[S]purious ca[T]egories that hide vast diff[E]rences in pheno[M]ena

One of the best branding efforts of the past thirty years, aside from Apple calling all of its products the iThing, is the National Science Foundation inventing STEM, for Science, Technology, Engineering and Mathematics. (STEM was originally SMET… good call on that change.) Policymakers might not be able to slow down enough to listen to a careful explanation, but “we need more STEM funding or we’ll fall behind the [Soviets/Japanese/Chinese/etc.]” is an elevator pitch that you can deliver even before the doors slide closed.

One of the signs that an idea has stuck is when others try to borrow it. So the arts want some attention, and they invent STEAM (or STREAM, if you like robotics). Add invention and entrepreneurship to your stem, and you get STEMIE, which sounds like a character on Family Guy.

But let’s take the simple version, the STEM we’re all familiar with. The problem is that those four terms aren’t strongly related. Mathematics is the logic and symbolic system that science works with. Science is the underlying principles and phenomena that engineers and technologists exploit. In academic terms, science and mathematics have always been part of the liberal arts [the exercise of judgment and investigation worthy of free citizens], whereas engineering and technology are vocational, ways for the rest of us to get jobs. They all offer problem sets with right answers at the end, but that’s nearly as far as the similarities go. In a way, they’re like the grandparent, parent, and child of the quantitative world: the careful and often ignored work of earlier generations enabling the pace and success of the later. Science and mathematics are speculative and risky fields of exploration. Engineering and technology offer knowable, nameable skills to sell on the job market.

We can see this in looking at the rise and fall of college majors. The National Center for Education Statistics shows the T&E components of STEM (computer/information sciences, engineering, engineering technology, and health professions) rising from 9.8% of all bachelor’s degrees in 1970 to over 20% in 2016, while the S&M (biological sciences, physical sciences, and math/statistics) has dropped slightly from about 10% to below 9%. Take the slight rise in biology out of that group, and we see that math alone has dropped from 3.0% to 1.2% of college degrees; physical sciences from 2.5% to 1.6%.

The public relations triumph of STEM without a paired understanding that we’re orphaning its science and math components represents the full victory of the vocational model of higher education. It’s the answer to “What are you gonna do with a degree in THAT?”

How Far I Have Left To Go

A couple of days ago, while not wanting to work on my current novel, and also while getting myself ready to do a few book talks once The Adjunct Underclass goes live next month, I re-read that manuscript for the first time since late January’s final copy edit. it’s a good book! Yay me!

But, now that it’s too late to make any changes, I (of course) found a few things I’d like to have changed. Trivial stuff, mostly, an inelegant punctuation or a word unnecessarily repeated within a couple of paragraphs. I’ll live.

But one of them made me sad, because it’s stupid. (It’s in Appendix A, if you want to look for it, the text within Table 11.)

One of the many arguments of the book is that we’re spending money on a ton of things in higher ed that we never did before: good things, real benefits that make colleges better than they ever were when I went for the first time in 1976. And one subset of those relatively new things are services—social, academic, and cultural—in support of a far more diverse student body. Diverse by gender, by age, by race and ethnicity, by national citizenship, by sexuality, by family status, by social class: you name the constituency, and most colleges serve a different student body than they historically have. And while we want that diversity, we also don’t seem to want to change the fundamental nature of the institution, so all the supports we offer are bolt-ons. They’re additions, not modifications. And so the money we spend on them, on their staffing and space and programming, is in competition with the money we might otherwise spend on faculty.

It’s a delicate argument, and one that I took pains to handle carefully, lest I ever seem to imply that our broader welcome is somehow a bad thing. It’s a terrific thing, but like all terrific things it has unintended consequences.

Anyway, one of the pre-publication reviewers asked for a chapter on the history of the shift toward adjunctification, with the labor struggles and the professional statements and the manifestoes and on and on. And I absolutely didn’t want to do that (nor did my editor), because a) it’s a different project b) being taken up right now by a trained historian, L. Maren Wood, that c) would have been a distraction from the more ecological argument I’d laid out. But as a sort of compromise, I decided to build an appendix that showed the vast number of ways that higher education has changed in recent decades. Since I’d already done most of the statistical research for that, I was able to bang it together—book tables 9 through 22—in less than two weeks.

Table 11 lays out a history of the changes in undergraduate demographics. From 47% female students in 1976 to 56% in 2015; from 16% students of color in 1976 to 42% in 2015; more undergraduate students over age 25; more high school students with learning disabilities going on to college. It’s all accurate, and important. The problem is in the row headings. Regarding gender, it says “More female students.” But under ethnicity, it says “Fewer white students.” First off, that’s just incorrect; there are more white students than ever, but they’re a smaller proportion of the total student body. But second, and far worse, the framing of “fewer white students” suggests an unfortunate change, something for us to lament, when in fact I’d intended exactly the opposite.

In trying to re-create my thinking at that hurried moment, I’m imagining that the longer version of that would have been “Given that the student body isn’t almost completely homogeneous in its whiteness any more, colleges have a lot more work to do to become more welcoming and more supportive of cultural difference.” But it doesn’t read that way.

So why DID I frame that heading the way I did? The one right above it—”more female students”—isn’t framed that way. Why did my hurried typing reflex go the direction that it did? Because I think I still have a long way to go to unlearn the racial patterns and attitudes that I grew up with. Because the problem with bias and privilege is that it’s invisible to the people who have the advantages. Because I’m a white person living in a state that’s 95% white, in a town that’s 98.5% white. Because “white students” is the category heading in my brain, and “students of color” is still them.

The problem I face as an individual is the same problem that we face all across higher education, and more broadly than that. I’m just not smart enough, not sensitive enough, not sophisticated enough, to make sure that I don’t automatically revert to old, ingrained habit. I’m way better than I once was, but every so often, I show myself how far I have left to go.

Being a decent person is not a status. It’s a daily project. It’s an aspiration, a gift perpetually just beyond reach.

Degree of Difficulty

We’ve been talking lately about how hard numbers are. Not to use them, but just to decide what they even mean.

Let’s shift for a moment to grades and the GPA. Grades have enormous communicative power, even though they mean almost nothing. They are symbols without significance.

Let’s take a single course, say Calculus 1. What does a B+ mean? Does it mean that a student got more than 83% but less than 87% of the homework and test questions correct over the course of the semester? (And does that calculation include partial credit for operations correctly done but with a trivial error somewhere? And do some questions have more points than others?) Or does it mean that the student was at the vaguely-better-than-average-but-not-at-the-top-of-this-semester’s-group level? Or does it mean we don’t hold out a lot of hope for you as a potential math major, but we aren’t quite ready to turn you away, so go ahead and try Calc 2? It’s quite likely that different faculty in the same department would calculate that grade a little differently, based on their interests and values—it could be a bookkeeping score, or a competitiveness score, or a communicative score.

Now let’s take two courses, Calculus 1 and Introduction to Racquetball. Does the same grade of B+ mean the same thing in both courses? They both weigh the same, 3.33 units…

Now let’s look at two different schools. I taught one online master’s course with ten students, three of whom at the end got what I considered to be reasonably justified grades of A or A–. But the work of those very best students—second-year master’s students, remember—at one school would have gotten them a B or B– in my first-year undergraduate writing courses at Duke.

The remarkable precision of the GPA, with all of its attendant stress, with its precise cutoffs for adequate or exemplary performance, is a ruse. It’s a nicely decorated cover for a complete conceptual shambles.

We can mess around with it, calculating “weighted GPAs” that offer more points for honors or AP courses, but that just shifts the artifice to a new location. How much harder is an honors class than a regular class? 18.4% harder? Should Organic Chemistry get an extra 24.91% grade boost over the far simpler Intro Chemistry in the same major sequence? If I transfer, should my A in my community college writing course be converted to a B on my University of California transcript?

Here’s the fact. When someone reads your college GPA, their thought process will look like this:

Hmm… degree from Smith College. Good school. Majored in economics, tough major. 3.34, pretty good student. We’ll call her in.

