I have a question about reporting Pearson Correlations. I have a total scale (EBPAS) with four subscales. I would like to report how the total scale and each subscale correlate with the independent variables. Is the following sentence a correct interpretation of the results:
The 3-point scale measuring "interest in research" correlated significantly with the EBPAS total scale (r = 0.31, p = .001). In particular the subscales' openness (r= .203, p= .024) and divergence (r= -.232, p= .012) significantly correlated with the EBPAS, whereas the subscales attitude and appeal did not show significant correlation with interest.
The results indicate that out of the EBPAS openness and divergence towards EBP have the strongest explanatory power for interest in research.
would it not be preferable to analyze the predictive power of each subscale using the multiple regression analysis with the subscales is IVs and the total scale as a DV?
Dear Tessa, yes - your description above is technically correct. However, since your total scale actually correlates with the IV more strongly than any of the subscales, I am not sure to what extent this split-up information is even useful. It seems to be the case that the construct that 'is the thing in common to, or that drives, all subscales' is actually conceptually closer to the IV than any of the subscales themselves.
Also, dear Yasser, we should expect strong collinearity between the subscales of a common scale - which is somewhat the purpose of the total scale -, so that the multiple regression could easily show non-meaningful results.
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Environmental Science