Self-regulation in Preschool Children: Hot and Cool Executive Control as Predictors of Later Classroom Learning Behaviors

Abstract

We examined associations between aspects of preschoolers’ executive control (both hot and cool), on-task involvement, and subsequent learning behaviors (attention/persistence, attitude toward learning, competence motivation) in kindergarten, taking into account potentially indirect effects of on-task involvement on relations between executive control and later learning behaviors. Moderating effects of gender and socioeconomic risk also were examined. Three hundred eighteen children were directly assessed regarding executive control (T1) and teachers rated their on-task involvement (T2); 108 kindergartners’ teachers evaluated children’s learning behaviors (T3). We modeled the hypothesized longitudinal relations, as well as separate models for moderators using partial least squares analyses (PLS). Findings often identified end-of-preschool on-task involvement as a mediator between earlier hot executive control and kindergarten learning behaviors, and sometimes indicated direct prediction of kindergarten learning behaviors by preschool cool executive control. Discussion centers upon interpreting the overall model and moderating influences, as well as potential applications.

Educational relevance statement:

When young children progressing through preschool and transitioning into kindergarten can manage their thoughts, emotions, and behaviors, become able to participate positively in a classroom, and ultimately feel competent and interested in learning, they are more likely to succeed in school. Our results show a progression across time in these abilities from early preschool to kindergarten. Specifically, preschoolers’ ability to delay gratification and inhibit behavior often supports their engagement in preschool classroom tasks, which in turn contributes to kindergarten learning behaviors. In contrast, their ability to focus, remember, plan, and solve problems plays a key role in kindergartners’ persistence and their ability to feel positive about learning and their own competence as they transition to elementary school. Understanding these distinct yet interconnected aspects of self-regulation can help inform strategies to prepare preschoolers for kindergarten and address the diverse needs of children from different backgrounds.

1. Introduction

Self-regulation is a broad construct representing the cognitive, motivational-affective, social, and physiological processes that modulate attention, emotion, and behavior to a given situation/stimulus, for the purpose of pursuing a goal (Zelazo & Carlson, 2020). As young children enter schooling, they are increasingly expected to be able to regulate their attention and behavior, along with their emotions, while engaging in learning experiences with teachers and classmates. New demands face them – entering the peer arena, independently responding to new adults, and attempting new developmental tasks. Given these new, taxing requirements, children who acquire self-regulatory abilities are at an advantage because they can sustain attention and resist distraction, tolerate frustration and refrain from acting impulsively, and consider the consequences of their behavior while keeping their goals in mind. In short, their developing self-regulation allows young children to adapt to changing demands, function effectively in their social milieu, and learn successfully (Zelazo & Carlson, 2020).

More specifically, self-regulatory abilities exert critical influence on several aspects of children’s early school success. For example, young children who demonstrate greater self-regulation abilities are better able to demonstrate classroom on-task involvement (OTI) and engagement – they are fully involved in classroom activities, using materials and teachers’ communications well (Coelho et al., 2019; Nelson et al., 2017; Veraksa et al, 2022). Aspects of self-regulation can even predict trajectories of later OTI (Pagani et al., 2012). Children with more developed self-regulation skills also exhibit more positive classroom learning behaviors (LBs), such as attention/persistence, attitude towards learning, and competence motivation (Acar et al., 2022; Guedes et al., 2024; O’Toole et al., 2020). Much research also has demonstrated links between early self-regulation and aspects of both concurrent and subsequent school readiness or academic success (for a meta-analysis supporting this assertion, see Robson et al., 2020; see also Nguyen & Duncan, 2019; O’Toole et al).

A few investigations also have found that OTI or LBs, examined singly, mediate the relation between self-regulation and specific aspects of school readiness or academic outcomes, albeit sometimes weakly (Nesbitt et al., 2015; O’Toole et al. 2020). These sparse results suggest value in investigating in a more fine-grained manner how self-regulation facilitates development of young children’s ability to become involved and engaged in classroom activities and indirectly fosters their later LBs. No previous research has to our knowledge discovered whether preschool classroom involvement can predict kindergarten LBs, given an earlier foundation of self-regulation. Substantiating such tripartite linkages among self-regulation, OTI, and later LBs is the purpose of our study. Thus, in this study we investigate the linkages between two aspects of self-regulation – “cool” and “hot” executive control – with preschoolers’ OTI as an intermediary outcome on the path to increasingly positive LBs in kindergarten. Further, we look into possible moderation of these linkages by children’s gender and socioeconomic risk. We now elucidate aspects of this model.

1.1. Self-Regulation

Self-regulation or executive control (EC) shows impressive development from toddlerhood to kindergarten age, moving from relative dependency on caregivers’ assistance to more independent abilities (Bernier et al., 2010). Significant gains occur in the third year of life reflecting this change, and continued rapid growth occurs from three to six years of age (Kochanska & Knaack, 2003; O’Toole et al., 2018).

Furthermore, across the preschool period, especially after age three, important distinctions emerge between two types of EC (Garon, 2016): (a) cool EC (CEC: evinced in more affectively neutral contexts); and (b) hot EC (HEC: exhibited in emotional or motivational contexts). During these years, both CEC and HEC show continuous improvement (Hongwanishkul et al., 2016; O’Toole et al., 2018), with HEC growth more gradual (Garon; Zelazo & Carlson, 2020). Although distinguishable, these two aspects of EC are both “top-down” processes involving activity of the prefrontal cortex and may work in concert, probably most discernibly in real-life situations. Thus, meeting behavioral expectations at school, interacting with others, and learning new skills require both CEC and HEC, via use of working memory, inhibition of prepotent responses while activating alternative subdominant responses, and flexible focusing or shifting of attention, in both emotional and less emotional situations (Garon; Garon et al., 2008; Zelazo et al., 2024). Despite normative changes, important individual differences exist in both CEC and HEC (as well as OTI and LBs), probably due to environmental variables such as poverty and characteristics of the child (Raver et al., 2011).

1.1.1. Cool Executive Control (CEC)

CEC, in which the dorsolateral prefrontal cortex (DL-PFC) plays a special role (Garon et al., 2008), encompasses a wide array of increasingly organized, flexible, goal-directed cognitive processes in response to relatively non-affective and novel situations, including complex cognitive tasks (Zelazo & Carlson, 2020). It involves attention, memory, monitoring and inhibiting behavior, planning, and problem solving (Hongwanishkul et al., 2016; Zelazo et al., 2024). As already implied, school tasks mandate that children use CEC abilities to follow rules and purposefully shift or focus their attention, flexibly responding to conflicting stimuli (e.g., playing “Simon Says”, manipulating puzzle pieces without letting a playmate interrupt, or disengaging from a task when the teacher says it is time to clean up).

1.1.2. Hot Executive Control (HEC)

Young children also need to demonstrate HEC, in situations requiring more affective and motivational processes and responses. These situations may involve delay of gratification, voluntary inhibition or activation of behavior, resisting negative or socially unpopular emotions, or reappraisal of the motivational significance of a stimulus (e.g., not touching a toy that belongs to someone else; waiting patiently for one’s teacher to provide drawing materials; refraining from becoming angry when tasks become difficult; see Garon, 2016; Hongwanishkul et al., 2016). Thus, guided by both emotional information from the limbic system and orbitofrontal cortical braking, HEC enables children to better control their impulses and balance their own self-defined needs with societal norms, promoting their motivation and engagement in schooling (Bassett et al., 2012).

Overall, it is important to consider both subtypes of EC because of their somewhat varying source and functions, as well varying relations with child outcomes, but as already noted, much overlap would be expected between them (Garon, 2016). It should also be noted that earlier-developing CEC engages and modulates HEC, indicative of the nature of their connection (Zelazo & Carlson, 2020). This relation will be reflected in our total model, in which we examine how each separate aspect of preschoolers’ EC relates to both their OTI and later LBs.

1.2. On-Task Involvement (OTI)

OTI can be defined as the ability to engage in appropriate classroom tasks. Behaviors supporting this engagement include interest in classroom activities, listening to the teacher’s instructions and requests, and seamlessly returning to activities if interrupted (Fitzpatrick & Pagani, 2013). Being involved in this manner within the daily classroom milieu is associated with preschoolers’ academic achievement and school attendance (Fitzpatrick & Pagani). Like EF, OTI develops over time—with kindergartners showing greater ability than preschoolers in this area (Cadima et al., 2015).

As already noted, OTI also is associated with EC (Robson et al., 2020), even longitudinally (Cadima et al.,2015). In some studies, HEC is the most strongly related to OTI (Miller et al., 2006). For example, children need to resist the attraction of an engaging activity because the class is currently engaged in another activity.

1.3. Classroom Learning Behaviors (LBs)

LBs are observable, teachable and malleable (McDermott et al., 2018); they are related to decreased risk of academic difficulties and predictive of later achievement and grades (McDermott et al., 2012). Unlike EC and OTI, developmental change in LBs has not been studied frequently. However, overall growth during the Head Start year has been found (Domínguez et al., 2010); in contrast, LBs of children transitioning from Head Start to kindergarten showed declines no matter their original level (McDermott et al., 2018).

For the current investigation, children’s LBs encompass a child’s motivation to acquire competence, their attitude towards learning new information, as well as their attention/persistence in acquiring new information (McDermott et al., 2002). If children have difficulties demonstrating these behaviors, school problems can persist or even increase over time (McClelland et al., 2000). Thus, LBs are of utmost importance, necessary to understand the impact that EC has on a child’s overall school success.

Again, there is evidence of a connection between children’s EC and their LBs both concurrently and across time (Acar et al., 2022; Guedes et al., 2024). Both CEC and HEC have been associated with classroom LBs (Acar et al., 2022). However, EC may indirectly relate to classroom LBs via OTI, with children who are not engaged likely to have difficulty maintaining motivation and capitalizing on learning opportunities; in contrast, early classroom involvement can set the stage for academic success by undergirding positive classroom LBs (Halliday et al., 2018). In this study, we take the approach that being productively involved in classroom activities, bolstered by already-acquired aspects of EC, may allow development of positive classroom experiences that fuel children’s motivation to learn, feeling of academic competence, and attention and persistence. Along with examining the relations of CEC and HEC with OTI and later LBs, we consider potentially important individual difference moderators of these relations, gender and socioeconomic risk.

1.4. Gender as Moderator

Previous research has found that there is significant variation in preschoolers’ EC across gender, favoring girls. These differences have been most often reported for HEC (Bassett et al., 2012; Heilman et al., 2008; Guedes et al., 2024; Silverman, 2021), behavioral EC (Matthews et al., 2009), or parent report of self-control/CEC (Eisenberg et al., 2010). These differences may emerge as early as 2 to 3 years of age (Eisenberg et al.) and last until at least kindergarten (Matthews et al.). Moreover, girls are more likely to demonstrate greater OTI than boys from preschool through elementary school (Halliday et al., 2018; Pagani et al., 2012; Veraksa et al., 2022). Similarly, gender differences favoring girls can occur in LBs (Domínguez et al., 2010; O’Toole et al., 2020).

Although these mean gender differences in EC, OTI, or LBs do not necessarily imply differing relations among these attributes, it is also true that differences in boys’ and girls’ developmental trajectories for EC have been uncovered – with boys more likely to show later gains than girls across the preschool period, persisting through kindergarten (Montroy et al., 2016). In contrast, girls showed sustained growth during preschool (Montroy et al., 2016; Pagani et al., 2012). In fact, in one study of EC boys did not catch up with girls, ending kindergarten at a level girls showed at the beginning of the year (Matthews et al., 2009). Further, girls have been found showed a higher rate of growth in LBs than boys across a preschool year (Domínguez et al., 2010). However, research on trajectories of OTI appears absent.