Or like this:

Degree from Wilton and Madison College? Never heard of it. Majored in business, GPA 2.81… naahh…

Or like this:

Degree from Michigan in philosophy? Wow, great program! But only a 3.15 GPA… Maybe he’s okay…

What we mean by a grade is this: within a specific context, this student was judged by a specific person to have been:

  • outstanding
  • strong
  • okay
  • disappointing
  • awful

It hardly seems warranted to average those across experiences, much less to imagine two places beyond the decimal. It’s a false precision that feels reassuring, like a stuffed bunny that can’t actually speak. Do I know my GPAs from college and from grad school? You bet I do. They have talismanic force to protect me in the face of a hostile world, an external validation that suffices, once in a while, in the absence of internal validation. Yes, you really were that good, they murmur, if only to me…

The Inelegance of Simple Numbers

Yesterday we talked about the difficulties of data management and data definition. Let’s look at a simple example, a college trying for some degree of fairness of workload across its faculty.

Most colleges have what they consider to be a standard “teaching load,” defined either in numbers of courses (a 3/2 like I had at Duke is three courses in the fall semester and two more in the spring) or in numbers of credits, like teaching 24 credits per year (the equivalent of eight three-credit courses). Different kinds of schools have different kinds of loads; let’s take a single and simple school as our example, and say the teaching load is 3/3. That sounds fair enough: everyone teaches three courses each semester. All good.

Au contraire, mon frère.

  • What if one of those courses has thirty students and another has twelve?
  • What if one of those courses is writing-intensive, with essays to read and mark up for every student every week, and the other has problem sets and a midterm and a final?
  • What if one of those courses is a greater number of credit hours than another?
  • What if one teacher has one section each of three different courses in a semester, requiring three different preparations for each class day, while another teacher leads three sections of the same course?
  • What if some of those courses are for grad students, or well-prepared upper-division students in the major, and other courses are for not-very-carefully selected first-year students of wildly differing abilities?
  • What if some of those courses are assigned teaching assistants and others are not?

To quote Mannix and Neale: what differences make a difference?

At the most recent college I worked at, we convened a task force to try to regularize the stipends we paid to our adjunct faculty. After some conversation, I developed a conceptual model that we called B.A.S.S., for Baseline, Adjustments, and Separate Stipends. The Baseline was simple: we determined that courses should be paid at $1,000 per credit hour (which already took some struggle, since different departments had developed wildly idiosyncratic payment patterns). In practice, the Separate Stipends boiled down to an extra $500 if someone was creating a new course; the course development was treated as a separately contracted service.

So B was simple, SS was simple… but we really got bit in the A.

Studio instructors, who had six to eight students in a section, didn’t believe that courses with forty students should be paid more. They didn’t believe that courses where students required weekly (or perhaps more frequent) review of homework should be paid more than courses in which teachers never reviewed work outside of class time. The argument to increase pay for long-term and more experienced instructors didn’t get much traction; we actually spent more time considering whether an instructor who had professional licensure should get a higher stipend, something that doesn’t affect either classroom experience or teacher workload. In the end, the only adjustment that got applied was that co-instructors should each receive 75% of base pay rather than 50%, because of the extra burden of coordination. [Oh, please…]

Studio arts are totally worth having in the university, but they skew the calculation of teaching load to be almost unrecognizable. Lots of studio courses are six credits, to reflect their symbolic importance in the major, and the amount of time students should expect to devote to the work. But they’re usually small courses; usually don’t require anywhere near the preparation of lectures of the art historian down the hall; and very rarely require the kinds of nightly homework markup common in writing courses. So a studio instructor gets twice the credit of her colleagues, while doing what amounts to half or less of the work required for a three-credit seminar with 25 students in the same department.

In this and innumerable other ways, the supposed objectivity of numbers actually reflects the culture from which they spring.

Still more on this tomorrow.

The Problem of Too Much Data

A friend of mine is a jack of all trades. From tree work to auto repair, from farming to light construction, he just does whatever presents itself at the moment, and has become a pillar of the community in doing so. He called this afternoon to talk about burning off our brush pile if it’s not too windy tomorrow, and I got an unexpected lesson in research design.

