Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Sep 1.
Published in final edited form as: Transl Issues Psychol Sci. 2014 Sep 1;1(3):203–216. doi: 10.1037/tps0000028

Inhibitory Control is Associated with Psychosocial, Cognitive, and Weight Outcomes in a Longitudinal Sample of Girls

Stephanie Anzman-Frasca 1, Lori A Francis 2, Leann L Birch 3
PMCID: PMC4583141  NIHMSID: NIHMS676669  PMID: 26417610

Abstract

Early self-regulation abilities have been highlighted as a robust predictor of adaptive development, but the extant literature has typically focused on outcomes in different developmental domains separately. The aim of the current study was to expand upon this research by testing pathways from girls’ inhibitory control at age 7, an aspect of self-regulation, to their psychosocial, cognitive, and weight outcomes from ages 9 to 15 (n=192). Results supported the hypothesis that greater inhibitory control is independently associated with better subsequent psychosocial, cognitive, and weight outcomes. These findings, combined with evidence that self-regulatory capacities are modifiable in early childhood, offer opportunities for interdisciplinary preventive interventions aiming to promote child health and well-being across domains and over time.

Keywords: inhibitory control, self-regulation, childhood obesity, cognition, well-being

Self-regulation is a broad term that includes self-control, will power, effortful control, delay of gratification, emotion regulation, executive function, and inhibitory control, overlapping constructs with their own specific definitions and measures. Self-regulation abilities emerge at the end of infancy, and marked improvements in purposeful self-regulation are evident across the toddler and preschool years, as children become increasingly able to execute effortful behaviors, such as suppressing activities when asked, lowering their voices, and paying attention (Kochanska, Murray, & Harlan, 2000). Individuals continue to improve their abilities to regulate their own emotions and behavior throughout childhood, adolescence, and even adulthood (Kopp, 1982). Self-regulation abilities can be modified by experience (Golan & Bachner-Melman, 2011; Johnson, 2000; Rothbart, Ellis, & Posner, 2004), and their implications are relative to context. Kagan (2003) provides the example of attention deficit hyperactivity disorder: the associated behaviors are problematic in current industrialized societies where children are expected to sit and learn for six hours per day, but this "disorder" may not have dire implications in a more traditional, agrarian society. Overall, the current environment is paradoxical, characterized by demands to restrain one's impulses but also by highly stimulating contexts, such as ubiquitous inexpensive, palatable foods and technological diversions. Thus, the current environment, built to promote convenience, efficiency, and value (Hill, Wyatt, Reed, & Peters, 2003), could make it difficult for many individuals to attain positive outcomes like a healthy weight status and psychosocial well-being. This paradox underscores the importance of research investigating inhibitory components of self-regulation (e.g., inhibitory control, or the ability to restrain a dominant behavioral response and execute a sub-dominant response; Rothbart et al., 2004) as predictors of healthy child development across domains.

There is a robust body of developmental research linking early inhibitory control to later psychosocial outcomes, including social competence and adjustment (e.g., Eisenberg et al., 2001; Eisenberg et al., 2005; Kochanska et al., 2000; Lengua, 2002; Mischel, Shoda, & Peake, 1988; Raver, Blackburn, Bancroft, & Torp, 1999; Rhoades, Greenberg, & Domitrovich, 2009; Shoda, Mischel, & Peake, 1990). In these studies, inhibitory control is typically measured via parent report or behavioral tasks that require inhibiting a pre-potent response (e.g., delay of gratification tasks). Using two laboratory assessments of inhibitory control, Rhoades et al. (2009) found that young children with greater inhibitory control had higher social skills and lower internalizing behaviors, which include symptoms of anxiety and depression. Further, parent-reported child inhibitory control was linked with multiple aspects of adjustment and self-competence within a single study (King, Lengua, & Monahan, 2013), with separate analytical models for each outcome.

Inhibitory control has also been linked with academic outcomes. For example, young children's performance on a composite of tasks, including waiting before touching a wrapped gift and waiting before eating a candy visible under a clear cup, was positively associated with early academic skills (Smith-Donald, Raver, & Hayes, 2007). These relationships may indicate cognitive and/or socio-emotional pathways. The process of inhibiting natural, dominant responses in order to enact more adaptive, sub-dominant responses in pursuit of long-term goals includes a cognitive component of managing one's attention (Mischel & Ayduk, 2004; Raver et al., 2011; Rothbart et al., 2004; Rueda, Posner, & Rothbart, 2005). Yet the literature linking inhibitory aspects of self-regulation to specific cognitive outcomes like working memory is mixed. Some studies have demonstrated positive relationships between inhibitory control and working memory in children (Mahy, Moses, & Kliegel, 2014) while others suggest that these constructs are orthogonal (Vuontela et al., 2013), without an overall relationship between them (Mahy & Moses, 2011).

Taken together, educational and psychological research supports the idea that children's inhibitory control levels have implications for multiple aspects of well-being. This literature could be expanded through additional assessments of objective and proximal cognitive outcomes like working memory and by examining this outcome in conjunction with other cognitive and psychosocial outcomes that have been linked to inhibitory control in previous, separate studies, including academic performance, adjustment, and self-competence, as well as indicators of physical health, such as weight status.

Compared to the literature linking inhibitory control and the aforementioned outcomes, investigations of inhibitory control's role in weight outcomes is a newer area of research. Childhood obesity rates have increased rapidly over the past four decades, and experts attribute this rapid increase in large part to environmental shifts (Mitchell, Catenacci, Wyatt, & Hill, 2011). The current environment in many countries is obesogenic, characterized by easily accessible, energy-dense, palatable foods and extensive opportunities to be sedentary (Hill et al., 2003). The health implications of this environment have been described as "passive obesity," where obesity is the result if one does not strive to maintain a normal weight status (Butland et al., 2007). Inhibitory control and broader self-regulation abilities have been highlighted as factors contributing to the likelihood of developing and maintaining healthy eating and physical activity habits in such environments.

