In the multiple regression model with 4 explanatory variables, with standard assumptions, estimated using ordinary least squares, all the coefficients turn out to be insignificant although the overall R-squared is high and the associated F-statistic is significant. Also, pair-wise correlations amongst the four explanatory variables are all low, and range between and , but are not statistically different from . Which of the following statements is false?
Tap an option to check your answer.
Checking your account…
In the multiple regression model with 4 explanatory variables, with standard assumptions, estimated using ordinary least squares, all the coefficients turn out to be insignificant although the overall R-squared is high and the associated F-statistic is significant. Also, pair-wise correlations amongst the four explanatory variables are all low, and range between 0.1 and 0.2, but are not statistically different from 0. Which of the following statements is false? (A) Even though the pair-wise correlations between the explanatory variables are low, since they are (individually) statistically different from zero, OLS coefficients are likely biased (B) This is likely a case of multicollinearity even though the pairwise correlations are low (C) If the analyst is only interested in making forecasts then the insignificance of coefficients is per se not a problem since the F-statistic is significant and R-squared is high (D) Dropping a variable may improve significance of remaining coefficients but they may be biased
Classical assumptions and their violations has come up 26 times in 7 exams since 2011.
This question is on Multicollinearity, part of Classical assumptions and their violations in Econometrics.
Highly correlated regressors: symptoms, consequences and the perfect case.
TestStop has 8 past-paper questions on this subtopic; the nearest in year are below.
Question from the DSE 2016 paper. Solutions are TestStop's own. Not affiliated with or endorsed by any examination authority.