If you go to those five schools, the percentage of minorities and the percentage of women is X and Y. Let's say it's 3% and 10%. If those companies hi

If you go to those five schools, the percentage of minorities and the percentage of women is X and Y. Let's say it's 3% and 10%. If those companies hire only 10% of the people they interview, then the number [of minority hires] you get is zero. Right? It's zero. And to the extent that they are unwilling to go outside of those five schools, it's going to stay zero. No, it's not zero, it should be a number that corresponds to 3% and 10%, respectively. Yes, those numbers suck, but it helps to get the math right. If the number is only 3% and 10% from those schools, then the problem starts at the funnel that feeds those schools [1]. More minorities and women need to apply to those schools, so that the percentage goes up. Should companies look beyond those five schools? Yes, but... I'm not going to speak for any other curriculums, but the truth is that most computer science curriculums in the US are absolute garbage. Many CS programs are just java schools [0]. What these companies have done is identified the schools whose programs have a reputation for not being java schools. Historically, you could figure out which programs are not java schools by looking for those that use books like SICP, K&R, CLRS, TaoCP, PAIP, AIMA, EOPL, the dragon book, etc. as the basis for their CS curriculum. If there are schools out there with a much greater percentage of minorities and women that use such books for foundational courses, then you should suggest those schools as places to recruit from. Simply put, the criteria for spending time, effort and money to recruit from other schools needs to be based primarily on the rigor of the curriculum, with demographics of the student body being a secondary goal. That all being said, a lot of the best software developers I've met that didn't go to those top schools didn't graduate with a degree in computer science at all. Pretty much all the others are auto-didacts. Assuming this is generalizable, the only actionable you can extract from it is not to recruit based on school at all. [0] http://www.joelonsoftware.com/articles/ThePerilsofJavaSchools.html http://www.joelonsoftware.com/articles/ThePerilsofJavaSchool... [1] https://news.ycombinator.com/item?id=8593277 https://news.ycombinator.com/item?id=8593277 [2] FWIW, I didn't graduate from a top school (NCSU) and double majored in textile technology and apparel mgmt and psychology, so I'm not even advocating a position I myself gain from.