Correspondingly, recent research suggests that individual differences in inhibitory control are associated with child weight status trajectories. Children with poorer performance on a delay of gratification task at age 4 were more likely to be overweight at age 11 (Seeyave et al., 2009), and young children with poorer performance on two delay tasks showed the most weight gain into early adolescence (Francis & Susman, 2009). Additionally, mother reports of girls’ inhibitory control levels at age 7 were associated with girls’ weight trajectories through age 15 (Anzman & Birch, 2009), and overweight school-age children showed evidence of poorer performance on a Stop Signal Task compared to normal-weight children (Guerrieri et al., 2007). Further, Epstein and colleagues have demonstrated that the combination of low inhibitory control and a high reinforcing value of food is linked with increased food intake and weight status (Feda, Roemmich, Roberts, & Epstein, 2015), and dietary interventions with a self-regulation component (e.g., instruction on problem solving in tempting situations; Israel, Guile, Baker, & Silverman, 1994) have demonstrated positive impacts on weight outcomes. This research has typically used body mass index (BMI) as an outcome measure; the accuracy of this measure as an indicator of adiposity has been shown to vary by weight status (Freedman & Sherry, 2009). Thus, the extant literature suggests that the promotion of inhibitory control skills could lead to improved energy balance and weight outcomes over time, with the potential for future research to augment these findings using additional, rigorous measures of adiposity.

To date, few researchers have studied benefits of inhibitory control or other components of self-regulation across multiple developmental domains simultaneously. One exception is a series of follow-up studies of a sample of preschoolers, in which the ability to delay gratification in early childhood has been linked with an array of subsequent cognitive and psychosocial outcomes (e.g., Mischel et al., 1988; Shoda et al., 1990), and in which it was recently demonstrated that preschoolers with longer delay times had a lower BMI 30 years later (Schlam et al., 2013). Similarly, Moffitt et al. (2011) demonstrated inverse associations between a composite index of childhood self-regulation and adult outcomes across domains and concluded that self-regulation is a malleable construct with widespread prevention implications.

The aim of the current study was to build upon this evidence, using a longitudinal dataset to investigate possible pathways from childhood inhibitory control to adolescent psychosocial, cognitive, and weight outcomes. Our baseline model follows research arguing that inhibitory control levels have a direct impact on well-being in each of these domains. We aimed to test the absolute fit of this model and to compare it with alternative models, to contribute to the state of the evidence about inhibitory control’s impact in multiple domains of child development. The current study contributes to the literature by including: (1) repeated measures of outcomes in multiple developmental domains, including objective assessments of working memory and of weight status (e.g., percent body fat, a direct measure of adiposity), (2) analysis of these outcomes within the same statistical model, and (3) a comparison of multiple models to explore whether relationships between inhibitory control and outcomes in each domain are best conceptualized as direct effects or whether some of these links may be explained by inter-relationships between outcomes.

Methods

Participants

Participants were 192 non-Hispanic White girls and their parents, who were recruited to participate in a longitudinal study about the health and development of young girls. A primary aim of the broader study was to study the emergence of dieting behavior prospectively (Balantekin, Savage, Marini, & Birch, 2014). Eligibility criteria for recruitment included living with both biological parents, the absence of severe food allergies or chronic medical problems affecting food intake, and the absence of dietary restrictions involving animal products; families were not recruited based on weight status or concerns about weight. Study visits took place when girls were 7 (n=192), 9 (n=183), 11 (n=177), 13 (n=168), and 15 (n=167) years old.

On average, parents were in their late 30s (mothers 37.4 ± 4.7 years; fathers 39.4 ± 5.4) when girls were 7 years old. Seventy-eight percent of reported family incomes were above $35,000. Parents were well-educated: mothers’ mean education level was 15 ± 2 years, and fathers’ was 15 ± 3 years. The prevalence of overweight (BMI percentile > 85) among girls in the sample at each time point was as follows: age 7: 21.4%; age 9: 31.2%; age 11: 29.4%; age 13: 25.6%; age 15: 20.1%. Fewer than three percent of girls were underweight (BMI percentile < 5) at each time point. The Pennsylvania State University Institutional Review Board approved all study procedures.

Measures

For the overarching constructs described below, latent variables composed of multiple observed measures were included as part of the structural equation models tested.

Inhibitory control

Mothers and fathers each reported on their daughters’ inhibitory control levels at age 7, using the 13-item inhibitory control subscale from the Child Behavior Questionnaire (Rothbart, Ahadi, Hershey, & Fisher, 2001). In this reliable and valid measure, temperament is measured using concrete examples of behaviors. Parents rate how well each statement represents their child on a scale of 1 (extremely untrue of your child) to 7 (extremely true of your child). Cronbach's alphas for mother-reported and father-reported inhibitory control in this sample were .81 and .74, respectively.

Weight outcomes

Weight outcomes at ages 9, 11, 13, and 15 included percent body fat and waist circumference; both were measured at each time point by trained nurses. Percent body fat was measured using dual x-ray absorptiometry (DXA) scans using the Hologic QDR 4500W (S/N 47261) in the array scan mode. Means and standard deviations in this sample were as follows: age 9: 26.7% (7.2); age 11: 27.6% (7.2); age 13: 27.0% (6.9); age 15: 27.9% (5.9). National DXA data from an overlapping time period showed the following mean percent body fat values for girls: age 9: 32.3%; age 11: 33.2%; age 13: 32.8%; and age 15: 31.7% (US Department of Health and Human Services, 2011). Waist circumference measurements were recorded in centimeters. Waist circumference is an excellent predictor of weight-related comorbidities, such as Type 2 diabetes and cardiovascular disease (Sharma, 2002). Means and standard deviations in this sample were as follows: age 9: 67.3 (9.3) cm; age 11: 73.7 (11.1) cm; age 13: 79.0 (11.5) cm; age 15: 79.1 (10.9) cm. National data from an overlapping time period showed the following mean waist circumference values for girls: age 9: 67.5 cm; age 11: 75.1 cm; age 13: 78.4 cm; age 15: 80.4 cm (US Department of Health and Human Services, 2008).

