Economics 521 Spring 2005 Multiple regression homework
The Fair data can be accessed through a link at the end of the course web site:
http://classes.maxwell.syr.edu/ecn521/syllabus.htm
Click on “Fair Vote data”; then go up to the top of the screen and click on “File”; then
click on “Save as”. Save the file “Fair3.xls” on your own drive.
(A) Using the LINEST command, replicate the regression (include a constant in all these
regressions) of Fair’s vote share variable on the book variables: growth rate, party, person, war,
duration, inflation, and goodnews. Remember to first change the vote data to what is in the
book, i.e., round to the nearest 1/10th. Also remember to “zero out” inflation and goodnews for
the years when the war variable takes the value 1. Calculate the t-statistics for each of the
explanatory variables and the errors for each observation. Compare your results with Fair’s
reports on pp. 47 and 48.
(B). Redo part (A) for a data table that differs from this one in that the vote share
variable is expressed not as a percentage but as a fraction, so that each value of that one variable
is one-hundredth the size it is in the original table. Compare the regression coefficients, errors, t-
statistics, and R you get with those you got in part (A).
2
(C). Redo part (A) for a data table that differs from this one in that the PERSON variable
is coded in the opposite way: each 0 is changed to 1 and each 1 is changed to 0. Compare the
regression coefficients, errors, t-statistics, and R you get with those you got in part (A).
2
(D). Redo part (A) for a data table that differs from this one in that there is one new
variable, namely the square of the growth rate (to allow for diminishing returns). You must
create the new column of data and then regress VOTE against 8 variables instead of 7. Compare
the regression coefficients, errors, t-statistics, and R you get with those you got in part (A).
2
(E). Redo part (A) for a data table that differs from this one in that there is one new
variable, the number of games won by the Red Sox in the election year. You must create the new
column of data and then regress VOTE against 8 variables instead of 7. Compare the regression
coefficients, errors, t-statistics, and R you get with those you got in part (A).
2
(F). Redo part (A) for a data table that differs from this one in that there is one new
variable, a dummy variable that takes the value 1 in those election years when the incumbent
party’s candidate was male and 0 for those when the incumbent party’s candidate was female.
You must create the new column of data and then regress VOTE against 8 variables instead of 7.
Compare the regression coefficients, errors, t-statistics, and R you get with those you got in part
2
(A).
(G). Redo part (A) for a data table that differs from this one in that you add the
observations for 2000 and 2004. Compare the regression coefficients, errors, t-statistics, and R2
you get with those you got in part (A).