He’s been chosen to be a test producer for a seed company this year, and he just got twenty packets of free seed, from tomatoes to lettuces to beans. There are seven hundred farmers in this program, which doesn’t sound like a lot, but due to the wonders of multiplication, we’re about to see how bad that can be.

Each farmer got somewhere between 15 and 50 different varieties of seeds in this program. Let’s say the average is 25. That’s 700 farmers with 25 varieties each, or 17,500 test plots around the country. And they’re asking each farmer to post a photograph every week, with a meter-stick in the photo, of each crop, to track date of emergence and rate of growth. Let’s say they average a twenty-week grow season, from plant to harvest. That’s 17,500 test crops times 20 photos… 350,000 photographs.

Posted onto the seed producer’s Facebook page.

Honest to god, they’re going to try to sort meaningful growth and production data out of hundreds of thousands of photos on Facebook, photos that will be completely undistinguishable from one another until the plants are at least somewhat mature. Photos from different lighting conditions, different skill levels and camera qualities, completely random file names. Just some guy or gal standing in a weedpatch with a measuring stick.

Just file names… It used to take me two or three weeks every semester to teach my Duke students how to name their homework files: section#.lastname.project#.draft#. It’s easy to do once you’ve gotten in the habit, to turn in a file labeled 35.childress.P1.D3.docx, but it takes practice and reminders even with Duke students. You think 700 farmers are going to label each photo with something like VT517.A9.0519.jpg and have that be a consistent protocol?

We spend tons of time in grad school on data analysis, learning statistical methods and applying them to artificially manufactured data sets to get better at doing the math. But none of that prepares you for how hard it is to collect and to manage the data in the first place. I had a job a long time ago, a tiny part of which was to figure out average length of juvenile detention in one county. I got to the probation office, and they wheeled in a cart of overstuffed file folders, random court and juvenile hall records in random order, and said, “Here you go. Let us know when you’re ready for the next batch.”

Data can be flawed by mis-definition, by mis-collection, by mis-transcription, by mis-categorization. It can be lost to a failed hard drive, lost to a programming error. It can also be artificially gained: the student records system PowerCampus creates a new record for a student who’s changed majors, for instance, or a faculty member who’s been promoted from assistant to associate professor. Every status change creates a new person, and it took hours to clean the duplicates before I could ever start doing the analytical work. (It took years to figure out how to write the queries in ways that reduced the duplications in the first place.)

Most of us aren’t Google or Cambridge Analytica, with teams of algorithmic designers. Most of us, most of the time, are trying to do relatively simple arithmetic—sums, averages, medians, quartiles, probabilities—with way too much data that we can’t always trust.

I’ll have a specific example of that tomorrow.

The Genre of the Shirt

I’ll be going to town meeting tonight, a hundred of my neighbors gathered together to deliberate our community’s business, and I’ll guarantee you that there will not be two identical shirts among those hundred (unless members of the volunteer fire company show up in uniform).

Go to a clothing store and ask for a shirt. You’ll be faced with dozens of questions that try to move you from generic to specific. Do you want a dress shirt or a work shirt or a sport shirt? Long sleeves or short? Pullover or button front? Solid or pattern? Collar? Size? I’m betting, at an order of magnitude, that there have been a hundred million identifiably different shirts made in world history, each iteration then made by the dozens or thousands or millions all on its own.

And yet we share a common understanding of the word shirt. It’s a garment that covers your torso and shoulders, generally symmetrical left to right, with a place for your arms to stick out. It’s not pants. It’s not a onesie. It’s not a dress. The shirt is the cultural norm that accommodates nearly infinite individual expression. The shirt is the genre.

Moving outside the genre is a risk that most consumers won’t take. The two-piece women’s swimsuit took almost twenty years to become common, forty years to become ubiquitous. And it’ll be quite a while before we see a more casual version of Billy Porter’s Academy Awards clothes when we go out to dinner locally, even though it’s just two genres put together.

