Investigating the developmental timing of self-regulation in early childhood

Abstract

Self-regulation often refers to the executive influence of cognitive resources to alter prepotent responses. The ability to engage cognitive resources as a form of executive process emerges and improves in the preschool-age years while the dominance of prepotent responses, such as emotional reactions, begins to decline from toddlerhood onward. However, little direct empirical evidence addresses the timing of an age-related increase in executive processes and a decrease in age-related prepotent responses over the course of early childhood. To address this gap, we examined children’s individual trajectories of change in prepotent responses and executive processes over time. At four age points (24 months, 36 months, 48 months, and 5 years), we observed children (46% female) during a procedure in which mothers were busy with work and told their children they had to wait to open a gift. Prepotent responses included children’s interest in and desire for the gift and their anger about the wait. Executive processes included children’s use of focused distraction, which is the strategy considered optimal for self-regulation in a waiting task. We examined individual differences in the timing of age-related changes in the proportion of time spent expressing a prepotent response and engaging executive processes using a series of nonlinear (generalized logistic) growth models. As hypothesized, the average proportion of time children expressed prepotent responses decreased with age, and the average proportion of time engaged in executive processes increased with age. Individual differences in the developmental timing of changes in prepotent responses and executive process were correlated r = .35 such that the timing of decrease in proportion of time expressing prepotent responses was coupled with the timing of increase in proportion of time engaging executive processes.

Keywords

Early childhood self-regulation developmental timing

The development of self-regulation in early childhood is essential for children’s competent functioning, including socioemotional development and school readiness (Blair & Raver, 2015; Moffitt et al., 2011). Longitudinal evidence suggests that even at this young age, individual differences in self-regulation are related to adolescent and adult outcomes (Casey et al., 2011; Watts et al., 2018). Models of self-regulation often conceptualize it as the influence of effortful engagement of cognitive resources on prepotent responses (Cole et al., 2019). From a developmental perspective, self-regulation emerges in early childhood as a function, in part, of the development of cognitive abilities—for example, memory, attention control—that can be recruited in the form of strategies that can help children regulate prepotent responses (Perry & Calkins, 2018). Thus, it is assumed that an age-related increase in the use of strategies that depend on cognitive abilities co-occurs with an age-related decrease in children’s expressions of prepotent responses, yet few studies actually test this proposition.

Studies of early childhood self-regulation often rely on paradigms that require children to cope with a delayed reward. The waiting task, commonly used to observe behavioral indicators of self-regulation in young children (Adrian et al., 2011), requires children to wait to open a gift (Vaughn et al., 1984). Specifically, these tasks are designed to (a) evoke specific prepotent responses, namely children’s interest in and desire for the reward (the gift) and anger about waiting (a blocked goal), and (b) provide an opportunity to observe what strategies children employ, particularly distracting themselves. Distraction involves the reallocation of attention away from the restricted gift and to other appropriate activities (e.g., play); it depends on the maturation of an executive attention network (Rothbart et al., 2011) that permits effortful control of attention and is consistently correlated with less anger (Calkins & Johnson, 1998).

The evidence establishes that individual differences in preschool-age children’s self-regulation skills are related to many concurrent and later outcomes, but few longitudinal studies investigate the development of self-regulation itself. Theoretically, the age-related decline in anger reactivity when goals are blocked is explained, in part, by increasing engagement of cognitive resources in the form of strategies. The steepest decline in anger reactivity frequency, duration, and intensity and increase in latency to anger occur during the second year of life (Barry & Kochanska, 2010; Cole et al., 2011). Overall, negative emotionality continues to decline between preschool and middle childhood (Blandon et al., 2008). In sum, the available evidence is consistent with the proposition of parallel developmental trajectories of prepotent responses and executive processes in early childhood. We used a series of nonlinear growth models to directly test the hypothesis that the timing of the decline in prepotent responding was associated with the timing of growth in executive processes between age 24 months and 5 years.

Prepotent responses are relatively automatic, highly probable responses that are elicited by a situation. Our study utilized a commonly used procedure in which children must wait to open a gift they want. The utility of the task is that children want to open the gift and often spend time focusing on the gift itself, asking to open it, or actively trying to open it. The functionalist view of emotion defines anger as an appraisal that a goal (e.g., opening the gift) is blocked but achievable and prepotent readiness to act to overcome the obstacle (e.g., trying to open the gift) (Barrett & Campos, 1987). Therefore, attention focused on the gift (a proxy for desire) and anger about waiting are considered indices of prepotent responses. Behavioral responses associated with desire and anger must be regulated to meet the demands of the waiting task. The common use of the waiting task with children between the ages of 2 through 9 years indicates that desire for the gift and frustration about waiting continue to occur but that children may come to regulate prepotent responses (Cole et al., 2011). The few longitudinal studies indicate a decline in anger expressions from toddlerhood to preschool age (e.g., Roben et al., 2013). Thus, although we expected to find individual differences in children’s trajectories of prepotent responses, we hypothesized, on average, a decrease in the proportion of task time children spent expressing prepotent responses between toddler and preschool-age years.

Executive process is a broad term referring to the engagement of various cognitive resources that can be recruited to regulate enacting prepotent responses. The ability to distract from a restricted reward that elicits prepotent desire and frustration relies on the executive attention network, which emerges during the third year of life (Rueda et al., 2005). Whereas infant focused attention is largely governed by novelty and is fleeting, the maturation of the anterior attention system in the preschool-age years enables more voluntary, sustained attention control (Rothbart et al., 2011; Ruff & Capozzoli, 2003). Previous research indicates marked age-related changes in frequency and duration of selective and focused attention during preschool years (Garon et al., 2008). In the waiting task, redirecting one’s attention via distraction is considered important for self-regulation (e.g., Gilliom et al., 2002). In waiting tasks, distraction is a prevalent strategy in which children shift their focus away from the object they desire and reallocate it to an appropriate activity. Distraction is thought to manage desire for the gift as well as the frustration and anger that emerges from the appraisal of a blocked goal (i.e., opening the desired gift). The frequency, duration, and relative use of distraction increases from ages 18 to months to 5 years, and this growth is related to independent ratings of children as better-regulated (Ratcliff et al., 2021). Thus, we hypothesized a developmental increase in the proportion of task time engaged in distraction (i.e., executive process) from toddler to preschool age.

In addition to the age-related changes in children’s executive processes and prepotent responses, there is evidence to suggest that the influence of executive processes on prepotent responses changes with age, becoming more effective as children advance through the preschool-age years (Cole et al., 2020). However, there is limited longitudinal evidence of a direct relation between developmental changes in executive processes and prepotent responses. Cole and colleagues (2011) showed that children engage in distraction more quickly and for longer as they reach preschool age and express anger less quickly and for a shorter duration. Furthermore, for older children (36 and 48 months), more time spent engaging in distraction was associated with longer latency to anger; however, for younger children (18 and 24 months), there was no association between duration of distraction and latency to anger. We expand on those findings by investigating a broader developmental question examining how changes in executive processes engagement and prepotent response expression are coupled in timing across early childhood. Based on the evidence, we hypothesized that executive processes and prepotent responses are inversely related, such that growth in time spent in an executive process-based strategy is associated with the decline in prepotent response-based behavior during this pivotal early childhood period.

The Present Study

In sum, we examined the development of two components of self-regulation (prepotent responses and executive processes) across early childhood. Using a longitudinal approach, we observed children’s behavior during the waiting task at ages 24 months, 36 months, 48 months, and 5 years. The waiting task is particularly well-suited for examining age-related changes in behavior because the task does not require substantial changes in administration as children age. Guided by prior conceptual models and evidence, we hypothesized that the proportion of time spent expressing prepotent responses (i.e., anger and desire) during an 8-min waiting task decreases with age and that the proportion of time engaged in executive processes (i.e., distraction) increases with age. Given that self-regulation is defined by the interrelation of prepotent responses and executive processes (Cole et al., 2019), we also examined whether individual differences in the decline in prepotent responses were accompanied by increases in the engagement of executive processes. Specifically, we hypothesized that the children who exhibited earlier declines in prepotent responses would also exhibit earlier increases in the engagement of executive processes. Results of nonlinear, generalized logistic growth models that extract timing information from available repeated measures provide one of the first empirical tests of whether developmental changes in prepotent responses are coupled with developmental changes in executive processes.

Method

Participants

Participants were 120 children (46% female) and their families from rural and semi-rural communities in the mid-Atlantic region of the United States, who completed the waiting task observations of a larger study. Participants were recruited via flyers distributed in community settings and letters sent to eligible families identified through birth announcements. Eligible participants had incomes below the national median income but above the national poverty threshold. Time 1 family income-to-needs ratio (INR), a measure of ability to meet basic needs relative to federal poverty standard, was 2.32 (SD = .87), indicating economic strain above the poverty line (INR = 1) but below middle class (INR = 3). Most mothers (64%) and approximately half of fathers (54%) had at least some college education. Mothers self-identified as working full-time (40%), working part-time (31%), or being unemployed (29%). Most fathers self-identified as working full-time (90%). Typical of the region, most children were identified by mothers as White (93% White; 7% biracial). While enrolled in the study, participants received annual feedback, newsletters, and compensation for participation. Informed consent was obtained from all adult participants in the study, and all study procedures were approved by The Pennsylvania State University’s institutional review board (Protocol #18993).

Longitudinal Procedure

Children were observed during a waiting task at ages 24 months (n = 111, 92%), 36 months (n = 120, 100%), 48 months (n = 120, 100%), and 5 years (n = 96, 80%). We use 24, 36, and 48 months to describe the first three observations because children were seen within 2 weeks of their birthday. At the fourth observation, children were seen when they were between age 5 years 0 months and 5 years 11 months. Children’s mean ages at each observation were 24.39 (SD = 1.3), 36.44 (SD = .80), 48.33 (SD = .67), and 68.20 (SD = 2.47) months, respectively. Of the 120 children, 115 participants completed all four assessments. Three children did not participate in the 36-month visit and two children had abbreviated task administration due to child distress. There was no evidence that missing data were systematically related to any of the measured variables (Fs < 1, ps > .250).

Waiting Task Procedure

Children and their mothers completed multiple tasks at each laboratory visit, including a waiting task (Vaughn et al., 1984). Mothers were informed in advance that they would complete questionnaires and instruct their children that they must wait until the work was done before they opened a gift. Mothers were asked to behave as they normally would when they are busy and their children must wait for something. The experimenter seated the mother and child at separate tables, gave the child one boring toy (e.g., toy car without wheels) and the mother questionnaires, and placed the shiny wrapped gift on the child’s table. Once the experimenter left the room, the mother told her child, “This is a surprise for you, but you have to wait until I am done with my work to open it.” Eight minutes (480 s) later, the experimenter returned, and the mother told the child to open the gift. The task was video recorded for later coding.

Measures

Independent teams coded children’s emotion expressions and behaviors. Coders were trained to at least 80% agreement with master coders and met weekly to resolve questions and receive feedback; 15% of cases were double-coded. Target emotion expressions and behaviors were aggregated into prepotent responses and executive processes indices that were measurement invariant across the four occasions.

