<p>Researchers like Hersch, Krueger & Dale, and Hoxby are trying to identify differences in inputs that correlate (or don’t correlate) with big differences in outcomes. UT Austin and Texas A&M don’t seem to be all that different either on the input or the output side. They are both in Barron’s selectivity category 3, with only a .07 difference in average GPA-W and a 26 pt difference in average SAT-CR. Their average mid-career salary difference is only $2,600 (per Payscale). These are small differences compared to the kinds of spreads that could be considered in the research literature. Rice and Auburn for example have a .54 difference in average GPA-W and a ~135 pt. difference in average SAT-CR. Their average mid-career salaries differ by over $21,000. </p>
<p>Researchers are asking:
given a difference that large in the earnings outcome, how much is attributable to differences in student ability (nature) v. college treatment effects (nurture)? If we control for ability (by considering only students admitted to both kinds of schools, or with similar high school stats, or who wound up in equally elite graduate programs) does the earnings difference disappear? Krueger & Dale say yes (pretty much), Hoxby and Hersch say no.</p>