<p>Like barrons, I’m confused by that second sentence, as well. I interpret that first sentence to mean that graduation rates from elite colleges are 1.06x those of grad rates from non-elite colleges. The second sentence seems to mean that equivalent students at the less selective colleges have a graduation rate 1/4 that of the students at the elite institutions. </p>
<p>To try to figure out what was going on, I went and looked at the paper itself, rather than the UM article. The first thing I learned is that I really need to learn more statistics. 
Maybe some knowledgeable person could go and take a look at Table/Figure 4 in the paper?</p>
<p>One interesting bit that wasn’t in the posted summary: “These results indicate that it matters who is obtaining an elite college education. While an elite college education may matter little in terms of career effects for those who actually attend an elite school, other college attendees with a lower propensity to attend may have benefited considerably had they attended an elite school.”</p>
<p>The authors return to this point at the very end of the paper: “Are there students, however, for which an elite college education would actually result in socioeconomic gains? We find that while students that actually attended elite schools would likely do just as well across their careers had they attended non-elite schools, randomly selected college attendees may have reaped significant rewards had they attended elite schools. This finding poses an interesting implication for the process of social stratification in American society.”</p>
<p>If I’m reading this correctly, it means that the biggest gain to individual students came when those who were least likely to attend an elite school (with likelihood being a combination of various socioeconomic factors) attended elite schools. Hoo boy! Imagine what fun someone could have dropping this little nugget in the Ivy forums. 
</p>
<p>It’s also interesting to read the summary of previous research on this subject. The results are mixed: sometimes the advantage goes to the elites, sometimes not. Researchers keep trying different methods of studying the issue. I imagine it’s hard to do a thorough longitudinal study on the topic when the environment of college admissions keeps changing. Certainly male graduates from Wisconsin in 1957 matriculating to college faced a very different admissions environment than that same group in 2008.</p>