From our replication of Biddle and Hamermesh's (JPE 1990)Links to an external site. Tables 3 and 4 in the programming HW last week, the standard errors in the paper likely did not address heteroscedasticity. Do so for Model 1 of Table 3 (i.e., estimate [sleep+naps] for the whole sample). Has your estimated coefficient on the main dependent variable (totwrk) changed? Has its standard error changed? If yes for either, state the direction of the change (i.e., increase vs. decrease)? For your convenience, the dataset can be accessed here:SLEEP75.dta.