These differing growth trajectories, with boys and girls differing in not only level of development but also rate of change, suggest that girls and boys should be examined separately in terms of how their EC predicts LBs as mediated by OTIs across the preschool to kindergarten transition. Given the greater affordances enabled by their earlier developing EC and its positive trajectory, when compared to boys, girls may be more ready to develop OTI and LBs, and ultimately more clearly demonstrate our total model. Thus, examining gender differences in our model will add to understanding of girls’ possible advantages in these areas, and perhaps more importantly, bring to the foreground the needs of boys. Discerning gender differences could help early childhood educators in targeting EC programming to those most in need and maximizing the development of those already on a positive trajectory.

1.5. Socioeconomic Risk as Moderator

EC, OTI, and LBs also are impacted by a variety of environmental factors, including socioeconomic status, and children living in low-income environments are at an increased risk of difficulties in these areas (Allee-Herndon & Roberts, 2019; Zelazo & Carlson, 2020; Zelazo et al., 2024). Low-income family life may include a series of vulnerabilities that may impact children’s EC, as follows: (a) stressful life events (e.g., death or illness in the family, family turmoil, separation from adult caregivers, loss of housing); (b) economic shocks such as job loss; (c) physical stressors such as substandard housing, crowding, loud noise, chaotic, unsafe environments, and inadequate nutrition; and (d) parents’ perceived strain (Baker et al., 2021; Buek, 2019). Further, parenting quality may be eroded by the experience of poverty, with loss of income, increases in community violence, and low wages making it difficult for parents to provide the scaffolding and learning resources so important to fostering EC (Hackman et al., 2015; Zelazo et al., 2024). Thus, the development of EC, and subsequently both OTI and LBs, may be disrupted for children experiencing the cumulative impact of these vulnerabilities.

Success in developing EC despite these odds may be crucial for later adjustment to the school setting as indicated by OTI and LBs. That is, early EC skills developed despite such conditions may especially matter to preschoolers living in low-income environments, buffering them from deleterious effects of their environment and subsequent negative outcomes (Miller et al., 2006; Smith-Donald et al., 2007; Zelazo et al., 2024). However, despite this possible buffering effect of EC, earlier research has not examined socioeconomic risk in relation to the full model studied here. Thus, in this study, we examine how the EC of children differing in socioeconomic risk differs in predicting OTI and LBs, expecting the model to more fully fleshed out for children at socioeconomic risk.

1.6. The Present Study

Little research has simultaneously focused upon the associations between CEC, HEC, OTI, and classroom LBs, especially taking into account the potentially indirect effects of OTI on the relation between EC and LBs. Thus, a cohesive model that could facilitate applications in early childhood education has not been created. Further, because gender and socioeconomic circumstances play a role in preschoolers’ classroom experiences, moderating effects of gender and socioeconomic risk must be taken into account. See figure 1 for our full model. Our model is informed by four research questions.

Figure 1.

Longitudinal Model: Executive Control Predicting On-Task Involvement and Learning Behaviors

1.6.1. Research Question 1: Is there a relation between children’s EC (CEC and HEC) in preschool and their later OTI and classroom LBs?

Based on previous evidence, we hypothesize that there will be direct positive associations between both CEC and HEC assessed during preschool and their OTI in the next preschool year. Further, we hypothesize that preschoolers’ CEC and HEC will be related to kindergarten classroom LBs (competence motivation, attitude towards learning, attention/persistence; see figure 1). Also included in the model is a path from CEC to HEC, as already noted.

1.6.2. Research Question 2: Does EC relate to positive LBs indirectly, via OTI?

Given evidence for relations between EC and OTI, and OTI and LBs, we expect that EC also is indirectly related to LBs via OTI.

1.6.3. Research Question 3: Are the relations of EC with OTI and LBs, and OTI with classroom LBs, moderated by gender?

Despite focusing only on main effects, earlier research raises the question of whether gender impacts the strength of relations between EC, OTI, and LBs. Thus, we explore the potential moderating effects of gender on the relations depicted in our model. We expect that relations of EC with both OTI and LBs, and the relations between OTI and LBs, will differ according to gender. Specifically, we expect that there will be more significant pathways in the model for girls; however, due to the exploratory nature of this research question, we do not hypothesize specific gender differences among the pathways.

1.6.4. Research Question 4: Are the relations between EC, OTI, and LBs, and OTI with LBs, moderated by living in poverty, operationalized by school type?

Despite prior evidence of diminished EC, OTI, and LBs for children living in poverty, further investigation is necessary to better understand how economic status plays a role in moderating the pathways among these variables. As a result, a central goal of the current investigation is to understand children’s EC and its relations with their later OTI and LBs in the context of socioeconomic risk. Given evidence of EC as a protective factor, we might expect more significant pathways in our model for children at socioeconomic risk. EC and OTI may “matter more” for children at risk in developing LBs. However, again, due to the exploratory nature of this research question, we do not hypothesize specific differences among the pathways. Socioeconomic risk will be indexed via child attendance in private child care versus Head Start.

2. Method

2.1. Participants

Data for this study are from a larger investigation focused on establishing valid and reliable assessment tools for social and emotional aspects of school readiness; the project was approved by our Institutional Review Board. Assessments were made at three time points: Time 1 (T1) during fall of the preschool year, after the children had become acclimated to the classroom; Time 2 (T2) during spring of the final preschool year; and Time 3 (T3) during the fall of the kindergarten year.

All children were recruited from private child care centers or Head Start programs in Northern Virginia, (N _total = 318, N _{Head Start} = 143). Children in the study had a mean age of 49.11 months at T1 and 50.2% were female. Among those whose parents reported race at T1 (n = 274), race was also somewhat evenly distributed (African American at 44% and White at 52%), with around 12 percent latine. About two thirds of Head Start attendees were African American, and about two thirds of children in private child care were White.

After meeting with directors, each preschool center was enrolled in the research project; next, we obtained consent from participating teachers in each facility. Then, parents in these teachers’ classrooms were recruited at child pick-up, information sessions held at the facilities, and/or through the help of facility social workers and directors. All teacher and parent participants gave written informed consent to participate, with parent giving consent for their children’s continued participation through kindergarten; children gave verbal assent for participating in direct assessment at T1. After receiving consent from parents for data to be collected on their children, preschool teachers were asked to complete several questionnaires; at T2, one of these teacher questionnaires is used in this study. Kindergarten teachers were recruited after permission was obtained: (1) from administrative personnel, to perform research in the public or private school system; and (2) from principals, to perform research in specific schools; they too completed several questionnaires, with one pertinent to this study.

We dichotomously categorized socioeconomic risk via Head Start and private child care centers. Head Start attendance was used as a proxy for high socioeconomic risk, given that Head Start uses consistent federal poverty guidelines for enrollment (Raver et al, 2009). Conversely, fewer than 10 percent of children attending private child care centers received income-based subsidies, allowing for overall classification of these facilities as serving children at low socioeconomic risk.

One hundred eight participants remained at T3 (N _{Head Start} = 60). This attrition was not related to child gender, ethnicity, race or socioeconomic risk. Because T3 data collection occurred in kindergarten classrooms, it was imperative but often quite difficult to track children’s whereabouts, and then obtain permission from both schools and teachers regarding data collection, especially because there had been no previous contact with our research team.

2.2. Measures

The measures in this investigation capture child attitudes and behaviors through a variety of data collection methods across the three time points (see figure 1). Children participated in direct assessment at T1, and received a sticker for their participation. Questionnaires were distributed to teachers at T2 and T3, to be completed at their convenience. A modest monetary incentive per completed child questionnaire was offered to the teachers to increase the probability of questionnaire completion. For each participating child in their classroom, preschool teachers were compensated $15 per child; because they completed more questionnaires (the preponderance of which are not the focus of this report), kindergarten teachers were compensated $25 per child.

2.2.1. Preschool Self-regulation Assessment (PSRA)

To measure EC at T1, we used the Preschool Self-Regulation Assessment (PSRA; Smith-Donald et al., 2007) a measure specifically designed to assess EC in behavior, emotion, and attention. This assessment consists of a structured battery of seven age-appropriate tasks to tap CEC and HEC (Bassett et al., 2012; Smith-Donald et al., 2007). We included three tasks (balance beam, pencil tap and tower turn-taking) to index CEC. The four tasks included to assess HEC were toy wrap (peek & wait), snack delay, and tongue tasks. The PSRA was administered by 12 trained and certified research assistants; coding reliability was moderately high to high on all tasks (intraclass correlations ranging from α = .57 to .97, with an average α = .87; Bassett et al., 2012).

2.2.2. Teacher Rating Scale of School Adjustment (TRSSA)

The TRSSA, a measure of classroom adjustment (Birch & Ladd, 1997), has been revised to improve psychometric adequacy (especially internal consistency) and better align with three basic domains of children’s school adjustment: (a) social competence or maturity in the classroom; (b) on-task classroom involvement; and (c) positive orientation to school activities (Betts & Rotenberg, 2007). For this study, we used Betts and Rotenberg’s On-Task Classroom Involvement factor to index OTI atT2. Six OTI items include those assessing how the child does the following: (a) follows teacher’s directions; (b) uses classroom materials responsibly; (c) listens carefully to teacher’s instructions and directions; (d) is interested in classroom activities; (e) responds promptly to teachers’ requests; and (f) if an activity is interrupted, can return to the activity. Cronbach’s α was .87 for this scale.

2.2.3. The Preschool Learning Behavior Scale (PLBS)

The PLBS is a 29-item teacher-report rating instrument measuring preschool children’s approaches to learning in a classroom environment (McDermott et al., 2002). It yields three learning behavior dimensions: (a) Competence Motivation (e.g., “The child takes initiative”); (b) Attention/ Persistence (e.g., “The child sticks to an activity for as long as can be expected for a child of this age”); and (c) Attitude Towards Learning (e.g., “Cooperates in group activities”). In this study all three scales were internally consistent (αs = .79 to .89). Earlier multi-method, multi-source validity analyses have substantiated the convergent and divergent validity for the PLBS dimensions, with similar reliability estimates for both White and non-White portions of the sample (Schaefer et al., 2004). For the current investigation, all three PLBS sub-scales (Competence motivation, Attention/persistence, & Attitudes Towards Learning) will be used to measure child LBs at T3, in kindergarten.

To clarify possible confusion, the PLBS ‘Attention/Persistence’ scale attempts to understand a child’s level of restlessness, concentration and distractibility, whereas the OTI factor of the TRSSA attempts to capture more of the child’s sense of responsibility, and ability to follow teacher instructions.

2.3. Analyses

Partial Least Squares (PLS; Esposito Vinzi et al., 2010; Ringle et al., 2005) analysis were performed to evaluate relations in the model proposed in Figure 1, as well as the possible moderation by gender and socioeconomic risk (enrollment in a private child care or Head Start classroom). PLS is a method for modeling relations between sets of observed variables by means of latent variables, where a measurement (outer) model and a structural (inner) model is specified.

For outer models, PLS estimates latent variables (LVs) based on the shared variance of manifest variables, using principal component weights to compute composite scores. This approach minimizes residuals and measurement error, improving psychometric reliability while reducing the data to a smaller set of theoretical LVs. These LVs retain the essential information from the manifest variables, enabling the investigation of relationships (inner model) and predictive validity, including direct and indirect pathways (Tsethlikai, 2010). Bootstrap procedures test path significance, and discriminant validity is assessed by comparing LV intercorrelations to the square root of their Average Variance Extracted (AVE).

