Playground Use and Executive Function Development During Preschool Years

The current research explores executive function (EF) development among the children at the University Of Rhode Island’s child development centers (CDCs). Specifically, the research explores the role of outdoor play in executive function development. Two samples of children from the Providence Child Development Center and the Kingston Child Development Center are compared twice (time 1 and time 2) across 5 months to assess the role of outdoor play spaces in executive function development. The independent variable, playground type has two levels: a playground structure (Kinsgton CDC) and open outdoor space (Providence CDC). Executive function is assessed using three tasks, the Day/Night task, the Backwards Digit Span, and the Standard Dimensional Change Card Sort Task (DCCS). Findings reveal no significant difference between the two samples at time 1 and time 2. Results suggest an outdoor playground does not provide greater benefits for executive function development than the use of an open space. Results can be used to inform educators regarding the use of outdoor space for playtime.


LIST OF TABLES
linked to physical activity in childhood, yet to date, little research explores the use of an outdoor play structure on EF development. Therefore, this research will explore the use of outdoor play space on the development of EF skills in early childhood.
Executive function, including inhibitory control, working memory, and cognitive flexibility, is used throughout life to organize and control goal-directed behavior (Banich, 2009). Specifically, inhibitory control enables self-control over one's behavior, emotions, and attention (Diamond, 2011). Working memory allows individuals to hold information in mind while actively manipulating the information, and cognitive flexibility refers to the ability to change perspectives and adjust to new or unexpected information (Diamond, 2011). These skills are important to ensure actions are not a result of impulse, challenges can be faced, and information can be built upon over time (Diamond, 2011). Moreover, these core components of EF lay the foundation for higher-order EF skills including reasoning, problem solving, and planning (Collins & Koechlin, 2012), making it important to foster such skills at a young age for success later in life.
Research supports the importance of EF for school readiness (Blair & Razza, 2007;Morrison, Ponitz, & McClelland 2010) and success in school across adolescence (Alloway & Alloway, 2010). Moreover, Blair and Razza (2007) found EF skills are more important for school readiness than IQ, as EF in preschool was related to measures of math and literacy ability in kindergarten. Additionally, previous research found EF is critical to career success (Bailey, 2007), making and keeping friends (Hughes & Dunn, 1998), and marital success (Eakin et al., 2004).
Understanding the effect of an outside play space on young children's EF development is important for educators and policy makers alike, as outdoor play is a consistent predictor of physical activity, which is linked to children's social, mental, and physical health (Veitch, Bagley, Ball, & Salmon, 2006 (Reznick et al, 2005), but the length of time representations are held and the number of representations that are retained develop after 6 months (Pelphrey & Reznick, 2002).
Particularly, Diamond and Doar (1989) found a 10 second increase in the number of seconds an infant can hold information in memory from 6 months to 1 year.
Additionally, several studies employing the Towers of London task, used to assess cognitive planning, have found the number of representations that can be retained over time differs from ages 3 to 5 (Bull, Espy, & Senn, 2004;Espy & Bull, 2005), and during this time period the ability to update this information also develops (Espy & Bull, 2005). Similar findings were reported by Davis & Pratt (1995) in their assessment of working memory using the backward digit span and forward digit span tasks. Carlson (2005) found between the ages of 3 and 5 the number of items a child can remember backwards improves, with 34% (n=29) of young three year olds completing 3 or more trials of the backwards digit span compared to 73% (n=65) of young four year olds completing 3 or more trials.
Similar age trends exist in terms of inhibitory control. In her sample of young children 24 months old to 4 years old, Carlson (2005) found age differences in the length of time children are able to delay a response. Specifically, 50% of 24-montholds were able to suppress eating a treat for 20 seconds, whereas 85% of 4-year-olds suppressed the urge for 1 minute and 72% of 4-year-olds suppressed eating the treat for 5 minutes (Carlson, 2005). In addition, multiple research studies have found age differences in children 3-to 5-years-old in tasks that require following an arbitrary rule and suppressing a dominant response (Carlson, 2005;Diamond, 1991). Research suggests that children develop this ability rapidly from young 3s to older 3s, yet more difficult versions of the initiating-suppressing tasks, such as Simon Says, are challenging for even 4-and 5-year-olds (Carlson, 2005). Research regarding strooplike tasks suggests children are able to solve tasks that involve greater conflict as they age. Specifically, Carlson (2005) found only 45% of 3-year-olds passed the grasssnow task, a task that requires children to inhibit a dominant response and follow an arbitrary rule, and not until 4.5 years of age did 80% of children pass the task. The increase in conflicting ideas is likely responsible for the difficulty children encounter during such tasks (Garon, Bryson, & Smith, 2008).
