Improved Executive Function and Science Achievement for At-Risk Middle School Girls in an Aerobic Fitness Program

Abstract

This study explored the effects of a middle school physical activity intervention for adolescent girls on the executive functioning involved in science learning. The girls, ages 11 to 14, were at risk for low self-esteem, sedentary lifestyle, and poor health outcomes. Executive function stems from interdependent cognitive control processes that influence goal setting and information processing, which complement higher order thinking required for acquiring scientific process skills. A 20-week informal triathlon training program served as the intervention for the treatment girls (n = 29). The comparison group of girls (n = 30) was randomly drawn from a matched sample of students of a similar demographic. Mean comparisons, ANCOVA, and Roy-Bargmann stepdown analysis were used to measure outcomes. The intervention contributed to significant improvement in several executive functions and science achievement. These results suggest that an afterschool program with a physical fitness component may improve the executive functions involved in science learning.

Keywords

academic achievement at-risk populations cognitive development health/nutrition/obesity middle school

Recent reports have called for improved science engagement and performance for middle school children to expand the science, technology, engineering, and mathematics (STEM) pipeline (National Research Council [NRC], 2007, 2012]. During middle school, students go through dramatic physical, emotional, behavioral, and cognitive changes, and gender differences in achievement and attitudes toward science widen (Jones, Howe, & Rua, 2000). There has been a recent focus on synthesizing findings from cognitive psychology and science education research to develop innovative treatments and assessments to diminish the gap in science learning (NRC, 2007). This study explored the effects of a physical activity intervention for at-risk middle school girls on the executive functioning involved in science learning. These young women were at risk for low self-esteem, experiencing a sedentary lifestyle, and/or being overweight.

Skills associated with the capacity to perform well in science are highly dependent upon abilities that fall under executive function (EF; Latzman, Elkovitch, Young, & Clark, 2010). EF is an umbrella term for higher order cognitive processes including working memory, inhibition, shifting, and the ability to plan, monitor, and carry out goal-directed actions (Skogan et al., 2016). Science process skills associated with EF involve the ability to understand and perform particular ways of observing, thinking, experimenting, problem solving, evaluating plausible explanations, validating conclusions, and storing newly acquired knowledge in long-term memory (Anderson, 1983). Purposeful behavior initiated to reach a goal parallels the scientific discovery process and informed skepticism. Research has suggested that EF improves with vigorous aerobic exercise (Hall, Smith, & Keele, 2001; Kramer et al., 1999); however, few studies have provided convergent evidence linking improved cognitive function with specific academic domains (Davis et al., 2011; Latzman et al., 2010). Structured, intensive fitness programs may provide the physical activity necessary to improve EF subcomponents related to science learning.

This study explored the effects of an afterschool fitness program on science learning and EF for at-risk middle school girls, as defined by the treatment program. Afterschool fitness programs can make a difference in building the requisite skills needed for engagement in learning by providing environments in which students can increase competency while developing personal resources that transfer to other realms of success (Lauer et al., 2006). This study addressed the following overarching research question:

Research Question 1: What is the effect of an out-of-school time aerobic fitness program on the EF and science achievement of at-risk middle school girls?

In exploring this research question, the relationship of physical activity to cognitive mechanisms that influence science achievement may be identified.

Theoretical Framework

The theoretical framework of this study incorporated aspects of EF associated with self-regulation, strategizing, and science learning. Anderson (2002) proposed a model of EF that stems from interdependent cognitive control processes that influence goal setting and information processing. Goal setting involves organizing actions to approach tasks efficiently, while information processing refers to response speed related to the ability to cope with switching tasks. Anderson (2002) further defined his model by identifying attentional control as the capacity to redirect automatic responses selectively and focus attention for long periods. Shifting and inhibition are considered subdomains of attentional control; shifting is the ability to navigate between tasks, and inhibition is the ability to suppress an automatic response in order to complete tasks (Miyake et al., 2000). Mature cognitive control involves self-regulation or resisting inappropriate or automatic responses, as well as adapting behavior to environmental changes (Davidson, Amso, Anderson, & Diamond, 2006), skills that contribute to reasoning and mathematics, and literacy ability in children (Blair & Razza, 2007). Research has suggested that problem solving, planning, and goal-directed actions are critical thinking skills necessary for science learning, often occurring through inquiry and collaboration (Schraw, Crippem, & Hartley, 2006).