This method, widely used by developmentalists (e.g., Tsethlikai, 2010, 2011), allows the exploration of hypothesized relationships among constructs with fewer restrictions than traditional structural modeling techniques. PLS is suitable for small sample sizes, provided a reasonable LV-to-participant ratio is maintained (Henseler et al., 2009), and offers advantages such as fewer assumptions about observational independence and data normality (Marjoribanks, 1997), error-free measurement (Tsethlikai, 2011), as well as robustness to missing data and multicollinearity (Garson, 2016). In this study, using Smart-PLS^™ (Ringle et al., 2005), and guidelines enunciated in Henseler et al. (2009), we evaluated outer and inner models for each Problem Question, as follows: (1) specified the manifest variables that met criteria for inclusion as indicators of LVs, and (2) calculated pathways in the inner model using these final LVs, to discern significant pathways. Across all models, significance levels were determined using a bootstrapping resampling algorithm (Esposito Vinzi et al., 2010). Missing data, especially for T3, were accounted for via multiple regression imputation; this method is recommended for PLS analyses, yielding stable path coefficient and loading estimates (Kock, 2018).

3. Results

3.1. Outer Model

In PLS, acceptable outer model fit must meet certain key criteria, including: 1) the manifest variables need to sufficiently load into the LV (i.e., have an AVE of .50 or above; Fornell & Larcker, 1981) and be internally consistent (i.e., composite reliability ≥ .60), and 2) the manifest variables need to represent enough average variance within the construct to demonstrate compelling levels of explained variance (i.e., loadings of at least .60; Esposito Vinci et al., 2010).

Within the current model, sufficient AVE, reliability, and factor loadings were demonstrated after removing several original manifest variables that did not meet one or more of the reliability and factor loading criteria (see Table 1 for final model). The following manifest variables were removed: (a) one from each original EC latent variable (balance beam from CEC and “tongue task” from HEC); (b) four from the PLBS competence motivation latent variable (“does not take refuge in helplessness”, “does not burst into tears when faced with difficulty”, “is not hesitant in talking about his/her activity”, and “shows a lively interest in activities”) ; and (c) three from the PLBS attitude towards learning latent variable (“shows desire to please you”, “is willing to accept help when an activity proves too difficult”, and “is willing to be helped”). No change occurred in the original six OTI or PLBS Attention/Persistence items’ inclusion. Two manifest variables for each EC latent variable, and 21 total for learning behaviors were retained.

Table 1.

Outer Model and Final R²s for Latent Variables (LV): Preschool to Kindergarten Classroom Executive Control, Involvement and Learning Behaviors

Manifest Variable	AVE	LVR2	Composite Reliability	Manifest Variable Loading
Hot Executive Control	.70	---	.82
Snack Delay Task				.838
Toy Peek Task				.831
Cool Executive Control	.66	---	.79
Pencil Tap Task				.922
Turn Taking Task				.688
On task Involvement	.61	.07	.90
Follows teachers’ directions				.855
If interrupted, goes back to activity				.613
Uses classroom materials responsibly				.782
Listens carefully to teachers’ instructions…				.882
Is interested in classroom activities				.691
Responds promptly to teachers’ requests				.835
Competence Motivation	.55	.05	.91
“Shows determination…”				.895
“Does not use headaches or other pains…”				.617
“Is not lacking in energy…”				.613
“Accepts new activities without fear…”				.750
“Is not dependent on adults for what to do…”				.774
“Does not say task is too hard without much effort…”				.813
“Is not reluctant to tackle new activity”				.791
“Does not adopt don’t care attitude…”				.775
Attitude Towards Learning	.61	.06	.86
“Cooperates in group activities”				.773
“Does not get aggressive or hostile…”				.738
“Pays attention to what you say”				.747
“Achieves goals even when mopey …”				.811
Attention/Persistence	.59	.07	.92
“Acts taking sufficient time…”				.845
“Cooperates in group activities”				.657
“Is not distracted too easily…”				.836
“Can settle into an activity”				.864
“Shows determination…”				.754
“Pays attention to what you say”				.810
“Tries hard and concentration does not fade…”				.679
“Sticks to an activity…”				.679
“Does not adopt a don’t-care attitude…”				.736

Thus, although there was theoretical justification for inclusion of all these manifest variables, a key assumption of PLS modeling is that, in addition to the inner model’s path structure, the outer model must demonstrate stability. With these modifications, our outer model LVs for CEC, HEC, OTI, competence motivation, attitude towards learning, and attention/persistence were acceptable in terms of reliability and factor loading adequacy. Given these LVs, the inner model can be evaluated next.

3.2. Inner Model

Examining both the Average Variance Extracted (AVE) and the correlations among LVs are the first steps in understanding the inner model. These results help determine discriminant and convergent validity. Discriminant validity indicates the extent to which an LV is significantly different from other LVs, which is determined by the AVE. The current model met the AVE criterion for each latent variable (see table 1). Additionally, the comparison of square root of AVE (see bolded diagonal values in table 2) compared with the intercorrelations among LVs indicated all LVs are more strongly correlated with their own manifest variables than with any other LV (Esposito Vinci et al., 2010), suggesting both discriminant and convergent validity.

Table 2.

Inner Model LV Correlations: Preschool to Kindergarten Classroom Executive Function, Involvement and Learning Behaviors

		1.	2.	3.	4.	5.	6.
1.	HEC (T1)	.83
2.	CEC (T1)	.45***	.81
3.	OTI (T2)	.27***	.15**	.78
4.	Competence Motivation (T3)	.08	.17**	.17**	.74
5.	Attitude Towards Learning (T3)	.12	.11	.22***	.68***	.78
6.	Attention /Persistence (T3)	.12	.13*	.25***	.80***	.79***	.77

Note. Square root of AVE (Average Variance Extracted) appears in bold on the diagonal; LV (Latent Variable) correlations appear below the diagonal.