Research regarding cognitive flexibility suggests success in this domain is related to other EF components (Garon, et al., 2008). Researchers have found children 3-years-old and younger have difficulty completing the Dimensional Change Card Sort, a common measurement of attention shifting for young children (Carlson, 2005).
In particular, Carlson (2005) found 3-year-olds are able to sort according to the first rule of the task but cannot shift to a new rule. After age 4, children are better able to shift to a new rule (Carlson, 2005). Many questions remain regarding shift-setting in the early preschool years (Garon et al., 2008).
Research suggests EF can be improved through various methods including social play, scaffolding, computerized training, and aerobic activity (Diamond & Lee, 2011). Best (2010) describes several studies that found both acute and chronic aerobic exercise aid in the development of EF among adolescents, with the latter being more beneficial (Budde et al., 2008;Davis et al., 2007). In addition, when 7 to 9 year olds were randomly assigned to 2 hours of fitness activities daily for a school year (70 minutes of aerobic activity, then motor skill development) compared to the control group, children who received fitness training showed more improvement in working memory (Diamond & Lee, 2011 The theory of embodied learning can be applied to physical activity and cognitive development. According to the theory of embodied learning, interaction between sensorimotor integration and the environment plays a pivotal role in the development of certain cognitive functions (Spencer et al., 2006). Although definitions of embodied learning vary across disciplines, there are several concepts that remain consistent throughout the literature. First, an individual's ability to interact with the environment influences cognition. As an individual explores his or her movement in a particular area, a framework for action control develops (Tomporowski, Lambourne & Okumura, 2011). Second, an organism is restricted in the type of cognitive processes possible based upon available physical structures (legs, arms, etc.). Lastly, physical structures influence the way the organism views the environment (Tomporowski et al., 2011).
According to the theory of embodied learning, the use of outdoor play structures may allow for cognitive processing that differs from processes used in the absence of structures. In addition, in line with the theory of embodied learning, the structures will influence the way children view the environment and explore their movement. The structures on the playground provide opportunities for children to jump, climb, explore space, and challenge their abilities. Furthermore, these structures may provide an opportunity for exploration and environmental feedback that differs from outdoor play in open space, such as the ability to power a swing through moving one's legs or exploring the distance and strength needed to reach across the monkey bars.
Additionally, motor development has been linked to EF (Koziol, Budding, & Chidekel, 2011). Koziol, Budding, and Chidekel (2011) explain the link between motor development and executive function, arguing that humans were designed to move and the fundamental purpose of an organism is to survive through environmental interactions. Koziol and colleagues (2011) posit that goal-directed action management requires the development of anticipatory control mechanisms to predict sensorimotor outcomes. This requires the development of "on-line" sensorimotor anticipation to adjust to the environment and "off-line" simulations to plan behavior. In this model, EF falls within "off-line" simulations. The authors argue that motor development and EF are inexorably linked and motor movements reflect action control, an early form of EF (Koziol et al., 2011). Supporting this argument is research linking motor development and executive function. For example, Piek, Dawson, Smith, and Gasson (2008) found a relationship between early gross motor problems and the later development of particular EF deficits in processing speed and working memory among school-aged children. In another study of adolescents, Westendorp and colleagues (2011) found children with learning disabilities performed more poorly on all motor tasks. Furthermore, rats engaged in exercise training that involved motoric climbing skills developed neural connections within the cerebellum, the area of the brain responsible for fine and gross motor skills, while rats engaged in aerobic activity improved cerebral blood flow (Black et al., 1990 Furthermore, motor activity is challenged when interacting with the natural environment which provides dynamic and rough playscapes (Fjørtoft, 2001).
Therefore, this exploratory study aims to discover if a difference exists among the different play areas in promoting EF development.
The playground structures available at the Kingston CDC include a swing set, a seesaw, pull-up bars, monkey bars, a climbing structure, a sliding pole, a slide, and a rock children often use for climbing and jumping. These structures allow the children to climb, jump, push, and hang. The children at the Providence CDC have access to outdoor space throughout the city. These areas allow for running but do not provide opportunities for climbing or jumping aside from a stage children use while playing.
Based on prior research, and utilizing the current play areas at each CDC, this exploratory study will aim to understand the differences and similarities in the use of outdoor space versus an outdoor play structure in fostering EF development in early childhood.

METHODOLOGY
The current exploratory study employs a longitudinal approach to understanding the influence of different outdoor play areas on EF development. Data were collected over a four month period, first in January, then again in May.