Previous factor analyses suggested that EF consists of three separate yet interrelated domains related to science achievement—inhibition, monitoring, and cognitive control (Latzman et al., 2010; Miyake et al., 2000). One limitation of previous literature was that few studies assessed physical fitness as a predictor of cognitive performance and science achievement (Davis et al., 2007). Latzman et al. (2010) were among the first to examine the direct relationship between EF and science achievement with male adolescents. Results indicated that shifting was related to science with process skills requiring flexibility with new information. Inhibition skills also predicted performance in mathematics and science. Monitoring, defined as “evaluating information in working memory” (Latzman et al., 2010, p. 456), predicted achievement in reading and social studies. Monitoring parallels metacognition, or knowledge about one’s cognition and learning process. Latzman et al. (2010) encouraged future research to explore the relationship between inhibition and science, postulating that inhibiting one’s conclusions until all steps of structured problem solving have been conducted improves science learning and process skills. Davis et al. (2011) reported that aerobic exercise improved the EF constructs planning and inhibition, as well as mathematics achievement, in overweight pre-adolescent children, evidenced by performance measures and increased prefrontal cortex activity. This study builds upon their work to explore the impacts of a physical activity treatment on EF domains related to science achievement in female adolescents.

EF Hypothesis

The EF hypothesis predicts that the largest improvements in cognition due to exercise will be on executive brain functions (Davis et al., 2007). There is evidence that EF develops throughout adolescence. Davis et al. (2007) reported that adolescents who participated in a 15-week physical activity program experienced significant improvements in EF. Other research has shown that acute and higher levels of aerobic fitness have been associated with increased attention span, concentration, and improved memory (Best, 2010; Davis et al., 2011; Smith et al., 2010), in contrast with research targeted at lower, less intense levels of physical exertion that did not show improvement (Etnier, Nowell, Landers, & Sibley, 2006). Research examining chronic physical activity on attention tasks in pre-adolescent children has suggested improved cognitive performance and warrants further study (Hillman et al., 2009).

Assessing how exercise influences children’s neurocognitive development may prove to be an important factor in understanding the role a physically active lifestyle plays in science learning (Tomporowski, Davis, Miller, & Naglieri, 2008). Detecting the effects of exercise on cognition may depend upon the particular EF assessed through selected outcome measures. Research has established that physical activity improves select aspects of EF related to academic achievement (Davis et al., 2011; Latzman et al., 2010), but more work is needed to relate cognition to performance in specific disciplines. It was hypothesized that the present intervention with an aerobic fitness component would improve select cognitive processes and science achievement in middle school girls. Evaluation of EF with appropriate age-based cognitive tests may provide understanding of how chronic aerobic exercise influences brain development and processing, and, consequently, science performance.

Research Design

The study employed a quasi-experimental research design with a sample of suburban public school adolescent girls in Grades 6 to 8 who elected to participate in the treatment, as well as a control group from a nearby public middle school selected for similar socioeconomic status. The researchers attempted to ensure that the participants in the two groups had similar backgrounds including grade, gender, ethnicity, weight, and height. They were also matched for science pretest scores to control for possible prior science achievement differences. The intervention participants (n = 29) met with the program facilitators 3 times per week from March through July. The matched comparison group (n = 30) completed the same physical activity, science achievement, and cognition instruments. The early springtime data collection ensured that the students were at a similar level of science learning within the same middle grade levels. All districts involved had standardized curricula for each grade in compliance with the New York State Intermediate Level Science Curriculum Grades 5-8 (New York State Education Department [NYSED], 2016). There were 59 students (age, 12.5 ± 0.9 years) who completed pre- and postassessments; the treatment group was somewhat younger although 70% of participants were 12 to 13 years of age (Table 1).

Table 1.

Participant Baseline Characteristics.

Physical characteristic	Control $\bar{X}$ (SD)	Intervention $\bar{X}$ (SD)
Age (years)*	12.93 (0.91)	12.03 (0.73)
Weight (kg)	52.47 (10.51)	56.91 (14.66)
Height (m)	1.56 (0.07)	1.54 (0.07)
PAQ-C Activity Score*	2.58 (0.63)	2.26 (0.57)
Ethnicity and socioeconomic status	n (%)	n (%)
Caucasian	19 (63)	16 (55)
Black	2 (7)	5 (17)
Hispanic	9 (30)	8 (28)
School socioeconomic status (% free/reduced lunch)	52	57
Age	n (%)	n (%)
11	3 (10)	6 (21)
12	4 (13)	17 (59)
13	15 (50)	5 (17)
14	8 (27)	1 (3)
EF and academic measures	Control $\bar{X}$ (SD)	Intervention $\bar{X}$ (SD)
Science Pretest	55.23 (18.52)	56.09 (21.86)
ISCI Pretest	28.83 (7.38)	28.73 (6.17)
MCI Pretest	32.60 (8.37)	33.11 (9.01)
CRI Pretest	66.10 (15.62)	72.61 (16.28)
ERI Pretest	27.10 (6.98)	27.81 (7.09)

Note. PAQ-C = Physical Activity Questionnaire for Children; EF = executive function; ISCI = Inhibitory Self-Control Index; MCI = metacognition index; CRI = cognitive regulation index; ERI = emotional regulation index.

p < .05.