^*p < .05.

^**p < .01.

^***p < .001.

The LV correlations found in Table 2 also serve as initial indicators of the hypothetical relations between LVs. Table 2‘s results suggest that HEC only demonstrated a significant relation to CEC and OTI and was not directly related to any of the LBs. However, CEC was related to OTI as well as the competence motivation LV. OTI at T2 had moderate, significant relations with all of the LBs assessed at T3.

In sum, results for outer and inner models demonstrated sufficient reliability and validity, after elimination of several ill-performing manifest variables (much like correcting for modification indices in other structural modeling methods). These sufficiently robust findings allowed us to evaluate the path models in Research Questions 1 through 4.

3.3. Overall Path Model

The final model for the entire sample (see figure 2) depicts the inner model with its path coefficients, which are equivalent to standardized beta regression coefficients. For Research Question 1, hypotheses put forward were only partially upheld; CEC in preschool predicted competence motivation in kindergarten (as well as contemporaneously measured HEC). These significant direct effects (see Figure 2) are similar to associations found in Table 2, where HEC was related only to T2 OTI, and no kindergarten LBs. For Research Question 2, hypotheses again were partially upheld. Preschool HEC at T1 was indirectly related to kindergarten all three LB LVs via OTI. The paths just described are for direct effects. For results where there was an indirect path to LBs via OTI , we calculated indirect effects via the Sobel test (Soper, 2013). Sobel test z-scores showed that OTI significantly mediated the effects of HEC on all three LB LVs.

Figure 2.

Longitudinal Findings: Executive Control Predicting On-Task Involvement and Learning Behaviors

3.4. Moderation of Path Model by Gender (Research Question 3)

We next examined gender differences in the overall path model. To confirm the significance of variation between boys and girls in this model, the pooled estimator for variance t-test was calculated for each pathway. Paths noted in red in Figures 3 and 4 illustrate significant gender differences in relations among study variables. For boys, the relation between HEC only and all three LBs was indirect via OTI (with two paths from OTI to LBs significantly different than girls’); However, for girls, that indirect role from HEC was diminished; their OTI directly predicted only their attention/persistence (LB). Thus, OTI predicted LBs more strongly for boys than girls.

Figure 3.

Longitudinal Finding, Boys Only: Executive Control Predicting On-Task Involvement and Learning Behaviors

Figure 4.

Longitudinal Finding, Girls Only: Executive Control Predicting On-Task Involvement and Learning Behaviors

In comparison, the direct effects of EC on LBs played a significant role for girls. Their CEC predicted all three components of LBs, with all paths significantly different from boys’. Further, HEC significantly predicted girls’ attitudes towards learning, as well as OTI.

Finally, we again tested whether the indirect effects of HEC on LBs indicated significant mediation via OTI. For boys, Sobel z-scores showed that OTI significantly mediated the effects of HEC on all three LB LVs; further, for girls the only possible pathway to test, from HEC to attention/persistence via OTI, also demonstrated significant mediation.

3.5. Moderation of Path Model by Socioeconomic Risk (Research Question 4)

We next examined socioeconomic risk differences in the overall path model. Pooled estimator for variance t-tests were once again conducted on each pathway. Paths noted in red in Figure 6 illustrate significant differences in paths’ significance. Many of the significant pathways found in the overall model were not present in the model for private child care (see Figure 5). For children in private child care, there were direct pathways from CEC to competence motivation and attention/persistence. Also, their HEC predicted OTI, which subsequently predicted attention/persistence. None of these pathways significantly differed from the model for children attending Head Start.

Figure 6.

Longitudinal Finding, Head Start Only: Executive Control Predicting On-Task Involvement and Learning Behaviors

Figure 5.

Longitudinal Finding, Private Daycare Only: Executive Control Predicting On-Task Involvement and Learning Behaviors

The model for children attending Head Start was quite different (see Figure 6). First, the path between CEC and OTI was significant only for Head Start children. Additionally, children in Head Start were found to demonstrate a stronger connection between OTI and all LB LVs. For both private child care and Head Start, CEC directly predicted competence motivation and HEC predicted OTI.

Finally, we again tested whether the indirect effects of HEC on LB LVs indicated significant mediation via OTI. For Head Start students, Sobel z-scores showed that OTI significantly mediated the effects of both HEC and CEC on all three LB LVs. For children enrolled in private child care the only possible pathway to test, from HEC to attention/persistence via OTI, also demonstrated significant mediation.

4. Discussion

Our study focused on building a model that demonstrated the linkage between EC and LBs, especially via OTI. It makes theoretical and empirical sense that aspects of EC attained in early preschool could facilitate OTI, the ability to participate fully in late preschool classroom activities, and that such classroom adjustment could pave a path to positive LBs in kindergarten. However, our comprehensive view of these interrelated developments has not been tested adequately. Further, there has been no study of potentially important distinctions in this model across gender and socioeconomic risk that should be considered in future investigations and applied endeavors. In pursuit of this nuanced view, our main findings and those for moderation, examined longitudinally and collectively, represent additions to the literature. Here we address the meaning of findings for each of these goals via our specific Research Questions.

