Durbin_Watson – The Durbin-Watson test statistic tests the null hypothesis that the residuals from an
1) Use the data in Data 4-3 and
generate the following variables:
lph = log (housing/population)
lpcgnp = log (gnp/population)
lr = log (intrate)
Estimate the model below using OLS:
= ? + ?1 lpcgnp + ?2 lr + ?t
Use your regression results to calculate the
residuals. (print the regression results but no need to print out your data).
Take a lag of the residuals.
Check for first order auto – correlation with
the residuals. Use the Durbin Watson test. Does the test show serial
the data in7-20 andrun the
regression below in Gretl. Make sure you open the data
as a time-series. (I know this is a cross section, but as stated in class,
there might be serial correlation based on how the data was organized.)
salary= ? + ?1height+
?2weight+ ?3points +?
Run the regression and print out the regression result and include a
copy on the homework. Check the
Durbin-Watson statistic. Do you have
serial correlation? Which kind?
Run a Whiteâs test to
check for heteroskedasticity. Based on
the Whiteâs Test, does your model have heteroskedasticity? Explain
out a copy of the Whiteâs test and the Durbin Watson test result and attach it
in your homework.