Multiple Regression Versus Multiple Correlation - Explained

[@tsegayewedajo949]

I was talking about the R square value, but you gave me an insight into the concept. Thank you

[@radtech2612]

Can I use pearson r to measure relationship of 3 variables? Also, which p value I could use in regression to determine if there is significant relationship among 3 variables? (Since in the model summary it shows r value but not p value)

[@jaykenarn6223]

So, if my research is a correlational design, can I use multiple regression analysis? (though the only things I need are the Multiple R and p-value?)

[@AOverload]

Thanks for the explanation! I was wondering whether there is any rule of thumb for wanting multiple correlation over multiple regression. It seems to me that you get the same information from both analyses, so is there any situation where one would prefer multiple correlation over multiple regression? And along those lines, can you even run a multiple correlation without running a multiple regression? Do they answer different research questions?

[@how2stats]

In practical terms, you might focus only on multiple R (multiple correlation) from a statistical power perspective, as a multiple correlation analysis requires less N, all other things equal. With multiple regression, where you examine the beta-weights for statistical significant, the N required for power = 80% is much higher.

[@AOverload]

@@how2stats Thanks for the clarification, that helps a lot!

[@jaykenarn6223]

What to do if there are two independent variables, two moderator variables, and one dependent variables in multiple correlation analysis? please help

[@dwilkins1189]

Thank you!!!

[@Preciousenglishworldwide]

Nice Nice Nice

[@taetae3298]

so the multiple regression and correlation analysis can be done together?? i'm having problem on choosing between those two

[@how2stats]

For multiple correlation, you're only interested in Multiple R. For multiple regression, you're interested in the multiple regression equation (i.e.,intercept and beta-weights). In practice, the vast majority of people are interested in multiple regression. For programs like SPSS, multiple correlation and multiple regression is conducted automatically together, so there's no choice to make from that perspective.

[@taetae3298]

@@how2stats okay thanks,,can i ask another question? im having problem in choosing analysis for one numerical dependent variable with 4 groups of independent categorical variable 😥 should i use one-way Anova?

[@couragee1]

Thanks!