4.1. Research Question 1

Instead of direct relations of both HEC and CEC with LBs, the only direct effect for the full sample was between CEC and the LB competence motivation. HEC only had a significant direct path to OTI, with mediation via OTI to later LBs. Findings are thus partially confirmatory of the scant research literature, which suggests that HEC may have a stronger relation to OTI than CEC (Miller et al., 2006): Preschool classrooms are replete with emotionally-charged challenges to OTI, so it makes sense that intentionally dealing with motivationally significant problems by inhibiting one’s desires could be useful for maintaining focus for following teacher directions and participating fully in classroom activities. At the same time, it also makes sense that CEC, with its focus on working memory, emotionally neutral focusing and response inhibition, cognitive flexibility, and planning, directly promotes LBs such as competence motivation – the ability to tackle and complete new activities even if they are difficult (Acar et al., 2022; Zelazo et al, 2024).

4.2. Research Question 2

Children who are not engaged/involved are likely to later have difficulty with motivation and negative attitudes towards learning (Halliday et al., 2018). Corroborating this earlier finding, we found that OTI played a supportive role in the development of later LBs. Children who are able to be productively involved in the preschool classroom milieu may be in a better position, through their positive, foundational educational experiences, to enter kindergarten motivated to tackle new and difficult activities, with a positive attitude towards participation in the classroom and ability to persist with difficult tasks. A firm foundation of involvement in the classroom in the penultimate preschool year – and responding positively toward teacher interactions – may stand children in good stead to demonstrate learning behaviors in kindergarten (Halliday et al.).

Taken together, our findings from Research Questions 1 and 2 have theoretical and practical implications. Specifically, theory is advanced by revealing longitudinal associations among three constructs that heretofore were almost always examined cross-sectionally/and or in pairs. Practically, seeing EC as a foundation for later OTI and ultimately LBs suggests that preschoolers and kindergartners could benefit from programming to enhance each of these abilities, especially EC.

4.3. Research Question 3

Overall, the pattern of findings from our gender moderation analysis partially confirmed our hypothesis that pathways between girls’ EC and LBs would be more apparent than those for boys. The greater number of significant pathways for girls suggests that EC plays a more important role for girls’ later LBs than it does for boys. Specifically, in contrast with boys, girls’ CEC directly predicted each facet of LB, and their HEC predicted one aspect of LB. Instead, boys’ OTI predicted each facet of LB.

How to interpret such findings? Perhaps girls’ greater baseline preschool EC, along with their more positive trajectories for its growth (Montroy et al., 2016), put them in a better position to use such skills in the development of kindergarten LBs. That is, girls’ early EC may be sufficiently developed and continuing to develop, and their OTI already more developed via cultural beliefs and socialization (Wanless et al., 2016), that their early EC can directly facilitate LBs. Instead, for boys, the relation between HEC only and all facets of LB is indirect, mediated by OTI.

Thus, boys’ OTI is pivotal for the development of later positive learning behaviors, and deserves study. Although HEC predicts later OTI for both girls and boys, this pathway could provide an entry point in ameliorating boys’ apparent needs; facilitating girls’ LBs might focus more on CEC. Given replication of these findings, educational programming could focus on these differences. The lack of research on the OTI to LB pathway (cf. Halliday et al., 2018), which seems so important for boys, also points to a gap that needs to be filled.

4.4. Research Question 4

We hypothesized that EC may have a larger impact on LBs of children at increased socioeconomic risk. Our findings supported this prediction, with children’s ability to participate productively in the Head Start classroom grounded in both CEC and HEC. A firm foundation in OTI facilitated the LBs Head Start children demonstrated in kindergarten. This mediation by OTI has been masked by its omission from previous research. Thus, our findings are in line with the “success against the odds” idea put forward earlier, suggesting that EC may be especially critical for children who are struggling with all the difficult circumstances associated with poverty, and that focusing on its promotion as a foundation for OTI is especially important for them (Zelazo & Carlson, 2020). Deeper investigation is warranted of the elements of the Head Start curriculum that could give children living in impoverished environments the opportunity to succeed despite their circumstances (Allee-Herndon & Roberts, 2019).

Although children in private child care shared a few of the same pathways as Head Start children to the ultimate outcome of kindergarten LBs, their model was more weakly demonstrated. Why might this be? Promoting EC and OTI would likely be beneficial for them, but could there be other routes to their LBs? As far as we can ascertain, there has been little or no investigation of such possibilities, especially given that much early childhood research examining the importance of EC for later educational success focus upon children at socioeconomic risk. Thus, attention could be profitably given to this issue uncovered in our results.

4.5. Implications and Further Research

Taken together, our results underscore the interrelatedness and importance of EC, OTI, and LBs – all of which have been documented to contribute to later school success; the specific elements of our longitudinal model could assist researchers and practitioners in the development of more effective programming. Regarding the promotion of EC, efficacious interventions specifically focused upon EC via play activities exist and are both short in duration and simple to implement (Schmitt et al., 2015; Traverso et al., 2015). However, transfer from such focused programs to other, nontraining, contexts is limited (Sankalaite et al., 2021; Takacs & Kassai, 2019).

Established curricula may better contextualize facilitation of EC, broadly focusing on creating predictable classroom structure, promoting improvements in language and training emotion regulation, and assisting teachers in their own EC (evidence for the efficacy of such curricula can be found in Bierman et al., 2008, and Raver et al., 2009). Such curricula, via their comprehensive approaches and focus on both structure and autonomy support, give children practice with scaffolded EC within optimal environments (Zelazo & Carlson, 2020). These programs have shown positive results on EC, especially for children at demographic risk (Sasser et al, 2017). Thus, they should not be overlooked, but it is important to acknowledge that they are complex in their application, expensive and often difficult to implement with fidelity.

Perhaps it would beneficial to refocus on more specific elements of early childhood programming that could assist educators in facilitating not only EC, but also OTI and LBs. These might include a holistic focus on structure, routine, verbal rules, with planned lessons, clearly communicated daily agenda and timing of activities, and dependable teacher-child relationships – that is, overall classroom management resulting in emotionally positive, less stressful environments (Muir et al., 2024; Raver et al., 2011; Savina, 2021). Although this approach does not necessarily preclude usage of such curricula as mentioned above, it does not focus on training specific skills and may increase generalization of gains (Muir et al., 2024).

