I am using Pearson correlation analysis to check the association between different variables, but now I want to apply the regression analysis as well. So is it appropriate to have these analysis together? If yes, please recommend me some studies/research articles.
Hello you can find the discussion here :
https://www.researchgate.net/post/Is_it_practical_and_logical_to_use_correlation_and_multivariate_multiple_regression_in_the_same_study2
Yes, you can use both. Correlation will be useful to show that the study constructs are not highly correlated (i.e. measure different things). You have to test your hypotheses using multiple regression analysis. When using CFA and SEM, discriminant validity is usually run. When the square root of each AVE value is greater than the absolute correlation value between a construct and other constructs, this provides evidence of discriminant validity.
It feels like you are using correlation in your exploratory analysis. Then, you can perform multiple linear regression and avoid including predictors that are strongly correlated (collinear).
I found some links: https://datascienceplus.com/multicollinearity-in-r/
https://cran.rstudio.com/web/packages/olsrr/vignettes/regression_diagnostics.html
Multiple linear regression
Multiple Linear Regression (MLR) method helps in establishing correlation between the independent and dependent variables.
So it is sufficient enough to be used for your analysis.
- Badges
- Science topic
- Similar topics
- Engineering
- Petroleum Engineering
- Gas