Cognitive outcomes

Overall school grades were calculated by averaging mothers' reports of girls' math, reading, science, and social studies grades at ages 9, 11, 13, and 15. At age 9, each grade was reported on a scale of 1–3 (below average, average, above average), and from age 11 and on, letter grades were used and were coded as follows before being averaged: an A was assigned a point value of 5, a B was 4, a C was 3, a D was 2 and an F was 1. A standardized measure of working memory and attention, the forwards and backwards digit span procedures from the Wechsler Intelligence Scale for Children-III, was administered by study staff at ages 13 and 15 (Wechsler, 1991). Study staff orally presented a series of number sequences, which the child was instructed to repeat verbatim, either forward or backward. Each part of the test (forwards and backwards procedures) contains seven items within which are two trials. Scoring is based on the child’s ability to pass both of an item's trials: 2 points are given if the child passes both trials; 1 point is given if the child passes one trial; 0 points are given if the child fails both trials. Scores are summed across trials and then summed across the two parts of the test to arrive at an overall digit span score. The maximum score for Digits Forward is 16 points, and the maximum score for Digits Backward is 14 points. Thus, the maximum total score is 30 points. At ages 13 and 15, the latent cognitive outcome variable was composed of both school grades and composite digit span scores. Cognitive outcomes at 9 and 11 reflect only school grades due to the unavailability of digit span scores; despite these differences, we observed strong, positive relationships between the cognitive variable at each time point in our structural models, including the relationship between school grades at age 11 and the latent cognitive outcome variable at age 13.

Psychosocial well-being

Psychosocial outcomes at ages 9, 11, 13, and 15 included self-reported depressive symptoms and perceived self-competence. At ages 9, 11, and 13, the depressive symptoms scale used was the Child Depression Inventory (CDI). The CDI is a widely used self-report measure that is suitable for 8 to 13 year old children (Kovacs & Beck, 1977). In this study, one item asking about suicidal ideation was removed as done previously (Smucker, Craighead, Craighead, & Green, 1986). Thus, total scores could range from 0 to 52. At age 15, the Center for Epidemiological Studies Depression scale (CES-D) was administered instead, in the interest of using a developmentally-appropriate measure. The CES-D is a widely used measure of depressive symptoms in non-clinical populations (Radloff, 1977). It is a self-report questionnaire consisting of 20 items. Depressive symptoms measured by the CDI and CES-D were used as continuous variables. Cronbach's alpha values at each time point were .80, .78, .83, and .92.

Girls' self-competence was assessed using Harter’s Perceived Self-Competence Scale. A version for older children was used when girls were 9 and 11 years old (Harter, 1982), and a version for adolescents was used at ages 13 and 15 (Harter, 1988). The format of the items in each version of the measure is such that children are given descriptions of two hypothetical children and are asked to choose the one who is most like them. Then they decide whether the child they chose is a lot like them or a little like them. The items are on a scale of 1 to 4, and higher scores indicate higher perceived self-competence. The purpose of this response system is to avoid socially-desirable answers and to avoid having children say that they are incompetent at a task. The subscale tapping global self-worth was used in this study, which consists of six items on the child version and five items on the adolescent version of the measure. Corresponding Cronbach's alpha values at each time point were .78, .82, .82, and .83.

Demographics

When girls were 7 years old, mothers and fathers reported their highest level of educational attainment, and mothers reported their family’s annual income.

Data Analysis

Means and standard deviations of variables of interest were calculated for girls with higher vs. lower inhibitory control (Table 1), and inter-correlations among the self-regulation measures and outcomes of interest at ages 9 and 15 were tested (Table 2). This information was intended to elucidate the bivariate associations between these variables and to justify their inclusion in the larger, more comprehensive models. We also tested correlations between girls’ inhibitory control and family income and maternal and paternal education levels. An advantage of a demographically homogeneous sample is that it could allow an investigation of links between self-regulation and outcomes without confounding due to associations between socioeconomic status and self-regulation. We tested correlations between income, education, and inhibitory control to verify this assumption.

Table 1.

Descriptive Statistics on Outcomes of Interest for Girls with Lower vs. Higher Inhibitory Control

Percent body fat Mean (SD) lower inhibitory control (n=82) Mean (SD) higher inhibitory control (n=110)
    Age 9 27.69 (7.21) 26.10 (7.16)
    Age 11 28.48 (7.19) 26.88 (7.17)
    Age 13 28.11 (7.22) 26.20 (6.55)
    Age 15 29.36 (5.67) 26.77 (5.90)
Waist circumference
    Age 9 69.15 (10.60) 65.96 (8.03)
    Age 11 75.73 (12.52) 72.14 (9.63)
    Age 13 80.20 (13.94) 78.02 (9.03)
    Age 15 81.06 (12.55) 77.58 (9.19)
Depressive symptoms
    Age 9 5.31 (4.88) 3.70 (3.61)
    Age 11 4.79 (4.06) 3.06 (3.15)
    Age 13 4.38 (4.41) 3.81 (4.17)
    Age 15* 14.37 (10.82) 12.00 (8.96)
Self-competence
    Age 9 3.33 (0.62) 3.51 (0.43)
    Age 11 3.38 (0.52) 3.61 (0.40)
    Age 13 3.50 (0.49) 3.56 (0.47)
    Age 15 3.32 (0.53) 3.44 (0.53)
School grades
    Age 9* 2.53 (0.46) 2.75 (0.38)
    Age 11 4.51 (0.47) 4.74 (0.38)
    Age 13 4.46 (0.62) 4.67 (0.48)
    Age 15 4.15 (0.73) 4.55 (0.55)
Digit span scores
    Age 13 15.01 (3.29) 16.30 (2.94)
    Age 15 16.41 (3.39) 17.61 (3.75)
Open in a new tab