But such programming could be profitably “unpacked” to zero in on the elements of early childhood practice that are pivotal to creating the emotionally positive, calm environments most relevant to the development of EC, OTI, and LB. Several kernels of practice have been enunciated by Zelazo and Carlson (2020; see also Buek, 2019; Muir et al., 2023, Saraç & Tarhan, 2020; Walker et al., 2020): (a) teacher scaffolding via sensitive calibration of challenges; (b) allowing children to practice EC – and also promote their OTI and LB – in autonomy-supportive classrooms (e.g., via encouraging independent exploration and curiosity and emphasizing both mediated and independent sociodramatic play); and (c) promoting children’s reflective thinking (i.e., helping them notice cognitive or emotional/motivational challenges, pause, consider options, and monitor ongoing solutions to the challenges). Other ways of promoting reflective thinking include (a) helping children become aware of how they learn; (b) drawing their attention to strategies they can use for classroom tasks; (c) letting them make decisions about how to work and evaluate that work, and (d) helping them be aware of their emotions while learning (Saraç & Tarhan).

Another kernel of practice focuses on teacher-child classroom-level interactions, especially classroom organization and instructional support (aspects of the holistic approach already elucidated) and dyadic teacher-child relationships (Kellens et al., 2023; Sankalaite et al., 2021). Further, positive dyadic relationships are associated not only with EC, but also OTI and LBs (e.g., Acar et al., 2021; Sankalaite et al; Saraç & Tarhan, 2020). Although classroom-level interaction is often studied as an intervention outcome, more evidence is needed that teacher-child dyadic relationships can be manipulated, especially because positive classroom-level interactions are not as conducive to EC promotion when dyadic relationships are less close or higher in conflict (Sankalaite et al.).

These techniques center upon top-down processes of CEC and HEC, sometimes with attendant facilitation of OTI and LB. However, mitigation of bottom-up processes involved in stress is important so that these kernels can be efficacious. Recent research has focused especially productively on practices that can lessen the pernicious effect of children’s stress upon EC, OTI, and LBs, and be easily incorporated into early child classrooms, including mindfulness practices (Takacs & Kassai, 2019); games (Bai et al., 2022; Kellens et al., 2023), and physical activity (Bai et al; Mulvey et al., 2018; Olive et al., 2024).

It is encouraging that early childhood educators can implement approaches and activities that can promote EC supportive of OTI and subsequent LBs, but more research will help to further pinpoint useful practices and assisting teachers in their implementation. Programming specifically targeting OTI and LBs also is missing. Furthermore, although many intervention studies have focused upon preschoolers at socioeconomic risk (Sankalaite et al., 2021), few have searched for effects for young children not at socioeconomic risk (an exception is Kellens et al., 2023).

Similarly, although gender differences in our models show that young boys’ development differs from girls’, few have investigated how to profitably promote these skills in boys; many intervention studies treat gender as a covariate, rendering recommendations for boys (and girls, for that matter) nonexistent. Some evidence suggests that boys’ developing EC profits more than girls’ from both parental factors such as sensitivity/responsiveness (Vrantsidis et al., 2022) and positive teacher interactions (Sankalaite et al., 2021). In addition, certain motor-based programs may be especially beneficial for boys (Otero et al., 2014). Given further evidence, these avenues might prove fruitful in pursuing ways to promote boys’ EC and its sequelae.

4.6. Limitations

Though this investigation has many advantages stemming from its multi-method, multi-source approach, there are a few potential limitations that remain. As noted earlier, substantial attrition between T1 and T3. The majority of the attrition at T3 was attributable to family mobility and/or difficulty in recruiting elementary school participation. A goal of future longitudinal work in this area obviously should attempt to control attrition, strategizing how to retain participants despite family mobility and other characteristics that could hamper continued involvement.

Further, we were constrained from asking parents about income by our private child care partners, and although we felt that our proxy for socioeconomic risk was adequate, it does pose potential issues. For example, Head Start teachers, compared to those in private child care facilities, often receive more professional development, are more likely to have college degrees, have more resources in their classrooms (e.g., Son et al., 2013), and these factors may impact their ability to support self-regulation. These differences suggest that a more fine-grained operationalization of socioeconomic risk might be preferred.

Another potential limitation pertains to sampling. Although the sample size of this investigation certainly meets the established requirements for PLS analysis, it would have been ideal if the sample were larger and even more diverse. With a more diverse, larger group of participants, child race/ethnicity could be considered in the model, to tailor educational response to findings.

4.6. Conclusions

Given the consequences for school readiness of young children’s difficulties with EC, OTI, and LB – particularly for children living in poverty – it is critical that researchers, early childhood educators, and policy makers better understand these abilities and their understudied interrelations (Raver et al., 2009). Toward this goal, we evaluated a model linking EC and LBs through OTI and varying across gender and socioeconomic risk. We found that the mediating influence of OTI predicting LBs, usually with respect to HEC, was notable across Research Questions. In contrast, CEC often was a more direct predictor of later LBs. Future research could build from this distinction to investigate how components of HEC and CEC could be leveraged to promote both OTI and LBs. In particular, highlighting the role of OTI in future research also might help avoid the decline in LBs at school entry discovered for children at socioeconomic risk (McDermott et al., 2018). Expanded examination is warranted to promote the skills in our entire model, including the important gender and socioeconomic differences we encountered.

Highlights.

• Modeled preschool self-regulation, on-task involvement and later learning behaviors

• On-task involvement often mediated path from self-regulation to learning behaviors

• Pathways from preK to kindergarten differed for cool and hot executive control

• Gender and level of socioeconomic risk moderated these longitudinal pathways

• Findings suggest need for early childhood educators to consider these pathways