First, although the scales used in this study were all mature scales that have been widely used, the reliability and validity of the scales were confirmed. Specifically, the reliability of each scale was analyzed by SPSS software. Then the confirmatory factor analysis (CFA) was used to assess the validity of each scale by Amos. Second, we conducted a correlation analysis between the demographic variables and the key five variables (parental involvement, parenting self-efficacy, parenting stress, emotional/behavioral problems and prosocial behaviors) to determine the control variables in the next analysis. Next, the Pearson correlation was performed to determine the relationships between the key variables, which is the prerequisite for mediation analysis. Then, to verify hypothesis 1, the hierarchy regression analysis for the emotional/behavioral problems and prosocial behaviors were conducted to determine the inferential relationship according to the sample data in this study.
Third, to verify Hypotheses 2–4, mediation analysis was used to examine the chain mediated effects of parenting self-efficacy and parenting stress between parental involvement and psychological adjustment in a holistic model, which had been widely used previously (Beeble et al., 2009 (link); Parkes and Sweeting, 2018 (link)). We characterized a significant mediated effect of parenting self-efficacy or parenting stress between parental involvement and psychological adjustment if the effect of the independent variable on the dependent variable was mediated by one mediating variable; a significant chain mediated effect if the effect of the independent variable on the dependent variable was mediated sequentially by two mediating variables.
Mediation analysis was performed with PROCESS in SPSS. As emotional/behavioral problems and prosocial behaviors represent two opposing aspects of psychological adjustment in children with ASD, two models were used to examine the effect of parental involvement on children’s psychological adjustment. In these two models, parental involvement was incorporated as the independent variable (X), and parenting self-efficacy and parenting stress served as the first-order mediator (M1) and second-order mediator (M2), respectively. Emotional/behavioral problems and prosocial behavior in children with ASD were the dependent variables (Y) (Lu et al., 2021 (link)). To determine the significance of mediation, we also used a bias-corrected bootstrap estimation approach with 5,000 samples in the study. A 95% confidence interval (CI) does not include zero, indicating that the mediating effect is significant (Yuan and Hayashi, 2003 (link)).
Third, to verify Hypotheses 2–4, mediation analysis was used to examine the chain mediated effects of parenting self-efficacy and parenting stress between parental involvement and psychological adjustment in a holistic model, which had been widely used previously (Beeble et al., 2009 (link); Parkes and Sweeting, 2018 (link)). We characterized a significant mediated effect of parenting self-efficacy or parenting stress between parental involvement and psychological adjustment if the effect of the independent variable on the dependent variable was mediated by one mediating variable; a significant chain mediated effect if the effect of the independent variable on the dependent variable was mediated sequentially by two mediating variables.
Mediation analysis was performed with PROCESS in SPSS. As emotional/behavioral problems and prosocial behaviors represent two opposing aspects of psychological adjustment in children with ASD, two models were used to examine the effect of parental involvement on children’s psychological adjustment. In these two models, parental involvement was incorporated as the independent variable (X), and parenting self-efficacy and parenting stress served as the first-order mediator (M1) and second-order mediator (M2), respectively. Emotional/behavioral problems and prosocial behavior in children with ASD were the dependent variables (Y) (Lu et al., 2021 (link)). To determine the significance of mediation, we also used a bias-corrected bootstrap estimation approach with 5,000 samples in the study. A 95% confidence interval (CI) does not include zero, indicating that the mediating effect is significant (Yuan and Hayashi, 2003 (link)).
Free full text: Click here