Prepotent Responses

Children’s prepotent response score was derived from second-by-second codes for interest in and desire for the gift and frustration and anger about waiting for it. Interest and desire for the gift were inferred from attention focused on the gift (e.g., staring at, trying to open the gift). Anger was inferred from angry bids to mother about waiting (e.g., “I want it!”) and disruptive behavior (e.g., throwing the toy, interfering with mother’s work) (all Cohen’s κ > .83). Anger expressions (e.g., brow furrowing, pursed/pressed lips, clenched jaw, square open mouth, harsh vocal quality) were coded on a 0–3 scale (absent to strong intensity; intra-class correlation [ICC] = 0.87). For our purposes, we did not consider the intensity of anger expressions, only the presence or absence of any of the above indices of prepotent responses. To obtain a summary score for each child that was measurement invariant across ages, prepotent responses (PR) proportion was calculated for each child at each assessment by dividing the total number of seconds a child engaged in any prepotent response by the total task time or total seconds the child was observed to account for missing data (480 s for most children).

Executive Processes

Children’s executive process score was a composite of second-by-second codes for distraction, defined as self-initiated, appropriate engagement in another object or activity (Cohen’s κ > .87). Distraction only included focused attention on an alternate activity (e.g., playing with the boring toy, looking at the posters on the walls, counting objects in the room, etc.). Distraction did not include attention focused on the gift or bids to the child’s mother. Parallel to prepotent responses, our measurement of executive processes (EP) proportion was calculated for each child at each assessment by dividing the total number of seconds a child engaged in distraction by the total task time or total seconds the child was observed to account for missing data (480 s for most children).

Age

Chronological age in months was a continuous variable that is used as a proxy of all age-related developmental processes. Exact age at each visit was calculated as the number of months between a child’s birth date and the visit date.

Data Analysis

Individual differences in the timing of age-related changes in prepotent responses and executive processes were examined using a multivariate nonlinear growth model based on prior work where sigmoid functions (e.g., logistic, Gompertz) were used to describe developmental processes (Grimm et al., 2017; Ram & Grimm, 2015). Age-related increases in children’s executive processes during the task were modeled using a generalized logistic growth function (Richards curve) that incorporated the assumption that all individuals’ engagement of executive processes would start at a minimum level (a lower asymptote) and end at a maximum level (an upper asymptote), whether or not the observation period captured the entire process (see also Grimm & Ram, 2009). More specifically, the four repeated measures of the proportion of time that individual i spent engaged in executive processes (EP) at occasion t were modeled as

E P_{t i} = β_{E P 0} + (β_{E P 1} - β_{E P 0}) {\frac{1}{{(1 + τ_{E P} e x p [- α_{E P} (a g e_{t i} - λ_{E P i})])}^{\frac{1}{τ_{E P}}}}} + e_{E P t i}

(1)

where β_EP0 defines the lower asymptote, in our case fixed = 0 (no executive processes engaged); β_EP1 defines the upper asymptote, in our case fixed = 1 (executive processes always engaged); λ_EPi is a person-specific coefficient that captures the relative timing of executive processes development, specifically the age at which an individual reaches an inflection point where development is moving most quickly; α_EP is a rate parameter that indicates the prototypical rate of development (i.e., rate governing change from the lower to the upper asymptote); τ_EP is an asymmetry parameter that indicates where the most rapid change (i.e., point of inflection) is located in relation to the lower and upper asymptotes; and e_EPti is residual time-specific error assumed normally distributed with a mean of zero and standard deviation σ_eEP. In parallel, the four repeated measures of proportion of time that individual i spent expressing prepotent responses at occasion t were modeled as

P R_{t i} = β_{P R 0} + (β_{P R 1} - β_{P R 0}) {\frac{1}{{(1 + τ_{P R} e x p [- α_{P R} (a g e_{t i} - λ_{P R i})])}^{\frac{1}{τ_{P R}}}}} + e_{P R t i}

(2)

where β_PR0 defines the lower asymptote, in our case fixed = 0 (no prepotent responses expressed); β_PR1 defines the upper asymptote, in our case fixed = 1 (prepotent responses always expressed); λ_PRi is a person-specific coefficient that captures the relative timing of developmental changes in prepotent responses, specifically considered here as the age at which an individual would reach an inflection point where development is moving most quickly; α_PR is a rate parameter that indicates the prototypical rate of development; τ_PR is an asymmetry parameter that indicates where the point of inflection is located in relation to the lower and upper asymptotes; and e_PRti is residual time-specific error assumed normally distributed with a mean of zero and standard deviation σ_eEP, and that may be correlated, ρ(e_PR, e_EP), with e_EPti.

In turn, the interindividual differences in the timing of executive processes and prepotent responses development were modeled as

λ_{E P i} = γ_{E P 00} + u_{E P i}

(3)

and

λ_{P R i} = γ_{P R 00} + u_{P R i}

(4)

where γ_EP and γ_PR indicate the timing of the development of executive processes and prepotent responses for the prototypical individual in the sample, and u_EPi and u_PRi indicate individual differences in executive processes and prepotent responses timing. Extent of differences in timing, which are assumed normally distributed, is given by the standard deviations σ_λEP and σ_λPR, and the correlation between those differences in timing is given by ρ(λ_PR, λ_EP). Specific to our hypotheses, we expected that the rate parameter indicating the prototypical trajectory for prepotent responses, α_PR, would be negative, indicating an age-related decline in prepotent responses; that the rate parameter indicating the prototypical trajectory for executive processes, α_EP, would be positive, indicating an age-related increase in executive processes; and that the coupling (used here as a general term for the cross-variable association) parameter, ρ(λ_PR, λ_EP), would be positive, indicating that the children who exhibited earlier declines in prepotent responses also tended to exhibit earlier increases in the engagement of executive processes.

For model fitting, we used a Bayesian analysis framework that allowed estimation of the complex, multivariate model in a manner that propels inferences about how individual-level change processes proceed across childhood. Specifically, all models were fit using the brms package in R (Bürkner, 2017) using four chains with 20,000 iterations, burn-in period of 10,000 with 10 step thinning; total of 4,000 post warm-up samples. We used default priors, a half Student’s-t prior with 3 degrees of freedom and scale parameter of 10 for standard deviations of random effects, a Lewandowski–Kurowicka–Joe prior with parameter of 1 for correlations among random effects, and a normal prior with weakly informative means for regression coefficients. Incomplete data were treated using standard missing at random assumptions. Convergence of the Markov Chain Monte Carlo (MCMC) algorithms was determined through graphical checks of the chains and inspection of posterior distributions and rhat values, all of which suggested that MCMC chains had converged. In model building, we compared generalized asymmetric (Richards) and standard logistic models using widely applicable information criterion (WAIC, Watanabe, 2010) and the leave-one-out information criterion based on posterior likelihoods (LOOIC, Vehtari et al., 2017). Across specifications, the bivariate Richards model (e.g., LOOIC = −571.9, SE = 43.5) always had better fits and predictions than the bivariate standard logistic model (e.g., LOOIC = −368.2, SE = 39.5, ΔELPD [expected log predictive density] = −101.8, ΔSE = 9.9). Relative fits and results from a full array of (univariate and bivariate linear, quadratic, exponential, logistic, and Richards) models indicated that the pattern of results was robust to differences in model specification.

Results

Descriptive statistics (Table 1) and corresponding box-and-whisker plots in the upper panel of Figure 1 show that the average proportion of task time children expressed prepotent responses decreased nonlinearly with age, from .51 at the age 24 months visit to .22 at the final age 5 visit. In complementary fashion, the average proportion of task time that children engaged executive processes increased nonlinearly with age, from .20 at the age 24 month visit to .48 at the final age 5 visit. The individual trajectories (lower panel of Figure 1) indicate the presence of individual differences in how children’s prepotent responses and executive processes changed with age.

Table 1.

Proportion of Waiting Task Children Exhibited Prepotent Responses and Executive Processes by Age.

Visit	Age in monthsMean (SD)	Prepotent Responses proportionMean (SD) [min, max]	Executive Processes proportionMean (SD) [min, max]
24 months (n = 111)	24.39 (1.30)	.51 (.24) [.02, .99]	.20 (.14) [.00, .54]
36 months (n = 117)	36.44 (0.80)	.30 (.20) [.02, .93]	.25 (.17) [.00, .67]
48 months (n = 117)	48.33 (0.67)	.26 (.18) [.00, .82]	.36 (.21) [.00, .87]
5 years (n = 96)	68.20 (2.47)	.22 (.14) [.02, .79]	.48 (.18) [.00, .87]

Note: Prepotent Responses and Executive Processes proportions calculated as the proportion of the total task time (480 s for most children) that the child exhibited prepotent responses or executive processes. SD = standard deviation, min = minimum observed score, max = maximum observed score.

Figure 1.

Upper Panel Shows Descriptive Statistics Using a Box-and-Whisker Plot of Prepotent Responses and Executive Processes (Distraction) Proportion of Task Time (480 s for Most Children) over Ages 24, 36, 48, And 5 Years. Lower panel shows individual trajectories of the proportion of task time (480 s for most children) that a child exhibited prepotent responses (PR; left) and distraction (right) at 24, 36, 48, and 5 years. N = 120.

These individual differences in prepotent responses and executive processes development were formally examined using a multivariate nonlinear growth model that specifically estimated how individual differences in the timing of prepotent responses development related to individual differences in the timing of executive processes development (see Table 2). In line with hypotheses, children’s prepotent responses (PR) decreased with age. The prototypical child’s development of prepotent responses from an assumed upper asymptote β_PR1 = 1 where they would continually exhibit prepotent responses pulled for by the task (i.e., frustration and anger) to an assumed lower asymptote β_PR0 = 0 was specifically described by three parameters: a rate parameter α_PR = −1.62 (95% credible interval [CI]: [−3.03, −0.52]) indicating how quickly prepotent responses decreases; an asymmetry parameter τ_PR = 68.12 (95% CI: [21.83, 122.75]) placing the sigmoidal inflection point (i.e., “middle” of the S) along the y-axis near the upper asymptote, and a timing parameter λ_PR = −5.00 (95% CI: [−13.97, 2.35]) locating that inflection point along the lower end of the x-axis. This prototypical trajectory is shown by the bold red line in the main panel of Figure 2, with the bold red circle indicating the projected point of inflection at age −5.00 months. Of note, the specific location of the inflection point is not meant for literal interpretation. This parameter merely provides optimal representation of the observed nonlinear developmental trajectories and the quantification of between-person differences in timing. Importantly for our purposes, there were, as indicated by the thin red lines, substantial (and inferentially meaningful) individual differences in the timing of prepotent responses development, σ_λPR = 5.64 (95% CI: [0.44, 10.24]). Some children’s prepotent responses declined earlier, and some children’s prepotent responses declined later.

Table 2.

Results from Bivariate Nonlinear Asymmetrical Growth Model Examining Age-Related Change in Prepotent Responses and Executive Processes.

