Multicollinearity occurs when predictor variables are highly correlated. What is a typical consequence in multiple regression?

Multicollinearity occurs when predictor variables are highly correlated. What is a typical consequence in multiple regression?

• A. It reduces sample size.
• B. High correlation among predictor variables; inflates standard errors and destabilizes coefficients.
• C. It increases the R-squared automatically.
• D. It changes the response variable distribution.

Answer: (B)

When predictor variables are highly correlated, the model has trouble separating their individual effects. This overlap in information makes the estimated coefficients more sensitive to small changes in the data, so their standard errors grow and the coefficients become unstable and hard to interpret. In practice, you’ll see wider confidence intervals for the individual effects, and the apparent importance of each predictor can flip with different samples even if the overall model fit looks decent. The other statements don’t capture this typical consequence. Multicollinearity doesn’t directly reduce the sample size, it doesn’t automatically raise R-squared, and it doesn’t change how the response variable itself is distributed.

When predictor variables are highly correlated, the model has trouble separating their individual effects. This overlap in information makes the estimated coefficients more sensitive to small changes in the data, so their standard errors grow and the coefficients become unstable and hard to interpret. In practice, you’ll see wider confidence intervals for the individual effects, and the apparent importance of each predictor can flip with different samples even if the overall model fit looks decent.

The other statements don’t capture this typical consequence. Multicollinearity doesn’t directly reduce the sample size, it doesn’t automatically raise R-squared, and it doesn’t change how the response variable itself is distributed.