Wash Monthly rankings out

Be careful of confirmation bias.

Or apply ED or SCEA. One-and-done. How many Stanford’s SCEA’s would use H as a ‘backup’? Or vice versa?

For the most part, cross-admit figures tend to track admit rates. And there’s a rational explanation for it. Applicants whose first choice is Harvard might also apply to Notre Dame as a backup. But students whose first choice is Notre Dame are much less likely to apply to Harvard as a backup—they’ll instead apply to schools with admit rates similar to or higher than Notre Dame, figuring if they don’t get admitted to Notre Dame, they’re unlikely to be admitted to schools with even lower admit rates. So the Harvard-Notre Dame cross-admit pool is skewed toward applicants who preferred Harvard from the outset, and many (most?) of those who preferred Notre Dame aren’t counted in Parchment-style cross-admit match-ups. Not surprisingly, Harvard wins the Parchment cross-admit match-up, 72% to 28%.

Within the state of Michigan, the University of Michigan and Michigan State both have intensely loyal, often multi-generational followings. Many students genuinely prefer Michigan State notwithstanding Michigan’s stronger academic reputation and rankings. But again, because Michigan has a much lower admit rate, in-state applicants whose first choice is Michigan might also apply to Michigan State as a backup, but fewer of those whose first choice is Michigan State will apply to Michigan as a backup, figuring if they’re not admitted to Michigan State they should look to schools of similar or lower selectivity. So Michigan swamps Michigan State in the Parchment cross-admit match-up, 85% to 15%. I’m quite confident, however, that if you talk to anyone in the state of Michigan, they’ll tell you that lopsided result is not an accurate reflection of overall student preferences among Michigan residents.

If you love your state school and you feel no compulsion or desire to apply to a top ranked private colleges, there is going to be no head to head choices between that state school and the private college. That comes into play only when you got into two colleges and you are choosing between them. Again, I am pretty certain that Parchment head to head stats for top 20 schools are pretty accurate. I think it reflects real preferences and is helpful. If your number one choice was UC Irvine, it would make a little sense to apply to Harvard, but this does not mean this person would have preferred UCI over Harvard had she gotten into both. My assumption of preference comes into play only after you actually got into both UCI and Harvard. But like I said, head to head stats might not be helpful for less similar colleges.

For example, if you got into Stanford REA and never applied to Harvard, then there is going to be no head to head decision between Stanford and Harvard. Isn’t that the way Parchment does head to head?

What (statistical methodology) do you base this on?