Pre-service teachers’ empathy and attitudes toward inclusive education—The chain mediating role of teaching motivation and inclusive education efficacy

PLOS One | https://doi.org/10.1371/journal.pone.0321066 April 29, 2025 7 / 18 representing strongly disagree to strongly agree. The higher questionnaire scores indicating more positive attitudes. The Cronbach’s alpha coefficients for scale was 0.74 [49]. This research questionnaire (Empathy, Teaching motivation & SACIE-R) was initially adapted from an English language version, which was then translated into Chinese by translator (who holds a master’s Degree in Chinese from Heze univer- sity) to ease respondents’ comprehension. The study also employed a bilingual English lecturer from Heze university for back translation, to check the accuracy and precision of the translation [50]. Data analysis This study used SPSS 29 and Smart PLS 4.1 as tools for statistical data analysis. The data were examined using Smart PLS software by Partial Least Squares Equation Modelling (PLS-SEM). The reason for using Partial Least Squares Regression (PLS) is because it is a commonly used multiple regression method. PLS can simultaneously consider the relationship between the independent, mediator and dependent variables, and it is often more suitable for building predic- tive models than other methods [51]. In this study, the data sample was analyzed in two steps using Smart PLS software. Firstly, the measurement model was evaluated, and the PLS Algorithm was chosen to test the reliability and validity of the measurement model. Second, the structural model was validated, and bootstrapping and PLS Predict were selected to

assess the significance of the path coefficients and the explanatory and predictive power of the model. Research results Common method bias In the study of the causal effect of independent variables on dependent variables, there may be a common method bias in observing the association between the two variables when the same measures are utilized [52]. Common method vari- ance (CMV) refers to the use of the same survey instrument that causes erroneous common variance among characteris- tics, and it is generally present in data collected using self-report scales [53]. The bias introduced by CMV is referred to as common method bias (CMB), a trait-independent systematic error that affects the measure’s validity [54]. In the statistical analysis, the researchers used the full covariance test proposed by Kock (2015) [55]. According to this method, a variance inflation factor (VIF) value less than 3.3 indicates that the model has no multicollinearity problem and the model is well constructed [56]. The VIF values of the present study are shown in Table 1. It can be observed that all the constructs had VIF values below 3.3. Therefore, this study is not affected by CMB. Measurement model The research model of empathy is a reflective first-order structural model [46]. Teaching motivation (TM) is a reflective second-order structural model, and educational interest (EI) and professional love (PL) are first-order structures [47]. Inclusive education efficacy (IEE) is a reflective higher-order structural model consisting of three first-order structures:

inclusive teaching efficacy (ITE), collaborative efficacy (CE), and behavior management efficacy (BME) [48]. Attitudes toward inclusive education is a reflective second-order model, and the three dimensions of attitudes, concerns, and Table 1. Full collinearity testing. PL EI Empathy ITE CE BME Attitudes Sentiments Concerns 1.29 1.29 1.00 1.53 1.51 1.46 1.32 1.27 1.40 Note: PL = Professional love; EI = Educational interest; ITE = Inclusive teaching efficacy; CE = Collaborative efficacy; BME = Behavior management efficacy. https://doi.org/10.1371/journal.pone.0321066.t001