Say we have a multiple regression model with three predictors and one outcome variable.

Each predictor has it's associated standardized beta, which tells you how many standard deviations the outcome variable increases given a 1 SD increase in the predictor variable.

Would each predictor's standardized beta be different from the Pearson's correlation coefficient for it's bivariate relationship with the outcome variable? Or would they be the same?

I've found unclear information saying that the standardized betas can be "thought of" as the correlation coefficient (r), but still struggling to understand whether they truly are the same thing.

I'm somewhat new to statistics so an intuitive explanation would be much appreciated.

  • $\begingroup$ If standardized $b$ (beta) coefficient were equivalent to Pearson $r$ no need for multiple regression would arise. In fact beta is related to the partial $r$ (see). $\endgroup$ – ttnphns Aug 28 '18 at 10:43

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.