It is always disappointing when those who can’t defend their arguments or “research reports” resort to name calling rather than engaging in an scholarly debate about the limitations of their work. I pointed out several weaknesses in a study recently released by Education Sector, and the response was name calling and character attack by one of its authors. Sadly, I believe that name calling is what you do when you can’t actually defend your position in more intellectually convincing ways. Yes, I am a lobbyist, but I was Princeton’s lobbyist before I was a for-profit lobbyist, and I was a professor and public-policy official before that, so some might say that I can see all sides of this debate perhaps better than almost anyone else.
As a scientist, I grew up in a tradition that prided itself on questioning the data presented by others to see if there might be another interpretation of the results or if, perhaps, there were confounding variables that may not have been considered in the experimental design or conclusions, or even if the study lacked adequate controls. I guess in the aloof land of higher-education think-tankery, some hold themselves beyond reproach and do not wish to engage in dialogue about the possible limitations of their work or alternative interpretations of the findings. Or maybe I’ve just pointed out the weakness in education research, in general, which is that we don’t have access to data that allow us to make legitimate comparisons and tease out the details that well-inform our positions.
Getting back to the point, however, your work has some serious holes in it and it would be a contribution to the field if you were willing to point them out and be honest about them.
You are correct that I was referring to the Graduate Research Survey data in IPEDS, which is, indeed, the reference I thought you were using in the study. In Note 6, for example, the text says only “Author analysis of 150 percent graduation rates for all four-year, for-profit institutions using the Integrated Postsecondary Education Data Survey.” Until just a few years ago, this was the definition used for the Graduation Rate Survey, so reasonable people might agree that it could easily have been understood that the data to which you referring was the GRS data. Perhaps some more precise writing on your part would have prevented the error on mine, but I do admit that I was referring to the GRS data, and therefore withdraw that element of my argument about one of the many weaknesses of the study.
I still believe that your methodology biases the denominator against non-selective institutions because institutions that serve a large number of low-income, non-traditional students do not have graduation rates that mimic those of elite, selective institutions. In other words, I still believe that your denominator is artificially deflated relative to your numerator since non-selective institutions have many more students who borrow than graduate. You made the point in your paper that excluding non-graduates from earlier studies misrepresented the truth about actual debt per credential, but I would argue that your methodology similarly misrepresents the truth for an individual student who might want to know how much he or she might need to borrow to complete a degree. For your study to be accurate and informative, you would have needed to break down borrowing averages by student demographic (dependent versus independent as well as family income level) as well as institutional type (selective, less selective, or non-selective). I do applaud you, however, for pointing out that student borrowing is highly dependent upon taxpayer subsidies and that the more the taxpayer contributes through Pell grants and direct budget appropriations, the less a student is required to borrow.
I’m not defending the lower graduation rates of non-selective institutions, but instead pointing out that for better or worse, the graduation rate among their students is lower, so dividing debt accumulated by all borrowers by the number of graduates artificially inflates the total for debt to degrees at non-selective institutions. All of us—for-profit institutions, community colleges, minority-serving institutions, public institutions, and even elite private institutions—wish we could find the answer for the tragedy of current college drop-out rates among low-income students, but those of us who believe that everyone has the right to earn an education aren’t willing to resort to the only proven solution to that problem, which is selective admissions. In other words, the number of students represented in the numerator and denominator of elite institutions is similar, where as the numerator and denominator of non-selective institutions can be vastly different.
On the issue of PLUS loans, I never said that upper-income parents don’t take PLUS loans, but instead that at places like Princeton, there is a great deal of borrowing on the part of wealthier parents outside of the PLUS system. Your study complete ignores the reality that parents of students at elite schools are likely to borrow significant amounts from non-federal sources, which artificially deflates your data regarding debt per credential at those institutions. Beyond that, it wasn’t until I read the interview with Erin that I could even tell that PLUS loans were included in your data set since the description in your report of your methodology is lacking appropriate detail.
