“The problem is not that humanities jobs are disappearing”
In an interview at Inside Higher Education, Jason Brennan (Georgetown) and Phillip Magness (American Institute for Economic Research), answer a question from interviewer Scott Jaschik about their view that universities are admitting too many PhD students.
Brennan and Magness, co-authors of Cracks in the Ivory Tower: The Moral Mess of Higher Education, respond:
Everyone likes to blame the poor state of the academic job market—especially in the humanities—on alleged cuts to faculty lines … The problem is not that humanities jobs are disappearing, but that many academic fields are graduating new Ph.D.s even faster than their full-time job market grows.
They cite data to support their answer:
U.S. Department of Education data (see, e.g., IPEDS tables 315.20 and equivalent in earlier reports) show that the total number of tenure-track assistant professors in four-year colleges has grown steadily since 2002, and is keeping pace with student enrollment … Data from the Bureau of Labor Statistics show that the total number of humanities professors (excluding part-timers) has not only increased by about 60,000 between 2000 and 2015, but that humanities professorship employment grew faster than any other field except all the health sciences.
The annual Survey of Earned Doctorates shows a similar pattern. In 2015, the humanities reported 1,383 full-time hires among newly minted Ph.D.s. The social sciences showed 1,215 hires (excluding psychology, which is sometimes categorized as a preprofessional discipline); life and agricultural sciences posted 920; math and computer science posted 441; engineering posted 399; and physical sciences posted 246 faculty commitments from the newest class of Ph.D. students …
The real problem is that while the humanities jobs are growing, Department of Education and other data sources show that the rate at which humanities departments graduate new Ph.D.s is even faster. So, the job market “shortage” is really job market glut of our own creation. Both administrators and faculty have perverse selfish incentives to churn out Ph.D.s. (For example, professors in doctoral programs get free grading, higher salaries and more prestige.)
You can read the full interview here.
Related: Against Reducing The Number of Philosophy PhDs
I’d be interested to hear more details from the authors, or anyone who knows this area.
For example, looks like some figures about faculty/student ratios didn’t make it into the IHE piece. What’s the background story on that 24-to-1 ratio, how has it changed or not? How is it calculated?
And what’s the story on those famous adjunctification stats we keep hearing about? How do those numbers square with the ones being offered here? What’s “supplemental instruction” and how does it figure into the authors’ account? How are the humanities like/unlike other areas wrt adjunct teaching? Etc.
Thanks in advance to anyone who can offer more info here.Report
I haven’t followed the many threads here for several days & just scanned the more recent comments, but wanted to offer a few concluding thoughts that will answer the comments and queries regarding tenure & adjunctification.
1. There are several sources that show that humanities faculty hiring is keeping pace with or exceeding the growth of university faculty as a whole. The clearest metric is the BLS occupational employment survey, although its design likely captures a few adjuncts on the lower tail of a sample that is almost entirely full timers. We conclude that this tail is far too small to account for the depicted trends as adjuncting also falls well below the tail boundary for the salary distribution. But we also see confirmation of full time humanities job growth in the HERI survey (which does differentiate full and part time hires) and in the Survey of Earned Doctorates (which records full time academic job commitments by discipline at graduation).
2. Data on Tenure/TT and non-TT full time faculty are difficult to disentangle due to a wide range of ways in which these are reported. For example, not all 4 year colleges/universities even have tenure systems and yet they do employ faculty on long-term contracts with job security for basically an entire career. On the other hand, there are many short term contracts that only last a year. And there are positions somewhere in between that are non-TT, but automatically renewing full time appointments that may occur yearly, every 2 years, every 5 years & so forth. It is therefore highly misleading to assume that all non-TT faculty are “contingent” and face some sort of non-renewing job precarity, even though some clearly do.
3A. There are some surveys on adjunct and other non-TT job trends. The two best are the 2012 Coalition on the Academic Workforce (http://www.academicworkforce.org/survey.html) and the 2012-13 Survey of Humanities Departments (https://www.humanitiesindicators.org/binaries/pdf/HDS2_final.pdf). The first of these actually contains strong evidence *against* the common depiction of adjuncts as “freeway fliers” living on tiny salaries and strings of contracts at multiple schools. The overwhelming majority of adjuncts only teach 1-2 classes and mostly at the same university. A sizable number of adjuncts are also part time by choice (moonlighting professionals with other jobs, retirees who teach a class or two) and do not reflect the precarity claims of the adjunct activist crowd.
3B. The Humanities Dept. survey similarly confirms that adjuncts are a distinct minority in *all* disciplines for which we have data. Depending on the discipline, they range from about 10% in the “best” cases to about a third of the teaching workforce in the “worst” (i.e. English/MLA & Communications). This survey also shows non-TT Full Time appointments, and again the numbers show that contingency is exaggerated in most discussions. Keeping in mind the caveats above (not all non-TT faculty are contingent and not all adjunct faculty want full time jobs), even the worst disciplines (English/MLA and Communications) still have clear majorities of their faculty on the full time payroll, and about half of their faculty on the tenure system. The rest are much higher – usually in the 65-80% range are full time.
4. These survey data come from 2012, which was right after the all time peak of the adjunct boom in the United States (fall 2011). Adjunct numbers have been dropping every year since that peak but full time hiring has been increasing at a normal historical pace (see https://nces.ed.gov/programs/coe/indicator_csc.asp). That means the full time faculty shares in these disciplines have likely increased since the survey was taken.
Combined, these data do in fact contradict the standard narratives of a humanities implosion stemming from adjunctification AND contingency. Where the stats are strong and cover several years, they also run strongly against this narrative. Where they are spotty or incomplete and limited to only a snapshot such as the tenure percentages, the partial data points we do have also cast doubt on the narrative. There are no data points, full or partial, that lend strong support for the adjunctification claim and those making a broader claim about contingency require several strong assumptions that are generally not supported by what we know about academic hiring (e.g. counting all non-TT as contingent, or the majority of adjuncts as the freeway flier type).
As a final point, there’s a strong case to be made that the optimal number of adjuncts or contingent faculty is not zero. I don’t know of any way we could determine an objective optimal level, but there are clear reasons to have faculty in these ranks (e.g. allowing professionals in other careers to moonlight in the classroom, continued academic activity for retired faculty, increased flexibility in course offerings and numbers of sections, more employment options for people PhDs who *only* want to teach and do not spend time on research). It also seems reasonable that this group might make up about 25% of the humanities workforce, which is also right around where we are at for adjunct faculty.Report
The 24:1 ratio is calculated from US Dept. of Ed stats on full time faculty and student enrollment by year. It’s more or less stable going back to 1970, hovering in the ~25:1 range and even slightly improving in recent years.
The adjunct statistics in most higher ed discussions are badly misinformed if not outright deceptive (as is the case with some of the AAUP numbers). Adjunct hiring started to boom around the 1990s and peaked in 2011, but has been on the decline – both in percentages and absolute numbers – since then. You can see the Dept. of Ed stats directly here: https://nces.ed.gov/programs/coe/indicator_csc.asp
There are several factors driving these trends, but the biggest one by far is for-profit higher ed. Adjunct use boomed with the growth of the for-profit industry, and it has contracted as that industry shrank after 2011. The reason is that for-profit schools use adjuncts to cover nearly their entire teaching loads. Adjuncts in traditional higher ed are still more prevalent today than they were 30 or 40 years ago, but in the supplemental sense. The full time faculty represented in the 24:1 ration has not changed, so adjuncts are being used as supplemental instructors beyond that – think more course offerings, course releases for senior faculty, lower average number of students in classroom etc. That’s where most of the adjunct work is taking place.Report
Thanks Phillip, that’s very helpful. I’m still a little unclear on “supplemental instruction”—if the full-time-faculty/student ratio is basically unchanged, is the idea that faculty are now getting more teaching relief via adjuncts? Are we teaching smaller classes or getting more frequent release from teaching than 40 years ago? Or are students taking more classes?Report
The measures of how faculty spend their time are deeply imperfect and largely dependent on self-reporting. But the short version is yes – full time professors appear to be teaching less today than equivalent-ranked professors did several decades ago. Think of a prof with a 4-4 in 1960 who now has a 3-3 or 3-2 load today. A lot of the slack is being picked up by adjuncts.
Also, universities are offering more classes at non-traditional times such as evenings and weekends. Adjunct use helps to facilitate more of that – especially with adjuncts who have full time day jobs in a related industry or profession and who moonlight one evening a week in the classroom.
Note that these are a very different type of adjunct than the commonly depicted “freeway flier” who teaches 4 classes at 4 different campuses. Again, stats are incomplete but they all suggest the long term career “freeway flier” types are actually pretty rare, and probably less than 10% of the adjunct population. Roughly a third are moonlighting professionals, and the remainder are a mix of retirees and others who only want part time work, grad students who pick up a teaching gig while finishing their dissertation, and people who adjunct for a few semesters while on the full time job market (who then either secure a full time job in academia, or go on to something else)Report
Again, very helpful, thank you. I’m interested to see how the discussion below of TT vs. full-time develops.Report
To some points others have made about full-time contingent faculty–commonly called ‘adjuncts’ if not called that under your operative definitions, and in any case definitely to be contrasted with tenure-track employment: this division of non-TT employment types into freeway fliers, moonlighters and retirees, and short-termers is not exhaustive. There are also lots of full-time and near-full-time contingent faculty for whom teaching is a long-term job. (Common at the 23 Cal State U. campuses, for example, where TT employment has shrunk in many departments even as enrollments have grown, are long-term lecturers.) These jobs can be more secure and pay better than the worst adjunct work, but they are still among the first to be cut in an academic downturn, they often come with less professional development support such as conference travel funding or leave time, they often come with less power to influence decisions that affect them no less than tenure-track faculty (such as who gets which teaching times or office spaces), and they often involve unpredictable workloads (such as not knowing how many courses they will get until the first week of the semester).Report
In no conceivable sense are full-time contingent faculty “commonly called ‘adjuncts’,” or at least they are not called that in the scholarly literature on the subject or any of the statistical accounting of faculty employment over time.
Even the title itself – adjunct professor – comes from a specific rank in the university system used to designate part time hires on a per-class basis. If you are hired to a contingent but full time position, you usually get classified as Instructor, Lecturer, or Visiting Assistant Professor. But never adjunct.
The distinction is important, as there’s a large gap between an adjunct making 3K a class with no benefits and a non-TT Instructor or Lecturer making 50K with benefits, albeit on a fixed contracted term. That gap is also much larger than the gap between a non-TT Lecturer at 50K and an entry level TT Assistant Professor at 60K, the main difference of course being length of contract (and even then the TT Asst Prof is only contracted for ~5 years and must earn promotion beyond that, or be out of a job).
Conflating the terminology as you do serves little purpose to this discussion other than the muddying of the waters around a complex subject.Report
You write, “In no conceivable sense are full-time contingent faculty “commonly called ‘adjuncts’,”” But there are at least three different people in this thread (including me) who are saying that in their idiolect ‘adjunct’ includes full-time contingent faculty. I could definitely be wrong about this, but I suspect that if we were to poll faculty across the country, we would find that at the very least a substantial minority (say, 30-40%) treat the term the same way that I do. That seems more than enough for that usage to be “common.” And I suspect that the number is much higher than 30-40%. Probably more like 60%. Given the conversation here, what would be your best guess at the percentage of faculty in the U.S. who use ‘adjunct’ in a way that includes full-time contingent work? What would you your low-point for taking this usage to be common? Should we conduct a survey to check?Report
I’m making a claim about how the term is used in the scholarly literature and in statistical reporting of higher ed data. It’s an entirely reasonable scholarly practice to expect conversation about that literature to follow its definitions.
Three random guys in a comments thread deciding that they want to use their own different definition that they made up on the spot, and then proceeding as if it has similar validity to the published literature, is not a constructive approach to continuing this conversation. That’s akin to saying “I don’t like what the GDP stats for Canada in 2017 show, so therefore here’s my own different definition of GDP that I just came up with and that I think is more accurate than the one that economists use.”Report
At my institution, people in 1 year contract non-tenure track positions are literally called adjuncts. I, frankly, don’t much care about whether you think this is the right terminology or not. For me, this problem is not a conceptual puzzle or a question of mere data analysis.
I’m genuinely concerned about the well-being of the faculty in this country (who form a clear majority of all faculty) who lack tenure-stream jobs. If your data don’t capture this very important problem then your data isn’t all that important to me. If you yourself don’t seem to think that the stress of year-to-year application for work, the low wages (in many places adjunct wages fall in the “low” or “very low” HUD income ranges), stress, unpaid labor (so much “voluntary” service), and delaying of life goals that this kind of employment produces then, again, I don’t think that the two of us could have a productive discussion on matters of academic labor.
This is very much a moral issue and no amount of semantic ambiguity over the term “adjunct” can hide the very unethical structure of academic labor nor can it hide how those structures have changed over time.Report
1. You are conflating your own normative position with our empirical claims about what the data say.
2. Your unverifiable anecdote about your own institution does not alter the statistical practices of an entire field of research.
3. If you wish to discuss the numbers and what they mean, our sources are there. But those sources also use a standard definition, meaning you should address them and our claims derived from them on those terms. Playing word games with your own alternative definitions is proving a fruitless exercise though, so forgive me for declining to indulge you any further.Report
At my institution, a large flagship state university, no faculty are called ‘adjuncts’. The distinction to which Phil is referring, with a puzzlingly unpleasant degree of defensiveness, is captured at CU Boulder as the distinction between lecturers (what Phil calls ‘adjuncts’) and instructors, who have three-year contracts, benefits, and a reasonable measure of job security (it is quite rare for an instructor not to be renewed, after the first couple of contract renewals). Given this disparity in naming conventions, which isn’t simply a matter of three anecdotes on this thread, I am genuinely interested in how the ‘scholarly literature’ on this topic makes sure that it is measuring what it claims to be measuring. Before you get all snarky and supercilious, Phil, I am not claiming that your claims are incorrect. I merely would like to know how researchers on this topic make sure that they have the correct statistics. If you simply looked at the available data for CU Boulder, you wouldn’t know that lecturers are faculty hired on a course by course, semester by semester basis, and instructors are faculty on three year renewable (usually) contracts. It took me several years to get used to that naming convention. At both my previous institutions, Rice and SMU, ‘lecturer’ referred to a full-time, usually renewable, appointment, and ‘instructor’ referred to a course by course, semester by semester hire. In other words, the exact opposite of how we name them at CU.Report
Alistair – You’re objection is a strange one. Anecdotal reports of individual universities using slightly different titles than the rest of academia do not affect the statistical classification of adjuncts in any way. That is done by annual reporting to the Dept. of Education, which applies a uniform definition and set of qualifications nation wide.
Neither is it “defensive” to insist on precision in such things. We are making an empirical claim about clearly defined statistics. It does no good to start throwing out alternative definitions when they do not match the stats sources we are using, and when they seem to have no other basis than a mixture of anecdote and ad hoc normative offerings. That leads me to conclude that those insisting upon them here are doing so only to avoid grappling with the implications of the stats and associated definition that we show.Report
The question is – given the wide variety of uses (and nonuses) of the word ‘adjunct’ at different institutions, what confidence should we have in the reliability of self-reported data from those institutions?
The reason you seem defensive is that you’re acting as though you can preserve certainty just by pointing to what information the data is *supposed* to capture as a way of avoiding the question of whether it in fact captures the information it claims to.Report
Derek – because the data we use comes from a Dept. of Education reporting system that also contains detailed instructions to universities on how to measure part time or adjunct faculty members as per the IPEDS reporting standards.
I suppose you could allege some degree of incompetence causes erroneous reporting in ways that diverge from the Dept. of Ed instructions. But speculating to that end is not enough to establish that it’s happening, let alone happening in a way that systematically skews the data toward our argument. The scholarly lit on both adjuncting and higher ed in general have long used the IPEDS data and treated it as a reliable source for stats on the US university system. If it is your belief that IPEDS is unreliable and should not be used to measure the data points we draw from it, it is also your burden to establish some evidence of that unreliability and, if you want to advance a different explanation for adjunct employment trends, propose a better alternative way to measure them.Report
Philip, you write: “In no conceivable sense are full-time contingent faculty “commonly called ‘adjuncts’,” or at least they are not called that in the scholarly literature on the subject or any of the statistical accounting of faculty employment over time. ” I take your word for it on how the scholarly literature and how the BLS use the term ‘adjunct.’ However, in the seven colleges and universities I have worked at, even those where ‘adjunct’ technically refers to a proper subset of non-TT positions, the term ‘adjunct’ is used colloquially to refer to all non-TT positions. (Again, in the CSU, this is common, even though the term doesn’t show up as a contractual term anywhere, as far as I can tell, as all contingent faculty are either Lecturers or Visiting (rank) Professors, regardless of part time or full time status.) This is the sense in which it is ‘conceivable’ that the term ‘adjunct’ is widely used: it’s widely used that way by the people who are being studied (faculty, administrators), even if it’s not used that way by the people studying them (scholars who research higher ed). It is certainly not “just made up” by “three random guys on the internet” (why in the world would one random person, let alone three random people, make up a definition of ‘adjunct’ that is just their own?!?) – not that that really matters to the substantive issue.
