Questions: F-Test and Joint Significance

5 questions to test your understanding

Score: 0 / 5
Question 1 Multiple Choice

You run a regression predicting annual salary using years_of_education and years_of_experience. Both variables are highly correlated and both have individually insignificant t-statistics (p > 0.10). What should you conclude?

ADrop both variables — individually insignificant coefficients mean the variables add no explanatory power
BDrop the less significant variable and re-run; the remaining variable may become significant
CRun an F-test first — correlated variables can be jointly significant even when individually insignificant
DBoth variables are statistically redundant, so keep only their average as a single predictor
Question 2 Multiple Choice

The F-statistic for overall model significance equals (RSS_restricted − RSS_unrestricted)/q ÷ RSS_unrestricted/(n−k−1). What does the term q represent in this formula?

AThe number of observations in the sample
BThe number of restrictions being tested — here, the number of slope coefficients jointly set to zero
CThe ratio of explained to unexplained variance in the full model
DThe degrees of freedom penalty for each additional regressor
Question 3 True / False

If most individual slope coefficients in a regression have p-values above 0.05, the overall F-test for joint significance will also fail to reject the null hypothesis.

TTrue
FFalse
Question 4 True / False

When testing a single linear restriction (q = 1), the F-statistic equals the square of the corresponding t-statistic.

TTrue
FFalse
Question 5 Short Answer

Why can't you simply run multiple t-tests — one for each coefficient — to determine whether a set of variables is jointly significant?

Think about your answer, then reveal below.