Article
European Physical Education Review
2021, Vol. 27(3) 574–594
Principals’ involvement ª The Author(s) 2020
Article reuse guidelines:
in comprehensive school sagepub.com/journals-permissionsDOI: 10.1177/1356336X20976687
journals.sagepub.com/home/epe
physical activity programmes:
A social-ecological perspective
Karie Orendorff Panayiotis Doutis
Montana State University, USA University of South Carolina, USA
Collin A Webster Brian Dauenhauer
University of South Carolina, USA University of Northern Colorado, USA
Diana Mı̂ndrilă David F Stodden
University of West Georgia, USA University of South Carolina, USA
Kathleen MW Cunningham
University of South Carolina, USA
Abstract
To support school physical education, United States (US) national organizations in medicine, health,
and education recommend that schools adopt comprehensive school physical activity programmes
(CSPAPs). An important factor in successful CSPAP implementation is the involvement of the school
principal. The purpose of this study was to examine relationships between principals’ self-reported
CSPAP involvement and their social-ecological-based perceptions about CSPAPs. We analysed survey
responses from a national sample of principals in the US who indicated their school had a CSPAP
(n ¼ 198). Exploratory structural equation modelling supported a single-factor solution for CSPAP
involvement (12 items) and three social-ecological factors, which we labelled ‘intrapersonal’ (nine
items mainly focused on expected outcomes of CSPAPs), ‘interpersonal’ (five items focused on par-
ents’ and teachers’ supportof the CSPAP), and ‘environmental’ (12 items focused on the support of the
overall school environment, the community surrounding the school, and public policy for the CSPAP).
Path analysis showed involvement factor scores are predicted by intrapersonal factor scores, which, in
turn, are predicted by interpersonal and environmental factor scores. The results of this study are
helpful in identifying targets for future research and practice related to CSPAP implementation and can
inform efforts to prepare and support principals with respect to their CSPAP involvement.
Corresponding author:
Collin A Webster, Blatt Physical Education Center, 1300 Wheat Street, University of South Carolina, Columbia, SC 29208, USA.
Email: websterc@mailbox.sc.edu
Orendorff et al. 575
Keywords
Head teacher, school leader, administrator, whole-of-school approach, child, adolescent, survey,
exploratory structural equation modelling, path analysis
Introduction
In the United States (US), the Physical Activity (PA) Guidelines for Americans (US Department of
Health and Human Services (USDHHS), 2018) recommend that children and adolescents ages 6–
17 should accumulate at least 60 minutes of mostly moderate-to-vigorous PA each day, as this can
lead to health benefits such as healthy bones and muscles, improved muscular strength and
endurance, reduced development of chronic disease risk factors, improved self-esteem, and
reduced stress and anxiety (Physical Activity Guidelines Advisory Committee, 2018). Unfortu-
nately, approximately three quarters of children and adolescents in the US are not meeting this
benchmark for daily PA participation (National Physical Activity Plan Alliance, 2018). Moreover,
children spend 80–93% of their waking hours in sedentary time (Turner et al., 2014) and it is
estimated that the majority of school-aged youth are not meeting national content standards in
physical education (Hastie, 2017). National organizations (e.g. Centers for Disease Control and
Prevention (CDC), 2011; Institute of Medicine (IOM), 2013 (now named the National Academy of
Medicine); Society of Health and Physical Educators (SHAPE) America, 2015) suggest that
schools should be a key focus of intervention efforts when attempting to increase and promote
youth PA engagement. Schools have access to nearly all youth in the US and school-based PA not
only can improve students’ health but also enhance their academic performance (IOM, 2013).
Furthermore, education settings can play a major role in public health initiatives related to PA,
given that schools directly reach approximately 25% of the US population (National Physical
Activity Plan Alliance, 2016).
Comprehensive school physical activity programmes
In 2008, the National Association for Sport and Physical Education (NASPE) (formerly an
association of the American Alliance for Health, Physical Education, Recreation and Dance
(AAHPERD), and now encompassed within SHAPE America) published a position statement
(NASPE, 2008), which was updated in 2015 (SHAPE America, 2015) called ‘Comprehensive
School Physical Activity Programs’ (CSPAPs). A CSPAP is described as a multicomponent
approach to enhancing physical education and promoting increased PA among school commu-
nities. The components of a CSPAP can include (a) physical education, (b) PA before and after
school, (c) PA during school, (d) staff involvement, and (e) family and community engagement
(SHAPE America, 2015). However, given limited descriptive research on CSPAPs, recent liter-
ature reflects a broader conceptualization of a CSPAP, which focuses less on the total number of
programme components and more on whether a school, through multiple possible approaches,
provides all students with sufficient opportunities to meet the programme goals (Webster et al.,
2020b, 2020c, 2020d). Implementing a CSPAP is intended to provide children and adolescents
with PA support and opportunities before, during, and after school as a means to ‘[developing]
physically educated students who are physically active for the nationally recommended 60-plus
minutes each day and who develop the knowledge, skills and confidence to be physically active for
a lifetime’ (SHAPE America, 2015: 3).
576 European Physical Education Review 27(3 )
A survey conducted by AAHPERD in 2011 found that CSPAPs existed in only 16% of elementary
schools, 13% of middle schools, and 6% of high schools. In 2013, the IOM endorsed a multi-
component, or ‘whole-of-school’, approach to youth PA promotion (IOM, 2013: 30) and the CDC
partnered with SHAPE America to develop a step-by-step guide for implementing a CSPAP (CDC,
2013). However, based on a recent survey by the CDC (Brener et al., 2017), a mere 3% of secondary
schools have implemented a CSPAP. Thus, while CSPAPs are recommended to increase students’
daily PA and support the educational goals of physical education (SHAPE America, 2015), and the
CDC has adopted the CSPAP framework as the national framework for physical education and PA in
the US (CDC, 2019), the apparent continued low prevalence of CSPAPs in the US underscores the need
to investigate the factors that are associated with programme implementation.
Principals’ involvement in CSPAPs
One factor that has consistently been cited as critical in school reform efforts and in creating and
maintaining an effective school is the principal (also referred to as the head teacher, headmaster/
headmistress, school director, and other titles in various contexts) and their involvement (Dow and
Oakley, 1992; Fullan, 2016). As a figurehead of school leadership, the principal is a key driver
in the school’s mission, vision, and culture (Bryk et al., 2010; Leithwood and Jantzi, 1999).
A principal’s involvement in a school programme, for example by actively assisting in its
development and taking part in its promotion, can lead to teachers feeling they are supported and
showing increased commitment to the programme (Leithwood and Jantzi, 1999). The principal is
the main reason programmes are or are not successful in schools; they hold the most influence for
implementation of programmes and policies (Datnow and Castellano, 2001).
Carson et al. (2014) proposed that support from administrators, such as the principal, is necessary
to the successful implementation of a CSPAP. However, the concept of ‘support’ may not fully
encompass the range of ways in which a principal might be involved with a CSPAP. According to
Webster et al. (2020e: 130), ‘Staff involvement with promoting PA can be broadly conceptualized in
terms of the direct and indirect actions staff members take to increase the concomitant or future
amount of PA in which youth engage’. A recent scoping review (Webster et al., 2020a) identified
three themes in the published recommendations for school administrators’ involvement in school-
based health promotion. These themes included collaboration (e.g. using a distributive leadership
style, being involved with defining health issues, identifying solutions, and monitoring progress),
advocacy (e.g. knowing about and valuing health promotion, and educating policymakers about the
importance of school health promotion), and support (e.g. providing resources for teachers, visiting
teachers’ classrooms, and offering feedback). Yet, a descriptive research base to inform the specific
actions school principals might take to be involved with a CSPAP is lacking.