A priori analysis indicated that a sample of 53 students would be necessary to detect medium to large effects with 80% power, using linear regression with 95% confidence (Faul, Erdfelder, Buchner, & Lang, 2009). Participant demographics and pretest scores for science achievement and EF were compiled during the baseline assessment phase and independent-samples t tests revealed no significant differences between the control and intervention groups regarding height, weight, science pretest, measures of EF, and ethnicity (Table 1). The Physical Activity Questionnaire for children (PAQ-C) score indicated their self-reported fitness levels at the start of the program (Crocker, Eklund, & Kowalski, 2000), with the treatment group reporting lower physical activity, as expected from the designated recruitment parameters. The PAQ-C measures frequency of moderate to vigorous sports, dance, physical education, and/or games reported over 7 days, with a score of 1 indicating “no activity,” and a 5 as “seven or more” days of moderate to vigorous physical activity. PAQ-C summary scores indicated relatively infrequent exercise with a range of 1 to 3 days for the intervention and control groups at baseline.

Context and Treatment

The program was founded in 2010 as a pilot for middle school girls, ages 11 to 14, identified by the treatment program as at risk for issues with low self-esteem, sedentary lifestyle, and/or being overweight or obese. Eligible participants were selected by a team of program coordinators, social workers, middle school teachers, instructional support teams, and nurses in identifying females struggling socially and/or academically in Grades 6 through 8. The program combined physiology education and physical activities such as triathlon training and yoga 3 times per week for 20 weeks. The fitness component of the program culminated with a timed sprint triathlon (300-yard swim, 7-mile bike ride, and 1.5-mile run). The girls were required to set goals, plan physical activities, and self-regulate to meet established benchmarks.

Science Achievement Assessment

Science achievement was measured using a 15-question pre-/post-multiple-choice assessment developed for middle school students from Project 2061 (American Association for the Advancement of Science [AAAS], 2014). These assessment items were reviewed by experts to identify specific knowledge and skills required to complete each task established from a normative sample of N = 1,000 middle school students from national field testing (AAAS, 2014). Fifteen questions were selected to measure middle school students’ science conceptual understandings and process skills, which two science education experts determined were aligned with the New York State Intermediate Level Science Curriculum Grades 5-8 (NYSED, 2016). Topics were selected from physical, Earth, and life sciences and were consistent with the science curricula learned by both treatment and control groups in their schools. Student scores were based on percentage correct. An independent-samples t test was conducted to compare science pretest scores between groups. Scores were normally distributed and responses demonstrated adequate measured reliability (α = .74). There was no significant difference in pretest science knowledge scores between the control and intervention groups, t(57) = –.16, p = .87.

Measurement of Cognitive Processes

Behavior Rating Inventory of Executive Function (BRIEF)

EF involves both cognitive and behavioral elements. The purpose of the BRIEF-Parent related scales questionnaire (Gioia, Isquith, Guy, & Kenworthy, 2000) is for parents to assess EF in their children aged 5 to 18. In responding to 86 items pre- and postintervention, parents were instructed to rate how often their children had problems with specific behaviors over the prior 6 months. Each item was coded as 1 for never, 2 for sometimes, and 3 for often, for a maximum global index score of 219. Higher scores on any of the constructs indicated an increase in problem behavior with lower scores indicating average to improved EF, therefore an inverse relationship between science performance and EF measures was expected. Informant report was used because research suggested that self-report of cognitive ability is only weakly related to neuropsychological test performance (Burgess, Alderman, Evans, Emslie, & Wilson, 1998).