Notes: For descriptive comparisons above, inhibitory control was split at the mean, similar to previous studies (Rothbart et al., 2001), but all subsequent analyses were conducted using continuous inhibitory control. All outcomes showed significant associ ations with continuous inhibitory control in repeated measures ANOVAs. Measures marked with asterisks were different at the indicated time points, in order to be developmentally appropriate. Thus, they can only be compared with values at other time points in an ordinal, not absolute, sense. The mean levels of mother-reported inhibitory control in each group depicted above were 4.43 (SD=0.66) vs. 5.48 (SD=0.48), and for father-reported inhibitory control, the means were 4.27 (SD=0.58) vs. 5.18 (SD=0.44). The differences between inhibitory control levels between groups is statistically significant, p<.0001.

Table 2.

Correlations between Parent-rated Inhibitory Control at Daughter Age 7 and Proximal and Distal Outcomes

Girls' outcomes at age 9 Girls' outcomes at age 15
Waist
circumference		 %
body
fat		 School
grades		 Depressive
symptoms		 Self-competence Waist
circumference		 %
body
fat		 School
grades		 Digit
span		 Depressive
symptoms		 Self-competence
Mother-rated IC r= −.29*** −.25*** .31*** −.25*** .21*** −.26*** −.27*** .30*** .11 −.20** .21**
n= 183 183 182 183 183 166 160 161 166 166 166
Father-rated IC r= −.18* −.15# .24** −.19* .11 −.16* −.30*** .25** .27*** −.13 .16*
n= 179 179 178 179 179 163 157 158 163 163 163
Open in a new tab

Notes: IC = inhibitory control.

#

p<.10,

*

p<.05,

**

p<.01,

***

p<.001.

Correlation between mother- and father-reported inhibitory control = 0.55, p<.0001. Sample sizes vary slightly due to missing data.

Next, structural equation models were tested using AMOS (Version 20, Meadville, PA). The initial hypothesized model depicted self-regulation as an independent predictor of developmental outcomes in all three domains. The analytic plan was to: 1) analyze this model, allowing error variances of the same measures to covary; 2) make modifications if the output indicated that variances of error terms, covariances between error terms, or residuals of the relations between latent variables should be fixed; and 3) to examine the absolute fit of the resulting model, focusing on the comparative fit index (CFI) and root mean square error of approximation (RMSEA). Chi-squared values were also examined, with the expectation that this statistic would be useful for assessing relative, not absolute, fit, as it tends to be a conservative test of model fit that is influenced by sample size (Byrne, 2010; Jöreskog & Sörbom, 1993).

After testing the hypothesized model, six additional structural equation models were tested (Figure 1a–f), retaining any changes to the error variances and residuals and making further modifications to them as necessary. In each of these models, the direct effect from self-regulation to outcomes in one domain was removed, positing instead that the impact on this domain occurred through one of the other domains. We assessed relative fit using the chi-square difference, determining if any of these alternative models showed a significantly better fit than the initial hypothesized model; models are shown in Figures 1 (alternative) and 2 (hypothesized). Although 13% of participants were lost to follow-up between ages 7 and 15, all families were retained in the structural equation models reported herein as these models handle missing data using full information maximum likelihood. Further, when examining differences between those who stayed in the study until age 15 vs. those who dropped out on all variables of interest available at age 7 (income, maternal education, paternal education, mother-reported inhibitory control, father-reported inhibitory control, waist circumference, depressive symptoms, and global self-worth), no significant differences were found (p>.10 for all).

Figure 1. Alternate Models Depicting Links between Inhibitory Control and Outcomes in Three Domains.

Figure 1

Open in a new tab

These models modify the baseline model (see Figure 2) by depicting prospective relationships between inhibitory control and each outcome domain through each of the other domains. Each model was fitted and compared with the baseline model. Only the latent constructs are depicted here, for interpretability, but measurement models were consistent with Figure 2. Abbreviations: WS=latent weight outcome construct; COG=latent cognitive outcome construct; PSYCH=latent psychosocial outcome construct.

Figure 2. Baseline Model Depicting Links between Inhibitory Control and Outcomes in Three Domains.

Figure 2

Open in a new tab

This model depicts direct pathways between girls' inhibitory control and their psychosocial, cognitive, and weight outcomes over time. Absolute fit of this model was assessed, and it was also compared with the six alternate models in Figure 1. Residuals and measurement error are not pictured, for clarity. Of all models tested, this model demonstrated the best fit and is shown here, with the resulting lambda values depicting relations between measured variables and latent constructs and beta coefficients depicting relations between the latent constructs (all significant at p<.001).

We also examined repeated measures ANOVA models in SAS Proc Mixed, testing relations between inhibitory control (averaged across mothers' and fathers' ratings when girls were age 7) and each measured outcome variable over time, to see whether results were consistent. For each of these models, four error structures (compound symmetry, first-order autoregressive, toeplitz, and unstructured) were compared, and the best-fitting one was specified in the final model. Independent variables were inhibitory control, as well as linear and quadratic trends representing changes in each outcome from age 9 to 15, and the interaction between these time trends and inhibitory control. Variance explained by inhibitory control in each of these models was estimated as recommended by Singer (1998). Ordinary least squares regression was used instead of repeated measures ANOVA for the outcome of digit span scores, given that these were only available at two time points. Variance explained by inhibitory control in this model was determined using the R2 value for the model. For the three outcome measures for which data were available at age 7 (waist circumference, depressive symptoms, and global self-concept), the ANOVA models were repeated with girls’ standings on the variables at age 7 included as a covariate.