I don’t believe that it is a conflation to point this out. The worry that I and others have had here might be re-described as a concern that the conflation happens, or appears to happen, in the interview you conducted with IHE. The selection captured here at DN starts with the quote, “Everyone likes to blame the poor state of the academic job market—especially in the humanities—on alleged cuts to faculty lines.” Now maybe by ‘everyone’ you just meant ‘everyone in the relevant scholarly literature.’ But I, and I suspect others in this thread, read that quote as trying to hook into the more popular conversation about this issue. And if that is what we mean here by ‘everyone’–if you are trying to address not just a few other scholars but the wide swath of higher ed rank and file professionals who are concerned with the state of the higher ed job market — then everyone I am aware of, i.e. the rank and file members of the profession (not people who analyze education for their work) and most reporters reporting on it, when they think about the “poor state of the academic job market” and “cuts to faculty lines” are referring to Tenure Track lines and the Tenure Track job market. What they are worried about, in other words, is the relative rise of *all* non-TT work in proportion to all TT work. Very many of us in this conversation colloquially refer to the non-TT work as ‘adjunct’ work (again, even if that’s not how the scholarly literature uses the term). If some attempt to address *this* problem then shifts the discussion to only reference a proper subset of non-TT work because that is how some scholarly literature operationalizes ‘adjunct,’ then when that shift is directed back to the popular discussion, that’s where the conflation (or as you put it elsewhere, “word game”) happens. And to keep our eye focused on the popular concern is not to conflate a “normative position” with what the empirical data say (as you accused Caligula’s Goat of doing). Instead it is to keep our eye on the issue that many people suspect is a real issue (even if there are other problems in the neighborhood), i.e., the loss of TT lines relative to enrollments. From the rank-and-file perspective, it looks not like the rest of us are using oddball anecdotes to try to weaken solid empirical research; rather, it looks like that solid empirical research isn’t tracking the conversation the rank and file are having, according to how you’ve told us the terms are defined in the scholarly empirical research literature. Which suggests to me, as someone who does not study higher ed for a living but who works in it and who cares about whether I should say advise my undergrads to go to grad school if what they ultimately want is–very specifically–a TT job one day, we need another study to determine whether the *TT* academic job market is in a “poor state.”
Of course, all of this is compatible with the possibility that PhD programs are producing too many PhDs, and I’m grateful to you and Jason for keeping our eye on that problem, as well.Report
To add late to this thread, it is also not strictly true that “If you are hired to a contingent but full time position, you usually get classified as Instructor, Lecturer, or Visiting Assistant Professor. But never adjunct.” When I was hired to a contingent full-time position at the University of Wisconsin-Milwaukee, I was classified as an Adjunct Assistant Professor; the term Visiting Assistant Professor was reserved for faculty with appointments at other universities.Report
Yes, everything Phil said is correct.
So a good summary is this: Universities are not replacing full-timers with adjuncts. The ratio of full-time faculty to students has been stable since 1970 and has actually been getting better since 1987. However, students at traditional four-year colleges are more likely to take a class with an adjunct today than in the distant past. Full-time faculty probably enjoy lower teaching loads now than in the past, and part of the slack is being taken up by poorly paid adjuncts. However, a good part of it is covered through other means, such as class size changes. We don’t have good enough data from the Department of Ed, HERI, CAW, AAUP, or other sources to know exactly what happened.Report
I am a bit unclear on why the distinction is being made between adjunct and full-time. Adjunct (and contingent) does not mean part-time but rather non-tenure track. There are plenty of adjunct/contingent jobs that are full-time but on limited contracts (1, 3, 5 years). The gov stats say nothing about adjuncts but rather about full-time/part-time.Report
A full time non-TT professor making 60K with benefits may be “contingent,” but is definitely not an adjunct.
Adjunct professor is a very specific definition and = a faculty member who is contracted to teach on a per-class basis, usually with minimal or no benefits or research and service obligations.
A large part of the confusion over the adjuncting issue is a result of the conflating of this definition. Some of the resulting imprecision is simple unfamiliarity with the data, but it is also the case that adjunct activists use the terms deceptively to foster a perception that “adjunctification” is still on the rise even though it peaked in 2011 and has been declining ever since.Report
That just isn’t true and does actually look at the conditions of contingent faculty. Even full-time are on a per-class basis. One year one might have 5 or 6 classes and another year maybe 1 or 2 and drop below full-time. Further, you only get payed if the course runs but with higher and higher class minimums that is no guarantee. All full-time signifies is teach/service load. For those non-tenured this can vary year to year.
The concern often is about tenure-track being replaced by non-tenure track positions. When one is ‘on the market’ they are searching for tenure-track jobs.Report
Can you explain what in particular you think isn’t true? Phil just means that the department of ed doesn’t classify adjuncts (no matter how many classes they teach) as full-time, so the charts and data we use won’t include them. He also means that a full-time long-term faculty members is not treated as contingent the way an adjunct is.
CAW surveys find that 70% of adjuncts teach 2 or fewer courses a semester. Regardless of which survey you use, you find that adjuncts teaching a full-time load are a minority.
However, HERI, CAW, and other data sources also find that adjuncts do not work as many hours as faculty teaching the same number of classes as those adjuncts. A full-time professor teaching six classes at a typical school works about twice as many hours over a year an adjunct teaching those classes (inclusive of travel time on both).Report
Yes. The department of ed classifies adjuncts differently and they appear on different tables. An adjunct who teaches 90 hours a week may indeed being doing more than a full-time’s worth of work, but will be classified as a part-timer (as a person working multiple part-time jobs) by the department of higher ed, by the BLS, or by any of the other major databases we used.Report
Woah. That was *not* the way I thought the reply would go. And now I feel like I’m missing something really big. Colloquially, in worried conversations that I’ve been involved in, the issue is about the increase of non-TT lines, generally. At least before today, I would have used “adjunct” to describe (at least) all non-TT labor by people with PhD in hand. Especially, full-time, fixed-term appointments, but also non-TT lines that are renewable or that fall into distinct categories, like our so-called “teaching professors.” But from what you’ve said here, it seems that for the BLS “adjunct” doesn’t mean what I would colloquially mean by that term. Specifically, the BLS wouldn’t count a full-time, non-TT person on a fixed 1-year contract as an adjunct. Is that right? Following up, then: Do the BLS numbers let us talk meaningfully about changes in TT versus non-TT instruction over the last twenty years? Do they tell us anything about how non-TT lines are being used in the universities? If not, then I’m not sure what exactly we’re supposed to be learning here.
For concreteness, what would the BLS numbers say about something like the following. In, say, 1970, a department had four TT lines and one retired professor teaching 1 or 2 courses each year on a per course basis. In 2010, that department had two TT lines and three full-time, non-TT lines (say, one renewable and two fixed term). Would that look like a decrease in adjunct teaching? How would BLS describe such a scenario?Report
They would describe your situation as follows:
1. In 1970, four full-time TT faculty. Whether the emeritus professor counts as an adjunct or not depends on the specifics on his contract.
2. In 2010, five full-time faculty, with the TT to NTT ratio you specify.
In case it wasn’t clear, the data indeed show that since 1970, the ratio of full-time TT to full-time non-TT faculty has fallen, though it has been stable (at about 33:17 or more currently 3:2) for two decades.Report
This conversation has been very interesting.
I worry about the tenure/non-tenure dichotomy being the main focus when discussing the job market.
I tend to distinguish three groups: 1) tenure/tenure track folks, 2) folks paid $3000 or so per class per semester, probably teaching at multiple institutions, and 3) non-tenure folks making $40 – $60k teaching at a single institution.
That tripartite distinction is important, because people who didn’t attend top-ranked programs are often aiming for jobs in class 3) (or full time at a community college), since the vast majority of TT jobs go to folks from top ranked programs, and the line between category 2) and category 3) jobs is often whether you can support your family and so whether you have to pursue a different career path.Report
I’m interested in the data on this third class. What are the salary statistics for non-tenure track full-time instructional faculty? Is it somewhere between $40k and $60k?Report
I’m undercaffeinated this morning, so perhaps I’m missing something, but the parenthetical here seems to be incorrect: “Data from the Bureau of Labor Statistics show that the total number of humanities professors (excluding part-timers) has not only increased by about 60,000 between 2000 and 2015.”
I assume this is supposed to come from the Occupational Employment Statistics; for example, Humanities Indicators reports an increase of about 60,000 humanities faculty over this period of time using OES data. (https://www.humanitiesindicators.org/content/indicatordoc.aspx?i=71) According to the survey coverage and scope: “The OES survey covers all full-time and part-time wage and salary workers in nonfarm industries.” (https://www.bls.gov/oes/oes_emp.htm#scope) HI notes that “The BLS estimate, based on a sample survey of employers, includes both full-time and part-time faculty.”