Parent rating of EF has been shown to be a good predictor of youth functioning in academic, social, behavioral, and emotional domains (Isquith, Gioia, Guy, & Kenworthy, 2015). Parents observed their children’s behaviors and problem-solving skills and offered insights that could not be observed by an independent investigator. Each pre-/postquestionnaire contained items in eight overlapping EF scales that were broken into four indices: inhibition index (ISCI), metacognition index (MCI), emotional regulation index (ERI), and the cognitive regulation index (CRI). Isquith et al. (2015) defined the ISCI as the ability to modulate actions, responses, emotions, and behavior via inhibition that is fundamental to metacognition and problem solving. An example item for inhibition is “Has trouble putting the brakes on his or her actions.” ISCI was calculated by adding the inhibition and emotional control scales. ERI measures the emotional response regulation, shifting, and adjustment to environmental changes. ERI was calculated by the sum of shifting and emotional control scales. MCI represents the ability to initiate, plan, implement, and sustain strategies. A sample item is “When given three things to do, only remembers the first or last.” MCI was calculated by the sum of working memory and plan/organize scales. CRI reflects the ability to control and manage cognitive processes and solve problems. CRI was obtained by the sum of the scales of planning, initiation, working memory, monitoring, and organization. Isquith et al. (2015) related CRI to the ability to solve problems in a variety of contexts and complete tasks such as schoolwork. A sample item is “Has trouble carrying out the actions needed to reach goals (saving money for a special item, studying to get a good grade).” Composite responses demonstrated high measured reliability (α = .97).

Parent-reported observations of EF-related behavior in the BRIEF were log transformed to normalize the distribution. An exploratory factor analysis with Varimax rotation for EF domains was conducted to identify factor loadings. The principal components analysis revealed three factors, accounting for 85% of the variance. Factor 1 included aspects of ISCI and MCI; Factor 2 was closely related to ERI; and Factor 3, CRI. The loadings support previous research that suggested EF is comprised of multiple factors rather than a single unitary factor (Anderson, 2002; Latzman et al., 2010; Miyake et al., 2000).

Measures of Physical Fitness

PAQ-C

A measure of physical activity was employed to assess whether the control and treatment groups had differences at the start of the program. The PAQ-C has been used successfully in longitudinal research to measure children’s general physical activity levels from childhood through adolescence (Crocker et al., 2000). The PAQ-C summary score is derived from nine items, each scored on a 5-point scale with “no activity” coded as a 1 and “7 times or more” coded as a 5. In a comparison of means between groups, an independent t test revealed a significant difference in baseline physical activity between treatment and control groups, t(57) = 2.04, p = .046, 95% confidence interval (CI) = [0.01, 0.64], d = .54, a medium effect; this was expected from a group classified as at-risk and established that the treatment girls were less active than the control group before the intervention.

Statistical Analyses

Several inferential techniques were employed to measure program outcomes. Independent-samples t tests were performed to compare the posttest scores on the science assessment for the treatment, control, and normative sample groups. ANCOVA was conducted to determine the effect of the intervention on science achievement, with science pretest as the covariate and group as the fixed factor. A MANCOVA was conducted to determine the individual and multivariate effects of the treatment on EF indices and science achievement, with science pretest as the covariate and group as the fixed factor. ANCOVA and MANCOVA were chosen to test the null hypotheses that there were no significant differences between the intervention and the control groups in science achievement and measures of three separate cognitive domains (inhibition, metacognition, cognitive regulation) while controlling for science pretest.

A Roy-Bargmann stepdown analysis was performed to examine the univariate significance of each dependent variable in the MANCOVA listed in a prioritized order. This procedure was selected to illuminate group differences among response variables in a hierarchical structure, so variables might be evaluated in order of importance. This allowed for a more nuanced interpretation of the MANCOVA results. In stepdown analysis, the highest priority dependent variable (science achievement) was tested through ANOVA and was then used as a covariate for the next dependent variable. Each successive dependent variable was then used as a covariate for the next dependent variable until all variance that contributed to significance was extracted (Tabachnick & Fidell, 2013). Prioritized dependent variables were based on the theoretical framework where the proposed model of EF stems from cognitive processes that contribute to science achievement (Anderson, 2002; Latzman et al., 2010; Miyake et al., 2000). Science achievement was the highest priority dependent variable, followed by cognitive regulation as it contributes to academic achievement and reasoning by way of the third dependent variable, metacognition. The final dependent variable was inhibition. This model was utilized to highlight the individual contributions of the dependent variables to the significance of the treatment. Assumptions of normality and homogeneity of variance were met.