A social-ecological perspective of principals’ CSPAP involvement
In addition to the limited research available to operationally define principals’ CSPAP involve-
ment, little is known about variables that may influence principals’ involvement with CSPAPs.
Van den Berg et al. (2017) found that principals perceived PA to have academic, physical, and
social-emotional benefits but also perceived barriers to implementing PA that were related to time
limitations, school priorities, space constraints, and financial challenges. A gap that remains in the
literature, however, is research on principals’ perceptions specifically pertaining to CSPAPs. It is
important to investigate these perceptions, particularly to understand how they may be related to
Orendorff et al. 577
principals’ CSPAP involvement. In the present study, we adopted a social-ecological perspective
to investigate principals’ perceptions of CSPAPs and how these perceptions relate to principals’
CSPAP involvement.
Based on the work of Bronfenbrenner (1976, 1977), a social-ecological perspective of human
behaviour considers the influence of a wide range of variables on targeted behaviours of interest.
Social-ecological models are used to organize these variables into multiple levels of influence and
illustrate relationships among individuals, groups, and their environments (Golden et al., 2015;
McLeroy et al., 1988; Sallis et al., 2006, 2008; Stokols, 1996). Numerous examples of social-
ecological models exist in theory and research specific to school-based PA promotion (e.g. Carson
et al., 2014; Langille and Rodgers, 2010; Webster and Suzuki, 2014; Webster et al., 2013). In
general, these models encompass five levels of influence: (a) intrapersonal, (b) interpersonal, (c)
organizational, (d) community, and (e) public policy. Intrapersonal variables may include beliefs
(e.g. self-efficacy in implementing an educational programme, expected outcomes of imple-
menting the programme), attitudes, and other psychological attributes (e.g. motivation to imple-
ment a new programme) of the individual whose behaviour is being targeted (e.g. school
principal). Variables at the interpersonal level tend to focus on relationships between the individual
and others with whom the individual interacts (e.g. other administrators, teachers, students, par-
ents, community stakeholders). The organizational level of influence is usually comprised of
variables that reflect institutional resources (e.g. school facilities, available educational materials),
norms (e.g. cultural customs commonly followed by students and staff, a school honour code),
practices (e.g. allocated curriculum time for different school subjects, instructional approaches
used by teachers), and policies (e.g. expectations for student conduct, whether or not teachers are
allowed to withhold recess as a punishment). At the community level, variables often centre on
factors, such as the support of other organizations (e.g. a facility joint-use agreement between the
school and a local YMCA), which surround the individual’s own organization. Public policy
variables may include the presence of any local, state, or federal policies that are pertinent to the
targeted behaviour (e.g. a state policy that requires schools to provide students with a certain
number of minutes of PA each week). In the context of a CSPAP, a social-ecological perspective
can help to identify key variables that should be the focus of interventions and professional
learning directed towards influencing the PA promotion efforts of programme implementors.
Previous social-ecological research in the area of school-based PA promotion has not con-
ceptualized or investigated the intrapersonal, interpersonal, organizational, community, or public
policy variables that may influence school principals’ CSPAP involvement.
Purpose of the study
Overall, although considerable research exists on the staff involvement component of the CSPAP
framework, particularly in regard to classroom teachers and afterschool programme staff, there is
currently an absence of research on principals’ involvement in CSPAPs (Webster et al., 2020e).
The purpose of this study was to examine relationships between principals’ CSPAP involvement
and their perceptions about CSPAPs from a social-ecological perspective. In accordance with
social-ecological theory and research related to school PA promotion (Carson et al., 2014; Langille
and Rodgers, 2010; Webster et al., 2013; Webster and Suzuki, 2014), we hypothesized that
intrapersonal variables would have the strongest and most direct association with the targeted
behaviour (i.e. principals’ CSPAP involvement), while variables at each subsequent level (inter-
personal, organizational, community, and public policy, respectively) would have an indirect and
578 European Physical Education Review 27(3 )
increasingly weaker association. The results of this study will help to advance research focused on
the role of principals in CSPAP implementation and may inform future directions in CSPAP
planning and professional development for school administrators.
Methods
Instrumentation
We developed an online survey to investigate principals’ involvement in CSPAPs. The survey
encompassed a range of variables hypothesized to be associated with principals’ CSPAP invol-
vement from two distinct theoretical perspectives (social ecology theory and social learning
theory). The present investigation, which focuses on social-ecological variables and their
relationship to principals’ CSPAP involvement, is the first study using data from the survey. The
final version of the survey, specifically the parts of the survey used in the present study and as
administered to the main study sample, is described in this section. The development of the survey
is described in the Procedures section that follows. All phases of survey development and
administration were facilitated via Survey Monkey.
The survey was organized into seven sections: (a) informed consent, (b) introduction, (c)
CSPAP implementation, (d) CSPAP involvement, (e) social-ecological factors, (f) professional/
context biography, and (g) participant demographics. The first page of the survey was an informed
consent form. Subsequently, in the introduction section, participants were provided with an
overview of the CSPAP framework and the purpose of the survey. The CSPAP implementation
section of the survey listed possible CSPAP components (e.g. physical education, PA during
school, PA before and after school) with examples and included a single dichotomous response
question (‘yes/no’) asking participants if their school has a CSPAP. A CSPAP was conceptualized
as follows, based on the recent work by Webster et al. (2020a, 2020b, 2020c):
[your school provides] OPPORTUNITIES, through any combination or variety of CSPAP components –
for all students at your school to: (a) receive standards-based physical education experiences designed
to prepare individuals for a lifetime of participation in physical activity, and (b) meet the national
guideline for school-aged youth to accumulate at least 60 minutes of mostly moderate-to-vigorous
physical activity each day (including lunch time activities, and before and after school activities).
A logic mechanism was built into the survey so that ‘yes’ respondents proceeded to the
subsequent two sections (involvement and social-ecological factors) with items written in the
past/present tense (e.g. ‘I am involved with supporting my school staff in their efforts to promote
physical activity’). ‘No’ respondents proceeded to two identical sections except that items were
written in the future tense (e.g. ‘I would be involved with supporting my school staff in their efforts
to promote physical activity’). The final two sections on professional context/biography and
participant demographics focused on the PA promotion strategies, within each CSPAP component,
currently used at participants’ schools, as well as participants’ school context (e.g. geographic
location, total student enrolment), educational background (e.g. highest degree earned, licensure
obtained), professional experience (e.g. years serving as a principal, years serving as a teacher),
age, gender, and race/ethnicity. Items focusing on PA promotion strategies within each CSPAP
component were based on recommended strategies from SHAPE America (2015) and the CDC
(2019).
Orendorff et al. 579
In all, 80 items from the final survey were used for this study. This item count is based on having
counted equivalent items for ‘yes’ and ‘no’ respondents as single items. One item was included to
assess CSPAP implementation; 12 items were included to assess the dependent variable (principal
involvement in a CSPAP); 52 items were included to assess social-ecological variables at five
levels of influence, including intrapersonal (29 items), interpersonal (five items), organizational
(nine items), community (five items), and public policy (four items); 12 items were included to
assess professional context/biography; and three items were included to assess participant
demographics.