	Estimate (SE)	Nonlinear age-related change in prepotent responses and executive processes
	Estimate (SE)	95% CI	PD	Rhat	Bulk ESS	Tail ESS
Prepotent responses (PR)
Fixed effects
α_PR	−1.62 (.62)	[−3.03, −0.52]	100%	1.00	1831	683
λ_PR	−5.00 (4.19)	[−13.97, 2.35]	89.50%	1.00	3758	3721
τ_PR	68.12 (26.04)	[21.83, 122.75]	100%	1.00	2207	789
Random effects
σ_λPR	5.64 (2.63)	[0.44, 10.24]	100%	1.00	2642	3408
Residual, σ_ePR	0.19 (0.01)	[0.18, 0.21]	100%	1.00	2874	3650
Executive processes (EP)
Fixed effects
α_EP	.79 (0.41)	[0.11, 1.67]	100%	1.00	1367	506
λ_EP	97.04 (4.68)	[85.91, 105.41]	100%	1.01	1494	521
τ_EP	38.94 (20.13)	[5.39, 82.34]	100%	1.00	1329	483
Random effects
σ_λEP	8.20 (2.28)	[2.81, 12.27]	100%	1.00	2946	2821
Residual, σ_eEP	0.17 (0.01)	[0.15, 0.18]	100%	1.00	3584	3902
ρ(λ_PR, λ_EP)	0.35 (0.37)	[−0.62, 0.92]	85.67%	1.00	2485	2514
ρ(e_PR, e_EP)	−0.44 (0.04)	[−0.52, −0.36]	100%	1.00	3296	3029
R ² (PR prop)	0.29 (0.05)	[0.20, 0.38]
R ² (EP prop)	0.36 (0.04)	[0.28, 0.43]

Note: N = 120. Unstandardized estimates, standard errors (SE), 95% credible intervals (CI) [lower, upper], and Probability of Direction (PD) from Bayesian fitting of nonlinear multilevel model; σ parameters are standard deviations; ρ parameters are correlations. Age scaled in months. Model based on up to four occasions nested within 120 participants, total of 441 observations.

Figure 2.

Model-Implied Developmental Trajectories of Prepotent Responses (Red) & Executive Processes (Blue) Across Early Childhood, with Specific Concentration on Interindividual Differences in Timing. Scatterplot insert shows interindividual differences in timing. N = 120.

Similarly in line with hypotheses, children’s executive processes increased with age. The prototypical child’s development of executive processes (EP) from an assumed lower asymptote β_EP0 = 0 where they exhibited no engagement of executive processes to an upper asymptote β_EP1 = 1 where they would engage executive processes for the entirety of the task was also specifically described by three parameters: a rate parameter α_EP = .79 (95% CI: [0.11, 1.67]) indicating how quickly executive processes increases; an asymmetry parameter τ_EP = 38.94 (95% CI: [5.39, 82.34]) placing the inflection point on the y-axis nearer the upper asymptote; and a timing parameter λ_EP = 97.04 (95% CI: [85.91, 105.41]) locating the inflection point at the upper end of the x-axis. This prototypical trajectory is shown by the bold blue line in the main panel of Figure 2. Again, as indicated by the thin blue lines, there were substantial (and inferentially meaningful) individual differences in the timing of executive processes development, σ_λEP = 8.20 (95% CI: [2.81, 12.27]). Some children’s executive processes increased earlier, and some children’s executive processes increased later.

Of primary interest were individual differences in the timing of age-related change in prepotent responses in relation to the timing of executive processes development, ρ(λ_PR, λ_EP) = .35 (95% CI: [−.62, .92]). In line with hypotheses, the bulk of evidence indicated this correlation was positive (probability of positive parameter = 85.67%). Although not a causal, within-person relation, this result provides evidence that between-person differences in development of prepotent responses and executive processes are coupled. Children with earlier executive processes increases were likely to have earlier prepotent response decreases.

Discussion

The results provide some of the first evidence that the developmental timing of age-related declines in prepotent responses is coupled with age-related increases in executive processes. The findings are consistent with the long-standing hypothesis that developmental growth in strategies that engage cognitive resources contributes to the successful regulation of prepotent responses. The nonlinear growth model used to capture and describe changes in waiting task behavior across early childhood revealed (a) age-related decreases in the proportion of time children spent engaging in prepotent responses and concomitant age-related increases in the proportion of time they spent engaging executive processes (distraction), (b) individual differences in the timing of developmental changes in both prepotent responses and executive processes, and (c) that the individual differences in the timing of decreased prepotent responding were correlated with individual differences in the timing of increased engagement of executive processes.

The proportion of task time spent expressing prepotent responses decreased with age, consistent with theoretical accounts of development (e.g., Kopp, 1982) and the limited longitudinal evidence that suggests anger expressions decline as children gain regulatory skills (Roben et al., 2013). For the purposes of our study, we examined all behaviors that must be regulated during the waiting task as prepotent responses (e.g., touching the gift, disruptive behavior, angry bids) and from these behaviors infer children’s desire for the gift and anger about the wait. An interesting question, however, is how desire and anger relate to each other and whether there are distinct trajectories of change in them during early childhood. Possibly, desires do not subside, but children learn to better regulate their anger.

Of note, our projections from asymmetric nonlinear growth models that assume children progress from exhibiting prolonged and continuous prepotent responses (upper asymptote) to exhibiting no prepotent responses (lower asymptote) located the period of the most rapid change (inflection point) very early in life. However, because the model was specifically constructed to accommodate nonlinearity in the available data (age 2 to 5 years), strong inferences about the shape of development during infancy are not warranted. Rather, we suggest further consideration of how age-related changes in children’s prepotent desires and discomforts during infancy may serve as a prelude to differences in the development of self-regulation in early childhood. For instance, anger reactivity increases as children gain locomotor skills (Campos et al., 2000; Roben et al., 2012). A future research goal is to identify multiple periods of marked change in prepotent responses as children are able to both act on the environment and purposively refocus their attention.

As hypothesized, the proportion of time spent engaged in a strategy that relies on an executive process increased with age. As Kopp (1982) postulated, the ability to engage executive processes emerges clearly for most children around 3 years of age. This finding aligns with evidence that the ability to refocus attention, which is integral to distraction, increases from 24 to 48 months (Ruff & Capozzoli, 2003) and that the use of distraction as a strategy increases and is quicker to occur during preschool-age years (Cole et al., 2011; Supplee et al., 2011).

