Some Info for the assignment

will have 7 questions – 4 will ask you to run a short Stata regression of the form:

. regress y x1 x2 x3

or . pwcorr x1 x2 x3

After running the regression or correlation, you will be able to answer the question based on the Stata output.

An example would be something like:

Given the following data, are y and rationing related to c. Where y = income c = consumption.

Run the regression and answer the question based on your Stata output.

Then the data would look like the following:

year y c rationing

1929 4091 3868 0

1930 3727 3569 0

1931 3534 3400 0

1932 3043 3081 0

1933 2950 3013 0

1934 3100 3088 0

1935 3359 3236 0

1936 3738 3523 0

1937 3836 3628 0

1938 3557 3517 0

……………………….

The following regression would look like:

. reg c y rationing

Source | SS df MS Number of obs = 32

————-+———————————- F(2, 29) = 213.37

Model | 18762774.9 2 9381387.45 Prob > F = 0.0000

Residual | 1275062.98 29 43967.6891 R-squared = 0.9364

————-+———————————- Adj R-squared = 0.9320

Total | 20037837.9 31 646381.867 Root MSE = 209.68

——————————————————————————

c | Coef. Std. Err. t P>|t| [95% Conf. Interval]

————-+—————————————————————-

y | .8086736 .0391749 20.64 0.000 .7285519 .8887952

rationing | -610.0382 114.895 -5.31 0.000 -845.025 -375.0515

_cons | 557.0569 185.5583 3.00 0.005 177.5476 936.5662

——————————————————————————

The important thing to note is whether or not the p-value of each of the independent variables indicates significance. Recall if the p-value is less than 0.05, the variable to which the p-value is attached is considered significant.

So for the above regression, y, rationing and the intercept _cons all have p-values less than 0.05 and thus are significant variables which have an effect on c – consumption.

Yes to the linear regression: You will only need to run: . regress Y X1 X2 X3 ….

and . pwcorr.

Chapter 12 is the key in Daniels and Minot – correlation and regression.

3 questions on correlation

4 questions on regression