Hi,
I am conducting a cross-sectional quantitative study, originally i was using 4 predictor variables and criterion and used multi-regressional analysis. This was significant and lead to three of my four predictors being supported.
pearsons correlation indicated these predictors were correlated the variance inflation factor (VIF) and tolerance indicated this was no an issue (low-moderate VIF and high tolerance) as the correlated variables were significant predictors.
predictors: professional self-doubt, developmental depletion and developmental growth (whether participants think they are growing in their careers). i also had years of experience that was non-significant. he criterion is well-being
I ran a serial mediation analysis using model 6. using a theoretical basis i added predictor variable 'self-doubt' as a predictor. developmental growth, developmental depletion and years of experience as mediators . well-being as the criterion.
the pathways between each variable and the criterion are the same as those found in multi-regression. indirect pathways were also found for developmental growth and depletion. depletion also influenced the link between growth and well-being. Is it reasonable to conclude these variables have direct and indirect effects on well-being?
The aim of my study is to see how these factors impact well-being.
If your objective is to determine the influence is more predictive, in mediating models therefore to obtain partial mediations all the significant values, or total mediations from which the direct effect is reduced and is not significant.
Kindly check https://academics.hamilton.edu/psychology/spsshmroho
it seems that you have ran appropriate analyses to test your research purpose (except for testing the relative significance of the mediators, you can test which mediating effect is stronger). However, it would be quite challenging to publish your findings into an academic journal due to the cross-sectional design (with the self-assessed measures only--it seems that all the measures were assessed by self-report).
I suggest you to test some potential moderating effects. For example, you can test whether years of experience can moderate some of the indirect effect or the relationship between professional self-doubt and developmental depletion/developmental growth. Also, you can test other variables in your data or demographic variables as a potential moderator.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology