I am working in a dissertation and I am using SMART PLS to test a model. The findings worked really well. I have a good R-square for the dependent variables and I was able to calculate the path coefficients to test the relationships between variables.
My question is should i use assumptions of regression before conducting SMART PLS test?
Yes, before running regression analysis through SMARTPLS you need to check multicollinearity between the IVs, linearity is another assumption, outliers need to checked, but most importantly, SMARTPLS yields good results even if the data is non normal. Normality is not a concern when estimation is done through SmartPLS.
It is always advisable that one must follow the assumptions of the regression. Besides this the prediction should also be in the range of the data. Outside the range of the data, the prediction should be used with cautious.
Validation of the model is always considered and would be good, if it predict well with independent data set (not used in model development).
I recommend that you meet the regression assumptions. If you do not, then your results may not be accurate based on the regression method used.
Thank you Wali, Rajiv,Freidoon, Edward and Emelina.
Your help is really appreciated
Check this out:
https://www.smartpls.com/resources/ebook_on_pls-sem.pdf
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size