Answer Happy

Suppose you want to use a Chow test to investigate whether the wage equations differ for males and females. Using your EAWE data set, you first regress the log hourly learnings (LGEARN) on years of schooling (S), work experience (EXP), dummy variable indicating black (ETHBLACK) and the dummy variable indicating Hispanic (ETHHISP). Then you repeat the regression using only the male respondents, and then repeat again using only the female respondents. The output of the three regressions are shown below:

reg LGEARN S EXP ETHBLACK ETHHISP Source SS df MS Model Residual 28.3500672 120.782971 4 495 7.08751679 .244006002 Number of obs FC 4, 495) Prob > F R-squared Adj R-squared = Root MSE 500 29.05 0.0000 0.1901 0.1836 .49397 Total 1 149.133038 499 .298863804 LGEARNI Coef. Std. Err. t P>t! [95% Conf. Interval] SI ΕΧΡΙ ETHBLACK ETHHISP cons .096657 .0434691 -.1072735 .0799981 1.072732 .0094076 0091028 .0718268 .0697128 1799187 10.27 4.78 -1.49 1.15 5.96 0.000 0.000 0.136 0.252 0.000 .0781732 .0255843 -.2483966 -.0569714 .7192336 .1151408 .0613539 .0338495 .2169676 1.426231 reg LGEARN S EXP ETHBLACK ETHHISP if MALE==1 Source 1 SS df MS Model Residual 16.4512168 58.8502579 4 245 4.11280419 .240205134 Number of obs = F 4, 245) Prob > F R-squared Adj R-squared Root MSE 250 17.12 0.0000 0.2185 0.2057 .49011 Total 1 75.3014746 249 .302415561 LGEARNI Coef. Std. Err. t P>It! [95% Conf. Interval] S EXPI ETHBLACK ETHHISP cons 1 .1076324 0507116 -.0096302 .1223439 9470023 .013383 0124896 .1066653 .104507 .248596 8.04 4.06 -0.09 1.17 3.81 0.000 0.000 0.928 0.243 0.000 .081272 .0261109 -.2197281 -.0835029 .4573443 .1339928 0753123 .2004678 .3281907 1.43666

reg LGEARN S EXP ETHBLACK ETHHISP if MALE==0 Source] SS df MS = = = Model Residual 15.1996261 57.1106317 4 245 3.79990652 233104619 Number of obs FC 4, 245) Prob > F R-squared Adj R-squared Root MSE 250 16.30 0.0000 0.2102 0.1973 .48281 = Total | 72.3102578 249 .290402642 = LGEARN I Coef. Std. Err. t P>t! [95% Conf. Interval] . ST EXP | ETHBLACK | ETHHISP cons | .1000626 .0362708 -.147189 0871414 .9802464 .0133134 0129693 .0948903 .0921421 .2596141 . 7.52 2.80 -1.55 0.95 3.78 0.000 0.006 0.122 0.345 0.000 .0738393 .0107253 -.3340938 -.0943503 .468886 .126286 .0618164 .0397158 .2686332 1.491607 Which of the following statements is correct? O We cannot reject the null hypothesis at 5% significance level, and conclude that we should use separate regressions to estimate the wage equation for males and females. We reject the null hypothesis at 5% significance level, and conclude that we should use a pooled regression to estimate the wage equation. O We cannot reject the null hypothesis at 5% significance level, and conclude that we should use a pooled regression to estimate the wage equation. We reject the null hypothesis at 5% significance level, and conclude that we should use separate regressions to estimate the wage equation for males and females.