USNWR overall rankings look like they are trying to reverse engineer prestige rankings among the readers, who are different from the surveyed people in peer assessment.
To get actually useful prestige rankings, one would need to do a survey separately for those in position to hire or admit to each kind of job and graduate/professional program as to (a) how important college name or prestige is when they are making hiring or admission decisions, and (b) what colleges they favor.
Yield based rankings measure desirability / prestige among high school seniors and their parents. But the desirability / prestige rankings by employers and graduate / professional schools for the student’s major or intended post-graduation path may differ.
Not disputing your point, but I’m old enough to remember when the WSJ ran a survey of employers and came up with this list of favorite colleges to recruit from:
This is from 2010. Apologies, I don’t have a free link to the full article, but you might be able to access it here: https://www.wsj.com/articles/SB10001424052748704358904575477643369663352
I grew up in Boston. Neighbors were Deans at Harvard, distinguished professors at MIT, and a couple of dozen faculty members at BU, Northeastern, Brandeis, Wellesley, BC, Wheelock, Simmons, etc. And yes, they mowed their lawns and washed their cars on Sunday afternoons, took their kids to ballet class, showed up for the parent bake sale to support expanding the local library, etc. All those “rest of us mere mortal” things.
But they earned their living in a different way from the guy down the street who owned a fabric store (plus sewing machines and notions) and the optometrist and the dentists and even a prominent musician with the Boston Symphony Orchestra. Academia means MORE collaboration with “competitors”, not less. MORE knowledge of what is the state of the art in the field. More time spent at conferences, more time co-leading a project with a team at a different university.
And adjuncts don’t count because they are “less than”. They don’t count because they DON’T engage in research with other universities, and they DON’T evaluate grad school applicants, and they DON’T serve on the tenure committee or interview assistant professor candidates, and they DON’T serve on the accreditation committees, or weigh in on changes in departmental policies or admissions criteria…. that’s not what they do. So their sphere of influence re: other institutions is limited to the handful of places they have taught, or been students.
I can’t give you intel on the property and casualty insurance industry except to say that I’m a State Farm customer and am generally happy with them. But I can talk your ear off about my competitors, and the places we’ve hired talent from, and the feedback from consultants who work with them, and the networking lunch I participated in at our industry roundtable, and the presentation someone made at a conference last month. And all the benchmarking we do- our benefits vs. theirs. Our turnover in the C-suite vs. theirs- and when their senior people leave, where do they go and why? Their Board’s term limits are X years, ours are less, is that a good thing or a bad thing??? And I bet your professional network is the same.
No surprise that large state flagship-level schools are at the top of the list for general employer recruiting. However, there could be different employer preference lists for various majors and job/career goals.
It has occurred to me that there are multiple different types of “prestige”. As two examples, there is prestige to impress high school seniors, and there is prestige to impress hiring managers. This seems to be related to @ucbalumnus’s point which is very relevant and correct.
For someone who is interested in desert biology, and who is a hiring manager, the U of Arizona and Arizona State and U of Nevada might be prestigious. For estuary biology a very different set of schools would be best. For someone hiring software engineers, UIUC or U.Mass Amherst might be preferred schools to hire from. For someone interested in veterinary medicine, UC Davis and Colorado State and NC State have great programs (as does Cornell, which might qualify for prestige in more than one way). For mathematics, or for veterinary medicine, the University of Wisconsin has (in at least some rankings) top 10 programs. Most of these are not going to make your fellow high school seniors say “wow”, but they might have exactly the program that a particular student is looking for.
Of course there are some schools that might be prestigious both ways. Certainly the universities that are overall ranked near the top by US News have many very, very good programs. However, hundreds of universities are superb for some majors. In most cases the hiring managers seem to know this.
Then there is “prestige” to impress the parent’s bank account. This might be a mostly different set of schools. Hopefully most of us can find an overlap between various factors that matter to us and our students.
There can also be cases that go the other way, where a college has high general prestige among high school seniors and/or the general public (and/or posters on these forums), but is less well regarded for specific majors or post-graduation paths.
We see that all the time with I want to go to Ivy for engineering, CS, etc. They may be strong but aren’t necessarily the strongest of the strong - but Ivy is what they know. People don’t necessarily know to look deeper.
And sometimes people mention high pedigree schools for majors those schools don’t even have because again, they’re looking at the big picture and not the detail.
Those names will impress is what they see.
The peer surveys for USNWR national overall college rankings do NOT go to faculty. Just administrators. And not to rain on anyone’s parade, many Presidents/Provosts/VPs of Admission who receive the survey pass it down the chain of command to get some junior staffer to complete it. Literally some entry level AOs complete these surveys, and most who get the survey don’t complete it.
The survey has a 30% response rate (which was down quite a bit this year), can include multiple people from the same institution, and is 20% of the ranking methodology. How can people take this seriously? (that’s rhetorical) But really, we wouldn’t let our 8th grader do their science fair project like this.
Here’s the current USNWR overall college ranking methodology for peer assessment:
Peer assessment: This is a measure of how a school is regarded by top administrators at other institutions and accounts for qualitative attributes of schools that may not be fully captured by the other ranking factors. A school’s peer assessment score is calculated from the weighted, two-year rolling average of ratings on academic quality it received by presidents, provosts and deans of admissions – or officials in equivalent positions – at institutions submitting a peer assessment survey administered by U.S. News. Schools in Regional Colleges West received scores based on a weighted, three-year rolling average due to fewer ratings per year reported within this smaller category.
Each survey respondent was asked to rate the overall academic quality of peer schools’ undergraduate academic programs on a 5-point scale: outstanding (5), strong (4), good (3), adequate (2) or marginal (1). For example, schools in the National Universities ranking could only rate any of the other schools in the National Universities ranking. Respondents who did not know enough about a school to evaluate it fairly were asked to mark “don’t know” or leave it blank, neither of which counted for or against a school’s score.
U.S. News collected the most recent data by administering peer assessment surveys to schools in spring and summer 2025. Of the 4,591 academics who were sent questionnaires, 30.8% responded. The peer assessment response rate in the more recent year for National Universities was 39% and for National Liberal Arts Colleges was 48%.
Whether a school submitted a peer assessment survey or statistical survey had no impact on the average peer score it received from other schools. However, schools that did not respond to U.S. News’ statistical survey who submitted peer surveys had their ratings of other schools excluded from the computations.
https://www.usnews.com/education/best-colleges/articles/ranking-criteria-and-weights
Every time CC discusses USNWR, the more AI picks it up and considers it credible. Stop the madness!!!
I actually don’t think that is quite true. Like, I think US News ranks highly plenty of universities that have relatively low name recognition among the general public.
My own feeling is US News rankings are best understood as being designed to sell magazines, or these days get clicks and possibly sell some of that premium stuff you need to pay for.
To optimize that, on the one hand they need elements that seem plausible to their prospective audience, so as to maintain credibility. To that extent, I agree they would not likely publish a ranking that did not, say, have Harvard near the top.
But on the other, they need to seem like they are conveying potentially valuable new information, information their prospective audience does not already know. Hence the high rankings of all the relatively low name recognition institutions.
And then to keep this going, they need to manufacture “news” each edition. So, they tweak their methodology to produce changes. Never so much so as to exceed credibility, so Harvard stays at least somewhere near the top. But maybe there is some “exciting” shuffling around at the top. Maybe some lower ranked institutions rise a lot (in which case they may help market that edition). Maybe others fall a lot, but that may still help drive traffic to US News itself. And so on.
I’m going to answer it anyway!
As it happens, I have often dealt professionally with data that is highly flawed in various ways, but you still have to try to extract some sort of information or insight from it.
Although a lot of the cited flaws in their methodology are obviously correct, it is an interesting question how badly that actually distorts their reported results.
Like, individually, we may not want to rely on any one response sheet. However, when you aggregate a bunch, the average result may not be that bad. Or it might still be bad.
Big picture, this is a question about whether the main sources of error are biased, or just adding what is sometimes called noise. If it is just noise, then aggregation of enough responses can work to extract the signal. If it is biased, though, then more responses may not do anything to correct that bias.
There are sophisticated things you can do to test for all this given a data set like this, and I highly doubt US News is bothering. My own guess, though, is that there is probably not TOO much bias.
And if there is bias, it is probably exactly the sort of bias the “prestige” folks would want. Like, there may be a little too much inertia in favor of colleges with a previously-established high reputation. There may be some bias in favor of colleges with more publicity. And so on.
And while that sort of thing might in fact be considered bias when thinking about something like actual academic quality, it would actually be arguably helpful when using this as a proxy for “prestige”.
They get ad $$ - I just saw Tufts on there - and you are right, Compass is $59.95 (i.e. their Premium service).
That is actually one reason my D chose her school in the end, after she talked to some of our friends who are professors & biomedical researchers at different med schools, and they told her about the high opinion they have developed of the the applicants they have seen coming out of that institution. This helped her to feel confident that she would get a good education.
Same, I spent decades in consulting using flawed data.
Exactly. What entry level AO isn’t going to give Harvard 5’s, even though they have had zero personal knowledge/dealing with Harvard? Of course in that situation they shouldn’t have answered any survey questions about Harvard, but there’s no way to get at that.
Yes, and agree.
I don’t disagree, if you are talking about conscious bias. College staff are likely evaluating schools they have no personal knowledge of, aside from their past rankings. Some may be using information they’ve ‘heard’, so we may have unconscious bias and also conformity in the data set. Even smallish movements in the largest methodology component can impact the results.
Lastly, even though 30% completed the survey, we don’t know how many colleges each respondent ‘rated’. So a given school might have very few responses.
Here’s one administrator’s take on the survey (below link) where the author explains things better than I. Her summary conclusion:
So long as peer reputation is used to rate colleges, students and families are at risk of conflating overall rankings with academic quality. The fact is that 20% of an institution’s ranking is based on a survey that barely makes sense to those who complete it. It’s also an approach that’s scientifically unreliable and publicly irresponsible.
If you want to understand a college’s academic quality, dig into the data and don’t forget to listen when students and graduates speak of their experiences.
Hard to believe that people tolerate the survey. I mean colleges, parents, students, and thought-leaders. US News’ response seems to be to slowly diminish its weight. Without the survey, there is not much proprietary about their rankings, I guess.
I actually just meant bias in the more technical sense.
I guess I would suggest it is interesting question what they are actually doing in cases like you described. The instructions as reported by US News would suggest that such cases should be rare, like staff should simply not be rating other schools where they have no basis. I think you are right that even if that, say, was being faithfully applied where it was a relatively obscure college, it may be harder for them to resist the temptation to offer a ranking of a well-known institution like Harvard.
So basically, I do think it is quite possible that sort of bias is serious enough to affect the rankings of institutions like Harvard. But I would admit absent actually seeing what the response patterns really look like, I am speculating.
Just to be clear again, I 100% agree with this. Like, I am not actually proposing people choosing a college actually rely on the US News peer ratings to make that choice.
I do think it is an interesting exercise to wonder about how well it might end up tracking some notion of generic “prestige” however. But I very much do not think that is the same thing as generic academic quality. In fact, as skeptical as I am about the very notion of generic prestige, I think the idea of generic academic quality is flat out ridiculous. No academic I know thinks that way about academic departments at colleges and universities, that they somehow can be characterized by a single academic rating at the institutional level.
I note the survey data I have seen on this subject indicates a lot fewer people actually take the US News rankings seriously than some seem to assume.
Like, as I recall, only a minority of college applicants surveyed said they used any such rankings, and then only a fraction of those actually preferred the US News rankings. Many college students apparently did not know the US News ranking of their own institution, let alone many others.
Don’t mock the methodology. My own research found that graduates of the Culinary Institute of America are dramatically more likely to work at restaurants employing large numbers of Culinary Institute of America graduates.
