An empirical investigation of physical literacy-based adolescent health promotion

The CEPS is a nationally representative large-scale follow-up survey designed and implemented by the National Survey Research Center of Renmin University of China. The project aims to reveal the effect of families, schools, communities and macrolevel social structures on the educational output of individuals and explore the process of educational output in an individual’s life course. With the 2013–2014 school year as the baseline and first- (grade 7) and third-grade (grade 9) junior high school students as the two study cohorts, data were collected from 28 county-level units (counties, districts, and cities) randomly selected as survey points. From the county-level units, 488 classes in 112 schools were randomly selected as the basis for survey implementation, and all students in these classes were sampled. For the survey, a multistage probability proportional to scale sampling method was employed: In Stage I, (primary sampling units), 28 counties (districts) were sampled from the county (district)-level administrative units across China; in Stage II (secondary sampling units), the seventh-grade and/or ninth-grade students of four schools were sampled from the geographic area under the jurisdiction of each sample county (district); in Stage III (tertiary sampling units), four classes (two seventh-grade classes and two ninth-grade classes) were sampled from each sample school; in Stage IV, all students in the sample classes made up the final survey sample. The questionnaires were self-administered, and the response rate was 91.9%.

CEPS follow-up data for 2014–2015 were downloaded and imported into Excel for sorting, and the sorted data were imported into SPSS version 23.0 for statistical analysis. The concept of wellness, whether it relates to PL or health promotion, is related to overall health. The health triangle is formed by physical, psychological, and social adaptation factors. Factors influencing mental health include stress, relationships with others, hobbies, psychological quality, and health education knowledge. Factors influencing physical health include physical exercise, family financial situation, sleep quality, physical condition, and health knowledge. Societal adaptation is influenced by factors such as self-confidence, interpersonal relationships, and emotional regulation ability. Based on these influencing factors, 19 of the most representative questions were extracted from the questionnaire: A29 (What do you think about this expectation?), A5 (Are you the only child in your family?), A8 (What is the marital status of your biological parents?), A17 (How would you describe the relationship between your parents?), C19 (Do you have the following sleep problems under normal circumstances?), B12 (Do you have any hobbies?), C26 (Do you agree that you can quickly adjust to the above emotions?), C21 (Have you ever attended health education classes in junior high school?), B8 (Do you usually complete homework assigned by teachers on weekends?), C20 (Did you attend health education classes in elementary school?), B16 (What is your means of transportation between home and school?), C13 (How often and for how long do you engage in physical activity: [] days a week, [] minutes a day?), A9 (What do you think of your family’s current financial situation?), C5 (Have you often fallen ill in the past year?), C1C2 (What are your height and weight measurements? [] cm, [] kg), B21 (Are you confident about your future?), A9 (What do you think of your family’s financial situation?), B6 (Do you agree that the people in this school are kind to you?), and C4 (How is your overall health?).

A total of 17 independent variables (or predictive factors) were included in the research model: whether a child is an only child, parental marital status, family financial situation, parents’ expectations regarding their children’s education, the degree of closeness to teachers or students, time spent doing schoolwork, whether a child has hobbies, means of transportation to school, degree of confidence about the future, whether a child has been sick in the past year, whether a child has sleep problems, whether a child attended health education classes in elementary school, whether a child has attended health education classes in junior high school, emotional adjustment speed, weekly physical activity frequency and time, and BMI.

A single-factor analysis was adopted to analyse the 17 factors that may affect adolescent health. Then, logistic regression was applied to identify the factors substantially affecting adolescent health. The factors identified were classified and summarised. A comparison of the two groups of measurement data was performed using a one-way t-test, and a comparison of the two groups of count data was performed using a one-way chi-squared test. Binary logistic regression was used to investigate the main factors affecting adolescent health, and a systematic cluster analysis was used to perform index classification.

As stated, 17 independent variables (or predictive factors) were included in the research model. A one-way chi-squared test or t-test was conducted to analyse these factors, and the results are shown in Table 1.

As illustrated in Table 1, preliminary statistics showed that of the 17 factors, 15 were statistically significant (P≤0.05). These factors were whether a child was an only child, parental marital status, family financial situation, relationship between parents, parents’ expectations regarding their children’s education, closeness to teachers or students, time spent doing schoolwork on weekends, whether a child has hobbies, degree of confidence about the future, physical condition in the past year, whether a child has had sleep problems, whether a child attended health education classes in elementary school, whether a child has attended health education classes in junior high school, emotional adjustment speed, and weekly physical activity time; as factors, means of transportation to school and BMI were not statistically significant (P>0.05).

Because the 15 influencing factors may interact with each other and become confounding factors, binary logistic regression was conducted to identify the true influencing factors. The results are shown in Tables 2–6.

The binary logistic regression results were statistically significant (χ² =1464.421, P<0.0005), meaning the model was appropriately constructed.

As illustrated in Table 3, the model explained 23.4% of the variance in participants’ answers in terms of whether they were healthy.

As illustrated in Table 4, the sensitivity, specificity, positive predictive value, and negative predictive value of the model were 89.1%, 39.5%, 73.36% and 65.87%, respectively.

As illustrated in Table 5, 10 variables were statistically significant (P≤0.05) in the model, and they were the main factors affecting adolescent health. These variables were parental marital status, family financial situation, closeness to teachers or students, whether a child has hobbies, degree of confidence about the future, physical condition in the past year, whether a child had sleep problems, whether a child attended health education classes in elementary school, emotional adjustment speed, and weekly physical activity time. In order of importance, the influencing factors are as follows: physical condition in the past year, whether a child has hobbies, degree of confidence about the future, family financial situation, and emotional adjustment speed.

As illustrated in Table 6, five variables were not included in the binary logistic regression equation: whether a child is an only child, relationship between parents, parents’ expectations regarding their children’s education, time spent doing schoolwork on weekends, and whether a child attended health education classes in junior high school.

Ten main variables were selected for the model, these are parental marital status, family financial situation, whether a child attended health education classes in elementary school, closeness to teachers or students, whether a child has sleep problems, whether a child has hobbies, weekly physical activity time, physical condition in the past year, degree of confidence about the future, and emotional adjustment speed. In the model, these factors appeared to be scattered with little connection and could not be explained on a practical level. For this reason, further index clustering of these 10 factors was performed through a cluster analysis, and an attempt was made to explain the results.

As illustrated in Figure 1, the factors affecting adolescent health can be divided into three categories: category I (parental marital status, whether a child has had sleep problems, whether a child has hobbies), category II (whether a child attended health education classes in elementary school, weekly physical activity time, physical condition in the past year), and category III (family financial situation, degree of confidence about the future, emotional adjustment speed, and closeness to teachers or students). Parental marital status affected the sleep quality and hobbies of adolescents; whether adolescents attended health education classes in elementary school affected their weekly physical activity time and physical conditions in the past year; and family economic conditions affected their confidence about the future and emotional adjustment speed, which in turn affected their adjustment in society (closeness to teachers or students). These three-factor categories can be summarised as endogenous (motivation), fundamental (physical ability), and permeable (confidence); these factors independently affect the health of adolescents. This categorisation is consistent with the main dimensions of PL proposed by Whitehead (Whitehead, 2010): motivation, confidence, physical ability, knowledge, and understanding.

Fig. 1

Based on an extensive review of the literature and expert argumentation, 17 factors were initially identified that may affect the health of adolescents. Empirical research revealed that of these factors, 15 were statistically significant in a single-factor statistical analysis and two were nonsignificant. Then, using a binary logistic regression equation, multifactor analyses were conducted to remove confounding factors; 10 factors remained that directly affected the health of adolescents. Next, a cluster analysis was performed for these 10 influencing factors, and they were categorised into three groups.

International scholars have attempted to identify the factors influencing PL from various perspectives. PL can be influenced by the self, parents and other family members, teachers or coaches, peers, and others. A child’s PL can be affected by their household, community, school, and other environments. PL is affected by diet, sleep, psychological status, lifestyle, physical activity, interpersonal relationships, physical fitness, motor skills, and other factors (Corbin, 2016; Robinson and Randall, 2017). Whitehead (2010) stated that motivation, confidence, physical ability, knowledge, understanding and effective interaction with one’s environment are the core elements that constitute PL, and they complement each other. Motivation promotes participation, which enhances a person’s confidence and physical ability. This in turn maintains or enhances a person’s motivation. Enhancements in confidence and physical ability can promote smooth interactions in various environments. Positive environmental interactions and taking on new challenges can in turn enhance a person’s confidence and physical ability. In this study, we classified the factors influencing PL through a cluster analysis, and the factors were consistent in terms of endogeneity, fundamentality, and permeability, as proposed by Ren (2018). Variables in categories I, II, and III were respectively the endogenous, fundamental, and permeable factors related to PL.

Canada is the first country in which PL was examined and where related research was applied in practice. According to the IPLA’s definition (International Physical Literacy Association (IPLA), 2016), PL has emotional (motivation and confidence), physical (physical ability), cognitive (knowledge and understanding) and behavioural (in terms of participation in physical activity for life) elements, and they are all interrelated (Canadian Sport for Life (CS4L), 2016; Ren, 2018). Through a binary logistic regression, 10 main factors affecting adolescent health were identified, and they can be classified based on their relation to these elements. Specifically, parental marital status, closeness to teachers or students, whether a child has hobbies, and the degree of confidence about future are emotional elements; family financial situation, physical condition in the past year, and whether a child has had sleep problems are physical elements; whether a child attended health education classes in elementary school and emotional adjustment speed are cognitive elements; and weekly physical activity time is a behavioural element. In the next section, we discuss these four elements separately.