Procedures
Phase 1: item construction, scale development, and participant instructions. To construct an initial set of
items for the survey, we completed a comprehensive literature search to find published records
(e.g. peer-reviewed articles, books, doctoral dissertations, public policies) in the areas of CSPAP
(e.g. Carson et al., 2014; Rink et al., 2010; SHAPE America, 2015), principals’ perceptions of
health-related school programming (e.g. physical education, recess, fitness, wellness) (e.g. Davis,
2012; Greaney et al., 2007; Morrison, 2006; Van den Berg et al., 2017; Weiler et al., 2003), and
social-ecological theory and research (e.g. Golden et al., 2015; McLeroy et al., 1988; Sallis et al.,
2008; Webster and Suzuki, 2014). We developed a six-point Likert-type response scale (see the
Instrumentation section) for the items based on recommendations for scale development (Fink,
2003; Johnson and Morgan, 2016; Krosnick and Fabrigar, 1997) and the treatment of categorical
variables (Finney and DiStefano, 2006). Based on recent literature, including research on physical
education teachers’ perceptions related to CSPAPs (Webster et al., 2020b, 2020c) and conceptual
work intended to advance CSPAP research and practice (Webster et al., 2020d), we incorporated
instructions for participants along with a definition and examples of a CSPAP (see Instrumentation
section).
Phase 2: pilot testing. Before data collection started, we obtained approval from the first author’s
university institutional review board (IRB) to conduct the study. There were two rounds of pilot
testing. The purpose of the first round of pilot testing was to enhance the content validity of the
survey and obtain feedback about the length, formatting, and clarity of the survey. The purpose of
the second round of pilot testing was to obtain feedback about the survey from a convenience
sample of principals and to conduct a preliminary assessment of the survey’s psychometric
properties.
For the first round, we emailed the survey to CSPAP authors (n ¼ 41), authors who have
published research using social-ecological theory (n ¼ 51), and university faculty in the area of
educational leadership (n ¼ 89). We identified CSPAP authors from an edited book about CSPAP
research and practice (Carson and Webster, 2020). For each section of the survey, we instructed
participants to rate the appropriateness of the items (completely inappropriate, mostly inap-
propriate, somewhat inappropriate, mostly appropriate, and completely appropriate) and provide
feedback about the content on each page. A total of 24 CSPAP authors (59% response rate)
provided responses. Using their ratings and comments, 24 items were removed (15 items intended
to measure intrapersonal variables, seven items intended to measure interpersonal variables, and
two items intended to measure organizational variables). Additionally, several minor revisions
were made: the term ‘CSPAP’ was replaced with ‘physical activity opportunities’ and ‘physical
activity programming’ in two of the items on the involvement scale, negatively worded items
580 European Physical Education Review 27(3 )
(e.g. ‘A CSPAP would lead to a chaotic school environment’) were either deleted or revised as
positively worded items, and items that included the term ‘parents’ were changed to read ‘parents/
guardians’.
To identify authors in the area of social-ecological theory, we located published articles online
(e.g. Golden et al., 2015; McLeroy et al., 1988; Sallis et al., 2008). We found email addresses for
51 authors and emailed the authors with instructions to match each of the social-ecological items
on the survey to its respective level of influence (intrapersonal, interpersonal, community, orga-
nizational, or public policy). We also asked the authors to provide feedback on the quality and
appropriateness of specific items and/or the items in general. A total of six authors provided
responses (12% response rate). Based on their feedback, items were found to match their intended
social-ecological levels, so no changes were made to the survey.
We compiled a list of 89 university faculty, representing 58 US institutions, who work in the
area of educational leadership and policy. These individuals encompassed all levels of academic
rank/seniority from assistant professor to retired professor, and many of the faculty had served as
former school- and district-level administrators. For each section of the survey, participants were
instructed to rate the appropriateness of the items (completely inappropriate, mostly inappropriate,
somewhat inappropriate, mostly appropriate, and completely appropriate), considering the roles
and responsibilities of a school principal, and provide feedback about the content on each page. A
total of 16 educational leadership faculty provided responses (18% response rate), which informed
several changes to the survey. Specifically, two items were removed (‘Others in my school
environment notice/would notice the impact of promoting physical activity’ and ‘I have a good
relationship with the teachers at my school’). Educational leadership faculty believed the princi-
pals would have low efficacy in judging the first item, and felt the second item was ‘leading’ and
that principals would not admit to having a bad relationship with their staff. Also, the introduction
to the survey was revised to use less academic jargon.
After making revisions to the survey based on the first round of pilot testing, we sent the survey
to a convenience sample of principals (n ¼ 42) from elementary, middle, and high schools and
asked them to complete the survey, as well as to provide general feedback. Twenty-one principals
responded (16 indicated they had a CSPAP at their school) for a 50% response rate. Using the data
from this second round of pilot testing, we employed Bayesian exploratory factor analysis (EFA) to
individually examine each survey scale. This served as a preliminary statistical analysis to identify
any additional items for potential removal from the final survey to be used with the main study
sample. Previous research showed that EFA can provide accurate results with samples as small as
20 or 10 observations when the data are well conditioned (i.e. high œ, low f, high p) (de Winter
et al., 2009; Preacher and MacCallum, 2002). Further, the Bayesian estimation method does not
rely on a large-sample theory, and therefore can provide accurate results with small samples, such
as our convenience sample of principals (Heerwegh, 2014; Muthén and Asparouhov, 2012). Items
with non-significant factor loadings were only removed for subsequent data collection in the main
study if the loadings were below 0.320 or if removing the items created convergence problems. In
all, five items were removed based on the results of the EFA. Additionally, in accordance with
respondents’ feedback about the survey, we revised the CSPAP implementation section of the
survey for brevity.
We also asked the respondents from the convenience sample to retake the survey one week after
their survey closed, to assess test–retest reliability. A one-week time limit was provided for them to
retake the survey. There were 13 respondents (11 indicated they had a CSPAP at their school) for a
Orendorff et al. 581
62% response rate. Thirteen items were removed from the survey due to having a Kappa value
lower than 0.2.
Phase 3: main study. The administration of our survey to the main study sample occurred in the
spring of 2019. Using information available on a federal website listing all public schools in the
US, we applied stratified random sampling to select 60 schools (20 elementary, 20 middle/junior
high, and 20 high) from each state (total of 3000 schools). Next, we visited the schools’ websites to
find email addresses (where available) for principals and compiled a list of 2941 email addresses.
With these addresses, we sent a blanket email briefly describing the study and inviting recipients to
complete the survey. The link to the survey was included in the email. Respondents for the main
study were 291 principals. For the purposes of this study, we only analysed data from participants
who indicated their school has a CSPAP (n ¼ 198).
Data analysis for the main study
Before conducting statistical analyses, we screened survey variables to examine the prevalence of
‘don’t know’ responses as well as the distribution of missing values. Missing responses represented
approximately 1.9% of the data. The ‘don’t know’ response option was included to prevent item
non-response, forced responses, or guessing, but was not included in statistical analysis because it
indicates insufficient information to provide an informed opinion. To avoid losing data, we
recorded ‘don’t know’ responses as missing values. Cases with more than 15% invalid responses
were listwise deleted. The resulting proportion of values that were missing or recoded as missing
ranged between 0.0% and 2.7% per item and represented approximately 2.3% of the data. Little’s
Missing Completely at Random (MCAR) test showed that these values were distributed com-
pletely at random (2(1825) ¼ 1800.867, p ¼ .652); therefore, we imputed them using the
expectation-maximization algorithm. The resulting sample included 198 respondents. We com-
puted the mean and standard deviation for all survey items to examine the distribution of survey
responses and identify the items with the highest ratings.
Exploratory structural equation modelling. Although social-ecological research exists in the area of
school PA promotion (e.g. Langille and Rodgers, 2010; Webster and Suzuki, 2014; Webster et al.,
2013), researchers in previous studies have not examined principals’ CSPAP involvement from a
social-ecological perspective. Since we developed the survey for our research and had not pre-
viously examined the factor structure of the final survey, the initial step of our investigation with
the main study sample was to identify the latent variables underlying the data. We used exploratory
structural equation modelling (ESEM) for this purpose because it allows the estimation of an
exploratory factor model with rotations and estimation of cross-loadings, thus yielding a more
realistic representation of the data and reducing estimation bias, while also allowing the compu-
tation of goodness-of-fit indices (Marsh et al., 2014; Morin and Maiano, 2011; Morin et al., 2013).
