The final result is a toss-up in a study comparing how for-profit and nonprofit colleges stack up in job market returns of their certificates and associate degrees. That finding is a big shift from the unflattering conclusion about for-profits reached in an earlier version of the paper.
The National Bureau of Economic Research released both iterations of the study , which was written by Kevin Lang and Russell Weinstein, two professors of economics at Boston University. A “working” version  of the research went live last summer. It found “large, statistically significant benefits from obtaining certificates/degrees from public and not-for-profit but not from for-profit institutions.”
Working papers, however, are works in progress. Researchers get feedback during the peer review process that can lead to revisions and improvements. That’s what happened with Lang and Weinstein’s study.
The two economists, who were not available for comment, apparently tweaked their methodology and came to a different conclusion about the relative value of credentials earned at for-profits.
“We find no statistically significant differential return to certificates or associates degrees between for-profits and not-for-profits,” they wrote in the paper, which was released last month.
Certificate holders from for-profits tended to fare slightly worse in the job market, according to the study, while associate degrees from for-profits were worth slightly more than those from nonprofit institutions. Hence no clear winner emerged.
The revised paper still included some worrisome findings about for-profits. Those colleges are typically more expensive than their nonprofit counterparts, particularly community colleges. For-profits charged an average of $6,300 more in annual tuition for certificate programs, according to the study’s sample, and $6,900 more per year for associate degrees.
“The return on investment is undoubtedly lower at for-profits,” the paper said.
However, the study’s most significant finding, its authors wrote, was the large variation in wages and labor market returns across majors and academic disciplines. Those program-specific comparisons are probably more valuable than comparing wage data at an institutionwide level.
“As policy-makers consider their response to the explosion of student debt,” the study said, “they may want to consider programs within institutions rather than simply the overall performance of the institution.”
Another potential takeaway of the study is the danger of drawing conclusions from working papers about politically charged issues.
Inside Higher Ed wrote about the paper last year. The article  cited several reasons to be cautious about the study’s findings, including hedges from its authors as well as concerns about the research from Anthony Carnevale, director of the Georgetown University Center on Education and the Workforce.
However, as the article noted, the study arrived amid controversy about the relative worth of credentials earned at for-profits. That means it was red meat to some for-profit critics and journalists, caveats notwithstanding.
For example, a headline on a Business Insider article  about the paper blared: “Students Get Absolutely Nothing Out of Attending For-Profit Colleges.”
Apples to Apples?
The study was based on a longitudinal data set from the U.S. Department of Education’s National Center for Education Statistics. It tracked the earnings of 16,680 graduates who first entered college in 2003.
The data were broken out by students who started in certificate and associate degree programs. The study did not distinguish which academic programs institutions offered. It did, however, control its results based on students’ majors and also attempted to factor in the socioeconomic backgrounds of students.
That was no easy task. The authors wrote that the controls they included to try to account for the relative disadvantages of students at for-profits were probably not enough for a fair comparison.
For example, for-profit certificate holders in 2009 earned $6,500 less than their peers from traditional colleges -- $26,059 compared to $32,514. But it might be wrong to draw a conclusion from that finding.
“It is by no means evident, however, that this difference is causal,” the study said. “Compared to those who start at not-for-profit institutions, students starting in certificate programs at for-profit institutions are much more likely to be black, Hispanic, female, younger, single and listed as a dependent at the time they enter college.”
Another caveat from the paper is that many students who enrolled in associate degree programs at nonprofits later transfer to earn a bachelor’s degree from a four-year institution without first earning the two-year degree. That scenario isn’t likely at for-profits, which might help the sector look relatively better in the wages of its associate-degree holders.
Part of the problem with such comparisons, Carnevale said, is that for-profits typically offer different academic programs than do community colleges or nonprofit four-year institutions.
“They’re really in different businesses,” he said. “Where there is overlap, they look comparable.”
Carnevale said certificates from for-profits do tend to lead to smaller salaries than those earned at community colleges. But for-profits offer plenty of different certificates with different results.
Furthermore, data sets are often too limited for apples-to-apples comparisons across sectors, according to Carnevale. And the results of studies that attempt to match up for-profits against nonprofits tend to be disappointingly muddled.
“Without better data, we don’t know the answer,” he said. “Comparing the two is not as fruitful as we all thought.”
Steve Gunderson, president of the Association of Public Sector Colleges and Universities, the for-profit sector's primary trade group, welcomed the revised findings. He said they reinforce the role for-profits play in job training.
"Due to cutbacks at the state and local level, our institutions are the only ones capable of responding to increased demands for a job ready workforce," Gunderson said in a written statement. "As the Education Department prepares to commence negotiated rulemaking, let this revision serve as a reminder for all parties that we should not develop regulations and conclusions based on incomplete, faulty or defective data."