Photo by Richard Shotwell/Invision/AP

A bookstore is a series of cultural expressions. Some of its genres have been around for hundreds of years, like romance and horror and mystery and cookbooks. Some have come within the last hundred or so years, like science fiction and popular business books. And some are even more recent than that, like graphic novels and literary YA. Like clothing genres, literary genres are simultaneously freeing and restricting, allowing vast individual expression within a knowable cultural frame.

All genres, whether clothing types or academic disciplines or types of cars, evolve slowly, and then one breakout period makes them broadly recognizable. The “crossover SUV” goes back to the Jeep Wagoneer of the ’40s, but the RAV4 and the Murano made them the most popular cars in America. The bikini had been around for fifteen years before it was made acceptable by Ursula Andress in Dr. No, by Annette Funicello’s beach movies, and by Goldie Hawn in Laugh In. And although we’d had comic books for decades, it took Will Eisner and Art Spiegelman to make the graphic novel a respectable category.

Creating a new genre needs pioneers, who will be mostly unknown. Later, it needs refinement by the polished pros who reveal its possibilities. And then, it will have always existed, will seem inevitable, will be its own cultural frame that enables infinite expression.

On Cooling the Mark Out

Although the term, mark, is commonly applied to a person who is given short-lived expectations by operators who have intentionally misrepresented the facts, a less restricted definition is desirable in analyzing the larger social scene. An expectation may finally prove false, even though it has been possible to sustain it for a long time and even though the operators acted in good faith. So, too, the disappointment of reasonable expectations, as well as misguided ones, creates a need for consolation. Persons who participate in what is recognized as a confidence game are found in only a few social settings, but persons who have to be cooled out are found in many. Cooling the mark out is one theme in a very basic social story.

Erving Goffman, 1953

In con games, the mark (victim of the con) is recruited by flattery. You’re smart enough to see the opportunity, you’re bold enough to do what others might not. And so the mark submits to the con, and loses. According to sociologist Erving Goffman, this moment at which the mark recognizes his loss represents a failure of an important self-concept, one that must be eased away from rather than simply broken and left behind.

For the mark, cooling represents a process of adjustment to an impossible situation — a situation arising from having defined himself in a way which the social facts come to contradict. The mark must therefore be supplied with a new set of apologies for himself, a new framework in which to see himself and judge himself.

Goffman identifies a couple of common ways of helping to ease the transition. The first is to offer “a status which differs from the one he has lost or failed to gain but which provides at least a something or a somebody for him to become.” The second is to offer “another chance to qualify for the role at which he has failed.”

In academia, the first strategy is called adjunct faculty or visiting scholar or professor of the practice, and the second is called postdoctoral fellow. The adjunct instructor is not the status that was hoped for, but at least it provides a role to play. The postdoc is also not the status that was hoped for, but the promise is that it represents merely a hold against payment sure to come.

We were all recruited by flattery, weren’t we. We were all separated from the herd, told we were special. We were given Greek terms like summa cum laude, mathematically demonstrated to approach or meet the 4.0 limit. We were welcomed to office hours, given special tasks, asked to speak at commencement. We were told by the undergraduate community that we were worthy, and that worth was affirmed as we were recruited by the doctoral community. Come to my school! No, no, come to MINE!

We performed well. No, not well. We were freaking awesome. We got straight A’s in the core, we killed the qualifying exams, we taught the intro courses and got the strong evals, we defended the proposal that allowed us to work independently, and then we defended the work we’d done. All five committee members agreed that we’d crushed it, they took us to dinner, told us we were the best ever. There’s never been another one like you…

And the phone never rang again. We were ghosted. We freaked out, asked our friends if we had spinach in our teeth or B.O. Or we went silent ourselves, hiding in shame, convinced of our failure. Or we got all needy at conferences, asking about job openings during the Q&A after the keynote, buttonholing a senior scholar over a drink as they desperately scanned the horizon for rescue.

And then we were offered a chance to be cooled: to adjunct, to be a postdoc. The time in the vacuum made us desperate for air, and we gasped “Yes!”