So adjuncts, by any definition, are included in the OES data.Report
Hi Dan – The measurement issue comes from the way the OES is collected, which is by survey instrument. It likely captures a small number of adjunct faculty while also missing others based on how they sample “part time” work. We can determine from the survey’s accompanying salary data however that these persons are on the extreme lower tail of the distribution, and thus only make up a couple percentage points at most. Or put another way, the surveyed salary level at the reported lower boundaries for these fields is well above what even a maxxed-out 4-4 load “freeway flier” adjunct makes, indicating that the survey scoops up at most only a few adjuncts on that bottom tail and not enough to skew the overall employment pattern.Report
Should also mention – there are several other surveys besides OES such as HERI that measure faculty percentages by discipline, and that can specifically exclude adjuncts. They confirm a similar pattern of humanities growth that OES reveals.Report
That inference seems shaky. Per the note explaining wage estimates, “Annual wages have been calculated by multiplying the hourly mean wage by a “year-round, full-time” hours figure of 2,080 hours; for those occupations where there is not an hourly wage published, the annual wage has been directly calculated from the reported survey data.” (https://www.bls.gov/oes/current/oes251126.htm#(2)) If a $3,000, 15-week adjunct position is counted as 10 hours/week, it annualizes to $41,600. The annualized wage is even higher when the adjunct position is counted as fewer hours per week. So, in the best case, you seem to be relying on very substantive assumptions about how adjunct salaries are annualized.Report
1. A $3K 15-week adjunct position would actually be counted as 3 hours/week, which is hours spent teaching and the only portion of allocated time reported from such contracts (yes – we all know adjuncts work outside of classroom time on prep and grading – it’s just not reported or standardized for employment purposes, and is largely variable based on how efficient the professor is at grading)
2. Even if you reject our inferences from BLS, you still have to get around the semi-annual survey data in other sources like HERI, which also show similar employment levels in the humanities after you exclude part-time survey respondents.Report
How am I to interpret data like the following:
Charts like these indicate that the problem is at least partly a decline in jobs.Report
We still don’t quite know what to make of charts like that. We agree in the book that many advocacy groups publish charts showing the number of advertised jobs has gone down, and when we fact check those charts, that seems correct. At the same time, the number of full-time and TT jobs in those fields, , both entry-level and senior, has gone up according to a bunch of other sources, even though advertised jobs have gone down. So, universities must be hiring more people in unadvertised jobs.Report
It doesn’t follow from what you’ve said here that universities are hiring more people in unadvertised jobs. Based only on what you’ve said here (maybe you address this somewhere else?), it could be that retirement rates have slowed more than hiring rates. If, for example, you go from a state of affairs in which the number hired is 10 per year and the number retiring is 10 per year to a state of affairs in which the number hired is 7 per year and the number retiring is 5 per year, then you will see the number of people employed going up while the number of job ads goes down. Concrete example:
Year . . . . . . Total Staff . . . . . . . Hired . . . . . . . Retired
1 . . . . . . . . . . . 100 . . . . . . . . . . . . 10 . . . . . . . . . . 10
2 . . . . . . . . . . . 100 . . . . . . . . . . . . 10 . . . . . . . . . . . 9
3 . . . . . . . . . . . 101 . . . . . . . . . . . . 10 . . . . . . . . . . . 9
4 . . . . . . . . . . . 102 . . . . . . . . . . . . . 9 . . . . . . . . . . . 8
5 . . . . . . . . . . . 103 . . . . . . . . . . . . . 9 . . . . . . . . . . . 7
6 . . . . . . . . . . . 105 . . . . . . . . . . . . . 7 . . . . . . . . . . . 7
7 . . . . . . . . . . . 105 . . . . . . . . . . . . . 7 . . . . . . . . . . . 6
8 . . . . . . . . . . . 106 . . . . . . . . . . . . . 7 . . . . . . . . . . . 5
9 . . . . . . . . . . . 108 . . . . . . . . . . . . . 7 . . . . . . . . . . . 5
10 . . . . . . . . . . 110 . . . . . . . . . . . . . 5 . . . . . . . . . . . 5
Here we have an increase of 10% in staffing over a ten year period in which the hiring rate decreases from 1 hire for every 10 staff to 1 hire for every 22 staff. Moreover, for this case, the total number of hires goes down, so even if we kept the the number of PhDs produced constant, the hiring rate out of PhDs on the market would also go down.Report
One big problem with charts like that from MLA & AHA & the other groups is that they only capture ads listed with those groups. They tend not to do very much in the way of systematizing whether those ads are for entry level positions, for senior hires, or even re-listings of failed searches and the sort. And they don’t reveal anything about ads that don’t use their listing service. So it ends up being at best an extremely imperfect measure of the number of jobs out there.
Survey of earned doctorates, by contrast, measures all PhD students at the time of graduation so it captures a more direct snapshot of who actually gets an academic job (vs. other types of jobs vs. no employment etc). You therefore end up with a decent picture of the number of entry-level faculty positions that were actually filled over the previous year, as opposed to those that were simply listed in some online job database.Report
And what’s the measure of retirements (or other ways of leaving the university)? Did you take into account both the inflow and outflow rates?Report
Full time hiring in US academia has outpaced full time retirement in every year since the mid 1980s when it flattened out for a couple years. And that includes the net addition of about 120,000 new full time faculty jobs since the 2007 financial crisis.Report
Could you tell me how retirement rate is being measured? Is this also in the BLS data or is your claim based on some alternative source(s)?Report
US Department of Ed statistics. They tabulate the number of full time professors in the US annually. That number is going up, ergo a net gain of new hires outpacing retirements.Report
I think you’re missing the point of my example. In my example, it is true that hiring outpaces retirements in most years. And it is true that the number of staff goes up over the entire period. That is consistent with the yearly rate at which staff retire and are hired both decreasing year-over-year. So let me ask again: Do you have a measurement of the rate at which people leave university employment? (Typically, I would think, this would be by way of retirement.)
Again, to be as clear as I can. The initial question was how to make sense of the claims that the hiring rate is going down while the number employed goes up. Jason said that it *must* be that there has been an increase in unadvertised hiring. To which I pointed out that if the rate at which retirements happen decreases faster than the rate at which hiring decreases, we will observe an increase in the number employed. So, it isn’t *necessarily* the case that there is a lot of unadvertised hiring. It might be what’s going on in the actual world. I don’t know: that’s why I was asking about the actual numbers! But it’s not an instructive reply to say that we’re seeing an increase in the number employed, so the retirement rate must be smaller than the hiring rate. Now, if we had a measure of the number hired every year and a measure of the number employed every year, then we could calculate the number that retired. That number would have as much error as in the two numbers used in the calculation, but it might be a very good estimate nonetheless. Similarly, if we have a direct measure of the number retiring and the number employed, we can calculate the number being hired. We could then compare that number to the numbers derived from looking at job ads. And we would have a better answer to the question of whether there are actually lots of unadvertised jobs.
After typing the above, Jason posted saying that the data are fine-grained enough to know that the total number of entry-level jobs has gone up. I had to stop and think about this for a bit, but that observation is *also* consistent with the hiring rates dropping. Here is an example that just divides out my first example.
Year . . . Jr Hire . . Jr Retire . . . Sr Hire . . . Sr Retire . . Total
1 . . . . . . 8 . . . . . . . . 1 . . . . . . . . . 2 . . . . . . . . . 9 . . . . . . 100
2 . . . . . . 8 . . . . . . . . 0 . . . . . . . . . 2 . . . . . . . . . 9 . . . . . . 100
3 . . . . . . 8 . . . . . . . . 1 . . . . . . . . . 2 . . . . . . . . . 8 . . . . . . 101
4 . . . . . . 8 . . . . . . . . 0 . . . . . . . . . 1 . . . . . . . . . 8 . . . . . . 102
5 . . . . . . 7 . . . . . . . . 0 . . . . . . . . . 2 . . . . . . . . . 7 . . . . . . 103
6 . . . . . . 7 . . . . . . . . 0 . . . . . . . . . 0 . . . . . . . . . 7 . . . . . . 105
7 . . . . . . 7 . . . . . . . . 1 . . . . . . . . . 0 . . . . . . . . . 5 . . . . . . 105
8 . . . . . . 6 . . . . . . . . 1 . . . . . . . . . 1 . . . . . . . . . 4 . . . . . . 106
9 . . . . . . 6 . . . . . . . . 0 . . . . . . . . . 1 . . . . . . . . . 5 . . . . . . 108
10 . . . . 5 . . . . . . . . 0 . . . . . . . . . 0 . . . . . . . . . 5 . . . . . 110
For simplicity, assume 36 junior faculty evenly split across six probationary years and 64 senior faculty at the start. (That fits pretty well with the distribution pictured here: https://www.aaup.org/NR/rdonlyres/6890A0F8-E43B-4B05-BEB4-C1A0720B1FC7/0/Fig1.pdf) Then assume that junior retirements include failures to make tenure. So, each year, six junior faculty are promoted to senior faculty. The result is that, at least for a while, we’ll observe an increase in the number of junior level positions, even though the hiring rate is decreasing. (Here is the junior/senior breakdown by year in the ten years above: 37/63, 39/62, 40/62, 42/61, 43/62, 44/61, 44/62, 43/65, 43/66, 42/67. How far off actual distributions are those?)
