Hello all,
I have applied the SMART-PLS method to test the relationship among 5 factors. I wonder if I have to mention all the relationships in the interpretation of R² even if it was not significant or only mention those that contribute significantly.
Example for the fatcor1:
42,1% of the variance in Factor 1 was explained by the following factors 2, 3, 4 and 5 (given that the relation "Factor3-Factor 1" was not significant as shown in the table below).
Or I have to cite only that contribute significantly, it means factors 2,4 & 5.
Dear Roumeissa Salhi
You need to check a number of papers published by Emaraled journals by searching the term like "Smart PLS", or "Structural equation modelling". You need to follow the interpretations of one of the paper published by emarald
Dear Zahid Bashir ,
I have consulted some papers on Smart-PLS methods but I have confused in this point. Logically I have to mention all the factors in the interpretation of R² even if its contribution is not significant because it contribute in the construct of the model, and in the interpretation of the effect size I will keep only that contribute significantly, but I want to confirm this point.
Best regards
@Salhi I don't how you have hypothesized your model as in your model, each variable is dependent variable thus you have five R^2 to report. Moreover, when it comes to reporting results from Smart-PLS, you need to report validity statistics of outer model (measurement model) i.e., loadings, Alpha reliability, CR, AVE and for discriminant validity as well. For structural model, you need to report Standardized regression weights, t-value, and p-value for each relationship.
The question is not about PLS so much as it is about statistical significance. And the advice from the American Statistical Association is to ignore the concept of statistical significance. See the editorial (Wasserstein, Schirm & Lazar 2019) in The American Statistician, "Moving to a world beyond "p < .05":
https://www.tandfonline.com/doi/pdf/10.1080/00031305.2019.1583913?needA
In short, report all of the variables in all of your equations.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis