What is the risk when two independent variables in a model are themselves highly correlated?

Correct Answer

C) Their individual coefficients become unreliable

Why this is correct: This describes multicollinearity. When independent variables (like living area and room count) are highly correlated, the regression model cannot isolate their separate effects on the dependent variable. The overall model fit (R-squared) may remain high, but the estimated coefficient for each correlated variable becomes unstable and unreliable for individual adjustment. Why the other choices are wrong: 'R-squared falls to zero for the whole model' is wrong because multicollinearity does not destroy overall model fit. 'The dependent variable can no longer be chosen' is incorrect; the dependent variable is defined by the problem. 'The regression cannot be computed at all' is false; computation proceeds, but the results for individual predictors are problematic. Exam tip: Remember, multicollinearity affects the reliability of individual coefficients, not the model's ability to compute or its overall explanatory power.