RobertNiles.com
About Robert
Reporting Help
Finding Data on the Internet
Questions for Candidates
Stats Lessons
Mean
Median
Percent
Per capita
Standard Deviation
Margin of Error
Data Analysis
Sample Sizes
Stats Tests
Help Board
Bookstore


A Mathematician Reads the Newspaper

Read about reporters who blew it when trying to write about stats. A mathematician leads you through the tricks, citing contemporary examples.
More information
-->

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Multiple Regression and Betas
Message posted by Suzanne on November 5, 1999 at 12:00 AM (ET)

I have three variables that I am analyzing: Predictor A, Predictor B, and a dependent variable. I am predicting that A and B both will be correlated with each other and with the dependent variable. I am also predicting that when I regress A and B simultaneously on my dependent variable, A will drop to nonsignificance, whereas B will remain significant. In other words, I am predicting that B will mediate the relationship between A and the dependent variable. In fact, this is what my data shows.

My question: When I looked at my standardized coefficients (betas) for A and B in the multiple regression, they BOTH were lower than they were for their respective bivariate regressions. What causes this drop in beta for both variables?


READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Multiple Regression and Betas
Message posted by JG on November 5, 1999 at 12:00 AM (ET)

I think that this means that A and B are positively correlated. I assume that adjuted R-squared got better.

Do stepwise regression . First regress on A and then regress the residuals on B, and then reverse the process. See what you get. Also calculate directly the correlation between A and B .


Re: Multiple Regression and Betas
Message posted by JG on November 5, 1999 at 12:00 AM (ET)

You may also wish to do path analysis if you model of the physical situation justifies this approach.


Re: Multiple Regression and Betas
Message posted by j-bamdad on November 5, 1999 at 12:00 AM (ET)

thank you


Re: Multiple Regression and Betas
Message posted by link@wild.com on March 2, 2000 at 12:00 AM (ET)

You may be experiencing something called "Beta Bounce" which occurs when you have high colinearity between you predictors. Becareful of this!!!



Your $5 contribution helps cover part the $500 annual cost of keeping this site online.

Niles Online Sites:RobertNiles.comTheme Park InsiderViolinist.com

RobertNiles.com™, the site, content and services 咀opyright 1996-2002, Robert Niles.
All rights reserved. Questions? Comments? Read my Privacy Policy, or E-mail me!