@suzyq7 - US News is not very transparent. What they do reveal is that high school rank counts as 3.25% of the total ranking and that the output of the formulas is translated to a 100 point scale - hence the 3.25 points that is mentioned in some earlier posts.
The straight class rank number is used, which can be deceptive. I assume this is why Lee was phasing out its use in the Tufts literature. Colleges that enroll students from weak high schools will do better than those that enroll students from strong high schools. What is interesting is that it tends to be the more competitive high schools that don’t report class rank (for fear of it hurting their student’s admissions chances at colleges trying to optimize rankings), so selective colleges may have lower percentages of applicants that report class rank.
For context, Tufts reported 90-91% in the top 10% in the years leading up to the year that the metric was dropped which is reasonably high. Most recently, Wake Forest reported 77-78%, Georgetown reported 89-91% and Harvard reported 95%. It is hard to attach meaning to the statistic though, because there is no way to know the competitiveness of the respective high schools that the applicants attended.
What US News does not reveal is how the score for each category is normalized before they combine the categories together (i.e. how does a 100 point difference in SAT score compare to a 1 point difference in class rank ?) Standard statistical practices are not always used. Back in the 90’s there was a big uproar when it was discovered that they used a logarithmic multiplier for student spending in order to keep HYP at the top of the rankings.
@aegis400 It should also be noted that US News does not allow a school to drop out of the rankings. If data is not supplied, then they make up a value (that tends to be punitive) and rank the school based on the made up value (this is what happened to Reed when they stopped supplying data).
Technically, since we don’t know what data value US News supplied on behalf of Tufts, and we don’t know the procedure they use to normalize it, we cannot know for sure what the exact impact on the rankings is.