With foxnewsdata, regress r_vote_share_ch on foxnews, male_ch, hisp_ch, black_ch, coll_ch, pop_ch, unemp_ch, group3, and group4. Restrict the sample to towns with votes1996 ≥ 2000 and weight the model...

With foxnewsdata, regress r_vote_share_ch on foxnews, male_ch, hisp_ch, black_ch, coll_ch, pop_ch, unemp_ch, group3, and group4. Restrict the sample to towns with votes1996 ≥ 2000 and weight the model by votes1996. Determine those observations with the highest-5% and lowest-5% residuals. What states do the observations from each of those groups tend to be from, relative to the sample? (Make sure to restrict the sample to those with votes1996 ≥ 2000 for the full sample and the highest- and lowest-residual groups.)