Here again, our asymmetric nonlinear growth model suggests the most rapid increase in executive processes may be occurring in later childhood. Although self-regulation emerges as early as 3 or 4 years, the effectiveness of self-regulatory attempts on altering prepotent desire and frustration at this age is limited (Cole et al., 2020). Executive processes development likely increases throughout childhood and perhaps even accelerates as children’s cognitive resources continue to develop and they experience greater expectations and more opportunities to control behavior (Blair & Ursache, 2011). Working memory and set shifting show the greatest gains after age 5 and into middle childhood (Best et al., 2009), and executive attention improves from age 7 to 10 years (Simonds et al., 2007). Although our ability to make inferences about self-regulation outside our observation window (age 2 to 5 years) is limited, our findings suggest the value of further investigating how these psychological processes contribute to children’s self-regulation given that, even at age 5 years, children in our sample engaged in distraction, on average, less than half of the total task time.

The waiting task mimics types of waiting children are expected to tolerate in school and at home; adults hope children will settle into an activity and stay focused on it until adults can attend to children’s desires, the equivalent of children remaining distracted for most of the waiting task. The available, if limited, evidence is that children engage in distraction briefly; the average length of a single bout of distraction is 1 min at 48 months of age (Cole et al., 2011). Future research may examine how children transition from brief to sustained distraction when they must wait for something they want.

Beyond average prepotent responses and executive processes trajectories, we found considerable individual differences in prepotent responses and executive processes trajectories. Differences in timing of prepotent responses decreases and executive processes increases spanned 13.83- and 27.66-month windows, respectively, providing clear evidence that the developmental timing of behavioral changes in self-regulation is quite heterogeneous (Bendezú et al., 2018; Supplee et al., 2011). Nonetheless, there is also some systematicity in relative timing. As hypothesized, the timing of age-related decline in expressions of prepotent responses was associated with the timing of age-related increases in executive processes engagement. Specifically, children whose prepotent responses declined earlier tended to be the same children whose executive processes increased earlier. Our findings add to evidence suggesting a relation between developmental changes in strategy use and age-related change in latency to anger (Cole et al., 2011) by providing direct evidence of a relation between age-related changes in strategy use and age-related changes in expressions of prepotent responses broadly. Although our finding is based on examination of between-child differences, the evidence of coupling between prepotent responses and executive processes development provides empirical support for the idea that the development of self-regulation lies in the interplay between prepotent responses and executive processes (Cole et al., 2019). This finding opens possibilities to investigate how a wide variety of individual and contextual differences (e.g., temperament, early life stress, caregiving, etc.) contribute to variability in the developmental timing of change in prepotent responses and executive processes.

Our findings must be viewed in relation to a number of limitations. First, the present findings were derived from repeated observations of children’s behavior in a single, albeit commonly used, laboratory task. Most parents reported that their children waited as they normally would, and children claimed they did not remember the tasks at previous visits (Cole et al., 2011). However, there is a need for evidence based on other tasks and situational factors. In addition, our sample represented children from rural/semi-rural, economically strained households who were predominately White. The evidence here suggests it would be valuable to investigate the relation between prepotent responses and executive processes in terms of varying sociocultural expectations, economic stress, and parenting goals that may influence self-regulation development. For instance, previous research indicates that anger regulation is poorer among children in low-income contexts if caregiving is inadequate (e.g., Raver, 2004); thus, future research with samples for which parenting is less adequate may benefit from addressing potential effects on developmental timing of executive processes in relation to prepotent responses. Another potential limitation is the use of executive processes and prepotent responses proportion measures which, although measurement invariant across age, make assumptions about the within-task temporal dynamics (e.g., how executive processes and prepotent responses are related second-to-second). Future work should delve further into the varied, ongoing dynamics between prepotent responses and executive processes during a task and how these within-task dynamics might change with age.

To our knowledge, this is the first study to use multivariate nonlinear growth models to examine the coupling of simultaneously occurring developmental changes. We formulated a model that matched the nature of the repeated measures data (proportion scores that range from 0.0 to 1.0) and could be used to describe developmental processes that span beyond the limited (4) number of repeated measures available. Acknowledging the nonlinearity of development, the sigmoid model allowed us to project from observations obtained in a relatively small window (age 2–5 years) to a larger developmental sequence that proceeded from developmental points prior to data collection (e.g., total absence of executive processes) to continuing developmental achievements (e.g., total regulation of prepotent responses) that likely occur at older ages. Model fitting was facilitated by Bayesian estimation and availability of the computational power now embedded in our computers. However, we still needed to impose some relatively strict assumptions on both the shape of the trajectories and the specific interindividual differences of interest (random effects in timing only). As indicated by the relatively wide confidence intervals for some parameters, we are pushing these N = 120, T = 4 data to the limit. Notably, results were quite robust in that we obtained very similar fits and results when using linear, quadratic, exponential, or sigmoid (logistic and Richard’s) models. Purposively pushing beyond simpler linear and polynomial models that ignore how the observed data are embedded in the measurement space (proportion scores range between 0.0 and 1.0) and a longer-term developmental sequence (prior to age 2 and after age 5 years), we were able to construct a multivariate and very general nonlinear model and obtain interpretable results from four-occasion data. Altogether, the success of the analysis approach suggests that new information can be fruitfully acquired from many researchers’ archived longitudinal data.