A group of 51 observed indicators was used to estimate latent variables. Our study met the
recommendation of 3:1–6:1 observations per variable (Cattell, 1978; de Winter et al., 2009).
Studies have shown that sample sizes as low as 50 or even 20 may be suitable for factor analysis
when the factor loadings are high, the number of factors is low, and the number of observed
variables is high (de Winter et al., 2009; MacCallum et al., 2001; Marsh et al., 1998). We used the
mean and variance adjusted weighted least squares (WLSMV) estimation method and geomin
rotation with the Mplus 8.2 software. This estimation method is recommended for small sample
582 European Physical Education Review 27(3 )
sizes and data that are ordinal or non-normally distributed (Finney and DiStefano, 2006). The
geomin rotation is an oblique procedure; oblique rotation methods are employed when factors are
expected to correlate (Browne, 2001).
The optimal number of factors was determined after examining the scree plot, the number of
eigenvalues larger than one, the interpretability of the factor structure, and goodness-of-fit indices.
The indices used to assess model fit were: (a) 2 and corresponding p-value, (b) 2/df, (c) the root
mean square error of approximation (RMSEA) and 90% confidence interval (CI), (d) the com-
parative fit index (CFI), (e) the Tucker–Lewis index (TLI), and (f) the weighted root mean residual
(WRMR). We sequentially removed cross-loading items and items with loadings lower than .320
(Costello and Osborne, 2005). When an optimal solution was identified, we estimated the internal
consistency of each factor using the Cronbach’s alpha coefficient.
Path analysis. Following the identification of the factors underlying the data, we computed factor
scores, which show the location of each individual on the identified factors (DiStefano et al.,
2009). Relationships between factor scores were then examined using path analysis. Relationships
in the path model were based on theoretical considerations and results from previous studies
(Langille and Rodgers, 2010; Webster and Suzuki, 2014; Webster et al., 2013), as well as on
specified scores on social-ecological factors as predictors of factor scores measuring principals’
CSPAP involvement. The indices used to assess model fit were: (a) 2 and corresponding p-value;
(b) 2/df; (c) the RMSEA and 90% CI; (d) the CFI; (e) the TLI; and (f) the standardized root mean
square residual (SRMR).
Results of the main study
Participant demographic and professional context/biographical information
The demographic and professional context/biographical information for participants in the main
study is presented in Table 1. Table 1 displays participants’ self-reported age band, gender
identification, race, education level, years of experience working as a principal, and grade level, as
well as participants’ self-reported school context information including total student enrolment,
number of students participating in free or reduced cost school lunch (a proxy for socioeconomic
status), school setting, and the region of the US where they are located. Overall, the percentages in
each of these categories were similar for principals who indicated their school has a CSPAP
compared to principals who indicated their school does not have a CSPAP. Table 2 displays the PA
promotion strategies, by CSPAP component, that the principals identified as being used at their
schools. These data clearly demonstrate that, based on principals’ self-reports, schools with a
CSPAP are using more PA promotion strategies across all CSPAP components than schools
without a CSPAP.
Factor structure
Initial exploratory analyses yielded six eigenvalues larger than one, while the scree plot indicated
solutions with 3–5 factors as optimal. We estimated models with 3–5 factors. Although the five-
factor solution had the best fit to the data, the fifth factor included only three items and all of them
were cross-loading. Based on goodness-of-fit indices (2 (557) ¼ 123.452, p < .001, 2/df ¼ 0.221,
RMSEA (90% CI) ¼ 0.079 (0.073–0.085), CFI ¼ 0.947, TLI ¼ 0.933, WRMR ¼ 0.950) and the
interpretability of the factors, we considered the four-factor solution optimal (Table 3). The 2 test
Orendorff et al. 583
Table 1. Participants’ self-reported demographics and school contexts for the main study.
Participants who Participants who
responded ‘no’ to their responded ‘yes’ to their
school having a CSPAP school having a CSPAP
(n ¼ 36) (n ¼ 147)
Age band 31–35 8% 1%
36–40 25% 12%
41–45 22% 20%
46–50 11% 23%
51–55 11% 14%
56–60 17% 18%
60þ 6% 12%
Gender identification Male 56% 58%
Female 44% 42%
Race Asian 0% 2%
African American or Black 8% 12%
Hispanic or Latino 6% 3%
White 86% 82%
No response 0% 1%
Education Masters 30% 25%
Masters þ30 42% 44%
Ed.D. 17% 24%
Ph.D. 11% 7%
Years of experience 0–5 31% 28%
working as a principal 6–10 31% 21%
11–15 8% 28%
16–20 11% 11%
Over 20 19% 12%
Grade level Elementary 19% 32%
Jr. high / middle school 42% 29%
High school 19% 26%
Grades K–8 6% 5%
Grades 7–12 14% 8%
Total student enrolment  400 36% 27%
401–600 31% 27%
601–800 19% 24%
801–1000 8% 8%
1001–1200 3% 4%
 1201 3% 10%
Percentage of students 0–20% 3% 16%
participating in free or 20–40% 14% 18%
reduced cost lunch 40–60% 36% 24%
60–80% 14% 15%
80–100% 33% 27%
School setting Metro urban 28% 17%
Suburban 25% 47%
(continued)
584 European Physical Education Review 27(3 )
Table 1. (continued)
Participants who Participants who
responded ‘no’ to their responded ‘yes’ to their
school having a CSPAP school having a CSPAP
(n ¼ 36) (n ¼ 147)
Rural 47% 36%
Region of US where Far West 3% 10%
employed Great Lakes 3% 12%
Midsouth 17% 19%
Midwest 13% 11%
Mountain West 25% 11%
New England 3% 9%
Northeast 6% 6%
Northwest 7% 3%
South Central 6% 6%
Southeast 17% 13%
Note: the sample sizes represented in this table are smaller than the total number of respondents because not all
respondents provided demographic/school context information.
CSPAP: comprehensive school physical activity programme.
is an overall measure of model fit; a non-significant 2 statistic indicates good fit (Barrett, 2007);
however, this statistic is sensitive to sample size and model size. Therefore, the 2/df is often used,
where values lower than 3 show good model fit (Finney and DiStefano, 2006). For RMSEA, values
larger than .10 show poor model fit, values between .08 and .10 indicate acceptable fit, values
between .05 and .08 indicate good fit, and values lower than .05 indicate excellent fit (Hu and
Bentler, 1999). CFI and TLI values above .95 indicate excellent model fit, while values larger than
.90 indicate good fit. WRMR values lower than 1 indicate very good fit (DiStefano et al., 2018; Yu
and Muthén, 2002). All items in the four-factor solution had statistically significant loadings,
above the cutoff of .320 (Costello and Osborne, 2005). Items with lower loadings and items that
were cross-loading were sequentially removed; the final solution retained 39 items.
Table 4 lists the items included in each factor along with their means, standard deviations, factor
loadings, standard errors, corresponding t statistics and p values. Descriptive analyses showed that
most respondents provided high ratings on the majority of the survey items. The item with the
highest average response was ‘A CSPAP enhances students’ physical development’ (M ¼ 5.31,
SD ¼ .755), followed by the item ‘A CSPAP promotes students’ social development’ (M ¼ 5.24,
SD ¼ .748). The item with the lowest average was ‘I am involved with setting performance
standards for my school’s CSPAP’ (M ¼ 3.47, SD ¼ 1.316). The strongest factor was labelled F1
(‘Involvement’) and included 13 items referring to CSPAP involvement. Item loadings ranged
between .463 and .851, and the item with the highest loading was ‘I am involved with CSPAP
planning at my school’. Cronbach’s alpha index of internal consistency for this factor was .900.
The second factor, F2 (‘Intrapersonal’), included nine items mostly referring to beliefs about the
outcomes of a CSPAP. Item loadings ranged between .578 and .914 and the item with the highest
loading was ‘A CSPAP promotes a whole-child learning approach’. Cronbach’s alpha index of
internal consistency for this factor was .898. The third factor, F3 (‘Interpersonal’), included five
items, which referred to personnel within the school community and their support of a CSPAP.
Orendorff et al. 585
Table 2. Physical activity promotion strategies by CSPAP component as perceived by respondents in the
main study.
Participants who Participants who
responded ‘no’ to responded ‘yes’ to
their school having their school having
a CSPAP (n ¼ 37) a CSPAP (n ¼ 148)
Physical education Standards-based instruction 49% 78%
Assessment of student learning 30% 55%
Opportunities to learn 30% 50%
Opportunities for moderate-to-vigorous 38% 60%
physical activity
Meaningful content 43% 42%
None 24% 4%
Physical activity Classroom-based physical activity 51% 74%
during school Recess 43% 55%
Physical activity assemblies 3% 23%
Physical activity drop-in opportunities 24% 35%
(e.g. keeping the gym open during lunch)
None 19% 7%
Don’t know 3% 3%
Physical activity Active transportation programmes/options 19% 20%
before/after Intramurals 16% 38%
school Interscholastic sports 57% 49%
Physical activity clubs 30% 52%
None 24% 11%
Don’t know 0% 3%
Staff involvement Staff wellness programming (e.g. fitness 46% 62%
programmes/events for teachers, health
screening for teachers)
Staff training for physical activity promotion 5% 18%
Administrators involved in promoting physical 19% 36%
activity
Classroom teachers involved in promoting 24% 57%
physical activity
None 32% 14%
Don’t know 3% 3%
Family and Facility joint-use agreements with outside 19% 43%
community organizations
engagement Physical activity programmes/events for families 8% 27%
Parents/guardians involved in physical activity 8% 18%
promotion
Community members/organizations 8% 37%
(e.g. universities, YMCAs, church groups)
involved in physical activity promotion
Active homework for students 3% 7%
None 57% 21%
Don’t know 8% 9%
Note: the sample sizes represented in this table are smaller than the total number of respondents because not all
respondents provided physical activity promotion strategies used at their school. Percentages for each component do not
add up to 100% because participants were asked to select all applicable response options.
CSPAP: comprehensive school physical activity programme.
586 European Physical Education Review 27(3 )
Table 3. Goodness-of-fit indices for estimated models.
Fit index Three-factor model Four-factor model Five-factor model
2 1806 123.452 995.142
df 592 557 523
p-value 0.000 0.000 0.000
2 /df 3.051 0.221 1.903
RMSEA (90% CI) 0.102 (0.096–0.107) 0.079 (0.073–0.085) 0.068 (0.061–0.074)
CFI 0.906 0.947 0.963
TLI 0.888 0.933 0.951
WRMR 1.342 0.950 0.772
RMSEA: root mean square error of approximation; CFI: comparative fit index; TLI: Tucker–Lewis index; WRMR: weighted
root mean residual.
Loadings on this factor ranged between .513 and .869 and the item with the highest loading was
‘Parents/guardians at my school are interested and willing to help our school toward our CSPAP
goals’. The internal consistency of this factor was alpha ¼ .842. The fourth factor, F4 (‘Envi-
ronmental’), included 12 items referring to organizational, community, and public policy variables
that may facilitate CSPAP implementation. Loadings on this factor ranged between .322 and .862,
and the item with the highest loading was ‘Policies in my school support physical activity’. The
internal consistency of this factor was alpha ¼ .840. As indicated in Table 5, all factor covariances
were statistically significant. The strongest relationships were F2–F1 and F2–F4, whereas the
weakest relationships were F3–F1 and F4–F1.
Path model
The estimated path model specified: (a) F4 factor scores as predictors of F3 factor scores; (b) F3
factors scores as predictors of F2 factor scores; and (c) F2 factor scores as predictors of F1 factor
scores (Figure 1). This model had an overall good fit to the data (2 (3) ¼ 7.031, p ¼ .071,
2/df ¼ 2.344, RMSEA (90% CI) ¼ 0.082 (0.000–0.164), CFI ¼ 0.982, TLI ¼ 0,964, SRMR ¼
0.040) and all path coefficients were statistically significant. The strongest relationship in the
model was F3–>F2 (estimate ¼ 0.610, S.E. ¼ 0.045, estimate/S.E. ¼ 13.693, p < .001), followed
by F2–>F1 (estimate ¼ 0.467, S.E. ¼ 0.056, estimate/S.E. ¼ 8.392, p < .001), and F4–>F3
(estimate ¼ 0.402, S.E. ¼ 0.031, estimate/S.E. ¼ 12.894, p < .001). The indirect effect of F3 and
F4 factor scores on F1 factor scores were, therefore, .285 and .114, respectively.
Discussion
Principals are a key part of the administrative support that is viewed as essential to the successful
implementation of CSPAPs (Carson et al., 2014). Using a social-ecological framework, this
study’s purpose was to examine relationships between principals’ self-reported CSPAP involve-
ment and their perceptions about CSPAPs. As this was the first study to investigate these variables,
a discussion of the descriptive data from the survey and the psychometric properties of our
measures is warranted before addressing the main results.
Orendorff et al. 587
Table 4. Descriptive statistics and factor loadings.
