

Discussion

Despite widespread agreement among developmentalists and educators alike that kindergarten classrooms are key contexts for not only academic skills but also the self-regulatory skills that facilitate a broad array of indicators of life success (Robson et al., 2020), it is still unclear just what features of the classroom environment or teacher-child interactions support the early development of these skills (Phillips et al., 2023). This study investigated a new measure intended to capture three key dimensions of the kindergarten classroom environment theorized to be foundational for children’s emergent social-emotional skills because of their likely relevance to self-regulation: classroom management, emotionally supportive interactions, and direct promotion of self-regulatory skills.

Our first research question asked whether these dimensions of the self-regulatory environment predicted change over the kindergarten year in self-regulatory skills, as assessed by a variety of methods, and our second research question asked whether these dimensions of the self-regulatory environment predicted change in three literacy/language and three math skills. We were surprised to find more associations between the self-regulatory classroom environment and academic skills than for self-regulatory skills. Indeed, the results demonstrated that only one of the three scales—Classroom Management—weakly predicted change in only one of the self-regulation outcomes: inhibitory control. Classroom Management did, however, predict change in all three assessments of literacy and language, and the assessment of conceptual counting. Further, scores on the Emotionally Supportive Interactions scale were modestly inversely related to gains in expressive language. However, this scale had poor psychometric properties. Teachers’ direct promotion of self-regulatory skills was not associated with any of the self-regulation or academic outcomes.

The Classroom Self-Regulatory Environment and Self-Regulatory Outcomes

Of the three scales tapping the classroom self-regulatory environment, only one was positively associated with self-regulatory skills, albeit weakly. Classroom Management predicted change over the kindergarten year in inhibitory control. However, it did not predict change in the other directly assessed measure of self-regulation—cognitive flexibility—or teacher- or assessor-reported self-regulation.

Although associations were small and did not reach statistical significance, higher scores on Emotionally Supportive Interactions were negatively associated with change over the kindergarten year in inhibitory control and cognitive flexibility. It is difficult to interpret findings for this scale because of its disappointing psychometric properties. It was designed to improve on past measures of the classroom environment by providing a richer representation of emotionally-laden teacher-student interactions aimed at both praising and correcting (including punishing) individual children’s behavior, assumed to carry the potential to support or undermine self-regulatory skill development. The innovation of this scale – the combination of items from three different instruments – may also have been its weakness. It displayed poor reliability and validity as evidenced in a low Cronbach’s alpha and several low factor loadings. To the extent that the scale was noisy, it may not have captured what was intended. It is also possible that by including multiple aspects of teacher-child interactions aimed at managing behavior, the scale overrepresented disciplinary interactions and under-represented other types of interactions that convey emotional support or a lack thereof. For example, it did not capture specific demonstrations of teachers’ warmth, sensitivity to individual students’ needs or contingency in responding to episodes of distress, all of which could provide key opportunities to scaffold self-regulatory skills (Murray et al., 2019). It is also possible that a measure of emotionally supportive interactions would be optimally measured at the dyadic student-teacher level rather than the teacher level, as ours was.

Nevertheless, there was reason to believe that teacher-child interactions aimed at addressing individual children’s behavior– and negative interactions in particular-- held the potential to predict change in self-regulatory skills, if not academic achievement. Past studies have found that behavior management strategies that rely on harsh affect or punishment are inversely associated with gains in self-regulation over the preschool (Fuhs, Farran & Nesbitt, 2013; Phillips et al., 2022) and kindergarten (Brendgen et al., 2006; L’Ecuyer et al., 2021) years. Our scale of Emotionally Supportive Interactions captured what are thought to be positive strategies, such as praise for desired behaviors and behavior approvals, as well as reverse-coded negative strategies. It is possible that classrooms that score higher on Emotionally Supportive Interactions are led by teachers who spend more time on discipline and are more directive overall of student behavior. Such an approach may provide fewer opportunities for students to practice their own self-regulation. Supporting this possibility, Degol and Bachman (2015) found that children whose preschool teachers engaged in more redirection of their behavior (assumed to be a positive strategy) experienced fewer gains in self-regulatory skill over the school year. Along these lines, it is possible that some behavior approvals, although typically viewed as positive, are expressions of frustration or imply comparisons between children (e.g., “I like how Reggie is being quiet”) and are hence perceived as censure by some or all students. For that matter, some behavior disapprovals may be expressed with positive affect, muddying the distinction between positive and negative approaches to managing children’s behavior. Further research is needed to explore the possibility that to understand early development in self-regulation, the distinction between teachers’ positive and negative forms of discipline may be less helpful than the distinction between strategies that grant more or less support for autonomy. There is a clear need for additional research that carefully observes and refines our understanding of teachers’ disciplinary strategies.

Surprisingly, there were no associations between classrooms’ scores on Direct Promotion of Self-Regulatory Skills and children’s change in self-regulation during the kindergarten year. This finding is consistent with Bardack and Obradovic (2019), who found that the T-DASEF, a scale assessing the frequency of elementary school teachers’ use of strategies for promoting students’ executive function skills, did not predict change in executive function scores over the school year. Both that study and ours appear to contradict evidence from the implementation literature that teachers’ direct instruction in identifying and managing unwanted emotions and impulses improves children’s ability to manage their attention and behavior in the classroom (Bierman et al., 2008; Diamond et al., 2007; Webster-Stratton et al., 2008). However, this evidence was yielded by evaluations of curricula designed for this very purpose, rather than observations of natural variations in these behaviors among teachers in “business-as-usual” conditions. It may well be the case that in the absence of a specific curriculum and accompanying professional supports for teachers to encourage self-regulatory skills, teachers’ naturally occurring range of such behaviors does not reach the level of effectiveness. Further, even within the validated curricula, some applications have failed to achieve their goals of improving self-regulatory behaviors (see, for example, Morris et al. 2014; Nesbitt & Farran, 2021).

It is also possible that the specific teacher behaviors captured by our Direct Promotion of Self-Regulatory Skills scale focused too much on strategies enabling emotional and behavioral regulation and too little on strategies enabling cognitive and attentional regulation. For example, no items focused on teachers’ assistance with children’s formulation, planning, and fulfillment of goals. Future research might benefit from a more expansive definition of the promotion of self-regulatory skills, or from efforts to link teachers’ supports for specific self-regulatory skills to improvement on those skills. In the meantime, troublingly little remains known about how best to foster self-regulatory skills in preschool and elementary school settings.

The Classroom Self-Regulatory Environment and Academic Outcomes

One feature of the self-regulatory environment, Classroom Management, was more consistently predictive of academic outcomes than it was of self-regulatory outcomes in kindergarten. Higher scores on this scale predicted all three literacy and language outcomes—letter-word identification, expressive vocabulary, and sentence structure—as well as conceptual counting. Although effect sizes were small (0.09–0.15), their consistency deserves attention. This scale closely resembles the Classroom Organization scale of the CLASS insofar as it taps teachers’ classroom-level strategies for improving classroom productivity, but the CLASS Classroom Organization scale also captures management of children’s behavior. The results here suggest that a measure exclusively tapping classroom-level management captures features of the early childhood classroom that have the potential to promote academic growth, perhaps even more so than one that also taps behavior management. Indeed, the features assessed by the Classroom Management scale (e.g., consistency, routine, preparedness, classroom awareness) should promote order throughout the school day, including during academic activities. This order may translate into both more time on and more effective instruction, but it is not immediately obvious why such an impact on instruction would be seen in all three literacy/language outcomes but only one math outcome (conceptual counting).

Unexpectedly, Emotionally Supportive Interactions had a negative (though weak) association with expressive vocabulary. As noted above, it may be that teachers in classrooms that score higher on this scale spend more time directing individual children’s behavior and less time on academic activities focused on the entire classroom. Taken together with the non-significant but negative associations between Emotionally Supportive Interactions and inhibitory control and cognitive flexibility, these findings suggest a need for further research on the role that strategies aimed at managing children’s behavior play in the kindergarten classroom. Again, the distinction between supportive and unsupportive strategies may be ambiguous, or these strategies be of secondary importance to the amount of time spent on discipline and other efforts to manage children’s behavior. Not only might time on discipline interfere with children’s autonomy, but it may also detract from instructional time.

The Self-Regulatory Environment as a Measure and Construct

Barnes et al. (2022) posited three pillars of the classroom that contribute to social and emotional learning: classroom management, emotionally supportive student-teacher interactions, and direct promotion of social and emotional skills. All three of these features of the environment seemed directly germane to self-regulatory skills, and, given the link between self-regulatory skills and academic ability, possibly to academic skills as well. The current study provides only weak support for this model, although emotionally supportive student-teacher interactions were measured suboptimally.

Interestingly, sensitivity tests revealed that the few associations found between the self-regulatory environment scales and child outcomes generally emerged only when models included all three scales simultaneously. This finding indicates that, for example, Classroom Management supports academic gains in language/literacy only when Emotionally Supportive Interactions and Direct Promotion of Self-Regulatory Skills are at their means (recall that all scales were standardized). Exploratory regression analyses did not find statistically significant interactions among the scales, but there is a need for a greater understanding of the way these aspects of the environment work together to facilitate the acquisition of academic skills.

Above we addressed the possibility that naturally occurring variation in the direct promotion of self-regulatory skills in our sample was overall too low to affect change in those skills. We should also consider the possibility that the use of such behaviors – like identifying emotions and strategies for self-soothing such as deep breathing – is most helpful in distressing situations that occur too rarely to be captured during a half-day classroom observation. The length of time observers spend in the classroom might simply not be enough to adequately assess the most impactful processes for self-regulatory growth in kindergarten. Alternatively, while attending to one dimension of the classroom, observers may miss others, such as key moments (e.g., instances of peer conflict) during which teacher responses are particularly potent predictors of developmental, and perhaps especially self-regulatory, outcomes (Murray et al., 2019).