So, I think I *still* want to know whether there is a measure for retirement and what that measure looks like year-over-year. But maybe I should just go looking for the data myself at this point.Report
Yes, the Department of Ed data is fine-grained enough to tell you how many faculty are at each rank. So we know that the total # of entry-level jobs have gone up, not just senior jobs.Report
I raised this issue on Facebook some time ago, and I still don’t for the life of me understand what the response to Jonathan (and my) query is supposed to be. Without knowledge of the rate of retirements, its simply impossible to estimate hiring from the number of positions. There are too many unknowns. Its like trying to infer the rate at which a faucet is running by looking at the level in a basin without knowing the rate at which water is draining out. You just cant do it. And you can’t tell just by looking at the amount of water that’s been in the basin for x amount of time without knowing the rate at which water leaves the basin prior to x years being up.Report
To borrow your faucet/basin analogy, it is indeed possible to determine from rising water in the basin that the faucet flow into the basin exceeds the rate that the basin is draining. And that is what we see when faculty numbers go up overall.
But even then your complaint is misplaced, because we also know the actual number of new hires for each year by group from the survey of earned doctorates.Report
“that the faucet flow into the basin exceeds the rate that the basin is draining.”
How on Earth do you continue to not get that that is irrelevant to the claim that the faucet flow is *not lower than it used to be*?????? What is the difficulty here???
You guys continue to insist that “universities *must* be hiring more people [than they used to ]in unadvertised jobs.” (Which is a prima facie implausible claim).
This is *not* a claim the rate of the faucet *compared to that of the drain*. It’s an *absolute* claim about the rate of the faucet. Please stop pretending otherwise.Report
It seems like your objection comes from insisting on not paying attention.
The number of new tenure-track assistant professor jobs and non-tenure-track but low-level/entry-level jobs in the humanities has been rising more or less steadily since 2000. In fact, it’s doing that faster than in other fields, even though demand for humanities degrees has fallen off the cliff. The total number of TT assistant professors in academia as whole has gone up dramatically over the past 20 years. The reason the job market is bad in English, but not economics, is not because English isn’t getting jobs (in fact, it’s adding jobs at a faster rate than econ despite the lack of student demand), but because English has added new PhDs at an even faster rate than the growth in jobs.
Demand for English BAs: way down, cut almost in half since 2000
Entry-level jobs for English professors: Up
Total jobs for English profs: Up
Number of English PhDs: Way up
Result: Bad market due to PhD glutReport
I’m paying attention. You keep equivocating between new and total. You’ve done it again here. You have data about the size of the pool and you keep using it to estimate the rate of inflow without knowing the rate of outflow. That’s true regardless of whether you are looking at total jobs or total assistant level jobs. Jonathan has explained one way this could be the case, but there are others. The rate at which people leave *any segment of the employment pool* can change over time and it almost certainly does. I addressed this point when I said “And you can’t tell just by looking at the amount of water that’s been in the basin for x amount of time without knowing the rate at which water leaves the basin prior to x years being up.” (x amount of time here being the 6-7 years it takes to get promoted.)
And in any case, at least you are making this slightly less missing-the-point point. Phil keeps *over and over* repeating that the data has to show that hiring-minus-quits is going up and that this proves the point. Which is clearly a case of him not paying attention.Report
I have no idea what you are asking.
The # of new entry level jobs of all types are up.
The # of senior jobs is also up.
The total number of individuals working as full-time faculty is up.Report
I don’t know how to slow this down enough.
you have an accounting identity:
hires-quits*=total number of jobs. that identity holds at every conceivable point. no other identity exists. but its an identity with 3 numbers in it, and you keep using 1 to estimate a 2nd without knowing the 3rd. What’s the confusion?
stop equivocating on # of jobs. you use it to sometimes mean # of people working, and sometimes to mean # of people being hired at a given time. they are not interchangeable without a quit rate. again, what is the confusion about this?
*understood to mean retirements, firing, tenure denials, quits, etc etcReport
I have to admit I don’t see what hangs on this issue. First, I can’t even see why the retirement rate would be changing enough to matter. This is because an increase in average retirement *age* isn’t sufficient for a decrease in retirement *rate*, only for a change in the composition (viz, the cohort) of the numerator. And, second, because it isn’t clear to me why a lowering retirement rate would exonerate PhD granting practices as they are.Report
Also, one thing that seems to be getting missed is that Jason and his co-author don’t *infer* the number of new hires from the growth rate of the total; they get it from surveys of recent phds.Report
Im not *claiming* that the retirement *rate* is doing anything. I’m pointing out that it is an *unknown*. you have an equation with two *unknowns*. you cant solve for anything. you certainly cant undermine another source of data (advertised positions) using a number you calculated from an equation with 2 unknowns.Report
You are mistaken. They undermine that data with survey data from new phds. They do not try to infer it from your equation.Report
I want to echo the comment by Full-Time Contingent Faculty. I appreciate both the data and the work y’all did to gather/present it. But I think the stats we need pertain to TT Faculty: what is the ratio of TT faculty to students, how has it changed, what is the ratio of TT professors to other full time professors…
Adjuncts are relatively rare at my University. But we increasingly rely on non-TT full-time faculty. And those faculty are in a pretty tough position:
-They have year-to-year contracts w/ no real job security (which matters for academic freedom)
-They get paid less to teach more
-Their salaries get slashed at the last minute if their classes are judged too small (which means we often can’t have them teach upper division classes unlikely to draw 25+ students)
-There’s all sorts of service they either can’t do or that it makes no sense for a non-TT faculty to do (which means the remaining TT faculty have to do more).
-They’re often, and understandably, on the job market year after year, which is time-consumptive and just massively draining
And so on. The point is that by emphasizing the adjunct/full-time distinction, y’all might painting an unrealistically rosy picture–both of the job prospects for newly minted PhD’s and of conditions at the University. I’m not claiming to know that things are worse than you represent, nor am I claiming that overproduction of PhD’s isn’t a major part of the problem. I’m saying data on the full-time/adjunct distinction doesn’t necessary tell us what we need to know.Report
Agreed. The tenure-track to full-time non-tenure-track ratio is lower now than in, say, 1960 to 1970. It seems to have stabilized over the past 20 years, though.
Here’s a good summary of the past 20 years:
Of 100 people who start a PhD program, half will get a PhD. Of that 50, about 20 will get a full-time academic job. 12-13 of that 20 will get a tenure track job. 7-8 will get a non-TT full-time job. Of the 12-13 who get a tenure-track job, 5 will get a job at a research-focused school.Report
Of course those are the overall numbers. It varies from field to field and school to school.Report
Thanks! I wouldn’t have guessed that the non-TT full time/TT ratio had stabilized. I suspect we’ve achieved that stability (at the level of overall numbers) by adding TT jobs in STEM related fields and taking them away from the Humanities. But that’s just a suspicion.Report
I’d be interested to know where the growth is occurring and where it isn’t. My general impression has been that there is a lot of job market growth in the humanities, but that it’s been restricted largely to either large private universities, which tend to have high enrollments in the humanities, and large public universities, which tend to have high enrollments in general. My general impression is that small private and public universities haven’t seen the same kind of growth and, in fact, have drastically reduced their humanities staff, but I might completely wrong.Report
Thanks for this discussion! I’m trying to wrap my head around this, but can you please try to explain the following phenomenon, in light of your data:
At my school (public, with graduate program), the faculty is half the size it was from 10 years ago. Retirements are not being replaced, and this is largely a pattern across all departments within the College of Arts & Sciences (i.e., all the departments are shrinking, including the humanities ones). Our overall faculty is stable, which is just to say that increased hiring in other colleges (e.g., engineering, business, etc.) are offsetting the CAS losses. Also, the % of our credit-hour production taught by adjuncts and other irregular (e.g., VAP) faculty is super-duper high, and higher than at any time since I’ve been here.
So far as I can tell, this is the pattern among similarly-situated universities, or at least my friends and I seem to agree on it when we talk about it. (I would think it also squares with the charts mentioned above.) How do we reconcile this with your view? Maybe some possibilities:
1. These anecdotal views aren’t supported more broadly by evidence.
2. It might be true in this setting (and others like it), but is counterveiled by other settings (i.e., it’s idiosyncratic, broadly speaking).