Estimate/ Two-tailed
Item M SD Estimate S.E. S.E. p-value
F1 (Involvement)
I am involved with CSPAP planning at my school 4.59 1.215 0.851 0.034 24.744 0.000
I am involved with evaluating my school’s CSPAP 4.60 1.208 0.751 0.043 17.279 0.000
I am involved with providing CSPAP professional 3.86 1.431 0.733 0.048 15.311 0.000
development opportunities at my school
I am involved with establishing physical activity 5.00 1.013 0.707 0.043 16.598 0.000
opportunities at my school
I am involved with setting performance standards 3.47 1.316 0.705 0.049 14.453 0.000
for my school’s CSPAP
I am involved with allocating resources for our 5.04 1.059 0.703 0.042 16.660 0.000
school’s CSPAP
I am involved with serving on my school’s CSPAP 3.83 1.581 0.645 0.050 12.846 0.000
committee or other related (e.g. school wellness)
board/task force
I am involved with building/maintaining partnerships 4.12 1.312 0.619 0.048 12.958 0.000
with community constituents to implement/
sustain our school’s CSPAP
I am involved with organizing physical activity 4.39 1.160 0.609 0.052 11.778 0.000
opportunities at my school
I am involved with staying up-to-date on best 4.03 1.160 0.602 0.057 10.499 0.000
practices for school physical activity programming
I am involved with advocating for our school’s CSPAP 4.76 1.086 0.556 0.044 12.547 0.000
I am involved with being a physically active role 4.56 1.160 0.463 0.054 8.511 0.000
model for others in my school
F2 (Intrapersonal)
A CSPAP promotes a whole-child learning approach 5.23 .772 0.914 0.021 47.608 0.000
A CSPAP facilitates student learning 5.21 .778 0.912 0.019 52.276 0.000
A CSPAP enhances students’ physical development 5.31 .755 0.874 0.028 31.184 0.000
A CSPAP promotes students’ social development 5.24 .748 0.863 0.039 21.967 0.000
A CSPAP is the ideal programme for our students to 5.21 .769 0.820 0.040 20.636 0.000
pursue a healthy lifestyle
A CSPAP improves cognitive performance 5.06 .707 0.815 0.057 14.390 0.000
A CSPAP fosters students’ attention to academics 5.02 .878 0.803 0.037 21.470 0.000
A CSPAP promotes increased school attendance 4.64 .928 0.610 0.064 9.605 0.000
A CSPAP helps our students pursue physically active 5.14 .724 0.578 0.051 11.391 0.000
lifestyles
F3 (Interpersonal)
Parents/guardians at my school are interested in and 4.08 1.094 0.869 0.032 26.803 0.000
willing to help our school towards our CSPAP goals
Classroom teachers at my school are able to 4.32 1.111 0.789 0.032 25.032 0.000
integrate physical activity into academic
instruction/learning
The majority of teachers at my school are capable of 4.36 1.002 0.711 0.037 19.361 0.000
contributing to developing and implementing a
CSPAP
(continued)
588 European Physical Education Review 27(3 )
Table 4. (continued)
Estimate/ Two-tailed
Item M SD Estimate S.E. S.E. p-value
Parents/guardians are engaged in helping our school 3.82 1.137 0.686 0.038 18.133 0.000
work towards our CSPAP goals
Parents/guardians at my school support our CSPAP 4.71 .863 0.513 0.043 11.876 0.000
F4 (Environmental)
Policies in my school support physical activity 5.14 .658 0.862 0.034 25.206 0.000
Policies in my state support school physical activity 5.01 .740 0.843 0.039 21.651 0.000
My school has sufficient funds to support a CSPAP 4.31 1.109 0.747 0.049 15.335 0.000
Adequate CSPAP resources are available for my 4.42 .946 0.706 0.043 16.542 0.000
school faculty/staff
Policies in my school support the promotion of 4.56 .974 0.704 0.043 16.297 0.000
students’ physical activity during school hours
every school day
The facilities at my school are adequate to 4.60 1.149 0.621 0.051 12.074 0.000
implement a CSPAP
Our school schedule can accommodate a CSPAP 4.83 .962 0.621 0.047 13.199 0.000
Our district superintendent supports CSPAPs 4.68 .943 0.569 0.053 10.763 0.000
Sufficient professional development for CSPAP is 3.88 1.177 0.510 0.054 9.531 0.000
available for my school staff
There are safe routes for active transportation (e.g. 4.08 1.349 0.465 0.061 7.646 0.000
walking, biking) to/from our school
Our school’s vision/mission includes the promotion 4.07 1.190 0.324 0.057 5.734 0.000
of physical activity
Policies at my school support the promotion of faculty/ 4.56 .974 0.322 0.057 5.649 0.000
staff wellness (e.g. health, fitness, physical activity)
CSPAP: comprehensive school physical activity programme.
Table 5. Factor covariances.
Estimate S.E. Est./S.E. Two-tailed p-value
F2–F1 0.491 0.049 10.014 0.000
F2–F4 0.418 0.048 8.633 0.000
F3–F2 0.376 0.048 7.821 0.000
F3–F4 0.313 0.051 6.110 0.000
F4–F1 0.245 0.058 4.232 0.000
F3–F1 0.232 0.054 4.317 0.000
Overall, principals indicated robust agreement concerning the benefits of a CSPAP for students,
but they expressed less agreement about certain aspects of their involvement in CSPAPs. Based on
these results, we suggest that while principals understand and appreciate the important contribu-
tions of a CSPAP to school goals, they may feel their ‘boots on the ground’ involvement with
CSPAPs is beyond their call of duty. Since principals’ active involvement in the development and
Orendorff et al. 589
Figure 1. Path model illustrating factors associated with principals’ involvement in comprehensive school
physical activity programmes (CSPAPs) from a social-ecological perspective.
Factor 1 (‘Involvement’) is directly predicted by Factor 2 (‘Intrapersonal’ level of influence), while Factor 2 is
directly predicted by Factor 3 (‘Interpersonal’ level of influence) and Factor 3 is directly predicted by Factor 4
(‘Environmental’ level of influence).
promotion of school programmes is critical to whether teachers feel they are supported for, and
whether teachers are committed to, such programmes (Leithwood and Jantzi, 1999), initiatives to
prepare principals for CSPAP implementation should emphasize the importance of their setting
performance standards for the programme, serving on a CSPAP committee, and providing CSPAP-
related professional development. Principals also agreed less overall that parents/guardians are
interested or engaged in supporting CSPAPs. Multicomponent interventions to increase youth PA
through school environments have been more effective when they include a focus on family and/or
community engagement (Russ et al., 2015). Therefore, strategies to increase family engagement in
CSPAPs should be another key focus of principal training.
Since we developed new measures to assess principals’ CSPAP involvement and perceptions,
we took several steps to increase the rigour of our measures, including an extensive literature
search, two rounds of pilot testing, and psychometric analysis. ESEM with data from the main
study sample indicated that the most robust and parsimonious model was a four-factor solution
containing a total of 38 items. This solution supports a single factor for principals’ CSPAP
involvement and three social-ecological factors. The items comprising the involvement factor
closely reflect the broad themes identified in published recommendations for administrators’
involvement in school-based health promotion, which include collaboration, advocacy, and sup-
port (Webster et al., 2020a). This suggests that many of the ways in which principals become
involved with CSPAPs may generalize to conceptualizations of school administrators’ involve-
ment in other school-based health initiatives, although additional research is needed to further
explore this proposition.
The three social-ecological factors each represent distinct levels of influence in relation to such
involvement, specifically intrapersonal, interpersonal, and ‘environmental’, which encompasses
items focusing on organizational, community, and public policy variables. While proposed social-
590 European Physical Education Review 27(3 )
ecological models often include five levels, it is not uncommon for researchers to collapse levels
into groupings that are similar to the factors retained in the present study. Fisher et al. (2018)
examined the relationship between the personal, social, and environmental factors regarding PA in
community-dwelling older adults, whereas Wilk and colleagues (2018) researched the PA levels of
grade 5 children by focusing on individual, interpersonal, and environmental levels of influence.
Other examples include studies in which individual, social, and environmental factors were used to
investigate young women’s participation in PA at transitional life stages (Craike et al., 2009) and
PA participation of adolescents in Spain (Devı́s-Devı́s et al., 2015).
Regarding the main focus of our study, we found significant associations between principals’
CSPAP involvement and all social-ecological factors (intrapersonal, interpersonal, and environ-
mental) assessed via the survey. Therefore, from a social-ecological perspective, variables across
all levels of influence appear to be important in the extent to which principals are involved with
CSPAPs. Furthermore, the strength of the relationships between factors and the results of the path
analysis support our hypothesis that intrapersonal variables have the strongest and most direct
association with principals’ CSPAP involvement, while factors at each subsequent level – inter-
personal and environmental (i.e. organizational, community, and public policy, respectively) –
have an indirect and increasingly weaker association. These results are consistent with those
reported in previous studies of school-based PA promotion that have used a social-ecological
perspective (Langille and Rodgers, 2010; Webster and Suzuki, 2014; Webster et al., 2013). For
instance, Webster et al. (2013) found that intrapersonal factors (i.e. domain-specific innovativeness
and perceived attributes of PA promotion) directly predicted elementary classroom teachers’ self-
reported PA promotion in the classroom setting, and that the organizational level of influence (i.e.
perceived school support for PA promotion) and the public policy level of influence (i.e. awareness
of a state PA policy for elementary schools) were indirect predictors. Given the extant research and
the results of the present investigation, we suggest that initiatives (e.g. professional development,
interventions) focused on supporting school-based PA promotion should include strategies that
target all levels of influence, but special attention should be given to the intraindividual level of
influence. To increase principals’ CSPAP involvement, it is important to ensure principals
recognize the many positive outcomes that a CSPAP can help to facilitate, such as increased PA
before and after school, decreased off-task behaviours, and increased academic performance
(Carson and Webster, 2020).
To the authors’ knowledge, this is the first study to examine principals’ CSPAP involvement.
Notable strengths of the study are its theoretical grounding within a social-ecological perspective,
systematic development of appropriate survey measures with sound psychometric properties, and
use of a stratified random national sample. However, this study is limited by the low response rate
to the survey, potential response bias, and reliance solely on principals’ perceptions. Future survey
studies of principals’ CSPAP involvement should aim to recruit proportionately representative
(e.g. by state) samples and use incentives to increase the response rate. It is also worth exploring
whether there are times during the year that principals may be more inclined to participate in a
survey. For this study, data from the main study sample were collected in the mid-to-late spring,
which coincides with end-of-year testing for many schools and may be a more hectic time for
principals than the beginning or middle parts of the school year. Finally, examining others’ (e.g.
teachers, parents) perceptions of principals’ CSPAP involvement and conducting studies that use
different methods (e.g. observation) and designs (e.g. qualitative) will allow for deeper analysis of
CSPAP implementation and help to build upon the results of the present study.
Orendorff et al. 591
Declaration of conflicting interests
The authors have no conflicts of interest to declare.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
ORCID iD
Collin A Webster https://orcid.org/0000-0003-1680-9149
References
American Alliance for Health, Physical Education, Recreation and Dance (2011) Comprehensive School
Physical Activity Program (CSPAP) survey report. Report, Reston, VA, Spring.
Barrett P (2007) Structural equation modeling: Adjudging model fit. Personality and Individual Differences
42(5): 815–824.
Brener ND, Demissie Z, McManus T, et al. (2017) School health profiles 2016: Characteristics of health
programs among secondary schools. Report, Centers for Disease Control and Prevention.
Bronfenbrenner U (1976) The experimental ecology of education. Educational Researcher 5(9): 5–15.
Bronfenbrenner U (1977) Toward an experimental ecology of human development. American Psychologist
32(7): 513–531.
Browne MW (2001) An overview of analytic rotation in exploratory factor analysis. Multivariate Behavioral
Research 36(1): 111–150.
Bryk AS, Sebring PB, Allensworth E, et al. (2010) Organizing Schools for Improvement: Lessons from
Chicago. Chicago, IL: University of Chicago Press.
Carson RL and Webster CA (2020) Comprehensive school physical activity programs: Putting research into
evidence-based practice. Champaign, IL: Human Kinetics.
Carson RL, Castelli DM, Beighle A, et al. (2014) School-based physical activity promotion: A conceptual
framework for research and practice. Childhood Obesity 10(2): 100–106.
Carson RL and Webster CA (eds) (2020) Comprehensive School Physical Activity Programs: Putting
Evidence-Based Research into Practice. Champaign, IL: Human Kinetics.
Cattell RB (1978) The Scientific Use of Factor Analysis in Behavioral and Life Sciences. New York: Plenum.
Centers for Disease Control and Prevention (2011) School health guidelines to promote healthy eating and
physical activity. MMWR. Recommendations and reports: Morbidity and mortality weekly report. Rec-
ommendations and reports 60(RR-5): 1.
Centers for Disease Control and Prevention (2013) Comprehensive school physical activity programs: A
guide for schools. Atlanta, GA: US Department of Health and Human Services.
Centers for Disease Control and Prevention (2019) Increasing Physical Education and Physical Activity: A
Framework for Schools. Centers for Disease Control and Prevention, US Department of Health and
Human Services, Atlanta, GA.
Costello AB and Osborne JW (2005) Best practices in exploratory factor analysis: Four recommendations for
getting the most from your analysis. Practical Assessment, Research & Evaluation 10(7): 1–9.
Craike MJ, Symons C and Zimmermann JAM (2009) Why do young women drop out of sport and physical
activity? A social ecological approach. Annals of Leisure Research 12(2): 148–172.
Datnow A and Castellano ME (2001) Managing and guiding school reform: Leadership in success for all
schools. Educational Administration Quarterly 37(2): 219–249.
Davis M (2012) Principals’ perceptions of factors associated with the implementation of school wellness
policies. Doctoral Dissertation, Mississippi State University, Starkville, MS.
Devı́s-Devı́s J, Beltrán-Carrillo VJ and Peiró-Velert C (2015) Exploring socio-ecological factors influencing
active and inactive Spanish students in years 12 and 13. Sport, Education and Society 20(3): 361–380.
592 European Physical Education Review 27(3 )
De Winter JD, Dodou D and Wieringa PA (2009) Exploratory factor analysis with small sample sizes.
Multivariate Behavioral Research 44(2): 147–181.
DiStefano C, Liu J, Jiang N, et al. (2018) Examination of the weighted root mean square residual: Evidence
for trustworthiness? Structural Equation Modeling: A Multidisciplinary Journal 25(3): 453–466.
DiStefano C, Zhu M and Mı̂ndrilă D (2009) Understanding and using factor scores: Considerations for the
applied researcher. Practical Assessment, Research, and Evaluation 14(1): 20. Available at: https://doi.
org/10.7275/da8t-4g52.
Dow II and Oakley WF (1992) School effectiveness and leadership. Alberta Journal of Educational Research
38(1): 33–47.
Fink A (2003) The Survey Handbook. London: SAGE.
Finney SJ and DiStefano C (2006) Non-normal and categorical data in structural equation modeling. Struc-
tural Equation Modeling: A Second Course 10(6): 269–314.
Fisher KL, Harrison EL, Bruner BG, et al. (2018) Predictors of physical activity levels in community-
dwelling older adults: A multivariate approach based on a socio-ecological framework. Journal of Aging
and Physical Activity 26(1): 114–120.
Fullan M (2016) The New Meaning of Educational Change. 5th ed. New York, NY: Teachers College Press.
Golden SD, McLeroy KR, Green LW, et al. (2015) Upending the social ecological model to guide health
promotion efforts toward policy and environmental change. Health Education & Behavior 42(Suppl 1):
8S–14S.
Greaney M, Hardwick CK, Mezgebu S, et al. (2007) Assessing the feasibility of a multi-program school-based
intervention to promote physical activity and healthful eating in middle schools prior to wide-scale
implementation. American Journal of Health Education 38(5): 250–257.
Hastie PA (2017) Revisiting the national physical education content standards: What do we really know about
our achievement of the physically educated/literate person? Journal of Teaching in Physical Education
36(1): 3–19.
Heerwegh D (2014) Small Sample Bayesian Factor Analysis. Pharmaceutical Users’ Software Exchange
annual conference. Available at: www.Lexjansen.Com/phuse/2014/sp/SP03.pdf (accessed 9 March
2020).
Hu LT and Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional
criteria versus new alternatives. Structural Equation Modeling 6(1): 1–55.
Institute of Medicine (2013) Educating the student body: Taking physical activity and physical education to
school. Report. Washington, DC: The National Academies Press.
Johnson RL and Morgan GB (2016) Survey Scales: A Guide to Development, Analysis, and Reporting. New
York, NY: The Guilford Press.
Krosnick JA and Fabrigar LR (1997) Designing rating scales for effective measurement in surveys. In: Lyberg
L, Biemer P, Collins M, et al. (eds) Survey Measurement and Process Quality. New York: John Wiley,
141–164.
Langille JLD and Rodgers WM (2010) Exploring the influence of a social ecological model on school-based
physical activity. Health Education & Behavior 37(6): 879–894.
Leithwood K and Jantzi D (1999) Transformational school leadership effects: A replication. School Effec-
tiveness and School Improvement 10(4): 451–479.
