We all know that Harvard is #1. Or is it? One helpful step in graduate school selection is researching the rankings for your particular field of interest. To help you get started, we've compiled an overview of the major ranking sources and their methodologies, strengths and weaknesses. Also, a word of caution: don't focus too much on rankings. We're not saying that they're unimportant. All we're saying is that they're just one way to think about school selection. So here's what you need to know about the (imperfect) world of rankings...
U.S. News & World Report
The granddaddy of American university rankings is U.S. News & World Report, a private publishing company. These folks have compiled rankings since 1983 and exert significant influence each year on applicants' school selection. While the publication's ranking methodology has it shortcomings, it establishes a useful basis for comparing the relative attractiveness of hundreds of academic programs. For purposes of graduate school, we'd point you in the direction of two of their rankings: "America's Best Graduate Schools" and "Best Colleges".
America's Best Graduate Schools is U.S. News' ranking of graduate schools. The first thing to note is that their rankings do not follow a uniform methodology for all programs. The five graduate programs with the largest enrollment -- business, education, engineering, law, and medicine -- are ranked using a combination of peer assessment and statistical data. The criteria and weight assigned to each varies by discipline, and a complete explanation can be found on U.S. News' web site. All other programs are ranked entirely based on peer assessment data from graduate school deans, program directors, and/or senior faculty in each field. Basically, U.S. News sends questionnaires asking respondents to rate the "academic quality" of programs in their field on a 5-point scale. Sounds imperfect, doesn't it?
U.S. News has also compiled information to let you view rankings based on sub-fields. For example, if you're interested in History, you can rank programs according to their strength in sub fields like European History or U.S. Colonial History. This is a nice feature, but you should take it with a grain of salt as it's probably less than exact.
Now, let's shift gears and talk about the Best Colleges ranking. This is simply a ranking of undergraduate programs, split into national and regional rankings. At this point, you might be asking yourself: "Why should I care about undergraduate rankings when I'm applying to graduate school?" Well, the reason is that a lot less effort has been put into creating and refining graduate school rankings. This is simply because the market is so much smaller – only a subset of college graduates apply to grad school. So, while an undergraduate ranking is not measuring what you need, it's still (somewhat) helpful because of the "halo" effect. Here's how the undergraduate ranking methodology works. It's based on the weighted average of 9 quantitative and qualitative criteria: faculty resources (20%), retention (20%), peer assessment (15%), student selectivity (15%), financial resources (10%), high school guidance counselor assessment (7.5%), graduation rate (7.5%), and alumni giving (5%). You can imagine how some dimensions, like faculty resources, peer assessment, or financial resources, can carry over into graduate programs.
National Research Council
The National Research Council (NRC) is an arm of the US National Academies. Its mission is to provide elected officials and the public with expert advice on public policy issues based on scientific evidence. Among its initiatives is its periodic publication titled A Data-Based Assessment of Research-Doctorate Programs in the United States. That's just a long-winded way of saying "doctorate rankings". So if you're interested in a doctorate, you should definitely review this. If you're thinking about a master's program, you should still look at it because of the potential "halo" effect. The current report is free after registering on the NRC's web site using your email address. You'll see both PDF and Excel files for download, and we would recommend focusing on the Excel file which has all the meat.
So what's special about the NRC rankings? Well, three characteristics distinguish it from U.S. News' rankings: focus, frequency of releases, and methodology. First, the NRC focuses exclusively on doctorate programs, whereas U.S. News considers both master's and doctorates. Second, they don't publish their rankings annually. Or even every other year. The last report was published in 2010 and is based on (gasp!) data from the 2005-06 academic year. And the last report was published in 1995! The NRC has stated its intention to release rankings every 2-3 years going forward, but that remains to be seen. Third, the NRC produces two rankings based on 20 criteria. The first type of ranking is survey-based ("S") and reflects the importance that faculty attribute to each of these criteria. The second type is regression-based ("R") and evaluates programs based on whether they share similar characteristics with top-tier programs. If that wasn't complicated enough, the NRC splits each "S" and "R" ranking into 5th and 95th percentile rankings. This is best explained through an example. Let's say that Stanford University's 5th percentile "S" ranking is #2 and it's 95th percentile "S" ranking is #8. What this means is that we can say with 90% confidence that Stanford "true" ranking is somewhere between #2 and #8. Exhale.
In addition to these rankings, the NRC produces rankings based on other criteria like department diversity. In doing so, the NRC is seeking to de-emphasize having one ranking system. Instead, they are (intelligently) reinforcing the idea that rankings are imperfect and schools should be examined from different angles.
Here's another nifty resource. Developed by mathematicians at Dartmouth College, PhDs.org is a ranking system for doctorate programs that piggybacks off of the work done at the NCR. Essentially, it lets you assign the importance of each of the NRC criteria, and spits out a custom ranking. You could achieve the same result by manipulating the raw NCR data, but this makes it a whole lot easier. Here's an example. Say that your main goal is to land a job in academia. Well, what you could do is assign an importance value of "5" (the max) to job placement (criteria #1) and academic jobs (criteria #2). The algorithm then spits out rankings of graduate schools. That's it. Easy as pie. Remember to expand all of the criteria, as there are lots of hidden layers.