Finally, the surprisingly weak links between our measures of the self-regulatory environment and children’s self-regulatory outcomes may reflect conceptual, rather than measurement, problems with the self-regulatory environment. It is possible that we overestimated the applicability of Barnes et al.’s (2020) three pillar model that was conceptualized in the context of social-emotional learning to classroom features that support self-regulation development. Or, as mentioned above, it may simply be that outside of well-designed interventions targeted to increase self-regulatory supports in the classrooms, with active and ongoing coaching and high fidelity of implementation, well-regulated “business as usual” elementary classrooms are just not powerful enough to affect development beyond the weak associations reported, as least as self-regulation was measured in this study.

It is also possible that prevailing conceptions of direct instruction in self-regulation tend to favor an overly intrusive approach to fostering self-regulation. For example, in some instances it may be more helpful for a child who is upset to be left alone to manage their emotions or behavior than it is to be given suggestions or directions by a teacher. Such efforts to promote autonomy may be key to developing the ability to regulate oneself rather than relying on cues from others. Future research should closely examine micro-interactions between students and teachers to determine optimal levels of teacher support; these are likely to vary by a multitude of circumstances such as the child’s temperament and status within the peer group, the situational demands placed on the child, the child’s history in managing emotions and behaviors, and the nature of the child-teacher relationship. In other words, there may be no such thing as a single ideal self-regulatory environment for all children. This conceptualization would be consistent with Bronfenbrenner’s (1996) person-process-context model, in which children’s own characteristics interact with those of their environment to influence development. While complicated, this web of factors may be key to our understanding just how the classroom environment can support, or at the very least not undermine, young children’s self-regulatory skills.

Limitations

This exploratory study has a number of limitations that pave the way for replication and next-stage research. First, while the measures used to capture features of the classroom self-regulatory environment–the ATSRS, COPG/TOPG, and Post–are finer tuned than earlier tools not designed specifically to assess the most theoretically meaningful aspects of the self-regulatory classroom, as with any new instruments, they are relatively untested. More data are needed to assess the reliability and validity of these measures, across different samples and in different locales. Future uses of these tools should also include more rigorous means of ensuring ongoing interrater reliability than were used here. Ideally, gold-standard coders should double-code observations intermittently rather than just review observers’ notes and scores.

Second, the three scales created here based on those measures yielded poor model fit and only weak correlations with the outcomes of interest. There is a clear need for more evidence of their construct validity in other settings. as well as additional exploratory work experimenting with other combinations of items across measures or developing new measures entirely.

Third, because our sample was entirely low-income and drawn from one school district in Tulsa, the extent to which our results may generalize to other contexts is unclear. Future studies should seek to replicate these results in different regions. They should also replicate these results with families that have a broader range of incomes than the low-income children in our sample. Children from low-income backgrounds experience chronic stress due to environmental adversity, which is in turn associated with poorer self-regulatory and academic skills (Noble et al., 2005; Raver, 2012; Rimm-Kaufman et al., 2000). There may be features of the classroom environment that are uniquely supportive for low-income children; even if the same features of the classroom matter for both lower- and higher-income children, they may require higher thresholds for lower-income children to feel their effects.

Conclusion and Future Directions

This study drew on three existing observational tools to create a comprehensive measure of the kindergarten self-regulatory environment comprising classroom management, emotionally supportive interactions, and direction promotion of self-regulatory skills. We posited that all three facets of the classroom environment would be pertinent to the development of self-regulatory skills, in particular, but found associations between only one scale (Classroom Management) and one skill (inhibitory control). We further found that only Classroom Management was consistently associated with academic outcomes. It is as yet unclear whether the null findings reported here are due to deficiencies in measurement or a broader misconceptualization of the self-regulatory environment.

These results, while disappointing, are entirely consistent with the field’s ongoing attempts to measure the instructional environment. Researchers have struggled to understand why long-used classroom observational measures like the CLASS with apparently solid theoretical foundations are only weakly predictive of growth in children’s academic skills (e.g., Burchinal, 2018). The Early Learning Network, a consortium of research teams across the country sponsored by the federal Institute for Education Sciences, is now testing a range of new classroom observational tools in an effort to identify the “active ingredients” that shape skill growth in early education and elementary classrooms. Thus far, these efforts have produced disappointing results, with no clear “winning” classroom observational measure that reliably predicts growth in children’s early academic skills (McCormick et al., 2022; Weiland et al., 2023). Nevertheless, it is incumbent on scholars to continue refining theoretical models and experimenting with methodological innovations to further our understanding of how the classroom context exerts its effects on children’s development during this early, critical period.

Highlights.

• A theoretically-based measure of the kindergarten classroom self-regulatory environment was developed to capture classroom management, emotionally supportive interactions, and direct promotion of self-regulatory skills.

• These classroom dimensions were tested as predictors of change over the kindergarten year in self-regulatory and academic skills in a sample of racially/ethnically-diverse low-income children in Tulsa, OK.

• Only classroom management showed positive associations with children’s skills, and those were small and primarily applied to language/literacy outcomes.

• Further research is needed to both conceptualize and operationalize how early classroom environments foster self-regulation.