In sum, we provide evidence supporting the relation between core components of self-regulation (i.e., prepotent responses and executive processes) over developmental time. Documenting this coupling in developmental timing can help us better understand how and when self-regulation efforts become effective and provide a basis for research that can inform and refine interventions aimed at fostering children’s self-regulation. We utilized a novel nonlinear growth modeling approach to simultaneously examine how children’s prepotent responses and executive processes changed over time. Prepotent responses decreased with age, and executive processes increased with age. Importantly, we found that the developmental timing of children’s decreases in prepotent responses was related to the timing of their increases in executive processes—empirical evidence that supports theoretical notions that self-regulation involves the interplay between prepotent responses and executive processes (Cole et al., 2019). The study capitalized on longitudinal data and nonlinear growth models to directly investigate developmental timing and if and how the coupling of changes in prepotent responses and changes in executive processes simultaneously, and together, contribute to the development of self-regulation in early childhood.

Supplemental Material

sj-docx-1-jbd-10.1177_01650254221111788 – Supplemental material for Investigating the developmental timing of self-regulation in early childhood

Supplemental material, sj-docx-1-jbd-10.1177_01650254221111788 for Investigating the developmental timing of self-regulation in early childhood by Lindsey M. Green, Breana G. Genaro, Kizzann Ashana Ratcliff, Pamela M. Cole and Nilam Ram in International Journal of Behavioral Development

Footnotes

Acknowledgements

Thank you to the study participants for providing data for such an extended period of time, and to the many research assistants who helped obtain such rich data.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute on Health (R01 HD076994, R24 HD041025, UL1 TR002014) and the Penn State Social Science Research Institute.

Informed Consent

Informed consent was obtained from all adult participants in the study, and all study procedures were approved by The Pennsylvania State University’s institutional review board (Protocol #18993).

ORCID iDs

Lindsey M. Green

Kizzann Ashana Ratcliff

Supplemental Material

Supplemental material for this article is available online.

References

Adrian

Zeman

Veits

(2011). Methodological implications of the affect revolution: A 35-year review of emotion regulation assessment in children. Journal of Experimental Child Psychology, 110(2), 171–197. https://doi.org/10.1016/j.jecp.2011.03.009

Barrett

K. C.

Campos

J. J.

(1987). Perspectives on emotional development II: A functionalist approach to emotions. In Osofsky

J. D.

(Ed.), Wiley series on personality processes. Handbook of infant development (pp. 555–578). Wiley.

Barry

R. A.

Kochanska

(2010). A longitudinal investigation of the affective environment in families with young children: From infancy to early school age. Emotion, 10(2), 237–249. https://doi.org/10.1037/a0018485

Bendezú

J. J.

Cole

P. M.

Tan

P. Z.

Armstrong

L. M.

Reitz

E. B.

Wolf

R. M.

(2018). Child language and parenting antecedents and externalizing outcomes of emotion regulation pathways across early childhood: A person-centered approach. Development and Psychopathology, 30(4), 1253–1268. https://doi.org/10.1017/S0954579417001675

Best

J. R.

Miller

P. H.

Jones

L. L.

(2009). Executive functions after age 5: Changes and correlates. Developmental Review, 29(3), 180–200. https://doi.org/10.1016/j.dr.2009.05.002

Blair

Raver

C. C.

(2015). School readiness and self-regulation: A developmental psychobiological approach. Annual Review of Psychology, 66, 711–731. https://doi.org/10.1146/annurev-psych-010814-015221

Blair

Ursache

(2011). A bidirectional model of executive functions and self-regulation. In Vohs

K. D.

Baumeister

R. F.

(Eds.), Handbook of self-regulation: Research, theory, and applications (pp. 300–320). Guilford.

Blandon

A. Y.

Calkins

S. D.

Keane

S. P.

O’Brien

. (2008). Individual differences in trajectories of emotion regulation processes: The effects of maternal depressive symptomatology and children’s physiological regulation. Developmental Psychology, 44(4), 1110–1123. https://doi.org/10.1037/0012-1649.44.4.1110

Bürkner

P. C.

(2017). brms: An R package for Bayesian multilevel models using Stan. Journal of Statistical Software, 80(1), 1–28. http://hdl.handle.net/10.18637/jss.v080.i01

10.

Calkins

S. D.

Johnson

M. C.

(1998). Toddler regulation of distress to frustrating events: Temperamental and maternal correlates. Infant Behavior and Development, 21(3), 379–395. https://doi.org/10.1016/S0163-6383(98)90015-7

11.

Campos

J. J.

Anderson

D. I.

Barbu-Roth

M. A.

Hubbard

E. M.

Hertenstein

M. J.

Witherington

(2000). Travel broadens the mind. Infancy, 1(2), 149–219. https://doi.org/10.1207/S15327078IN0102_1

12.

Casey

B. J.

Somerville

L. H.

Gotlib

I. H.

Ayduk

Franklin

N. T.

Askren

M. K.

Jonides

Berman

M. G.

Wilson

N. L.

Teslovich

Glover

Zayas

Mischel

Shoda

(2011). Behavioral and neural correlates of delay of gratification 40 years later. Proceedings of the National Academy of Sciences of the United States of America, 108(36), 14998–15003. https://doi.org/10.1073/pnas.1108561108

13.

Cole

P. M.

Lougheed

J. P.

Chow

S. M.

Ram

(2020). Development of emotion regulation dynamics across early childhood: A multiple time-scale approach. Affective Science, 1, 28–41. https://doi.org/10.1007/s42761-020-00004-y

14.

Cole

P. M.

Ram

English

M. S.

(2019). Toward a unifying model of self-regulation: A developmental approach. Child Development Perspectives, 1(2), 91–96. https://doi.org/10.1111/cdep.12316

15.

