Hello,
I am running a multiple linear regression in Prism for an example I will teach in class. When I calculate the correlation coefficients, they come out different than I get in SPSS. They also are different to the values calculated when Prism calculates the correlation matrix outside of multiple linear regression. Does multiple linear regression in Prism use a different calculation than in SPSS, or in Prism when calculating the correlation matrix? See attached. Thank you!
Hello Hannah,
The correlations you show (among the 5 variables) appear to match between the two applications. What's labeled "correlation matrix" in the prism output represents the correlations among the estimated regression coefficients, not among the original variables. So, it's not apples to apples.
Good luck with your presentation.
Thank you so much! I thought it must be looking at something different, but I just couldn't figure out what it was.
Hello Hannah :
Try the correlation in MINITAB program and do a comparison with SPSS results
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS