Amos software

The data was extracted from a structured questionnaire in which the students had to value the attributes on a scale of 0–7, in which 0 represented “of lowest value” and 7 “of highest value”. The questionnaire contained 14 items grouped as four latent variables. The four latent variables were system characteristics (CAR), effectiveness of online teaching (EFIC), relationship with teachers and evaluation (PROF), and satisfaction with online teaching (SAT). The items were selected from the literature on e-learning and quality of teaching (Table 1). Attributes representing four dimensions were analysed by means of a structural equations model.
The structural equations model was based on the literature (Figure 1). This model posits that the system’s characteristics have a direct effect on the efficacy of virtual training and on students’ relations with their teachers and aspects of the evaluation. Likewise, these latent variables have a direct effect on student satisfaction with online teaching. The structural equation model was estimated using the AMOS software (International Business Machines Corporation IBM, Armonk, NY, USA).

AMOS software (version 7.0; IBM Corp) was used for confirmatory factor analysis. SPSS software (version 24.0; IBM Corp) was used for exploratory factor analysis, common method bias, descriptive statistical analysis, and correlation analysis. SPSS PROCESS macro (version 3.5) was used to verify the moderated mediation models [40 ]. All regression coefficients were tested using the bias-corrected percentile bootstrap method. The theoretical model was tested by estimating the 95% confidence intervals of the mediation and moderating effects with 5000 repeated samples. An effect was considered significant if the confidence interval did not include 0.

Cross-tabulation was performed on age group and the items of job attractiveness, and the factors dental hygienists feel would improve the work environment. Correspondence analysis was performed with this cross-tabulation. To visualize the relationships, the results were illustrated graphically as biplots [13 (link)]. A three-parameter logistic model with item response theory (IRT) analysis was applied to calculate item discrimination, item difficulties, and item guesses for job attractiveness and satisfaction [1 (link),13 (link),14 (link)]. Item response and information curves are graphically illustrated. The analyses were carried out using R software version 3.50 (Institute for Statistics and Mathematics, Wien, Australia) with the LTR and irtoys packages using the following formula: $P_i\left(\theta \right)=\frac{\left(1-c_i\right)}{1+e^{-D{a}_i\left(\theta -b_i\right)}}$
where a_i: discrimination, b_i: difficulty and c_i: guessing.
Factor analysis with varimax rotation was performed to determine the latent variables for structural equation modelling (SEM). The structural relationship between job attractiveness and job satisfaction was calculated using AMOS software (24.0, IBM, Tokyo, Japan).

This questionnaire focuses on the contents of entrepreneurship education curriculum, entrepreneurship practice, and college students’ satisfaction with entrepreneurship education on five-point Likert scale (1 = strongly disagree; 2 = disagree; 3 = average; 4 = agree; 5 = strongly agree), with a total of 3 dimensions and 17 items. The dimensions of students’ entrepreneurship education in universities are shown in Table 4. IBM SPSS and Amos software were used to test the reliability of the scale (see Table 5).

The data was stored in Microsoft Access (Microsoft, Seattle, WA, USA), and we used the software of SPSS (Version 24.0; IBM Corp., Armonk, NY, USA) to perform descriptive statistical analyses and scale reliability tests, and calculate Cronbach’s coefficients. Exploratory factor analysis was performed to examine the structural validity of the scales. AMOS software (Version. 24.0; IBM Corp.) was used for the confirmatory factor analysis and construct validity verification of the scale. Due to the lack of a consistent scale, we have carried out a unified transformation of these variables. Spearman correlation coefficients were estimated for observed variables, including PE fit, job satisfaction, turnover intention, and professional efficacy.
We also constructed a structural equation model to verify the research hypotheses and assess the impact of intermediating variables. Finally, a hierarchical regression model was used to examine the relationship between demographic variables (e.g., age, occupation, and education) and PJ and PG fit. We also checked the multicollinearity among independent variables and the equal variance of linear models used in this study. We found no issue of the multicollinearity among independent variables and heterogeneity of the models.