Chapter Summary

Large correlations between explanatory variables in a regression model produce collinearity. Collinearity leads to surprising coefficients with unexpected signs, imprecise estimates, and wide confidence intervals. You can detect collinearity before fitting a model in the correlation matrix and scatterplot matrix. Once you have fit a regression model, use the variance inflation factor to quantify the extent to which collinearity increases the standard error of a slope. Remedies for collinearity include combining variables and removing one of a highly correlated pair.

Key Terms

Objectives

  • Choose explanatory variables for a multiple regression through a combination of substantive insight and ...

Get Statistics for Business: Decision Making and Analysis, 3rd Edition now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.