Back in May I shared a projected ranking, and in December I shared an updated projected ranking. And perhaps there’s no better place to begin than my introspective post, “Could Yale Law really be projected to drop in the next USNWR rankings?”:
A rankings projection requires me to put myself “out there” in a public way. I project the rankings and then later find whether my projections are accurate. My projections for the Spring 2026 release of the rankings put Yale around 5th. This, by far, has led to the most disbelief (and even outrage).
To be clear, I could be wrong. I projected Yale to drop to 3d last year, and it instead stuck at 1st (in a tie). This is not so remarkable, because, as I emphasize, very slight errors that I make in the employment stats, in rounding, in what other schools do, and the like can alter the positioning of schools in relation to one another. Because the rankings are now so compressed, they are increasingly volatile and more responsive to even modest changes in some data. 3d-to-1st is really a rounding error coupled with a tie.
But 5th-to-1st would be a more dramatic error on my part (barring some change to the methodology). So, why might it happen?
Well, Yale didn’t end up 5th, but it ended up 2d. So the very thing I projected most dramatically was in error. That said, we know the reason. Yale’s peer reputation score—the results of the survey from law school deans and faculty—rose this year from 4.5 to 4.7. Yale’s score had been sagging, in the 4.4 to 4.5 range for a few years, and my projections assumed they’d stay in that range. But that peer score helped them significantly. In my own rankings, I projected them 5th, but if I swapped the 4.5 for a 4.7 (which is one-eighth of the total rankings formula), I’d have projected them 3d—and close enough given that they ended up 2d.
But we’re not just here for Yale. I put myself out there. How did I do? Okay, with room for improvement.
I project all 194 schools, although I only publicly share the top 100ish.
I got 24 schools exactly right (including 16 in the top 50), 70 schools within +/-1 spot, and 109 within +/-2 spots.
On the whole, that’s pretty good. As I mentioned, very small differences in things like peer reputation score or lawyer/judges reputation score, along with small mathematical or data errors, can jostle schools around, and given how close the schools are, moving from one tied position to another can move many schools around. And those differences tend to grow the lower down the rankings one moves. So 109 within +/-2 spots is pretty good.
When I back fill with the actual reputation scores, 149 of the 194 schools improved or stayed the same in their rankings projections. I would have had 40 exactly right, 72 +/-1, and 125 +/-2.
Yale was probably my biggest miss, but pulling it out of 1st place was still a noteworthy projection. I also underestimated Washington University (13, not 17), Boston College (20, not 23), BYU (24, not 30), and Wisconsin (26, not 30). I underestimated both Wisconsin schools (including Marquette) once again, a sign I am doing something wrong with bar passage statistics.
Schools I overestimated included USC (26, not 23), Wake (30, not 26), Florida (34, not 29), and Florida State (34, not 30).
As usual, I’ll try to dig back into the rankings to see where I went amiss.
One small (and good) methodology change: USNWR is now giving greater weight to the verbal score than the quantitative score in its equivalency of the GRE to LSAT scores.
And as for the rankings themselves? They are, as usual, noisy without a lot of substance behind them. Some things are, of course, interesting. Harvard continues to sag under the new methodology, and to a lesser extend Yale, calling into question the “endgame” of the original “boycott” that instigated the new methodology.
The new USNWR employment metric is heavily weighted in the ranking and highly manipulable by schools, as I’ve highlighted in trying to assess employment categories. Many changes arise solely because of year-to-year vagaries in bar passage or employment.
Prospective law students understandably and increasingly rely on a variety of alternative metrics, and the granular employment data is also helpful for efforts to distinguish between schools in meaningful ways.
And perhaps the most valuable heuristic of movement in the USNWR law school rankings? It’s a symbol of law school management. The rankings are simply something to be managed. If a school is managing the rankings well or poorly, one can observe and attempt to assess why that might be the case.
Honest question: do you think your projected rankings may have influenced YLS’s peer scores?