Multiple choice Flashcards
Does a biased estimator always have a higher MSE than an unbiased estimator?
No
Are OLS residuals correlated with the exogenous variables?
No
Why do OLS coefficient estimates follow a t distribution instead of a normal distribution?
Because error variance is unknown
If we want to test the hypothesis that a coefficient is positive, do we use a two tailed test?
No
Does the standard error of the regression always lie in the range 0 to 1?
No
The smaller the standard error:
The better the fit
The f test for joint significance is distributed as F with degrees of freedom equal to:
T-k-1, k
T = number of observations
k = number of slope coefficients
Would serial correlation of the errors in a regression model lead us to believe that the OLS coefficient estimates are biased?
No
If the Durbin-Watson statistic lies between the upper and lower critical bounds, should we accept the null that there is no serial correlation?
Yes
Is heteroskedasticity most often found in time series regression models?
No
How can you deal with heteroskedasticity?
Scaling the data appropriately
In the bivariate regression model, regression residuals can be estimated as:
Yi - a(hat) - B(hat)Xi
If the covariance between X and Y is negative then this implies that their correlation coefficient is:
Also negative
In the bivariate regression model, To get an unbiased estimator of error variance, we use the formula:
RSS/N-2
A model is said to be “misspecified” if:
- It excludes a relevant explanatory variable
- It includes an irrelevant explanatory variable
- The functional form used is incorrect