Results

Demographic and anthropometric variables were similar in the control and intervention group with respect to height, weight, and ethnicity, though the treatment group was slightly younger and less physically active at the baseline (Table 1). To test the effects of the intervention as it related to science achievement, two initial steps were taken. First, independent-samples t tests compared the posttest science scores by item for the 15-item science assessment and revealed significantly improved change scores for the treatment group when comparing (a) the treatment and control groups, t(14) = 3.14, p = .007, 95% CI = [2.4, 12.6], d = .58, a medium effect; and (b) the normative sample and treatment groups, t(14) = 3.48, p = .004, 95% CI = [3.5, 13.8], d = .54, a medium effect. These results suggest that the intervention contributed to an increase in posttest achievement on the science assessment. Second, ANCOVA was conducted with group as the fixed factor, science achievement as the dependent variable, while controlling for the science pretest. There was a significant effect with the treatment group scoring higher gains with a medium effect size accounting for 26% of the variance, F(1, 54) = 7.67, p < .01, adjusted R² = .26, $η_{p}^{2}$ = .08.

Change scores for EF were calculated for ISCI, MCI, CRI, and ERI constructs to preserve power in the N = 59 sample and elicit inferences from quasi-experimental data. In nonequivalent group designs where an intervention occurs in the treatment group that does not happen in the control, change scores allow for a more direct, intuitive evaluation of the models and their implications (Allison, 1990). Independent-samples t tests were conducted for changes in EF domains to determine whether treatment and control groups differed. Significant differences were observed with the treatment group exhibiting higher gains in inhibition (ISCI), t(54) = 2.18, p = .03, 95% CI = [0.39, 9.58], d = .59, a medium effect; metacognition (MCI), t(54) = 2.86, p = .006, 95% CI = [2.26, 12.88], d = .79, a large effect; and cognitive regulation (CRI), t(54) = 2.27, p = .03, 95% CI = [1.27, 20.53], d = .62, a medium effect. The change in ERI was not significant when comparing groups, t(54) = –.40, p = .69, 95% CI = [–.06, .05], and was eliminated from inclusion in multivariable analysis.

To test the effect of the intervention on index measures of EF as they related to science achievement, a MANCOVA was conducted with group as the fixed factor, and science achievement, ISCI, MCI, and CRI as dependent variables while controlling for the science pretest. There was a significant multivariate effect with the treatment group reporting higher gains on cognition measures and science achievement, F(4, 50) = 3.75, p = .01. The adjusted R² for the overall model was .24, indicating moderate accounting for the variance in science performance and EF. There was a significant medium main effect with more improvement by the treatment group on science achievement, F(1, 53) = 4.84, p = .032, $η_{p}^{2}$ = .08, and a significantly higher gain with a medium effect for inhibition, F(1, 53) = 4.59, p = .037, $η_{p}^{2}$ = .08. The treatment group showed significant improvement on metacognition with a medium effect when compared with the control group, F(1, 53) = 7.97, p = .007, $η_{p}^{2}$ = .13, as well as cognitive regulation, F(1, 53) = 5.01, p = .029, $η_{p}^{2}$ = .09, with a medium effect.

ANCOVA was used after MANCOVA in Roy-Bargmann stepdown analysis to examine the contributions of dependent variables across groups (Tabachnick & Fidell, 2013). To test the null hypothesis that there was no effect across groups when the dependent variables were listed in a specific order, science achievement was given the highest priority, followed by CRI, MCI, and ISCI as the final priority; this was based on the theoretical framework that EF consists of three separate yet interrelated domains that contribute to science achievement. Three dependent variables—science achievement, cognitive regulation, and metacognition—made unique contributions to the composite dependent variable that distinguished between groups. After adjusting for differences in covariates, science achievement had a significant difference by group stepdown, F(1, 54) = 4.62, p = .04, $η_{p}^{2}$ = .08, a medium effect. With differences due to science achievement already entered, cognitive regulation made a significant contribution by treatment group with a medium effect, stepdown F(1, 53) = 4.99, p = .03, $η_{p}^{2}$ = .09. Metacognition, adjusted by science achievement and cognitive regulation also made a unique contribution to the composite dependent variable with a medium effect, stepdown F(1, 52) = 4.46, p = .04, $η_{p}^{2}$ = .08. Univariate analysis revealed that inhibition did not make a contribution, stepdown F(1, 51) = .10, p = .75. The results for MANCOVA and stepdown analysis are summarized in Table 2.

Table 2.

Multivariable Results for Science Achievement, Inhibition, Metacognition, and Cognitive Regulation Change Scores by Treatment Group.

Change score measure (N = 59)	MS	F	$η_{p}^{2}$	Adjusted R²	Stepdown F	Stepdown $η_{p}^{2}$
Science achievement	1,492.77	4.84*	.08	.240	4.62*	.08
Cognitive regulation	1,639.36	5.01*	.08	.053	4.99*	.09
Metacognition	791.84	7.97**	.13	.099	4.48*	.08
Inhibition	341.89	4.59*	.09	.047	.099	N/A

Note. MS = Mean Squares

p < .05. **p < .01.