Results

Descriptive Statistics and Correlations

The mean mother-reported inhibitory control score was 5.03 (SD=0.77; range=2.0–6.8), and the mean father-reported inhibitory control score was 4.79 (SD=0.67; range=2.2–6.3). The descriptive statistics for the outcome variables of interest and correlations between inhibitory control and these variables are reported in Tables 1 and 2 respectively. Mother- and father-reported inhibitory control scores were significantly associated with each of the outcomes of interest at the first (age 9) or last (age 15) outcome measurement (or both). Mother-reported inhibitory control was not associated with family income (r=.05, p=.49) or mothers' (r=.01, p=.88) or fathers' education level (r=−.02, p=.76) at the age 7 assessment, nor was father-reported inhibitory control associated with income (r=.03, p=.71) or mothers' (r=.05, p=.53) or fathers' education level (r=.05, p=.48).

Structural Equation Models

The hypothesized model demonstrated an acceptable fit based on the following: RMSEA = .06 (less than .08 is considered acceptable); CFI = .95 (greater than or equal to .95 is considered excellent). Additionally, all of the standardized beta coefficients representing relations between the latent constructs had an absolute value greater than .56, and all were statistically significant (p<.001; Figure 2). In contrast, chi-square indicated significant differences between the hypothesized model and the data: Χ2(df=230)= 351, p<.0001; yet as previously mentioned, this is typical and has led to recommendations to consult the other fit indices when assessing absolute fit.

When assessing relative fit by comparing the hypothesized model (Figure 2) to alternate models depicting indirect, rather than direct, effects of self-regulation on each domain in turn, each of these alternative models demonstrated a significantly poorer fit than the hypothesized model. Specifically, results from chi-square difference tests were as follows: for the alternate model depicting inhibitory control’s impact on psychosocial outcomes through weight (Figure 1a): ΔΧ2(df=2)=13.0, p<.01; for the model depicting impact on weight outcomes through psychosocial outcomes (Figure 1b): ΔΧ2(df=2)=23.4, p<.001; for the model depicting impact on weight outcomes through cognition (Figure 1c): ΔΧ2(df=1)=20.7, p<.001; for the model depicting impact on cognitive outcomes through weight (Figure 1d): ΔΧ2(df=9)=664.9, p<.001; for the model depicting impact on cognitive outcomes through psychosocial outcomes (Figure 1e): ΔΧ2(df=9)=670.4, p<.001; and for the model depicting impact on psychosocial outcomes through cognition (Figure 1f): ΔΧ2(df=2)=6.3, p<.05. The change in degrees of freedom varies, because in some cases, additional parameters needed to be modified in order for model identification to occur. For example, in testing the models specifying impacts on cognitive outcomes through other outcomes (1d–e), numerous additional parameters were fixed based on output showing negative residuals and error variances.

Repeated measures ANOVA results were consistent with findings from the structural equation models. Girls with higher parent-reported inhibitory control at age 7 had: a lower percent body fat (F(1,182)=12.39, p<.001), smaller waist circumference (F(1,184)= 13.24, p<.001), lower depressive symptoms (F(1,184)=7.24, p<.01), higher self-competence (F(1,184)=12.77, p<.001), and higher school grades (F(1,184)=24.61, p<.0001) across ages 9 to 15, as well as higher average digit span scores across ages 13 to 15 (β=0.38, p<.001). The explained variance attributed to inhibitory control in these models was estimated to be: 1.14%, 6.92%, 5.66%, 7.52%, 7.84%, and 6.67% respectively. There were significant linear (F(1,184=4.98, p<.05) and quadratic (F(1,184)=5.56, p<.05) changes in waist circumference, such that there is an overall increase across ages 9 to 15, with a leveling off between ages 13 and 15, as well as significant quadratic changes in depressive symptoms (F(1,184)=7.15, p<.01) and school grades (F(1,184)=55.98, p<.0001), such that depressive symptoms were lowest and grades were highest at age 11 (see Table 1 for means). None of these overall trends interacted with inhibitory control levels: thus, while inhibitory control was associated with levels of each outcome across ages 9 to 15, the patterns of changes in these outcomes did not differ for girls with higher vs. lower inhibitory control. All repeated measures ANOVA results were consistent when adjusting for mothers' and fathers' education levels and annual family income at the initial assessment. Some, but not all, results shifted when including age 7 variables in the models as additional covariates where possible: The relationship between inhibitory control and waist circumference from age 9 to 15 was no longer significant when adjusting for waist circumference at age 7 (F(1,182)=2.10, p=.15). However, relationships between inhibitory control and depressive symptoms (F(1,183)=4.48, p<.05) and global self-concept (F(1,183)=12.24, p<.001) remained significant when adjusting for status at age 7.

Discussion

This study provides evidence that inhibitory control is associated with multiple indicators of well-being from middle childhood into adolescence. Our hypothesized model, depicting these pathways as direct effects from inhibitory control to outcomes in each of three domains, demonstrated an acceptable absolute fit, and was superior to a series of alternate models. These findings are consistent with recent research highlighting self-regulation as a predictor of adaptive development across domains (Moffitt et al., 2011), and they extend this work through repeated assessments of multiple outcomes over time and a consideration of mechanisms through which inhibitory control may impact these outcomes.

There is a robust literature demonstrating the importance of early inhibitory control and related constructs, but for the most part, studies have focused on impacts on outcomes within a single domain. A rare example of a related study examining multiple domains simultaneously comes from Moffitt and colleagues, who demonstrated that a composite indicator of childhood self-control was associated with adulthood outcomes related to health, wealth, and crime (Moffitt et al., 2011). In the current study, we aimed to delve into more proximal developmental outcomes, demonstrating that childhood inhibitory control was linked with cognitive and psychosocial well-being and weight-related outcomes from middle childhood into adolescence, and that structural equation models fit best when these links were conceptualized as three direct pathways from inhibitory control to outcomes in each domain.