I understand why you couldn’t include Perkins loans in your calculation, but wouldn’t it have strengthened your paper to point out that at elite institutions, students and families have many more financial resources at their fingertips, including Perkins loans? You also erroneously characterized Princeton’s “no loan” policy by failing to explain that this policy does not apply to the Estimated Family Contribution or families above a certain income. Princeton and other institutions like it make very careful calculations to determine the income cap for those who qualify for the no loan program, so it isn’t as if every student who benefits from the no loan policy.
Perhaps a better question to ask is not why more institutions can’t be like Princeton, but instead why an institution like Princeton with $14 billion in the bank needs to participate in the Federal Student Aid program at all, especially in these austere times. The taxpayer already makes a huge contribution by allowing Princeton to keep the profits of their investments without paying taxes on that money (or property taxes on their enormous real-estate holdings), so should those whose own kids will never have the chance to attend Princeton also be expected to support this wealthy, elite institution through federal student-aid dollars that your data demonstrate its students really don’t need? Maybe students at Princeton SHOULD borrow more than $3,000 so that precious Pell resources can go to poorer students at other institutions who don’t have the advantages conferred by a Princeton diploma, the Princeton alumni network or the Princeton endowment.
You also pretty much gloss over the reality that dependent students are far more likely to have parents footing the bill—through savings, payment plans or non-federal borrowing —than are independent students. In reality, a better way to look at these data might be through the lens of borrowing by dependent students versus independent students.
I was at the NACIQI hearing I referenced in my blog, and if I remember correctly, you were not there for the full testimony since the session began earlier than scheduled and you were returning from another meeting. I was in the room and heard Dr. Tilghman testify about who is and who is not in her peer group—in defense of her proposal that elite institutions should be moved out of regional accreditation and into a different system that includes only their peers. I completely agree with Dr. Tilghman that Princeton and Mercer Community Colleges are not peers and I am glad that she had the guts to say so. These are different institutions with different faculty, different students, different missions and different levels of resources, so to include them in the same analysis of almost anything is bound to be misleading.
While you say that you weren’t comparing elite institutions to for-profit institutions, your narrative and figures tend to focus on those two types of institutions to the exclusion of others (and you failed to give more than a passing nod to the fact that even among for-profit institutions, there is a great deal of variability among missions, students served, degrees conferred and everything other than tax status). Perhaps the reader might have been interested to see debt-per-credential figures for other institutions, such as public institutions that serve a large number of out-of-state students, and the non-flagship state schools you describe in your notes. Moreover, recent data show that the students who leave college with the largest amount of debt per credential include those who attend four-year HBCUs. I make this point not to disparage HBCUs (I believe HBCUs and other MSIs serve an incredibly important role in our higher-education system), but instead to say that many HBCUs serve a student population that is low-income and independent (even if not by the federal definition, by the historical reality of their circumstances). These institutions, too, are between a rock and a hard place in trying to balance their access mission with today’s completion agenda, and a closer look might tell us that it is time to focus our resources on those students who need them most as opposed to those who have other means, including the non-taxed resources provided by well-endowed institutions.
You distributed a policy brief that will have tremendous policy implications, and in it you failed to point out the various weaknesses in your methodology, biases in the data, and potential sources of error as well as alternative interpretations. Policy makers could jump to erroneous conclusions based on the omissions you made.
Simply said, it is time to look at institutions in the context of the students they serve and degrees they confer rather than through the lens of metrics that favor traditional students and wealthy, selective institutions or that focus almost exclusively on tax status. The metrics we have today do not work in evaluating outcomes for the largest growing population in higher education, which are non-traditional students attending non-selective institutions. What we really need to do is look at outcomes versus total cost of education (including tuition, government appropriations and taxpayer subsidies, including tax-exempt status) and student demographics so that we can really see what is going on—which is that the students who need it least (students with perfect SATs have a pretty good shot at life no matter where they go to school) get tremendous financial support from a variety of sources and those who need it most are left to scramble for the crumbs.