The first step of the IMP was to establish a planning group and conduct the needs assessment. Using a community participatory research approach, the planning group brought together university researchers from multiple disciplines (engineering, health education, and kinesiology), teachers, students, and community partners to co-design the Project SMART game (Figure 1). The advantage of using a community participatory research process is that it involves researchers and stakeholders working together to understand a problematic situation and change it for a better community (Baum et al., 2006; Kemmis et al., 2013). This process usually focuses on social change that promotes behaviors and challenges inequality. Since the approach is context-specific, it is focused on the needs of a particular group—in this case, elementary school children. Through this process, the school community members and research team share in the decision-making, instead of the researchers making all decisions and asking the school community members for input after decisions have already been made. Needs assessment data came from one student focus group interview, subjective and objective PA data, and the Fitness Education Index. All needs assessment data would diagnose what barriers are encountered in the current school settings to promote PA using the SMART Project.

The second step was to identify expected outcomes of the intervention and formulate developmentally appropriate performance objectives (i.e., sub-behaviors within the desired behavior) at each level for behavior change. For example, to achieve expected outcomes, an individual would need to take many actions, including moving more in the classroom, monitoring PA level using technology, and avoiding prolonged periods of sedentary behaviors. As mentioned earlier, in the aims of this project, the desired outcomes were to increase PA levels in and out of school and the surrounding community. However, it was recognized that PA could arise in several different contexts (e.g., during classroom activity, physical education lessons, and before/after school) and have many environmental influences (e.g., personal, interpersonal, and environmental). Therefore, researchers conducted adult focus group interviews to identify each context’s expected outcomes and each level of influence aligned with CSPAP components. From these data, we created the conceptual model for Project SMART.

The third step was designing a program aligned with theoretical methods and practical strategies. A theoretical approach is a general process intended to change behavioral determinants (e.g., self-efficacy). Practical application refers to the way such a theoretical method is put into practice in the actual classroom. Here, researchers chose the Self-Determination Theory (SDT) constructs to connect Project SMART game modules to learning activities intended to increase intrinsic motivation for participation in PA. Many interventions use the SDT to improve PA levels by motivating school-aged children (Gillison et al., 2019; Vasconcellos et al., 2020). This motivational theory has been widely used in recent years to develop interventions aiming to improve student motivation in multidimensional contexts such as physical education lessons (Cheon and Reeve, 2015), classrooms (Sergis et al., 2018) and extracurricular programs (González-Cutre et al., 2018). However, few studies undertook an intervention with students from minority groups, and the program was disconnected from their needs and interests (Fahlman et al., 2015). Project SMART engaged children, teachers, parents, and staff from socioeconomically disadvantaged neighborhoods as co-researchers in the community participatory research. The SDT (Ryan and Deci, 2000) provides a theoretical framework that plays an essential role in motivating students to have a physically active lifestyle. Our study applies this framework in the elementary school settings to help readers utilize the theory in real-world scenarios with students. For example, relatedness is addressed by encouraging students to ask questions, engaging in authentic co-design and gameplay, and continuously providing student with peer and role model support. Challenging yet achievable goals have been established, and students are encouraged to self-monitor to increase competence regarding PA and game participation (Teixeira et al., 2020). Also, the co-researchers examined the self-reported PA data across 20 physical education lessons. A time series analysis was used to understand the nature of the observations and to explain the PA monitoring. Findings from the ARIMA modeling analysis (Dixon, 1992) of the repeated measure PA data collection are presented.

Refining the program structure and identifying outcomes for program implementation were part of the fourth step. This step includes defining and describing the learning activities of the intervention and program protocol. Together the inputs, activities, and short-, intermediate- and long-term outcomes for Project SMART were identified and conceptualized in the logic model.

To establish a participatory planning group, researchers engaged the community, springboarding off the WCWH initiative and establishing a research team of community members, teachers, and university researchers to serve as co-investigators. This section presents an overview of the needs assessment data through four different resources.

The student focus group found the 4th grade students highly engaged and motivated by the potential of contributing to the design of an interactive, educational game. The students reported that they wanted gamification elements of competition, levels, leaderboards, and avatars, while the teacher suggested that the game be cooperative and driven by the aggregate class PA so that no student could be singled out as a non-contributor or identified as falling behind the others. Once the idea of cooperation was introduced to the students, they were agreeable to the recommendation brought forward by the teachers. As a result, the research team decided to have a cooperative but individually propelled PA game. The teacher suggested that the game had to be directly aligned with the grade-level educational learning outcomes in Science, Technology, Engineering, and Math (STEM) and those outcomes for social-emotional learning (SEL), cultural studies, and physical education. Several teachers and school administrators would later confirm, the intervention would be a non-starter if the educational standards were not included because they were under pressure to use all academic time to focus on achievement.

Objective PA data revealed that none of the 20 observed physical education lessons achieved the recommended standard of 50% of the class time or at least 23 min in MVPA [Office of Disease Prevention and Health Promotion (ODPHP), 2019]. Only three specific lessons resulted in a class average of approximately 40% or higher of time in MVPA. Lesson content significantly influenced the amount of MVPA during a class session (Health-related fitness lesson M MVPA = 12.56 min; SD = 8.00; Low organizational games lesson M MVPA = 12.32 min; SD = 1.99; Motor skills lesson M MVPA = 17.67 min, SD = 6.08; F(2, 139) = 10.09, p < 0.001). Surprisingly, the health-related fitness lessons that employed fitness stations and a circuit training format did not produce the highest MVPA. Instead, the lessons focused on motor skills, where each student had their own equipment and space to repeatedly practice resulted in the highest MVPA. Although contrary to previous research (Nader et al., 2018), the type of school or teacher was not significantly related to the amount of MVPA. This is likely because we only worked with physical education teachers who were already known to be conscientious professionals. Therefore, differences across the schools were not in the teacher’s approach but instead potentially related to contextual variables (e.g., high and low academically performing students).

Self-reported PA intensity data collected using an RFID badge can be found in Table 1. Initially, the 4th graders were only accurate approximately one-third of the time and tended to overestimate the intensity and total volume of PA (58%) but did improve once given feedback from the accelerometers. Students in the 5th grade underestimated (46%) their PA intensity and volume and instead reported if they followed the rules and had listened to the teacher. Students’ self-reported PA was related to behavioral compliance and listening to the teacher more than participating at moderate to vigorous intensity. These developmentally appropriate responses are rarely considered during the design of a PA intervention. PA intensity is related to the magnitude of health benefits. As such, it should be introduced as an educational outcome of the intervention (e.g., jumping rope made me sweat, but walking on the balance beam did not). A Chi-Squared analysis X² (2, 90) = 12.83, p = 0.002 revealed that 5th graders were significantly more accurate than 4th graders, as would be developmentally expected. There was no significant difference between time one and time three concerning the accuracy of reporting of PA intensity for all students. However, a paired sample t-test of just 5th graders revealed a significant improvement in accuracy for time one and time three, p < 0.05, suggesting a potential learning effect. For example, between time one and time two assessments of PA reporting accuracy, we added a traffic light poster that included descriptions defining each color (Figure 3). Further, the teacher stood nearby during the check-out, so the students had independent time to carefully consider their selection. Finally, as a co-investigator, the teacher suggested offering an SEL lesson focused on honesty in PA settings.

The Fitness Education Index analysis was based on the sum score from the 21 teachers representing the degree of readiness (M = 40.15; SD = 9.58). The index revealed that eight schools were already engaged in the delivery of CSPAP components. However, with 13 schools still a work in progress on implementing PA opportunities across the school day (e.g., before/after school, recess, PA in the classroom, daily physical education), the research team needed to plan on providing more support for the schools for successful implementation of a PA intervention. The resulting decision was to provide teacher professional development, familiarization to Project SMART, and have the students run the game while the teachers concentrated on delivering content. Further, to engage multiple components and levels simultaneously, the researchers mapped the CSPAP component, performance objectives, SDT constructs to help the school personnel identify who was best positioned to integrate that portion of the intervention (Table 2).

The current intervention’s primary outcome was to increase the student’s PA level during the school day. Here, the researchers formulated the expected outcomes within CSPAP components and identified the performance objectives to develop a logic model. As the first process, a focus group interview was conducted with the community and school stakeholders.

The adult focus groups resulted in the formulation of three themes. First, the researchers learned that the community, made up of four neighborhoods, was unincorporated and relied heavily on schools for governance and guidance. “We trust our teachers and the nurses at the clinic,” stated one parent (which was later corroborated across multiple data sources). This is a common problem in suburban and rural communities with no main street or large chain stores as its centerpiece (e.g., Walmart). The second theme that emerged before, during, and beyond the pandemic was the lack of reliable high-speed Internet access. The school district reported through attendance records and weekly teacher reports that in Spring 2020, one-third of total district enrollment did not have access to the Internet, and therefore thousands of school-aged children were not able to participate in remote learning activities. Of particular concern were families with children in grades PreK-2, who did not initially receive a school district device (i.e., Chromebook).

Further, as reported in a five-question survey of 76 heads of households in this community, families only had one device, typically a cell phone, which inhibited them from accessing health information and resources, like time and locations for food bank distributions. Since reporting these findings, mobile hotspots and Chromebooks have been deployed to all families with at least one child registered with the school district. The last theme was a concern about the lack of PA and obesity among children. In general, there were misconceptions about how much PA children get during a school day when on campus. When asked how much PA their children were participating in, the consensus was that the schools took care of that on the weekdays and that the family was responsible for engagement on the weekends. As a result, families did not have access to PA opportunities or information about being active as a family.

Based on the themes, the research team decided that the expected outcome would be to “increase PA” within the CSPAP framework during school hours. It was determined that focusing on PA within the home and school environment at the onset of the intervention would be overwhelming. It is currently recommended that children participate in 60 min or more of MVPA daily (Pate and O’Neill, 2008); however, it was recognized that it was an unrealistic goal for physical education classes alone. Therefore, it was decided that each CSPAP component would play an independent but valuable role in the intervention. For example, a school should provide at least 20 min of recess per day, in addition to a physical education class that allows participating in free-time PA and practice skills learned in physical education lessons. Also, recess is typically delivered in an outdoor, unstructured environment, which may favor engagement of at least 50% of recess time in MVPA (Stratton and Mullan, 2005; Tran et al., 2013). The goals were relative to the individual and the class—overall, PA should significantly increase from baseline levels.

To meet our objectives, the students need to act on intrinsic motivation to help them consider how and when to move more and sit less (Gardner and Lally, 2013). Therefore, the expected outcomes focused on increasing the rate of PA participation required meeting the three psychological needs of students’ motivation (i.e., autonomy, relatedness, and competence) within CSPAP components. Next, performance objectives for each of the program outcomes should be specified. This step requires what needs to happen to influence the expected outcomes (Table 3). Finally, the team aligned the performance objectives in the theoretical determinants of motivation with each CSPAP component.

A review of the literature confirmed that the SDT was likely the most appropriate for Project SMART. This intervention requires motivation, and because the SDT is focused on that construct and the principles of three innate psychological needs for autonomy, competence and relatedness (Ryan and Deci, 2000), it was selected. Autonomy refers to behavior that is self-endorsed that people agree with and find congruent within themselves. Students have autonomous motivation when they are interested and enjoy the activity. The autonomous experience is a full set of volition, willingness and choice about what students are doing at the moment. Relatedness is feeling cared for and connected to others. It has to do with a sense of belonging and a feeling that one’s matters to others. Relatedness is enhanced not just by people treating a person warmly, but also by one’s giving to them, and one can matter in their lives that is part of what helps us feel connected. So, it is not one-way, but it has to do with bidirectional connections. Competence is essential to physical wellness to feel useful in people’s environment to have some sense of mastery of essential things to people. In contrast, when these psychological needs are not fulfilled, people regulate their behavior based on controlled reasons (Karagiannidis et al., 2015).

Self-Determination Theory also postulates that human behavior in any context can be intrinsically motivated, extrinsically motivated, or amotivated. Intrinsic motivation is evident when individuals freely engage in activities they find interesting and enjoyable and offer the opportunity for learning or task accomplishment (Pelletier et al., 1995). In contrast, extrinsic motivation is apparent when individuals perform an activity because they value their associated outcomes, such as public praise and extrinsic rewards, more than the activity itself. A lack of motivation refers to the shortage of intrinsic and extrinsic motivation and represents a complete lack of self-determination and volition concerning the target behavior (Deci and Ryan, 1985; Ryan and Deci, 2000). For example, an unmotivated student may feel that physical education does not serve any purpose and may exhibit boredom, low attendance, or passive participation in the lessons. Therefore school-based PA interventions should consider how to stimulate intrinsic motivation rather than unmotivated students surrounding the school. Supporting SDT with three psychological needs is strongly correlated with autonomous student motivation in PA environments (Vasconcellos et al., 2020). Project SMART applies motivation and behavior change techniques (Teixeira et al., 2020), encouraging students to explore and share their perspectives, provide meaningful rationales to be physically active, and urge students to self-initiate PA and gameplay to increase intrinsic motivation (Table 3).