3. A bunch of humanities jobs are being offloaded to other colleges (e.g., sort of like Jason’s, for example).
Or? It just seems like your analysis flies in the face of everything I actually witness, so trying to debunk my observations. Thanks!Report
(By “overall faculty is stable”, I mean university-wide, not departamentally; I think the context indicates that, but it’s ambiguous.)Report
The answer is 1, your anecdote may be right and may even apply to many other colleges, but the overall there is not only growth in humanities jobs (including TT asst prof and other entry-level jobs), but faster growth than in nearly all other fields, despite most other fields having higher student demand.
As for 3, the Department of Ed would classify my job as a business job, not a humanities job, even though I have a PhD in philosophy and even though I often teach social science classes at the business school. They classify you by department, not content of the class. A physicist who only teaches and researches philosophy of science in a physics department would not be counted as doing humanities, but he would if his tenure home were in philosophy.Report
If data doesn’t differentiate between 100 tt faculty and 100 faculty on 1-year contracts (i.e., if both jobs would paint the same picture of the market if the data were used) then we need better data. It’s absolutely disingenuous to say that the labor market hasn’t changed since the 1970s for the worse and ironically, the problem we’re exposing is a really standard philosophical one requiring that we make sure we’re using the same terms.
NTT = adjunct on my own understanding of those terms because they both lack the kind of job protections that used to mark off academic labor as free in an important sense (i.e., tenure provides a significant kind of freedom for speech, research, and to advocate for improvements in one’s workplace without the concern of losing one’s job).
It *might* be useful to notice trends between “paid-by-the-course” adjuncts and “yearly-contract” adjunct (or even “multi-year-contract” adjunct) but I take it that when most of us warn our promising students away from going to graduate school it’s because good jobs (i.e., tt jobs) are fewer and fewer in number, applicants are greater and greater, and the stress of moving around year-over-over from one “full-time” adjunct position to another is not only heartbreaking and stress-inducing but inhumane.Report
I don’t get it. The data does differentiate and we are up front about the numbers whenever we can be. For instance, in 2003, there were 152688 TT assistant profs in the US; in 2016 (the last year of data out) there were 176347 TT assistant profs.Report
Where do the incentives for having Ph.D. programs come from and could they be addressed from the ground up, or are top down solutions the only plausible way to reduce those incentives?Report
They cover that in the book. I can’t do it full justice, but here’s a brief answer:
1. Having a Ph.D. Program makes the university look good and gives it extra money from legislatures (if it’s s state university).
2. It gives the department extra money.
3. It gives the professors graders.
4. It lets the professors teach more motivated students.
5. It gives the chance for professors to spread their own views through the field.
6. When administration threatens to cut programs they get a lot of push back.
It seems hard to imagine how a top-down approach would change those incentives.Report
Every department and even admin units have weird selfish incentives to keep grad programs around. Even shitty grad programs mean extra money for admins and faculty.
Just as it’s useful to model firms in a competitive market as profit maximizers, it’s useful to model individual units in non-profits as budget maximizers. This model has a great deal of predictive and explanatory power in university settings.Report
Isn’t prestige-maximisation also pretty important?Report
Naively one might expect that more data transparency about prospects would help. At the moment, your prospects of a successful career in academia (albeit the metrics for that are contested) vary radically with your department, but that’s not always obvious to students. I am more optimistic about reducing the number of PhD programs than about across-the-board reductions in their size, not least because – in B&M’s terms – it relies in part in harnessing the incentives of potential PhD students, not on persuading faculty to act against their own incentives.Report
I am quite puzzled by the data and even more by Brennan and Magness’ inferences about it. It’s not just a matter of what “precise” or “detailed” definition the department of education uses (which they could clarify to us). It’s that a lot of the reporting seems, on the face of it, difficult to understand. Say, ” In 2015, the humanities reported 1,383 full-time hires among newly minted Ph.D.s.” – sounds easy but, I am really wondering how reliable the reporting of “new minted Ph.D.s is – who counts as newly minted PhD and who as full-time? Are ABD’s among these? Are people 2 years after completing PHD but getting the first job counted? Or is it just people who upon finishing, got a job in the next 12 months? Or started the job? Is a full time lecturer hired for a yeat by his own department after PhD counted in? Unless one knows way more the information is almost useless. How well are the staff who report these things versed in the the data required of them? I often find myself utterly puzzled by various forms and I have studied statistics for a couple of years. When one looks at the “increase” in jobs – well, of course, there are more jobs, there are way more people in US than 20 years ago. It’d be bizzare if there were not. In any case, my feeling is – B&M are making a certain interpretation of the data – data given in a certain not entirely intelligible form (intelligible to the very community about which it is and for which it should be useful) – to make a political claim. Not sure what the political agenda might be, but it seems a bit of the sort – we need to cut the funding to humanities, they are overproducing…Report
Yeah, whom is this for? I just don’t quite see what could be salutary about taking a hard line to the effect that universities are admitting too many PhD students. I mean, I’m not even sure that I get why it matters whether B&M are correct in their interpretation of the data. Suppose the number of humanities jobs is increasing – that’s quite consistent with thinking that it should be increasing at a faster rate.Report
There is a fairly big difference between “we need more TT jobs because, relative to UG population size, there aren’t as many as there used to be” and “we need more TT jobs because, relative to UG population size, we are educating more PhD students than we used to”. The latter, but not the former, has a reasonably obvious alternative solution (albeit Brennan and Magness demonstrate why implementing that solution is harder than it looks).Report
Keep in mind that it’s an interview, and they also cut it down, removing some of our numbers and sources. You can read chapters 7 and 8 of the book for a full picture.
We have multiple sources from the Department of Ed, HERI, CAW, AAUP, BLS, and others.
Each tell the same story: The ratio of full-time faculty to students is stable. The ratio of full-time non-TT to full-time TT faculty is stable. Entry-Level hiring of all types has generally kept pace with enrollments, except in the following sense: English and the other humanities are up despite falling enrollments, while economics is not keeping pace with its massive increase in enrollments. Senior jobs are stable too, but no one cares about those, since they are decrying the state of the market. The humanities are not simply surviving, but actually increasing their relative presence on campus despite losing students. However, the ratio of job-seekers to humanities jobs has gone up because the increase in job seekers is bigger than the increase in jobs.
The various advocacy magazines simply don’t report these numbers. They don’t look at things like department of ed data and instead rely on numbers from advocacy groups like the MLA. The MLA’s advertisements are indeed down, though the number of TT and non-TT full-time jobs in English is up, and up despite plummeting enrollments.
My suspicion, of course, is that people want to resist our story for self-interested reasons. If we are right–and we clearly are if you look through the data we cite in the book–then the “Humanities are dying so give us more money” story looks bad. But people want more money and it’s in their interest to push that story.
I understand that inclination. I love money, too. As an analogy, in chapter 5, we describe how the practice of grading is unreliable, harmful to students, and often involves incoherent math, but after describing things which would help students more, I say that I sure hope my own university doesn’t follow my advice. If it did, it might cost me thousands a year in time that could have been spent giving paid talks or writing articles which lead to increased pay raises.Report
P.S. Check out this chart here:
The charts give figures for degrees awarded in the two subjects, but is there evidence of a similar decline in total number of students taught in English, including non-majors and students taking service courses? That’s the more relevant figure for estimating faculty needs. My admittedly anecdotal sense is that humanities departments are doing a higher percentage of service teaching these days, as against teaching majors.Report
Focusing on fine-grained details of which jobs are in which category risks missing the bigger picture.
As far as I can see, no-one is contesting B&M’s claim that humanities PhDs granted have increased significantly more rapidly over the last few decades than the undergraduate population has (and the claim looks fairly solid on the basis of the data they present in their book, albeit that data is noisy in some places). If that’s so, it’s pretty close to analytic that the job situation will have worsened for humanities PhDs even if the number of jobs has kept pace with the UG population. There just isn’t a mystery here. Of course, *maybe* the job situation has also worsened (though again B&M’s data looks pretty persuasive at establishing that it at least hasn’t changed much in the last 20 years). But there’s no grounds for inferring that from the increasingly demanding job market when there is a much more straightforward causal mechanism in play.Report
It’s irresponsible to conflate adjuncting with non-TT positions. The idea that tenure is equivalent to a good (e.g. non-exploitative) position is simply wrong. Most professors in most countries don’t have tenure. More to the point, the overwhelming majority of jobs don’t have anything like a tenure track. Even one year contracts represent a level of job security that most people would find incomprehensibly generous. My sister is a highly competent and highly educated civil engineer. She’s also an At Will employee, and could be fired tomorrow if her firm runs into financial difficulty. My brother works at a hedge fund – or at least he did until the partners decided to take the fund in another direction and his whole desk was let go without warning. I’m not TT, but I have way more job security than either of them. And my family is incredibly fortunate in our educations and careers. There’s no such thing as a tenure track human resources administrator or a tenure track plumber.
It should also be borne in mind that overproduction of PhDs is responsible for the degradation in tenure. Tenure was a midcentury innovation that was developed during the tight academic job market that arose after the GI Bill caused a huge surge in the demand for college professors. It’s a perk. Universities feel no incentive to keep offering that perk when every job offering has hundreds of applicants regardless of the salary or perks on offer.Report
The ‘most professors in most countries’ comment needs to be treated with a bit of caution. Compared to Europe (the other academic job market I know), the US has extremely weak employment protections. In the UK, if you want to fire someone you need a fairly extensive paper trail, and unless you can establish lack of competence or a disciplinary problem (which takes a lot of time and effort to do) you have to provide a substantial redundancy settlement. The protection given by tenure looms so large in the US because it contrasts so sharply with the employment default.Report
I have to agree that the main problem today is that there are way too many PhD programs granting way too many PhDs. Many of these programs have horrible placement rates and don’t care enough about what will happen to their graduate students. I have said before on other forums that probably about half of the PhD programs should be shut down. This is just off the top of my head but seems reasonable. We need to stop flooding the job market with PhDs, as this is destroying the value of the degree. In fact, it’s doing more than this: it’s making the humanities look bad in the public’s eyes too. Many of these PhDs who can’t get jobs don’t speak fondly of academia. The internet is flooded with negative articles about academia written by people who are very angry. What effect do you think these pieces have on society? Well, for one, it probably effects undergraduate majors in the humanities, which from what I’ve read are in decline. It’s a good rule not to educate a bunch of people and then make their job prospects so poor by overproduction that they end up hating you.Report
This is just fascinating–maybe I’m weird but I kind of like finding out I’ve been totally wrong about something. As far as I can tell there’s just two things that are hard to account for still. One is the apparent decline in advertised full time jobs in places like English and Philosophy. If those are falling while the total number of FT jobs is increasing something weird is going on–why the dramatic increase in unadvertised jobs? And then secondly, how is the supply of Ph.D.s so high (and growing?) in the face of flat or only modestly growing demand? I get why departments have incentives to keep growing their graduate ranks despite placement problems, but it’s surprising that the market isn’t doing its ruthless thing in shrinking supply, particularly given the decline in majors in the humanities. Maybe that’s coming. Or maybe the only people majoring in English these days are people who want to get a Ph.D.Report
Supply would only matter if there weren’t enough (in the eyes of the department) qualified PhD applicants. I expect that for most universities their applicants would have to be pretty abysmal before they decided that taking no one was a good option.Report
I think Cautious Until Tenured and Norcross’s comments among others here reflect an important consideration that really gets overlooked: There is a huge variety in non-TT jobs. I’ve been an adjunct, a year to year lecturer, and am now full time faculty at a community college and, while all those jobs are non-TT they’re all vastly different in pay, job security, workload, and working conditions. The proper adjunct job had no benefits, low pay, and absolutely no security; in fact we didn’t even know how many classes the powers that be would dole out to us until right before the semester started. The lecturer job had low pay and little to no formal job security (though I never heard of anyone being let go) but a low workload and good benefits. My current job has almost as much job security as a TT job, good pay, and good benefits, but also a higher workload. My point is that instead of using blunderbuss terms like non-TT or worse yet misclassifying everyone as “adjuncts” we need to develop some categories that capture the real complexity on the ground and use them to study the real state of faculty employment. One thing we certainly can’t trust here are the titles colleges use. Not only are they not standardized but they have every reason to obfuscate what’s really going if they use a lot of adjuncts. This seems like a prerequisite to an intelligent discussion of these issues. Getting clearer on these categories and how many people fall in each would also help newly minted PhDs to better navigate the job market. I know from my own experience that a lot of them wrongly dismiss all non-TT jobs and pursue some really awful jobs because they’re nominally tenure track. It would also stop some of the more dishonest claims about the effectiveness of adjuncts versus full timers as teachers. The studies that claim to show adjuncts are as effective as full time faculty, which if I’m not wrong Brennan has made a big deal of elsewhere, actually follow full-time lecturers rather than proper adjuncts
Another statistic that we really need to know is what percentage of students taking various classes are taught by adjuncts. Even if one buys Brennan and Magness’ claims or trusts their interpretation of the figures (and I certainly don’t for the reasons that Livengood and Eric mention among others) it seems to me pretty much undeniable that many more students are taught by adjuncts than previously. Now I could be wrong about that but I’d bet my next paycheck I’m not. If that’s true then there could be all sorts of reasons for it, but I think it’s very disturbing. One thing to keep in mind on this figure is that we must be careful to distinguish this figure from the total number of classes taught by adjuncts. At the school I adjuncted at for a few years the typical class taught by adjuncts was 60-90 students while no full time faculty ever taught a class of over 30 (and many of their classes were way under-enrolled due to the fact they were more often major classes rather than Gen Eds). The school could and did claim then that 2/3 of their classes were taught by full time faculty, which was true, but it obscured the more important fact that at least in philosophy 50% or more of the students in philosophy classes any given semester were being taught by adjuncts.Report
Thanks to the authors of this study. If I followed the discussion here, one of the take-aways is that a lot of people are getting jobs from unadvertised positions. The obvious follow-up, then, is: how does one find about about and apply for these unadvertised jobs?Report
I base my understanding of the humanities employment situation on data sources at the Department of Education and NSF.
But I reach a conclusion that the PhD “glut” is overstated.
Published data includes:
*Total number of humanities full and part-time faculty: nearly 120,000
*Humanities PhDs produced per year: about 5300 (with some 420 in philosophy),
*How many of these new PhDs take academic jobs: a little over 75%.
*Median age of a humanities PhD recipient: 34.
Suppose, for the sake of an estimate, that there is no increase in overall employment (which would work in favor of the “glut” if true).
Then for every humanities PhD recipient who goes into academia, there should be a job that lasts for a career of 30 years.
I don’t claim this is the final analysis. But it suggests the “glut” is overstated.Report
This doesn’t quite work. Yes, the number of available jobs in the humanities matches the number of PhDs who take academic jobs: that’s pretty close to an accounting identity. You need to compare the number of jobs to the number of PhDs who *want* academic jobs. (You also need to look at whether the increase in supply drives the creation of more part-time and / or less attractive positions.)Report
I’ll take this as agreeing with my point. All I am doing is citing public data, after all.
Is there a survey of humanities PhDs, to see how many *want* academic jobs? The NSF only has data on how many take such jobs.
If every humanities PhD wants an academic job, then there is a “glut.”
According to the NSF data, the US supply of philosophy PhDs is not increasing.Report
I have not yet read B & M’s book, but it should be noted that their overall incentive story holds even if they were wrong about the source of the job competitiveness.
Look at the placement data at mid and near-top tier PhD programs in philosophy. A good numbers of graduates are placing into outside-of-ed occupations, adjunct positions, or administrative positions (see, e.g., Penn); even who are placed in TT positions are often placed at the most random of colleges. And yet the admissions committees refuse to change admissions. Why is this? *Even if everyone knew* the reason jobs were competitive was because of an overall reduction in humanities jobs, there seems to have been very little change in practice at these programs.
The only possible logic I could see is that it’s in the institution’s interest to have so many PhD students–whether faculty see it as rewarding, or they need TAs and lecturers, or they have the money for the grad program and know they’d lose it if they didn’t use it. Whatever the story, it seems the only possible reason to sustain a program at a constant size with a consistently crappy or subpar placement rate is because the institution cares *comparatively* little about the placement rate when assessing their admissions process.Report
Do any departments consider their placement rates at all when deciding how many new PhD students to admit? I admit I had never thought that might even be a factor in their decision-making.Report
Other disciplines do, because they consider is a failure if a PhD lands in a non-academia/non-research job. E.g. top business school PhD programs do. The reason they give PhD students stipends and resources is to produce scholars. Why shouldn’t *whether they actually become scholars* be a factor in every part of PhD program administration?Report
I saw several empirical questions in this thread about what might or might not be the case with regard to tenure, full time/part time, contract length, class sizes, and so forth. I think a good number of these can be gotten from the data, for anyone that wants to have a look:
You could access the IPEDS data center here: https://nces.ed.gov/datalab/index.aspx
And you can download the data set here: https://nces.ed.gov/ipeds/use-the-data/download-access-database
For more information on what these categories mean, you can look at the instructions given to people who enter the data here: https://surveys.nces.ed.gov/ipeds/visresults.aspx.
In particular, the instructions for the Human Resource Survey which includes data for tenure/non-tenure, part-full time, etc, is here: https://surveys.nces.ed.gov/IPEDS/Downloads/Forms/package_1_43.pdf#page=43
Also, if you search for the name of your own institution and IPEDS you will likely find whatever data has been reported. I would hope that at least in some cases this can help translate whatever terminology is being used in your institution to the language used in this reporting instrument. Or you might find things that look off. In any case, it will always be someone at your own university that sends these numbers to the Department of Education, but it’s a huge mess of numbers so I wouldn’t be surprised if things get mis-classified at times.Report
I actually tried digging into this yesterday but hit a wall trying to separate humanities faculty from the rest. Did I miss something, or does the IPEDS data indeed not track faculty by field?Report
BLS has data by field, but you need to then break it down by part-timers vs full-timers. There are other data sources for that which corresponds to BLS data. Remember, the interview above was just an interview, and because they abridged it, they didn’t list even all the sources we gave them.Report
Right but then BLS doesn’t have TT/non-TT right? (Maybe also not for-profit vs. non-profit?) I couldn’t find a source that had all the main variables at issue here.Report
That was the point I made up here:
The initial reply was that, based on the distribution of incomes, the survey probably didn’t include many people working part-time for a few thousand dollars a class. I pointed out that, because of the way incomes are annualized, this argument is questionable. Philip Magness then appealed to a different survey, HERI (). You can see the most recent questionnaire from this survey here: It does ask respondents about T/TT status and discipline (questions 44a and 44b).
However, at first glance the sampling frame for the HERI survey isn’t clear to me, especially between rounds. Schools pay to participate () and schools seem to drop in and out between rounds of the survey (). I don’t see a link to any discussion of HERI’s sampling frame on their website. And I would also want to read Brennan and Magness’ account of their analysis.Report
(Whoops, the links all disappeared. Let me try that again.)
That was the point I made up here: http://dailynous.com/2019/05/24/problem-not-humanities-jobs-disappearing/#comment-323287
The initial reply was that, based on the distribution of incomes, the survey probably didn’t include many people working part-time for a few thousand dollars a class. I pointed out that, because of the way incomes are annualized, this argument is questionable. Philip Magness then appealed to a different survey, HERI (https://heri.ucla.edu/heri-faculty-survey/). You can see the most recent questionnaire from this survey here: https://www.heri.ucla.edu/PDFs/surveyAdmin/fac/INSTRUMENT/2016FAC.pdf It does ask respondents about T/TT status and discipline (questions 44a and 44b).
However, at first glance the sampling frame for the HERI survey isn’t clear to me, especially between rounds. Schools pay to participate (https://heri.ucla.edu/heri-faculty-survey/) and schools seem to drop in and out between rounds of the survey (https://ucla.app.box.com/v/FAC-Participation-History). I don’t see a link to any discussion of HERI’s sampling frame on their website. And I would also want to read Brennan and Magness’ account of their analysis.Report
Thanks, but looks like HERI charges for access to their data (and IRB approval may be required too?).
To be clear, I was hoping for a single, freely available data set that would allow one to create a reproducible verification of these two claims:
1. The ratio of TT humanities faculty to undergraduate students has held steady, for the last 20 years let’s say.
2. The ratio of humanities PhDs minted annually to undergraduates enrolled has increased over the same period.
That would, it seems to me, resolve many of the worries raised above. I would love to post some R code on GitHub that downloads the data and produces the visualizations and figures verifying these claims. That way we could all get on the same page, and skeptics could play with the code to test out their concerns.
Unfortunately, it doesn’t look like the requisite data is publicly available.Report
For #2 we can use IPEDS data. I did exactly that in this post in 2017: http://dailynous.com/2017/12/18/underproduction-philosophy-phds-daniel-hicks/ The code is here: https://github.com/dhicks/overproduction_PhDs
NB I looked at the inverse ratio (BAs to PhDs, not PhDs to BAs). So when I find that the ratio in humanities is historically high, that means we’ve been producing relatively *more* BAs than PhDs.Report
Thanks, Dan! I had forgotten about your post, will revisit it now. (I’m delighted you made your code available btw, and that is uses RMarkdown + tidyverse!)
I’m curious incidentally how the picture changes if we compare to total undergraduate enrolments, instead of to majors. As others pointed out both in this thread and the one on your post, philosophers often do a lot of service teaching consistently, even as majors rise and fall. (This is certainly true in my department: we teach gobs of logic to STEM students who will never take another philosophy course.)Report
A simple question for the authors. You are making a statistical argument. Did you conduct statistical analyses? If so, which?Report
I don’t think I understand what this means.
Sometimes you do a statistical analysis because you’re working with a sample from a larger dataset and you want to know the error bar in the sampling process. (Example: opinion polls.) But B&M are mostly reporting census-type data, for which that issue doesn’t arise.
Sometimes you determine, e.g., that some quantity has increased or decreased, but you want to check whether it’s just noise. But B&M are reporting double-digit percentage changes in values in the middle thousands, so it’s immediately obvious that it’s not noise. (If you want a quick rule-of-thumb estimate, take the square root of the dataset size. Except in very rare setups, that gives you an order-of-magnitude estimate for the size of random fluctuations. If the actual change is way larger than that, no need to worry about whether it’s noise. (So in a sample set of ~1000, expect fluctuations around ~30, or ~3%). This doesn’t give you a sharp estimate but it’s a good quick-and-dirty rule.)
If you don’t mean either of these, then I don’t know what you have in mind. ‘Statistical analysis’ isn’t a magic spell that has to be cast before you’re allowed to appeal to quantitative data in your argument, after all. It’s a set of specific tools to handle specific questions one might have about data in relatively-ambiguous circumstances.Report
This is primarily a US debate, conditions being different in other parts of the world. So as an outsider I am going to express my views with what I hope is a becoming diffidence. However, it seems to me that there is a plausible moral argument for cutting back on the admissions to philosophy PhD programmes in the US and perhaps cutting back on the number of such PhD programmes *whether or not* there has been a significant drop –off in demand and /or funding for undergraduate teaching in the humanities in general and philosophy in particular.
1) There has been a significant drop off in demand and/or funding for undergraduate teaching in philosophy, leading to a reduction in the number of decent, more-or-less permanent jobs.
2) There has been no significant drop off (or maybe some increase) in the output of Philosophy PhDs.
3) Thus there is an oversupply of philosophy PhDs seeking work, leading to unemployment, heartbreak and a highly exploitable work force which is regularly and often ruthlessly exploited.
4) This is an awful situation and the philosophy profession should do something about it if there is a practicable policy available.
5) It is not within the power of the profession (at least in the short to medium term) to increase demand/and or funding for undergraduate teaching in philosophy.
6) Thus the only practicable policy *for the profession* is to cut back the admissions to philosophy PhD programmes and perhaps to cut back on the number of such programmes.
7) Therefore the profession should collectively cut back on the admissions to philosophy PhD programmes and perhaps should cut back on the number of such programmes.
Case B [Brennan & Magness])
1’) There has been no significant drop off in demand and/or funding for undergraduate teaching in philosophy, hence no reduction in the number of decent, more-or-less permanent jobs.
2’) However there has been a significant increase in the output of Philosophy PhDs (relative to the demand for their services)
3) Thus there is an oversupply of philosophy PhDs seeking work, leading to unemployment, heartbreak and a highly exploitable work force which is regularly and often ruthlessly exploited.
4) This is an awful situation and the philosophy profession should do something about if there is a practicable policy available.
5’) It is not within the power of the profession (at least in the short to medium term) to increase demand/and or funding for undergraduate teaching in philosophy to absorb the oversupply of PhDs.
6) Thus the only practicable policy *for the profession* is to cut back on the admissions to philosophy PhD programmes and perhaps to cut back on the number of such programmes.
7) Therefore the profession should collectively cut back on the admissions to philosophy PhD programmes and perhaps should cut back on the number of such programmes.
So either way, the profession should cut back on the admissions to philosophy programmes.
The contentious assumption in this argument is 4), namely that the bad situation described at 3) [which is widely acknowledged] is something that *the profession* should do something about if a practicable policy is available. If you don’t think that the profession has a duty to fix the problem (if a practicable solution is available) then you can escape the dilemma. I suppose somebody with delusions of political efficacy might contest assumptions 5) and 5’). But an increase in funding or demand sufficient to mop up the oversupply of philosophy PhDs, would require a slew of humanities-loving governments coming to power at both the state and federal levels and implementing a range of radical educational and cultural reforms. And they would have to do it pretty quickly to solve the problem in the short-to-medium term. That’s not very likely and it certainly not something that it is within the power of the philosophy profession to bring about.Report
PS. I suppose that there is another solution. PhD programmes should make it clear to their clients that they cannot expect to get a job teaching philosophy at the end of it. . If a sufficient number of new-minted PhD’s graduated with the assumption that they were NOT going to become professional philosophers, then the oversupply problem would evaporate. The programmes would have to have a virtual sign over their doors saying something like ‘Abandon hope of a philosophy job HALF of ye who enter here!’) If enough entrants actually believed the sign and acted on it on graduation then I guess the problem MIGHT go away. But I am not sure how many people would be willing to undertake a North –American style taught PhD (as opposed to a European style research-only PhD) if they really believed that there would be probably be no philosophical job at the end of it.Report