Cole

P. M.

Tan

P. Z.

Hall

S. E.

Zhang

Crnic

K. A.

Blair

C. B.

(2011). Developmental changes in anger expression and attention focus: Learning to wait. Developmental Psychology, 47(4), 1078–1089. https://doi.org/10.1037/a0023813

16.

Garon

Bryson

S. E.

Smith

I. M.

(2008). Executive function in preschoolers: A review using an integrative framework. Psychological Bulletin, 134(1), 31–60. https://doi.org/10.1037/0033-2909.134.1.31

17.

Gilliom

Shaw

D. S.

Beck

J. E.

Schonberg

M. A.

Lukon

J. L.

(2002). Anger regulation in disadvantaged preschool boys: Strategies, antecedents, and the development of self-control. Developmental Psychology, 38(2), 222–235. https://doi.org/10.1037/0012-1649.38.2.222

18.

Grimm

K. J.

Ram

(2009). Nonlinear growth models in M plus and SAS. Structural Equation Modeling, 16(4), 676–701. https://doi.org/10.1080/10705510903206055

19.

Grimm

K. J.

Ram

Estabrook

(2017). Growth modeling: Structural equation and multilevel modeling approaches. Guilford.

20.

Kopp

C. B.

(1982). Antecedents of self-regulation: A developmental perspective. Developmental Psychology, 18(2), 199–214. https://doi.org/10.1037/0012-1649.18.2.199

21.

Moffitt

T. E.

Arseneault

Belsky

Dickson

Hancox

R. J.

Harrington

Houts

Poulton

Roberts

B. W.

Ross

Sears

M. R.

Thomson

W. M.

Caspi

(2011). A gradient of childhood self-control predicts health, wealth, and public safety. Proceedings of the National Academy of Sciences, 108(7), 2693–2698. https://doi.org/10.1073/pnas.1010076108

22.

Perry

N. B.

Calkins

S. D.

(2018). A biopsychosocial perspective on the development of emotion regulation across childhood. In Cole

P. M.

Hollenstein

(Eds.), Emotion regulation: A matter of time (pp. 3–30). Routledge.

23.

Ram

Grimm

K. J.

(2015). Growth curve modeling and longitudinal factor analysis. In Overton

W. F.

Molenaar

P. C. M.

Lerner

R. M.

(Eds.), Handbook of child psychology and developmental science: Theory and method (vol. 1, 7th ed., pp. 758–788). Wiley. https://doi.org/10.1002/9781118963418.childpsy120

24.

Ratcliff

K. A.

Vazquez

L. C.

Lunkenheimer

E. S.

Cole

P. M.

(2021). Longitudinal changes in young children’s strategy use for emotion regulation. Developmental Psychology, 57(9), 1471–1486. https://doi.org/10.1037/dev0001235

25.

Raver

C. C.

(2004). Placing emotional self-regulation in sociocultural and socioeconomic contexts. Child Development, 75(2), 346–353. https://doi.org/10.1111/j.1467-8624.2004.00676.x

26.

Roben

C. K. P.

Bass

A. J.

Moore

G. A.

Murray-Kolb

Tan

P. Z.

Gilmore

R. O.

Buss

K. A.

Cole

P. M.

Teti

L. O.

(2012). Let me go: The influences of crawling experience and temperament on the development of anger expression. Infancy, 17(5), 558–577. https://doi.org/10.1111/j.1532-7078.2011.00092.x

27.

Roben

C. K. P.

Cole

P. M.

Armstrong

L. M.

(2013). Longitudinal relations among language skills, anger expression, and regulatory strategies in early childhood. Child Development, 84(3), 891–905. https://doi.org/10.1111/cdev.12027

28.

Rothbart

M. K.

Sheese

B. E.

Rueda

M. R.

Posner

M. I.

(2011). Developing mechanisms of self-regulation in early life. Emotion Review, 3(2), 207–213. https://doi.org/10.1177/1754073910387943

29.

Rueda

M. R.

Posner

M. I.

Rothbart

M. K.

(2005). The development of executive attention: Contributions to the emergence of self-regulation. Developmental Neuropsychology, 28(2), 573–594. https://doi.org/10.1207/s15326942dn2802_2

30.

Ruff

H. A.

Capozzoli

M. C.

(2003). Development of attention and distractibility in the first 4 years of life. Developmental Psychology, 39(5), 877–890. https://doi.org/10.1037/0012-1649.39.5.877

31.

Simonds

Kieras

J. E.

Rueda

M. R.

Rothbart

M. K.

(2007). Effortful control, executive attention, and emotional regulation in 7–10-year-old children. Cognitive Development, 22(4), 474–488. https://doi.org/10.1016/j.cogdev.2007.08.009

32.

Supplee

L. H.

Skuban

E. M.

Trentacosta

C. J.

Shaw

D. S.

Stoltz

(2011). Preschool boys’ development of emotional self-regulation strategies in a sample at risk for behavior problems. The Journal of Genetic Psychology, 172(2), 95–120. https://doi.org/10.1080/00221325.2010.510545

33.

Vaughn

B. E.

Kopp

C. B.

Krakow

J. B.

(1984). The emergence and consolidation of self-control from eighteen to thirty months of age: Normative trends and individual differences. Child Development, 55(3), 990–1004. https://doi.org/10.2307/1130151

34.

Vehtari

Gelman

Gabry

(2017). Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Statistics and Computing, 27(5), 1413–1432. https://doi.org/10.1007/s11222-016-9696-4

35.

Watanabe

(2010). Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular learning theory. The Journal of Machine Learning Research, 11, 3571–3594.

36.

Watts

T. W.

Duncan

G. J.

Quan

(2018). Revisiting the marshmallow test: A conceptual replication investigating links between early delay of gratification and later outcomes. Psychological Science, 29(7), 1159–1177. https://doi.org/10.1177/0956797618761661

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB