Beta Coefficient

The beta coefficients are the regression coefficients you would have obtained had you first standardized all of your variables to a mean of 0 and a standard deviation of 1. Thus, the advantage of beta coefficients (as compared to B Coefficients that are not standardized) is that the magnitude of these beta coefficients allows you to compare the relative contribution of each independent variable in the prediction of the dependent variable.

The Regression Summary spreadsheets of Multiple Regression, GLM, GRM, Experimental Design, etc. will produce both the raw regression coefficients (B Coefficients) and the beta coefficients. The beta coefficients should be reviewed in order to evaluate the relative contribution of each predictor to the overall prediction of the dependent variable, and their interpretation is similar to that of partial correlations. To apply the regression equation to new observations, i.e., in order to compute predicted values (in the metric/scaling of the original variables), use the B coefficients.

See also, B Coefficient, partial correlations, and the Multiple Regression Overviews.