<p>
</p>
<p>xiggi: the “scores” as you call them (i.e. the coefficients of the various columns in the US News table) can be inferred from the table itself. That US News publishes the relative weights of the different factors is decisive in being able to do this; it increases the accuracy available from the somewhat incomplete data to the point where a reliable result can come out. It is not as simple as running a regression to extract the coefficients, which may be why the other poster’s attempt didn’t work. I can advise if anyone here wants to reverse engineer the US News ratings; my to-do list is too long to code it myself before the next ranking comes out.</p>