Growing evidence supports the implications of SDT for health behavior change. For example, SDT-based research has shown that more self-determined regulations can predict PA engagement adherence (González-Cutre et al., 2018). In addition, in educational settings, gamification and gaming elements can help students change behavior by supporting the basic psychological needs for autonomy, competence and relatedness (Kim and Castelli, 2021). In the previous step, the performance objectives were created seeking answers to such questions as, “What PA should I choose to do on my own time?” and “What activity do I like and feel comfortable playing?” These questions target the SDT determinants of autonomy, competence, and relatedness. After identifying the performance objectives and determinants of behavior change, the research team developed the logic model–what students have to learn to accomplish those objectives through change objectives. To guide the research team through this step of the process, members created a conceptual highlighting of the causal pathway between gameplay and increased PA and increased academic skills mediated by self-determination (Figure 4).

The research team tested how Project SMART gameplay can be applied in an SDT-based framework as an example of this conceptual model’s practical application. The participants were asked to self-report PA intensity when they participated in a physical education class. The students could report data from their class and were encouraged to report PA beyond that experience (e.g., recess, walking home from school, PA in the home environment) as part of the gameplay. Because this request produced a set of 20 repeated observations or the same variable over 10-weeks, SPSS v.26 was used to conduct a time series of measures to forecast stability of responses and potential regression toward the mean through Autoregressive Integrated Moving Average (ARIMA modeling: Dixon, 1992; Cromwell et al., 1994). No single reported score was more than two standard deviations away from the mean, suggesting no outlying scores were included in the final analysis. In Figure 5, when the value is closer to one (d = 1), it means the series is not stationary, which is reflected in the data across the first 6 weeks, but from that point forward, you see a d = 0, which means that the series is stationary and that children have increased and are currently maintaining their PA intensity to report participation as “active” or “very active” on these days. Autocorrelations were calculated using the time series observations and observations in the previous step, called lags. The analysis revealed no significant serial correlations in 15 of the 20 observations, suggesting that self-reported PA is a stable, reliable construct predicting intensity. Further, these data indicate that after PA increased from baseline, they were maintained over 10 weeks following, regardless of physical education content or the number of opportunities at school.

Given these results and the application of SDT, a team of teachers, university students, and faculty began building the learning modules. Aggregate class PA unlocked learning activities associated with SDT constructs and STEM and SEL grade-level outcomes. The team attempted to increase perceived competence in STEM learning that includes moving. It was hypothesized that increased perceived competence would be a transferable academic achievement (Ntoumanis, 2001).

Each psychological need is relevant from the perspective of SDT and was integrated into each learning module. Specifically, autonomy was targeted through activities that students found interesting and enjoyable. The modules gave students a choice in what actions to take and when—virtually traveling across the country gave students ownership of their learning. The student-centered and strength-based approach brought a new level of relatedness to healthy decision-making (e.g., Should we stop and complete the module that will provide us with water for our journey?). Many of the learning activities are scenarios and web-based applications to increase relatedness. Competence is enhanced through progress bars reflecting achievement of goals, completion of modules and rewarding of badges. The opportunities to practice healthy decision-making within a supportive environment have merit. Unconventionally, on their journey, 5th graders will encounter an outbreak of malaria in Jamestown, Virginia, while 4th graders will compare census data from 1850 and 2010 to explain the environmental impacts of a growing population. The contextual and pedagogical approaches of addressing all STEM disciplines simultaneously include using class PA and location data, assessing for learning, alignment with learning state and national standards, and integrating historical, cultural, and geographic facts, making this intervention both novel and timely.

During this step, Project SMART intervention pilot testing was performed in selected elementary schools. Educational game elements structure the Project SMART to optimize the motivation to increase the PA level for children. Previous research has shown that games have a notable motivational factor; they utilize several mechanisms to encourage students to engage with them, often without any reward, just for the joy of playing (Cronk, 2012; Beça et al., 2020). This is also aligned with intrinsic motivation. Participation in the game teaches and reinforces knowledge and develops essential skills such as problem-solving, collaboration, and communication (Hamari et al., 2014; Dicheva et al., 2015). Recent studies have adopted game elements such as exergaming and virtual PA using technologies to promote children’s PA. The emerging game element using technologies is called gamification. Gamification has been used to support user engagement and enhance positive service use patterns, such as increasing user activity, social interaction, or quality and productivity of action (Hamari, 2013). The process of gamification is an innovative pedagogical approach employing game design elements in non-game contexts is a relatively new and rapidly growing field (Hamari et al., 2014; Chou, 2019). This approach has been integrated into an educational context to address social behavior and student motivation problems. Therefore, in Project SMART the researchers combined both gamification and SDT framework to build a motivating educational module system to increase the PA level.