How Money Rated the Best Graduate School Programs

It’s no secret that earning a college degree in America is an expensive endeavor. Graduate degrees, in particular, often leave students with large amounts of student debt. A recent report from the U.S. Department of Education shows that graduate school borrowing has soared since 2000 — the average debt load now sits at $66,000 — even though the typical earnings premium that comes with a graduate degree has been stagnant.

With that in mind, it’s more important than ever for prospective students to approach their graduate school search with return on investment in mind. To help students look at graduate programs through the lens of value, Money partnered with College Factual to produce our inaugural ranking of post-baccalaureate programs. Here’s how we did it.

Table of contents:

• Ranking methodology, in brief
• General methodology notes
• Quality score, in detail
• Affordability score, in detail
• Value score, in detail

Graduate school programs ranking methodology, in brief

Initial programs list

We started our analysis with a list of all schools in the United States that reported degree completions in 2022 for the relevant program and award level. We used completions for the Classification of Instructional Program (CIP) codes listed below, as reported in the federal government’s Integrated Postsecondary Education Data System (IPEDS).

CIP codes are organized into three levels: 2-digit umbrellas that cover more granular 4-digit and 6-digit program classifications. We tallied all the completions that fell under the codes list above.

From there, we filtered the list to include only the following:

• Must have more than 20 students in the degree program(s).
• Must be a public or private, not-for-profit school (for-profit schools excluded).
• Must be a regionally accredited school.
• Must not be a mostly online school, defined as a school with 80% or more of its students entirely online.
• Must have sufficient, reliable data to be analyzed.

Ranking factors summary

We defined a good value as a program where you will receive a quality education for a low cost, in relation to what you are getting. Broadly, that means our ratings look at the quality of a program against its affordability.

Quality score

For our assessment of the quality of a program, we grouped our data into seven high-level categories. These category scores were then combined to calculate an overall quality score.

The following table lists the categories, their relative importance to the overall ranking, and the question they aimed to answer. (You can find more details on these in the in-depth methodology below.)

Affordability score

In an effort to highlight where your money will go farthest, our rankings seek to identify quality programs that cost less than their similar quality peers. We calculated the cost of attending the school for that program, including tuition and fees, living expenses, books and supplies, and any other costs the school reports.

For tuition and fees, we used the in-state tuition figure reported by the school. These costs were then totalled and scored based on their affordability.

Value score

Since value is a mix of quality and affordability, we multiplied those scores to create a new value score. We put double the weight on quality to ensure quality schools are favored over simply inexpensive ones.

The programs were then grouped into half-star ratings from 0 to 5 to form the star rating you see. Our final lists feature all schools with a 4 or higher, indicating that they are among the best values.

General methodology details

Aggregating hierarchical data

Several of the program-specific measures we use are reported by either IPEDS or the Department of Education’s College Scorecard, arranged by CIP code. This Classification of Instructional Programs is organized in a three-level hierarchy to categorize educational programs at a school. Much of this data is reported at either the second or third level of this hierarchy — these are the more narrowly defined program classifications that are grouped underneath a larger umbrella at the first level.

All of our ranking lists are programs that are associated with CIP codes at the first or second level, commonly called 2-digit and 4-digit CIPs. To get relevant measures at each higher level, we take a weighted average of the values at all the levels below it.

Multi-cohort averages by year

Much of our data, including student loan default rates, along with debt repayment rates and earnings, is reported for multiple cohorts of students in a single year’s data release. In these cases we leveraged all the years’ worth of data and created a weighted average of the combined cohorts as our final measure. This creates a robust measure that is not subject to fluctuations by individual year.

To do this we first calculated a normalized score for each cohort so we can assess how each school did relative to other schools for that cohort. We then took a weighted average of those scores. This has the benefit of isolating each school’s relative performance within a given year so external events (ie. COVID) don’t have an outsized impact across years.

Broader school- and program-related factors

For each of our quality factors, we look first to the measure that directly corresponds to the college, major and award level in question. This is given the most weight.

However, we think that all else being equal, a school that has a strong broader graduate school offering, and/or undergraduate school offering in that same major, has an advantage and demonstrates it’s a real center of excellence for that major. We therefore also look at how the school performs in a particular major for all graduate-level programs, as well as for the school as a whole. While this is given less importance than program-specific data, it serves as an additional indicator of strength and an alternative measure in cases where we lack more granular degree-level data.

For some measures we also look at performance for the school as a whole across all majors. These factors are always given a low weight relative to the other major-specific factors.

Score calculation

Each of the “scores” mentioned are typically a normalized number from 0 to 1. This allows us to compare data in different forms, like dollars and percentages, on a single scale.

To transform a score into a star indicator, we use k-means clustering to group the scores into half star increments from 0 to 5.

Missing data

There are a few data points that we required for inclusion, including regional accreditation and some program level accreditations. We also require some level of major-specific earnings and employment data given how heavily those are weighted.

If we are unable to calculate a score for one of the 7 high-level factors in the quality section, that school will receive 0 points for that factor and the school’s final score will be diminished. However, we do not outright exclude the school, because there are cases where even with that drawback a school is so good in the other factors that it can outrank its peers.

When it comes to calculating the individual data points that go into the 7 high-level scores, we only score data we have. If a data point is missing then we effectively ignore that piece when calculating the score for the related high-level factor. If a school were to be missing all of the sub-factors for a particular high-level score, it would score zero for that factor, which would hurt its rankings as noted above.

Quality ranking factors, in detail

Each of the 7 high-level factor scores previously noted are themselves a combination of multiple data points, which we cover in more detail below. Educational quality is difficult to measure on a large scale, as even nationally used entrance exams measure where a student begins, not what they learn while enrolled. And while graduation rates are the most commonly used indicator of quality at the undergraduate level, program-specific completion rates are not available at the graduate level. While they are imperfect, we feel the factors below are the best available data to get an idea of a program’s quality.

We’ve noted the importance of each broad section in parentheses below, based on the relative weight that factor carries in our scoring.

Accreditation score (up to 2x weight)

Regional accreditation is considered the standard bearer for quality assurance in higher education, so we require all schools on our lists to be regionally accredited at the institutional level.

We made an exception for a few law schools that were entirely focused on law and were accredited only by the American Bar Association, the relevant accrediting body for that program. Due to their narrow focus on law and being sufficiently accredited for that program, we thought it reasonable to include them.

All of the programs we ranked also have at least one program-specific accrediting body. For some fields, getting a degree that is not accredited by the relevant topical accrediting body would be a serious disadvantage. We therefore required program-specific accreditation for inclusion on seven of our nine lists. For these programs, accreditation carries no additional weight.

For two of our program lists, we chose not to force program-specific accreditation as a requirement for inclusion, as several well-known programs were not covered by the corresponding accrediting agency. We still believe accreditation is a mark of quality, so those schools that had a program-specific accreditation received positive points.

The table below lists each program area covered, the relevant accrediting agencies we used and whether being accredited by one was a requirement for inclusion.

Most of the accreditation data was obtained from the Database of Accredited

Postsecondary Institutions and Programs. Where that was lacking, we obtained the accreditation data directly from the corresponding accrediting agency.

For computer science, we gave credit if the school was accredited by ABET for any computer related program, including those related to:

• Computer Science
• Computer Software and Apps
• Computer Programming
• Computer Information Systems (CIS)
• Computer Engineering

Focus score (2x)

A program that is very small compared to the overall size of the school may be less likely to garner the resources needed for that program to be a success. A school that graduates a significant number of students in that program and major, though, is demonstrating an area of concentration that is more likely to come with resources.

Our focus score uses the following factors.

Demand score (1x)

A school that is in high demand in the marketplace of higher education is a mark of that program’s success. While larger doesn’t necessarily mean better, a sufficiently large program may be more likely to have adequate resources dedicated to it, and the pool of fellow alumni provides greater networking opportunities.

Our demand score uses the following factors:

Peer score (1x or 4x, depending on the program)

A diverse student body that excels academically can be an advantage while you’re enrolled, and also after you’ve left school. You’ll learn alongside high-achieving peers, and then you’ll be able to tap those connections while in the workforce.

For programs where we had more robust data because there is a specific entrance example that is typically used for admittance, this overall factor was weighted higher. (These programs were law, business and medicine.)

Our peer score includes the following measures:

Debt Score (4x)

Student loans are a critical resource that opens up access to graduate school for students who otherwise may not be able to attend. But while taking on debt is often necessary, we believe quality programs should not burden students with an unaffordable amount of debt, as measured by typical debt loads, default rates and repayment rates.

Our debt score looks at that following factors to determine this:

The default rate data by college is obtained from the Department of Education’s Federal Student Aid (FSA) Cohort Default Rate (CDR) data. The figure we use is the weighted average of three different annual cohorts, based on the most recently available default rate data.

The rest of the data for this factor comes from the Department of Education’s College Scorecard. The final default rate data is calculated as the weighted average of each of the four years of cohort data reported.

For the repayment rates, we take the weighted average of four cohorts of borrowers: those who graduated 1, 3, 5 and 7 years prior.

Employment Score (8x)

With our employment score, we seek to get an estimate of your likelihood of getting a job after graduation. We used the Department of Education’s College Scorecard data for this.

The employment rate is calculated using two figures: the number of graduates reported to be working and not enrolled, and the number of graduates reported to be not working, nor enrolled elsewhere.

The employment score is calculated using the following factors:

Earnings Score (16x)

Our earnings score measures how much you are likely to make once you do get a job. A higher earnings rate indicates a degree that is in demand by employers and results in a greater return on your investment. We recognize that, particularly in some fields, not everyone earns a graduate degree with the expectation of a significant earnings boost. But we believe a graduate degree should, at the minimum, lead to a job with livable wages, so we also look at how likely students are to go on to earn low salaries.

The earnings and poverty rate data we use are obtained from the Education Department’s College Scorecard data, and both represent a weighted average of the two cohort years provided. The earnings data indicates the overall median earnings, while the poverty rate is a measure of the number of graduates whose earnings are less than 150% of the poverty line for that year.

This score is made up of the following specific factors:

Affordability score, in detail

The following costs were considered when calculating our affordability score:

• Tuition and fees: Graduate level tuition and fees as reported by IPEDS. For law and medical rankings, the program specific costs reported by IPEDS are used. Business school costs were sourced by our research team, typically from the school’s website.
• Living expenses: We used the average of the off-campus, not-living-with-family estimate and the on-campus estimate, as reported to IPEDS. Where only one was reported, that was used, and if neither existed, we set this to the average reported for all schools, which was $15,741.
• Books and supplies: Books and supplies, as reported to IPEDS.
• Other expenses: Any other expenses reported by the school.

For tuition and fees, we used in-state rates for public institutions. We recognize this makes programs at public colleges appear more affordable than they may actually be for students who are enrolling from out of state. But there is no publicly available enrollment data that captures the residence status of graduate students, which we would need to create a cost that more accurately balances in-state and out-of-state prices.

Value score, in detail

Before calculating our final score, we removed all programs that scored in the bottom 20% for quality. We then multiplied the overall quality score by the affordability score to create a value score, which is what our final rating is driven by.

We use two parts quality to one part affordability for this calculation, which serves as an additional measure to ensure quality schools are favored over ones that are simply inexpensive.

The programs are then grouped into half-star ratings from 0 to 5, to form the star rating you see. Our final lists feature all schools with a score of 4 or higher.

Questions about our methodology? Email us at college@money.com