Discussion

The present study suggests science achievement and EF components were significantly improved for at-risk students who participated in the aerobic fitness intervention. Multiple comparisons showed that the intervention group of adolescent middle school girls demonstrated statistically higher science achievement, with a moderate effect. Inhibition, metacognition, and cognitive regulation were significantly improved compared with the control; data revealed medium effect sizes for inhibition and cognitive regulation, and a large effect size for metacognition. MANCOVA revealed that 24% of the variance in cognition and science achievement between the control and experimental groups could be accounted for by the treatment. These data support the use of physical activity interventions to improve aspects of EF, which may also be related to improvements in science achievement.

Significant lower scores on the post-BRIEF survey suggested improved behavior and EF for the intervention group. They had higher science achievement and improved cognitive regulation and metacognition, supporting the model whereby an exercise intervention related improvement in mutually dependent EF domains that influence cognitive control and goal setting. The results provide evidence that the fitness treatment is a promising innovation for young, at-risk girls to improve science learning along with associated cognitive processes. The relationship between EF and science proficiency is supported by data indicating increased self-regulation and problem solving that led to more efficient learning.

While the multivariate analysis revealed a significant finding for inhibition, further analysis is needed to determine the specific contributions of inhibition to science achievement. In the stepdown analysis, the effects of criteria variables were examined. If previous steps extracted all variance that contributed to significance, later steps would show no further differences. As the ISCI was composed of inhibition and emotional control indices, it may be necessary to examine them separately with additional measures, independent of the moderate to large effects found with cognitive regulation and metacognition. Although Latzman et al. (2010) and Davis et al. (2011) found inhibition to be related to academic achievement, they did not explore its impact in a hierarchical model. Inhibition may be important in formulating scientific conclusions after all evidence is evaluated. However, metacognition and cognitive regulation may supersede inhibition in terms of foundational importance in the scientific method. The ERI, comprised of shifting and emotional control composites from the BRIEF, was not significantly changed. It may be necessary to measure shifting with time dependent processing tasks and emotional control in order to detect notable changes in ERI.

A particularly interesting finding was the large effect size for metacognition as it relates to science achievement. The improved MCI demonstrates the long-term effects of physical activity on working memory, contradicting earlier inconclusive findings (Smith et al., 2010). Our work provides evidence that chronic physical activity can enhance cognitive control abilities including increased attention span, concentration, and improved memory. These improvements enable individuals to maintain the control necessary to inhibit automatic responses, adapt behavior to new demands, and maintain goal-directed behavior, ultimately leading to self-regulation for improving academic success. These cognitive functions contribute to the critical thinking skills necessary for making scientific inferences. Consequently, these impacts have implications for narrowing the achievement gap in science and promoting STEM aspirations early in the lives of at-risk young women.

In the present study, adolescent girls engaged in increased physical activity while setting goals and planning performance training for a youth triathlon. These efforts required them to organize actions to approach tasks efficiently, monitor performance, and evaluate new information while changing previously sedentary lifestyles. The results suggest treatment group participation had significant effects on cognition and science performance, providing evidence for the importance of aerobic exercise and physical activity in improving specific EF domains that support science learning.

Limitations

A limitation of this research was that observational fitness data were not used. Self-reported physical activity data were used as a proxy for fitness to test for group differences at the start of the program. Although there were different levels of physical activity at the start, posttreatment fitness measures were not available for both groups. The treatment participants were much more physically active at the conclusion of the program as they shifted their sedentary lifestyles to the new demands of triathlon training. Due to the intensive nature of the fitness measures (i.e., running, biking, swimming), it was difficult to collect these data on the control group. We assumed the control group would not change their physical activity significantly during the 20-week duration. It should also be noted that this study did not include assessment of motivation for science learning or exercise. A separate qualitative research analysis investigated intrinsic and extrinsic motivation and related constructs to understand the potential for improved fitness, cognition, and attitudes toward science (Authors, 2017). In addition, the young women who participated self-selected for this program and were not randomly assigned to treatment and control groups. Randomized trials using chronic aerobic exercise training in adolescents are needed with larger sample sizes and more specific measures associated with EF to elicit causal inferences. Our research is a first step in contributing to increased understanding of the effects of physical fitness through the mechanisms of inhibition, metacognition, and cognitive regulation on science performance.

Conclusions and Implications

Afterschool informal aerobic fitness programs can play a key role in engaging youth in science by providing opportunities to increase physical activity and improve health, which may facilitate EF functions related to goal setting and problem solving. In addition, advances in the understanding of brain development and mechanisms of learning have substantial implications for science education. As EF develops with increasing age, physical activity may be particularly beneficial for a student’s ability to reason, monitor academic behaviors, and learn new concepts. This fitness intervention may be easily modified to scale up mechanisms for increasing the amount of physical activity during or outside of the school day. This might include a recommended minimum threshold for physical fitness training during the academic year, broader opportunities for triathlon training for middle school students, and more fitness activities geared toward holistic wellness to promote lifestyle shifts. The stimulation of physical activity holds tremendous promise for improving EF and achievement in related academic domains such as science.

Self-esteem, school achievement, attitudes, and motivation, particularly for girls, often decline in the middle school years (Jones et al., 2000). Advances in the understanding of brain development and mechanisms of learning have substantial implications for academic achievement. Evaluation of EF with appropriate age-based cognitive tests may provide a more reliable understanding of how chronic aerobic exercise interventions influence brain development and processing skills related to science achievement. Furthermore, different levels of physical fitness may have varying effects on cognitive and academic outcomes. Future research is needed to explore the impact of dose response as well as causal mechanisms, building upon research that suggests increased vigor and frequency of physical activity results in proportional increases in mental capabilities related to specific academic disciplines (Davis et al., 2011).

The treatment program in the present study promoted student learning and has particular significance for at-risk female students. Sustained increases in physical activity show promise for improving cognitive function such as planning for goal accomplishment, the evaluation of alternatives, and revision of hypotheses so they are consistent with new information; these skills are foundational in scientific reasoning. Young at-risk women who develop fitness habits that counteract the effects of sedentary lifestyles may improve cognitive processes, EF, science achievement, and ultimately interest in science coursework and careers.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Angela M. Kelly

Author Biographies

Jennifer Gatz is a doctoral candidate in science education at the Institute for STEM Education at Stony Brook University. She is a biology and science research teacher at Patchogue-Medford High School on Long Island, NY. Her research interests include the impacts of aerobic fitness on executive function, motivation, and science learning.

Angela M. Kelly is associate director of science education in the Institute for STEM Education and associate professor in the Department of Physics and Astronomy at Stony Brook University. She was awarded a PhD in science education from Columbia University. Her research interests include improving opportunities for science learning for children in underserved communities.

Sheri L. Clark is a proposal development specialist in the Office of Proposal Development at Stony Brook University. She was awarded a PhD in psychology from Stony Brook University and served as a postdoctoral associate in the Center for Inclusive Education. Her research interests include psychosocial pathways to STEM engagement.

References

Allison

P. D.

(1990). Change scores as dependent variables in regression analysis. Sociological Methodology, 20, 93-114.

American Association for the Advancement of Science. (2014). AAAS Project 2061 science assessment. Washington, DC: Author.

Anderson

O. R.

(1983). A neuromathematical model of human information processing. Journal of Research in Science Teaching, 20, 603-620.

Anderson

(2002). Assessment and development of executive function during childhood. Child Neuropsychology, 8, 71-82.

Best

J. R.

(2010). Effects of physical activity on children’s executive function: Contributions of experimental research on aerobic exercise. Developmental Review, 30, 331-551.

Blair

Razza

R. P.

(2007). Relating effortful control, executive function, and false belief understanding to emerging math and literacy ability in kindergarten. Child Development, 78, 647-663.

Burgess

P. W.

Alderman

Evans

Emslie

Wilson

B. A.

(1998). The ecological validity of tests of executive function. Journal of the International Neuropsychological Society, 4, 547-558.

Crocker

P. R. E.

Eklund

R. C.

Kowalski

K. C.

(2000). Children’s physical activity and physical self-perceptions. Journal of Sports Sciences, 18, 383-394.

Davidson

M. C.

Amso

Anderson

L. C.

Diamond

(2006). Development of cognitive control and executive functions from 4 to 13 years: Evidence from manipulations of memory, inhibition, and task switching. Neuropsychologia, 44, 2037-2078.

10.

Davis

C. L.

Tomporowski

P. D.

Boyle

C. A.

Waller

J. L.

Miller

P. H.

Naglieri

J. A.

Gregowski

(2007). Effects of aerobic exercise on overweight children’s cognitive functioning: A randomized controlled trial. Research Quarterly for Exercise and Sport, 78, 510-519.

11.

