# Stata Assignment Harvard Case Solution & Analysis

Question 1

a).

The results of the regression show that the percentage explanation of all the independent variables selected on the total gas production in months 2 to 12 is 37.71%. The probability of the regression model is 0.000, which is lower than the level of significance. Therefore, we can say that this regression model is fitted. If we analyze the coefficients of all the independent variables, then we can see that the drill direction, the length of the fractured section of the well and the log of the production of the gas in the first year less first month is also significant. The relative coefficients are 200498, 215995 and 299365 respectively for these 3 independent variables. The regression results are reported in exhibit 1 in the appendix.

b).

If heteroskedasticity is present in the data, then the results of the regression analysis would be invalidated and it invalidates all the statistical tests of significance, which assume that all the modeling errors are uniform and uncorrelated. Therefore, the errors should not be correlated otherwise it will distort the regression analysis results.

Stata Assignment Harvard Case Solution & Analysis

Question 2

a).

The first test for heteroskedasticity has been performed in stata, which is shown in exhibit 2. The results of this test show a p value of 0.000, which means that there is no constant variance and heteroskedasticity exists in the data.

b).

The second test for heteroskedasticity has been performed in stata, which is shown in exhibit 2. The results of this test show a p value of 0.000 for lgasperfoot and lperf_dist variables, which mean that there is no constant variance and heteroskedasticity exists in the data.

c).

The third test for heteroskedasticity has been performed in stata, which is shown in exhibit 2. The results of this test show a p value of 0.000 for heteroskedasticity, which means that there is no constant variance and heteroskedasticity in the data.

Question 3

Using the predict command we have created the yi variables and the estimated residuals. The regression model has been estimated as shown in exhibit 3. It could be seen that the model has a p-value of 0.000, which again shows that heteroskedasticity is present in the data.

Question 4

The plot is shown in exhibit 4 in the appendix, the fitted lines show that the fitted values are highly correlated between y variable and the x variable. This shows that all the previous regression results are distorted and invalid.

Question 5

a).

The regression model in part one has been re-run using the vce (robust) and vce (Cluster County) commands in the regression model. Through this, we assume the effects of heteroskedasticity have been now removed.

b).

First, with the vce (robust) command, the coefficients remain same as in part one and their significance remains same. However, the F test value for this model is 7117.44. For the second regression model we used the command vce (cluster county), then the coefficients remain same and the regression model is significant.................

This is just a sample partical work. Please place the order on the website to get your own originally done case solution.

Other Similar Case Solutions like