Exam 2: Simple Regression Analysis

You can conduct a positive sign test on a negative coefficient, but the results will be the same every time: Do not reject Ho.

The constant term is biased in a regression missing an important explanatory variable.

A) Criticize stepwise regression B) What criteria are important for determining if a particular explanatory variable belongs in a regression. (4pts.)

If a regression is underspecified, then the estimators will not average out to equal the true values of the structural parameters in repeated sampling.

The Schwarz Criterion is monotonically decreasing with respect to the number of explanatory variables.

If a test of significance results in "Do not reject Ho", then there is absolutely no chance that a TYPE II error has been committed

A 95% confidence interval for a given coefficient can never be larger than a 90% confidence interval.

If a regression is over specified then the estimators should change markedly once the extraneous variables are removed.

Raising the critical level of a test from 5% to 10% decreases the probability that the test will result in a TYPE II error.

Dependent Variable: WHY Method: Least Squares Sample(adjusted): 19652006 Included observations: 30 atter adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C -19.60486 123.5064 -0.158736 0.8747 EX1 20.04112 12.56987 1.594377 0.1189 EX2 0.517140 0.007738 66.83540 0.0000 R-squared 0.991948 Mean dependent var 2775.633 Adjusted R-squared 0.991536 S.D. dependent var 1915.813 S.E. of regression 176.2591 Akaike info criterion 13.25054 Sum squared resid 1211624. Schwarz criterion 13.37466 Log likelihood -275.2613 F-statistic 2402.403 Durbin-Watson stat 0.289425 Prob(F-statistic) 0.000000 Use the Eviews output above to help with A) through I) below. A) Interpret the coefficient on EX1. B) Interpret R² for this regression. C) Replace the "?"'s with the appropriate values. An 80% confidence interval around the coefficient on EX1 is: 20.04112 +/- (?)(?) D) Perform a positive sign test on the coefficient attached to EX2. Show the 5-step procedure. E) IF you performed a test of significance on the constant term at the 87 percent critical level, would you reject the null hypothesis or not? Explain. F) Perform a test to determine if the coefficient attached to EX2 is greater than 7.47. Show the 5-step procedure. G) Based on the information in the table above (and ignoring Ramsey's Reset test which we will perform in a moment) do you think the regression above is under or over specified? Explain. H)1.Perform Ramsey's Reset Test on the regression above using the Eviews printout below. Show the 5-step procedure. 2. What is the difference between "FITTED^2" and "FITTED^3" in the Eviews printout above? Do the two terms appear in the restricted or unrestricted regression?

The standard error of the regression (SER) is monotonically decreasing with respect to the number of explanatory variables.

If the prob-value is less than .05, then the hypothesis test will reject Ho at the 5% critical level.