Participants
The study utilized the University of Rhode Island's Child Development Centers (Providence and Kingston). Children ages 3 to 5 years old, enrolled in the fullday program, and their parents, were invited to participate in the study. The letter inviting parents to join the research was sent to 60 families, with 40 returning the consent form, i.e., 20 families refused to participate, resulting in a sample size of 40 children.
Three children refused to play at least one of the three games with the researcher, therefore, these children were removed from analyses. A fourth child was no longer with the CDC at the time of the second data collection, thus, this child was also removed from analyses.

Response Inhibition
Day/Night task (Gerstadt, Hong, & Diamond 1994): The researcher engaged the children in a brief conversation about when the sun comes up (in the day) and when the moon and stars come out (in the night). The researcher then presented a card with a yellow sun drawn (to represent the day) and a card with the moon and stars (to represent the night). Next, the experimenter explained this is a "silly" game and when the day card is shown, children should say night. Children were instructed to say day when the night card was shown. After practice trails children were tested on 16 consecutive trials. The number of correct trials (out of 16) was recorded for each child.
If children completed 12/16 trials, or more, they were counted as passing the task.
Although the psychometric properties of the Day/Night task have not been extensively reviewed, research indicates the task possesses high internal reliability (α=.91) (Rhoades, Greenberg, & Domitrovich, 2009), high test-retest reliability (Thorell & Wåhlstedt, 2006), and high predictive reliability for academic achievements (Duncan, 2012). In her review of executive function measures for preschool children, Carlson (2005) found that 27% of their sample size of old 3-year-olds (n=45) passed the Day/Night task and 68% of older 4 year olds (n=19) passed.

Cognitive Flexibility
Standard Dimensional Change Card Sort (DCCS) (Zelazo, 2006): Children were introduced to two boxes with target cards (a red bunny and a blue boat) glued to each box. The experimenter introduced the game as a sorting game and asked the children to place all of the red and blue bunny cards into the red bunny box and the red and blue boat cards into the blue boat box. This was considered the "shape game," bunny vs. boat. After five consecutive trials the experimenter switched to the "color game," red vs. blue, and explained to the children that all red cards were to be placed in the red bunny box and all of the blue cards were to be placed in the blue boat box, regardless of object shape. Based on previous research, there were five post switch trials-two were compatible with the old rule and three were incompatible with the old rule. The total number of correct incompatible trials was recorded.  (Duncan, 2012). Twenty-five percent of Carlson's (2005) sample of 3- year-olds (n=79) passed the DCCS task and 76% of older 4-year-olds (n=38) passed.

Working Memory
Backward Digit Span (Davis & Pratt, 1996): The experimenter introduced the children to a puppet and explaining that the puppet is silly, and whatever she says the puppet says backwards. The experimenter demonstrated saying "1, 2" and the puppet followed "2, 1." Then, the children were invited to try using the example. The task began with two digits and the number of digits increased by 1 until the child made an error on three consecutive trials. The highest level of completion was recorded (two, three, four, or five). To pass the backwards digit tasks children must complete the task using 3+ digits. Gathercole (1995)  Based on research suggesting associations between specific play activities and EF, parents were also asked to indicate the amount of time their children spend engaged in particular play activities outside of school. Specifically, parents were asked to estimate the typical amount of daily time (in 15 minutes increments) children typically engaged in running games (Tuckman & Hinkle, 1986), organized movement activities (Brown, 1967), informal sports with friends/family (Davis et al., 2007), imitation games (Carlson, 2005), aerobic activities, computer activities, martial arts, and mindfulness activities (Tomporowski et al., 2007) outside of school. Teachers were also asked to complete a questionnaire addressing the time spent within the classroom on each of the aforementioned activities.

Procedure
The researcher collected data in a private space within each of the CDCs. The research followed a pre-test and post-test design in attempt to gather data regarding development over 4 months. The test was first administered to each child in January 2015 and again in May 2015. Parents and teachers were asked to complete the questionnaire in January. A letter was sent home to parents explaining the project and procedures. The letter indicated a date and time the researcher was available at each center to answer any questions regarding the research and to distribute and collect signed informed consent forms. After parental consent and parental permission were given, children were asked to give assent prior to participation in the three tasks. The researcher explained to the children they had the option not to play, or to stop playing the games at any time. Parental informed consent, parental permission, child assent, and teacher consent were obtained and all research complied with URI's IRB.