Davis

C. L.

Tomporowski

P. D.

McDowell

J. E.

Austin

B. P.

Miller

P. H.

Yanasak

N. E.

. . . Nalgieri

J. A.

(2011). Exercise improves executive function and achievement and alters brain activation in overweight children: A randomized, controlled trial. Health Psychology, 30, 91-98.

12.

Etnier

J. L.

Nowell

P. M.

Landers

D. M.

Sibley

B. A.

(2006). A meta-regression to examine the relationship between aerobic fitness and cognitive performance. Brain Research Reviews, 52, 119-130.

13.

Faul

Erdfelder

Buchner

Lang

A. G.

(2009). Statistical power analyses using G*Power 3.1: Test for correlation and regression analyses. Behavior Research Methods, 41, 1149-1160.

14.

Gatz

Kelly

A. M.

(2018). Afterschool school triathlon training for 11-14 year old girls: Influences on academic motivation and achievement. Health Education Journal, 77, 156-168.

15.

Gioia

G. A.

Isquith

P. K.

Guy

S. C.

Kenworthy

(2000). Test review: Behavior Rating Inventory of Executive Function. Child Neuropsychology, 6, 235-238.

16.

Hall

C. D.

Smith

A. L.

Keele

S. W.

(2001). The impact of aerobic activity on cognitive function in older adults: A new concept based on the concept of executive control. European Journal of Cognitive Psychology, 13, 279-300.

17.

Hillman

C. H.

Pontifex

M. B.

Raine

L. B.

Castelli

D. M.

Hall

E. E.

Kramer

A. F.

(2009). The effect of acute treadmill walking on cognitive and academic achievement in preadolescent children. Neuroscience, 159, 1044-1054.

18.

Isquith

P. K.

Gioia

G. A.

Guy

S. C.

Kenworthy

(2015). Behavior Rating Inventory of Executive Function, second edition: Parent form interpretive report. Lutz, FL: Psychology Assessment Resources.

19.

Jones

G. M.

Howe

Rua

M. J.

(2000). Gender differences in students’ experiences, interests, and attitudes towards science. Science Education, 84, 180-192.

20.

Kramer

A. F.

Hahn

Cohen

N. J.

Banich

M. T.

McAuley

Harrison

C. R.

. . . Colombe

(1999). Ageing, fitness and neurocognitive function. Nature, 400, 418-419.

21.

Latzman

R. D.

Elkovitch

Young

Clark

L. A.

(2010). The contribution of executive function to academic achievement among male adolescents. Journal of Clinical and Experimental Neuropsychology, 32, 455-462.

22.

Lauer

P. A.

Akiba

Wilkerson

S. B.

Apthorp

H. S.

Snow

Martin-Glenn

M. L.

(2006). Out-of-school-time programs: A meta-analysis of effects for at-risk students. Review of Educational Research, 76, 275-313.

23.

Miyake

Friedman

N. P.

Emerson

M. J.

Witzki

A. H.

Howerter

Wager

T. D.

(2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cognitive Psychology, 41, 49-100.

24.

National Research Council. (2007). Taking science to school: Learning and teaching science in grades K-8. Washington, DC: National Academies Press.

25.

National Research Council. (2012). A framework for K-12 science education: Practices, crosscutting concepts, and core ideas. Washington, DC: National Academies Press.

26.

New York State Education Department. (2016). Intermediate level science core curriculum grades 5-8. Albany, NY: Author.

27.

Schraw

Crippem

K. J.

Hartley

(2006). Promoting self-regulation in science education: Metacognition as part of a broader perspective on learning. Research in Science Education, 36, 111-139.

28.

Skogan

A. H.

Egeland

Zeiner

Øvergaard

K. R.

Oerbeck

Reichborn-Kjennerud

Aase

(2016). Factor structure of the Behavior Rating Inventory of Executive Functions (BRIEF-P) at age three years. Child Neuropsychology, 22, 472-492.

29.

Smith

P. J.

Blumenthal

J. A.

Hoffman

B. M.

Cooper

Strauman

T. A.

Welsh-Bohmer

. . . Sherwood

(2010). Aerobic exercise and neurocognitive performance: A meta-analytic review of randomized controlled trials. Psychosomatic Medicine, 72, 239-252.

30.

Tabachnick

B. G.

Fidell

L. S.

(2013). Using multivariate statistics (6th ed.). London, England: Pearson.

31.

Tomporowski

P. D.

Davis

C. L.

Miller

P. H.

Naglieri

J. A.

(2008). Exercise and children’s intelligence, cognition and academic achievement. Educational Psychology Review, 20, 111-131.