Repeated measures ANOVA models bolstered these findings, as inhibitory control was linked with overall levels of each measured outcome. While associated with psychosocial, cognitive, and weight outcomes, inhibitory control did not interact with time, suggesting that these outcomes are relatively stable throughout middle childhood and early adolescence, with children retaining their rank order across levels of inhibitory control. These findings suggest the potential promise of early interventions aiming to impact inhibitory control while these abilities are developing. Further, follow-up analyses adjusting for age 7 status on outcome variables where available supported hypotheses that inhibitory control temporally precedes these outcomes in some cases (depressive symptoms, global self-concept). In other cases, relationships between inhibitory control and outcomes were non-significant when making such adjustments (waist circumference) or were not testable with the current dataset; thus, future research in this area should continue to examine directionality.

While the current results add to the growing body of evidence highlighting inhibitory control as an important predictor of well-being across domains, there are limitations to consider. First, inhibitory control was parent-reported in this study and was only available at one time point. Laboratory tasks such as the delay of gratification task could offer a more objective assessment of inhibitory control. Yet, parent-reported inhibitory control was linked to multiple outcomes, just as a more complex self-control composite was in past studies (Moffitt et al., 2011). This finding highlights the utility of the Child Behavior Questionnaire in characterizing children's self-regulation in a meaningful way. Additionally, some outcome measures changed over time (e.g., the depressive symptoms measure changed from the CDI to the CES-D at age 15), but this limitation is balanced by the strength of including developmentally appropriate measures. It is also a strength that outcome measures were from multiple modalities, including objective laboratory assessments, child report, and parent report, though the results could be extended by examining more objective measures of cognition/achievement, given potential biases present in school grades data, and also through examinations of other, more general aspects of psychosocial functioning (e.g., mental health symptoms checklists). The use of an array of methods and multiple informants decreases the likelihood that the current results were due to parental biases or shared method variance. The current sample size was also relatively small for the purposes of structural equation modeling; we added the individual mixed models to bolster the results. Finally, the study sample was homogeneous and included only girls. While confounders are less of a concern when samples are homogeneous, as evidenced by the nonsignificant correlations between demographic factors and inhibitory control, it is important to underscore that the current results cannot be generalized beyond non-Hispanic, White girls in families with relatively high incomes and education levels. Self-regulation has been cited as an important resilience factor in disadvantaged populations (Blair & Diamond, 2008; Diamond & Lee, 2011; Evans & Rosenbaum, 2008; Raver et al., 1999; Raver et al., 2011), and parallel analyses among underserved children are an important next step to shed light on the extent to which improved inhibitory control may simultaneously attenuate problems like the achievement gap and health disparities. Yet, it can also be argued that in current environments, all children could benefit from efforts to promote inhibitory components of self-regulation.

The results of this study highlight associations between inhibitory control and child outcomes across domains over time, and as such, they highlight opportunities for prevention. Experimental research with undergraduates highlights the possibility of strengthening inhibitory control of eating behavior through practice (Houben, 2011; Veling, Aarts, & Papies, 2011), and the developmental psychology literature provides examples of successful prevention programs that have impacted outcomes like school readiness via inhibitory control (Raver et al., 2011). This and other evidence that self-regulation is modifiable (Johnson, 2000; Raver et al., 2011; Zelazo & Carlson, 2012) suggest the utility of consolidating self-regulation research across disciplines, in order to address the well-being of the whole child. Future prevention research targeting inhibitory control should be interdisciplinary, facilitating the assessment of impacts across developmental domains, and should explore the extent to which effects differ across socio-demographic groups, as well as sensitive periods during which effects are greatest. Such efforts have the potential to equip young children with capacities that could promote overall well-being as they develop in contexts that provide challenges to enacting healthy “sub-dominant” responses. These efforts could be coupled with efforts to modify the environment, so that self-regulation capacities are not taxed as much (Baumeister, Vohs, & Tice, 2007). Macro-level changes like this operate over a longer time scale, so it is important in the meantime to harness knowledge about self-regulation to equip children to succeed in current environments, particularly those children at greatest risk.

In sum, results from this study provide support for our hypothesized model, with direct paths from childhood inhibitory control to subsequent psychosocial, cognitive, and weight outcomes. Dominant behavioral responses elicited by current environments can be viewed as maladaptive, highlighting the potential of promoting early self-regulation abilities as a way to decrease the likelihood of negative behaviors and promote healthy child development across domains. Interdisciplinary collaboration between researchers studying self-regulation from different perspectives may result in opportunities to simultaneously increase the likelihood of healthy psychosocial, cognitive, and physical development in childhood and beyond.

Acknowledgments

The authors would like to thank the families who participated in this longitudinal study. Funding was provided by NIH HD 32973, NIH HD 46567-01 and M01 RR10732.

