Diagnose and fix multicollinearity in income regression

You want to model the association between gender and income with a regression, adjusting for other covariates such as:

• age
• education (for instance, highest degree or years of education)
• years_since_degree

Questions:

1. Define multicollinearity. Among these predictors, is it likely to occur? Explain why or why not.
2. What checks or metrics would you use to detect multicollinearity in practice?
3. If multicollinearity is present, what are practical ways to handle it while still addressing the gender–income question? Compare the tradeoffs between interpretability and prediction.

Overview: This item tests knowledge of multicollinearity and regression diagnostics in the Statistics & Math area, aimed at applied regression modeling skills for an intermediate data scientist.