Study Design and Data Collection
A total of 730 undergraduate students from the College of Medicine and Allied Medical Sciences at Federal University Dutse, Nigeria (FUD), were recruited for the current study, which was a cross-sectional survey, between April 2023 and June 2023. The convenience sample approach was used for selecting the participants because of their proximity and because it was convenient for the researchers. Additionally, it is more accessible, easier, and lays a strong foundation for generalizability (i.e., making sure that the data collected are representative of the population from which the sample was obtained) [30]. Both a paper-based approach and a Google Form link were shared with the students who met the study's inclusion criteria to complete the survey questionnaire. The study inclusion criteria were: the College of Medicine and Allied Medical Sciences, FUD, undergraduate students who were in their first to final year, registered students during the data collection period, and those who gave consent to participate.
Ethical Approval
Prior to the commencement of the study, the Human Research Ethics Committee, Ministry of Health, Jigawa State, Nigeria granted ethical approval [JGHREC/2023/151], and the study was carried out in conformity with the Helsinki Declaration.
Items generation
The new Environmental Determinants of Health (EDH) scale was developed through a multi-step questionnaire development process that included gathering information from literature reviews, including the environmental health promotion model [31], the European environmental questionnaire for physical activity participation [32], and the Meikirch model of health [23, 24]. Subsequently, we sought content ideas from experts with experience in health psychology, psychometrics, public health, and questionnaire development. Also, An in-depth interview with 12 undergraduate students was conducted. In the initial stage of development, the EDH scale contained 18 items with two (2) hypothesized factors: the natural environment (8 items) and the built environment (10 items). All the items were evaluated using a five-point rating ranging from 1 (strongly disagree) to 5 (strongly agree).
Content Validity, Face Validity, and Pre-Testing of the EDH Scale
After generating the items, we conducted content validity by inviting six experts from health psychology, psychometrics, public health, and questionnaire development to score the relevance of each item to its domain using a non-face-to-face approach. Based on the recommended guidelines [33-36], the content validity was calculated to estimate the item content validity index (I-CVI) and scale content validity index (S-CVI). The S-CVIs were 1.00 for both the natural environment and the built environment, and the I-CVI for all the items was 1.00. Hence, these CVI values met the standard cut-up values [33, 36].
The face validity index (FVI) was also computed to evaluate the items' clarity and comprehension. Each item was graded on its clarity and comprehension by ten undergraduate students from the College of Medicine and Allied Medical Sciences FUD. Based on the standard guidelines [37, 38], the item face validity index (I-FVI) and scale face validity index (S-FVI) were computed. An online survey was used to collect all of the responses. The EDH scale's I-FVI varied from 0.90 to 1, and its S-FVI were 0.99 (natural environment) and 1 (built environment). Thus, the EDH had an acceptable FVI values [38].
Data Analysis
The data were pre-screened to investigate incorrect data entry and missing values, and only the items with complete responses were used in the analysis. The Statistical Product and Service Solution (SPSS) version 27 was used to conduct exploratory factor analysis (EFA). Subsequently, the EFA model was further examined using a Confirmatory Factor Analysis (CFA) with Mplus 8. The MLR Estimator was utilized for this study during the CFA due to its robustness to non-normal data distributions [39].
There were 300 participants in the EFA sample. Principal axis factoring with Promax rotation was used to test the 18 completed items on the EDH scale in order to identify the primary contributing factors. After investigating the factors with eigenvalues greater than one, those with factor loadings greater than 0.40 were deemed to be statistically relevant and kept for further CFA [35, 40]. The EFA model will be re-specified if an item is deleted, and the factor loadings will be investigated after each re-specification. For each factor, an acceptable Cronbach's alpha value of 0.60 or above was considered acceptable [35].
With 430 participants, the CFA was applied to further test the EFA model. For the purposes of the current study, the recommended standardized factor loading of equal to or greater than 0.40 was applied as the threshold for maintaining or removing an item [41, 42]. According to Hair et al., [43], the acceptable fit indices for a sample size greater than 250 with 12 items and higher were: comparative fit index (CFI) or Tucker and Lewis index (TLI) greater than 0.92; standardized root mean square residual (SRMR) less than 0.08; and root mean square error of approximation (RMSEA) less than 0.07. In order to improve the model fit indices, the model was re-specified based on the CFA modification index after taking adequate theoretical guidance into account.
Average variance extracted (AVE) and composite reliability (CR) were computed in order to further examine the convergent validity of the EDH scale. For CR and AVE, respectively, the acceptable cut-off values were equal to or greater than 0.70 and 0.50 [35, 40, 44]. If the AVE values are less than 0.50 but the CR values are above 0.60, convergent validity is still considered acceptable [44]. The discriminant validity—or how much one factor varies from another—was determined by looking at the correlation between the factors [40]. A correlation coefficient of 0.85 or less between two factors was considered sufficient for discriminant validity [40]. Also, Fornell and Larcker [42] noted that the AVE of the constructs must be greater than the squared of the correlation coefficient (i.e., the shared variance between the factors) for discriminant validity to be confirmed. In order to assess test-retest reliability, a sub-sample of 70 respondents completes the EDH questionnaire twice over a period of 7 days. An intra-class correlation coefficient (ICC) value greater than 0.70 indicates sufficient stability [45, 46].