References

  1. Anzman SL, Birch LL. Low inhibitory control and restrictive feeding practices predict weight outcomes. The Journal of Pediatrics. 2009;155(5):651–656. doi: 10.1016/j.jpeds.2009.04.052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Balantekin KN, Savage JS, Marini ME, Birch LL. Parental encouragement of dieting promotes daughters’ early dieting. Appetite. 2014;80:190–196. doi: 10.1016/j.appet.2014.05.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baumeister RF, Vohs KD, Tice DM. The strength model of self-control. Current Directions in Psychological Science. 2007;16(6):351–355. [Google Scholar]
  4. Blair C, Diamond A. Biological processes in prevention and intervention: The promotion of self-regulation as a means of preventing school failure. Development and Psychopathology. 2008;20(03):899–911. doi: 10.1017/S0954579408000436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Butland B, Jebb S, Kopelman P, McPherson K, Thomas S, Mardell J, Parry V. Foresight. Tackling Obesities: Future Choices - Project Report. London: Government Office for Science; 2007. [Google Scholar]
  6. Byrne B. Structural equation modeling with AMOS: Basic concepts, applications, and programming. New York: Routledge; 2010. [Google Scholar]
  7. Diamond A, Lee K. Interventions shown to aid executive function development in children 4 to 12 years old. Science. 2011;333(6045):959–964. doi: 10.1126/science.1204529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Eisenberg N, Cumberland A, Spinrad TL, Fabes RA, Shepard SA, Reiser M, Guthrie IK. The relations of regulation and emotionality to children's externalizing and internalizing problem behavior. Child Development. 2001;72(4):1112–1134. doi: 10.1111/1467-8624.00337. [DOI] [PubMed] [Google Scholar]
  9. Eisenberg N, Zhou Q, Spinrad TL, Valiente C, Fabes RA, Liew J. Relations among positive parenting, children's effortful control, and externalizing problems: A three-wave longitudinal study. Child Development. 2005;76(5):1055–1071. doi: 10.1111/j.1467-8624.2005.00897.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Evans GW, Rosenbaum J. Self-regulation and the income-achievement gap. Early Childhood Research Quarterly. 2008;23(4):504–514. [Google Scholar]
  11. Feda DM, Roemmich JN, Roberts A, Epstein LH. Food reinforcement and delay discounting in zBMI-discordant siblings. Appetite. 2015;85:185–189. doi: 10.1016/j.appet.2014.11.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Francis LA, Susman EJ. Self-regulation and rapid weight gain in children from age 3 to 12 years. Archives of Pediatrics & Adolescent Medicine. 2009;163(4):297. doi: 10.1001/archpediatrics.2008.579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Freedman DS, Sherry B. The validity of BMI as an indicator of body fatness and risk among children. Pediatrics. 2009;124(Supplement 1):S23–S34. doi: 10.1542/peds.2008-3586E. [DOI] [PubMed] [Google Scholar]
  14. French SA, Story M, Perry CL. Self-esteem and obesity in children and adolescents: A literature review. Obesity Research. 1995;3(5):479–490. doi: 10.1002/j.1550-8528.1995.tb00179.x. [DOI] [PubMed] [Google Scholar]
  15. Golan M, Bachner-Melman R. Self-regulation and the management of childhood obesity. Journal of Community Medicine & Health Education. 2011;1:107. [Google Scholar]
  16. Harter S. The perceived competence scale for children. Child Development. 1982;53:87–97. [PubMed] [Google Scholar]
  17. Harter S. Manual for the Self-Perception Profile for Adolescents. University of Denver; 1988. [Google Scholar]
  18. Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity and the environment: Where do we go from here? Science. 2003;299(5608):853–855. doi: 10.1126/science.1079857. [DOI] [PubMed] [Google Scholar]
  19. Houben K. Overcoming the urge to splurge: Influencing eating behavior by manipulating inhibitory control. Journal of behavior therapy and experimental psychiatry. 2011;42(3):384–388. doi: 10.1016/j.jbtep.2011.02.008. [DOI] [PubMed] [Google Scholar]
  20. Israel AC, Guile CA, Baker JE, Silverman WK. An evaluation of enhanced self-regulation training in the treatment of childhood obesity. Journal of Pediatric Psychology. 1994;19(6):737–749. doi: 10.1093/jpepsy/19.6.737. [DOI] [PubMed] [Google Scholar]
  21. Johnson SL. Improving preschoolers' self-regulation of energy intake. Pediatrics. 2000;106(6):1429–1435. doi: 10.1542/peds.106.6.1429. [DOI] [PubMed] [Google Scholar]
  22. Jöreskog KG, Sörbom D. LISREL 8: Structural equation modeling with the SIMPLIS command language. Skokie, IL: Scientific Software International; 1993. [Google Scholar]
  23. Kagan J. Biology, context, and developmental inquiry. Annual Review of Psychology. 2003;54(1):1–23. doi: 10.1146/annurev.psych.54.101601.145240. [DOI] [PubMed] [Google Scholar]
  24. King KM, Lengua LJ, Monahan KC. Individual differences in the development of self-regulation during pre-adolescence: Connections to context and adjustment. Journal of Abnormal Child Psychology. 2013;41(1):57–69. doi: 10.1007/s10802-012-9665-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kochanska G, Murray KT, Harlan ET. Effortful control in early childhood: continuity and change, antecedents, and implications for social development. Developmental Psychology. 2000;36(2):220–232. [PubMed] [Google Scholar]
  26. Kopp CB. Antecedents of self-regulation: a developmental perspective. Developmental Psychology. 1982;18(2):199–214. [Google Scholar]
  27. Kovacs M, Beck AT. Depression in Childhood: Diagnosis, Treatment, and Conceptual Models. New York, NY: Raven Press; 1977. An empirical-clinical approach toward a definition of childhood depression. [Google Scholar]
  28. Lengua LJ. The contribution of emotionality and self-regulation to the understanding of children’s response to multiple risk. Child Development. 2002;73(1):144–161. doi: 10.1111/1467-8624.00397. [DOI] [PubMed] [Google Scholar]
  29. Mahy CE, Moses LJ. Executive functioning and prospective memory in young children. Cognitive Development. 2011;26(3):269–281. [Google Scholar]
  30. Mahy CE, Moses LJ, Kliegel M. The impact of age, ongoing task difficulty, and cue salience on preschoolers’ prospective memory performance: The role of executive function. Journal of Experimental Child Psychology. 2014 doi: 10.1016/j.jecp.2014.01.006. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
  31. McClelland MM, Acock AC, Piccinin A, Rhea SA, Stallings MC. Relations between preschool attention span-persistence and age 25 educational outcomes. Early Childhood Research Quarterly. 2012;28:314–324. doi: 10.1016/j.ecresq.2012.07.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mischel W, Ayduk O. Willpower in a cognitive-affective processing system. In: Baumeister RF, Vohs KD, editors. Handbook of Self-Regulation: Research, Theory, and Applications. New York, NY: Guilford Press; 2004. pp. 83–105. [Google Scholar]
  33. Mischel W, Shoda Y, Peake PK. The nature of adolescent competencies predicted by preschool delay of gratification. Journal of Personality and Social Psychology. 1988;54(4):687. doi: 10.1037//0022-3514.54.4.687. [DOI] [PubMed] [Google Scholar]
  34. Mitchell N, Catenacci V, Wyatt H, Hill J. Obesity: Overview of an Epidemic. Psychiatric Clinics of North America. 2011;34:717–732. doi: 10.1016/j.psc.2011.08.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Moffitt TE, Arseneault L, Belsky D, Dickson N, Hancox RJ, Harrington H, Ross S. A gradient of childhood self-control predicts health, wealth, and public safety. Proceedings of the National Academy of Sciences. 2011;108(7):2693–2698. doi: 10.1073/pnas.1010076108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Radloff LS. The CES-D scale a self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1(3):385–401. [Google Scholar]
  37. Raver CC, Blackburn EK, Bancroft M, Torp N. Relations between effective emotional self-regulation, attentional control, and low-income preschoolers' social competence with peers. Early Education and Development. 1999;10(3):333–350. [Google Scholar]
  38. Raver CC, Jones SM, Li-Grining C, Zhai F, Bub K, Pressler E. CSRP’s Impact on Low-Income Preschoolers’ Preacademic Skills: Self-Regulation as a Mediating Mechanism. Child Development. 2011;82(1):362–378. doi: 10.1111/j.1467-8624.2010.01561.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rhoades BL, Greenberg MT, Domitrovich CE. The contribution of inhibitory control to preschoolers' social—emotional competence. Journal of Applied Developmental Psychology. 2009;30(3):310–320. [Google Scholar]
  40. Rothbart MK, Ahadi SA, Hershey KL, Fisher P. Investigations of temperament at three to seven years: The Children's Behavior Questionnaire. Child Development. 2001;72(5):1394–1408. doi: 10.1111/1467-8624.00355. [DOI] [PubMed] [Google Scholar]
  41. Rothbart MK, Ellis LK, Posner MI. Temperament and self-regulation. In: Baumeister RF, Vohs KD, editors. Handbook of Self-Regulation: Research, Theory, and Applications. New York, NY: Guilford Press; 2004. pp. 441–460. [Google Scholar]
  42. Rueda MR, Posner MI, Rothbart MK. The development of executive attention: Contributions to the emergence of self-regulation. Developmental Neuropsychology. 2005;28(2):573–594. doi: 10.1207/s15326942dn2802_2. [DOI] [PubMed] [Google Scholar]
  43. Schlam TR, Wilson NL, Shoda Y, Mischel W, Ayduk O. Preschoolers' Delay of Gratification Predicts their Body Mass 30 Years Later. The Journal of Pediatrics. 2013;162(1):90–93. doi: 10.1016/j.jpeds.2012.06.049. doi: http://dx.doi.org/10.1016/j.jpeds.2012.06.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Seeyave DM, Coleman S, Appugliese D, Corwyn RF, Bradley RH, Davidson NS, Lumeng JC. Ability to delay gratification at age 4 years and risk of overweight at age 11 years. Archives of Pediatrics & Adolescent Medicine. 2009;163(4):303–308. doi: 10.1001/archpediatrics.2009.12. doi: [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sharma A. Adipose tissue: a mediator of cardiovascular risk. International Journal of Obesity and Related Metabolic Disorders. 2002;26:S5–7. doi: 10.1038/sj.ijo.0802210. [DOI] [PubMed] [Google Scholar]
  46. Shoda Y, Mischel W, Peake PK. Predicting adolescent cognitive and self-regulatory competencies from preschool delay of gratification: Identifying diagnostic conditions. Developmental Psychology. 1990;26(6):978. [Google Scholar]
  47. Singer JD. Using SAS PROC MIXED to fit multilevel models, hierarchical models, and individual growth models. Journal of Educational and Behavioral Statistics. 1998;23:323–355. [Google Scholar]
  48. Smith-Donald R, Raver CC, Hayes T. Preliminary construct and concurrent validity of the Preschool Self-Regulation Assessment (PSRA) for field-based research. Early Childhood Research Quarterly. 2007;22(2):173–187. [Google Scholar]
  49. Smucker MR, Craighead WE, Craighead LW, Green BJ. Normative and reliability data for the Children's Depression Inventory. Journal of Abnormal Child Psychology. 1986;14(1):25–39. doi: 10.1007/BF00917219. [DOI] [PubMed] [Google Scholar]
  50. US Department of Health and Human Services. Anthropometric reference data for children and adults: United States, 2003–2006. Washington, DC: National Center for Health Statistics; 2008. [Google Scholar]
  51. US Department of Health and Human Services. Smoothed percentage body fat percentiles for US children and adolescents, 1999–2004. Washington, DC: National Center for Health Statistics; 2011. [PubMed] [Google Scholar]
  52. Veling H, Aarts H, Papies EK. Using stop signals to inhibit chronic dieters’ responses toward palatable foods. Behaviour research and therapy. 2011;49(11):771–780. doi: 10.1016/j.brat.2011.08.005. [DOI] [PubMed] [Google Scholar]
  53. Vuontela V, Carlson S, Troberg AM, Fontell T, Simola P, Saarinen S, Aronen ET. Working memory, attention, inhibition, and their relation to adaptive functioning and behavioral/emotional symptoms in school-aged children. Child Psychiatry & Human Development. 2013;44(1):105–122. doi: 10.1007/s10578-012-0313-2. [DOI] [PubMed] [Google Scholar]
  54. Wechsler D. Manual for the Wechsler intelligence scale for children, revised. San Antonio, TX: Psychological Corporation; 1991. [Google Scholar]
  55. Zelazo PD, Carlson SM. Hot and cool executive function in childhood and adolescence: Development and plasticity. Child Development Perspectives. 2012;6(4):354–360. [Google Scholar]

RESOURCES