Data Analysis
All collected data was entered into SPSS V.21. Data was checked for normality, skewness, missingness, and distribution, with recoding of variables occurring as needed. The researcher then used cross tabs and t-tests to test for demographic differences between the two groups. Using cross tabs, a chi-square analysis was used to explore if the variables were independent, i.e., if the child development center was associated with specific demographics that may influence results. Next, correlational analyses were used to test for confounding variables; specifically if mean EF scores varied by demographic differences, which may influence findings. Then, t-tests were used to determine if mean scores on the Time 1 (T1) EF tests and the Time 2 (T2) EF tests varied by group status. A change score variable was created by subtracting T1 EF mean scores from T2 EF mean scores. The research question was then addressed using independent groups t-test to compare change scores by group status. Finally, a hierarchical regression was used to assess the strength of the association between variables and the power of the independent variables in predicting T2 outcomes.

FINDINGS
Crosstab analyses revealed the two samples were statistically similar based on gender, age, household income, and parental education (Table 1). Furthermore, the two samples engaged in similar activities throughout the day, with few exceptions.
The greatest difference between the two groups was in time spent on an outdoor Exploratory analyses revealed little variability among children in results of the DCCS task, at both pre and post test. Specifically, 84% of the sample completed the task successfully at T1 and T2. Therefore, the DCCS task was removed from analyses.
Using the Backwards Digit Span task and the Day/Night Task, overall, results revealed no significant difference in executive function scores between the two centers. Independent samples T-tests revealed no significant difference between the two groups on each task, at T1 and T2. Table 2 shows the mean and standard deviations for each sample at pre and post tests. .14) again, the mean scores were not significantly different. It is likely these findings are due to the small sample size. Change score variables were created to compare the two samples over time (     (Carlson, 2005).
Mixed findings regarding executive function and gender are found throughout the literature. For instance, Raaijmakers et al, (2008) found boys exhibit greater deficits in overall EF than girls. Similarly, Diamond and Lee (2010) suggest young boys typically benefit more than young girls from interventions aimed at improving EF.
These findings support the results of the Backwards Digit Task found in this study.
However, some research suggests no significant gender differences in EF during young childhood (Thorell & Wåhlstedt, 2006). Thus, the association between gender and executive function development is an area that warrants further research.
Although previous research has not explored the association between playground structures and EF development, the results provide insight into the application of several theories previously discussed. To begin, using the theory of embodied learning, the results suggest an outdoor play structure designed for young children may not offer environmental feedback more beneficial for children than the natural environment. Additionally, the finding that gains were made across both centers, on each task, likely highlights the typical development of EF throughout young childhood. Specifically, an abundance of research suggests children's EF skills rapidly increase between the ages of 3 and 5 (Bull et al., 2004;Carlson, 2005). It is possible the increases in EF outcomes were not significant because 4 months does not allow enough time for significant gains in development. However, the increase in mean scores across the two EF tasks aligns with previous research (Carlson, 2005).
Several limitations must be noted. To begin, the homogeneity of the sample limits the generalizability of the results to white, middle class, educated families.
Secondly, this was an exploratory study; therefore, the study did not use a randomized control group. A controlled study should be designed to further investigate the influence of play space on executive function development. For example, a study employing an elementary school with multiple classrooms that can be randomly assigned to a controlled play environment vs. free play on a play structure may provide further insight into the differences, if any, between play spaces in fostering EF development. Furthermore, an observational approach may be beneficial, to observe the activities children engage in when on the playground. Next, the research was conducted throughout the winter months, possibly influencing the amount of time children spent engaged in outdoor play. A recommendation for future research is to lengthen the time of the study to incorporate all four seasons and a longer developmental span. Additionally, previous research suggests high test-retest reliability for all of the measures used. Therefore, it is possible 4 months was not adequate time to show significant developmental changes with these measures. The small sample size also serves as a limitation as it reduces the power to reject the null hypothesis, i.e., when the sample size is small, small effects will not be statistically significant.
Lastly, the results of the DCCS task suggest this task may not be an accurate measurement of working memory in ages 3 to 5. This finding aligns with previous literature that suggests the standard DCCS task minimizes inhibitory demands (Best & Miller, 2010). Future research should explore possible measurements for this age group that allows researchers to explore the variability of working memory between ages 3 to 5. Specifically in this sample, 57.1% of three year olds successfully completed the task, 89.5% of four year olds successfully completed the task, and 88% of five year olds successfully completed the task. It is likely little variability was found in this study because the majority of the sample was four and five year olds. Thus, these results suggest a more accurate measurement or modification of the current DCCS task is needed to explore differences in working memory between the ages of 3 and 5 years old.