MacCallum RC, Widaman KF, Preacher KJ, et al. (2001) Sample size in factor analysis: The role of model
error. Multivariate Behavioral Research 36(4): 611–637.
McLeroy KR, Bibeau D, Steckler A, et al. (1988) An ecological perspective on health promotion programs.
Health Education Quarterly 15(4): 351–377.
Marsh HW, Hau KT, Balla JR, et al. (1998) Is more ever too much? The number of indicators per factor in
confirmatory factor analysis. Multivariate Behavioral Research 33(2): 181–220.
Marsh HW, Morin AJ, Parker PD, et al. (2014) Exploratory structural equation modeling: An integration of
the best features of exploratory and confirmatory factor analysis. Annual Review of Clinical Psychology
10(1): 85–110.
Orendorff et al. 593
Morin AJ and Maiano C (2011) Cross-validation of the short form of the physical self-inventory (PSI-S) using
exploratory structural equation modeling (ESEM). Psychology of Sport and Exercise 12(5): 540–554.
Morin AJS, Marsh HW and Nagengast B (2013) Chapter 10. Exploratory structural equation modeling. In:
Hancock GR and Mueller RO (eds) Structural Equation Modeling: A Second Course (2nd ed.). Charlotte,
NC: Information Age Publishing, Inc, 395–436.
Morrison LE (2006) Principals’ perceptions regarding recess in Georgia elementary schools. Doctoral
dissertation, Georgia Southern University, Statesboro, GA.
Muthén B and Asparouhov T (2012) Bayesian SEM: A more flexible representation of substantive theory.
Psychological Methods 17: 313–335.
National Association for Sport and Physical Education (2008) Comprehensive School Physical Activity
Programs [position statement]. Available at: https://files.eric.ed.gov/fulltext/ED541610.pdf (accessed 9
March 2020).
National Physical Activity Plan Alliance (2016) U.S. National Physical Activity Plan. Available at: https://
physicalactivityplan.org/docs/2016NPAP_Finalforwebsite.pdf (accessed 9 March 2020).
National Physical Activity Plan Alliance (2018) The 2018 United States report card on physical activity for
children and youth. Report. Washington, DC: National Physical Activity Plan Alliance.
Physical Activity Guidelines Advisory Committee (2018) 2018 Physical Activity Guidelines Advisory Com-
mittee scientific report. Washington, DC: US Department of Health and Human Services.
Preacher KJ and MacCallum RC (2002) Exploratory factor analysis in behavior genetics research: Factor
recovery with small sample sizes. Behavior Genetics 32(2): 153–161.
Rink J, Hall TJ and Williams LH (2010) Schoolwide Physical Activity: A Comprehensive Guide to Designing
and Conducting Programs. Champaign, IL: Human Kinetics.
Russ LB, Webster CA, Beets MW, et al. (2015) Systematic review and meta-analysis of multicomponent
interventions through schools to increase physical activity. Journal of Physical Activity & Health 12(10):
1436–1446.
Sallis JF, Cervero RB, Ascher W, et al. (2006) An ecological approach to creating active living communities.
Annual Review of Public Health 27(1): 297–322.
Sallis JF, Owen N and Fisher EB (2008) Ecological models of health behavior. In: Glanz K, Rimer BK and
Viswanath K (eds), Health Behavior and Health Education: Theory, Research, and Practice. San Fran-
cisco, CA: Jossey-Bass, 465–485.
Society of Health and Physical Educators (SHAPE) America (2015) Comprehensive school physical activity
programs: Helping all students log 60 minutes of physical activity each day [Position statement]. Available
at: www.shapeamerica.org/uploads/pdfs/2018/advocacy/position-statements/CSPAP-final.pdf (accessed
9 March 2020).
Stokols D (1996) Translating social ecological theory into guidelines for community health promotion.
American Journal of Health Promotion 10(4): 282–298.
Turner L, Johnson TG, Slater SJ, et al. (2014) Physical activity practices in elementary schools and associa-
tions with physical education staffing and training. Research Quarterly for Exercise and Sport 85(4):
488–501.
US Department of Health and Human Services (2018) Physical activity guidelines for Americans (2nd ed.).
Report. Washington, DC: US Department of Health and Human Services.
Van den Berg V, Salimi R, de Groot RH, et al. (2017) “It’s a battle . . .You want to do it, but how will you get
it done?”: Teachers’ and principals’ perceptions of implementing additional physical activity in school for
academic performance. International Journal of Environmental Research and Public Health 14(10): 1160.
Available at: https://doi.org/10.3390/ijerph14101160.
Webster CA and Suzuki N (2014) Land of the rising pulse: A social ecological perspective of physical
activity opportunities for schoolchildren in Japan. Journal of Teaching in Physical Education 33(3):
304–325.
594 European Physical Education Review 27(3 )
Webster CA, Caputi P, Perreault M, et al. (2013) Elementary classroom teachers’ adoption of physical activity
promotion in the context of a statewide policy: An innovation diffusion and socio-ecologic perspective.
Journal of Teaching in Physical Education 32(4): 419–440.
Webster CA, Glascoe G, Moore C, et al. (2020a) Recommendations for administrators’ involvement in
school-based health promotion: A scoping review. International Journal of Environmental Research and
Public Health 17(17): 6249. Available at: https://doi.org/10.3390/ijerph17176249
Webster CA, Mı̂ndrilă D, Moore C, et al. (2020b) Exploring the role of physical education teachers’ domain-
specific innovativeness, educational background, and perceived school support in CSPAP adoption.
Journal of Teaching in Physical Education 39(1): 36–47.
Webster CA, Mı̂ndrilă D, Moore C, et al. (2020c) Measuring and comparing physical education teachers’
perceived attributes of CSPAPs: An innovation adoption perspective. Journal of Teaching in Physical
Education 39(1): 78–90.
Webster CA, Rink JE, Carson RL, et al. (2020d). The comprehensive school physical activity program model:
A proposed illustrative supplement to help move the needle on youth physical activity. Kinesiology
Review 9(2): 112–121.
Webster CA, Weaver RG, Carman M, et al. (2020e). Staff involvement. In: Carson RL and Webster CA (eds)
Comprehensive School Physical Activity Programs: Putting Research into Evidence-Based Practice.
Champaign, IL: Human Kinetics, 127–142.
Weiler RM, Pigg RM and McDermott RJ (2003) Evaluation of the Florida coordinated school health program
pilot schools project. Journal of School Health 73(1): 3–8.
Wilk P, Clark A., Maltby A, et al. (2018) Examining individual, interpersonal, and environmental influences
on children’s physical activity levels. SSM-Population Health 4: 76–85.
Yu C and Muthén B (2002) Evaluation of model fit indices for latent variable models with categorical and
continuous outcomes. In: Annual meeting of the American Educational Research Association, New
Orleans, LA, 1–5 April 2002.
Author biographies
Karie Orendorff is an assistant professor in the Department of Health and Human Development at Montana
State University in Bozeman, Montana, USA.
Collin A Webster is a professor in the Department of Physical Education at the University of South Carolina
in Columbia, South Carolina, USA.
Diana Mı̂ndrilă is an associate professor in the Department of Educational Leadership, Research, and School
Improvement at the University of West Georgia in Carrollton, Georgia, USA.
Kathleen MW Cunningham is an assistant professor in the Department of Educational Leadership and
Policies at the University of South Carolina in Columbia, South Carolina, USA.
Panayiotis Doutis is an assistant professor in the Department of Physical Education at the University of South
Carolina in Columbia, South Carolina, USA.
Brian Dauenhauer is an associate professor in the School of Sport and Exercise Science at the University of
Northern Colorado in Greeley, Colorado, USA.
David F Stodden is a professor in the Department of Physical Education at the University of South Carolina
in Columbia, South Carolina, USA.