I am doing a research project for my thesis. After running the analysis, I get the results attached. From chi-square the model doesn't fit good the data, what about other fit indexes? RMR and RMSEA are permanently high. Can you suggest modifications to improve the model?
I see regression coefficients are significant, can I assume some of my hypothesis are confirmed
?
Hello Margherita,
The results suggest that:
1. The overall model does not do a very good job of reproducing the observed correlations among the variables in your data set.
2. Some of the proposed path coefficients are significantly different from zero (which, for purely manifest / observed variables, you could have verified with ordinary regression equations).
3. As there are no modification indices shown, there is no way to offer advice on what was given.
In general, relying on modification indices ignores the relevant theory/rationale which supported evaluation of the proposed model in the first place. These offer tremendous opportunity to both overfit the data and to arrive at a model that isn't congruent with the one you originally justified.
Since your query doesn't reveal your hypotheses, the specific manifest variables (and how each is quantified), the sample and data collection method, or the basis for the model, it would be a challenge for anyone to offer focused advice instead of generalities. Perhaps you could elaborate your post with this information?
Good luck with your work.
@Margherita I would agree with @David that your model shows considerably poor fit indices while most of the paths are found to be significant however, looking at squared multiple correlations, it seems that your final outcome variable is probably purchase frequency and for explanatory of the model for purchase frequency is really low (.062). I would suggest you to share screenshot of your structural model in order to get some good advices. Wishing you best with your work.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling