<p>Mini, thanks for posting that Winston study. I had heard about it, but never read it. I also agree completely with your UCLA/Berkeley/Dartmouth/Brown comment.</p>
<p>A few comments:</p>
<p>The degree of underrepresentation changes a lot if you look at the SAT figures only: 1% (10%) at a 1420 cutoff, and 4% (40%) at a 1300 cutoff, which I think is probably the realistic cutoff absent significant other factors. Obviously, this is an explanation, not an excuse, and it is a good argument for using the ACT instead of, or alongside, the SAT. It also suggests that there is a regional element to the underrepresentation -- it may not be lower income people who are underrepresented so much as lower income people from certain regions.</p>
<p>Winston's data are a powerful argument for better marketing by elite schools to lower income kids and families. However, one of the marketing challenges may be that the available high-quality low income kids are diffused somewhat in populations of lower-ability kids, especially in rural communities, so that it becomes more difficult to find them, even if they're there. It's easy to send a rep to Andover or Scarsdale High, and you'll find a whole bunch of qualified largely high-income kids. It's easy to send a rep to my kids' school, too, or other urban public magnets, and you'll find a whole bunch of mixed-income qualified kids. It's not so easy to service a bunch of small rural high schools in Nebraska or Mississippi, which may or may not produce a qualified kid or two in any particular year.</p>
<p>I stand by my belief that in my community, the qualified kids are relatively concentrated, but clearly that isn't true elsewhere (maybe not even in NYC, per Sybbie). </p>
<p>Obviously, there are a ton of cultural factors which a marketing effort must address. Somehow, I doubt that eliminating SCEA does a whole lot to address them, but I know that isn't the only thing Harvard (and others) will do.</p>
<p>Finally, there is obviously a significant issue lurking here where race and income level diverge. I don't think it's any secret that the elite colleges care about race more than income, and that most of their marketing to low income students has been to low income URMs. Also, the real scandal in Winston's data is the enormous underrepresentation of kids in the middle income quadrants relative to ability.</p>