Thursday, June 13, 2013

Michelle Meyer: Miller Interacted, Intervened

Michelle Meyer had more to say about my comments on Geoffrey Miller's fatshaming troll-tweet than would fit in Blogger's comment box. I post her comments here, along with some responses.

Meyer's Comments

Hi Zach,

For your readers' convenience, here's what I said “re: latest in #matingmind flap”:

(1) I remain skeptical he intended this as research but intent is (alas) key under regs & he claims it was.

(2) If he did intend to contribute to generalizable knowledge, his tweeps were hum subjects b/c involved interaction (& maybe intervention).

(3) Only real loophole available to him is if his investigation wasn’t “systematic,” but most IRBs have low threshold for that (or ignore).

(4) Lots of reasons to dislike research regs/IRB system, but this isn’t an especially good case for deregulating "just talking" research.

(5) Even if you think a study 2 test reactions 2 provocative tweets shouldn’t be subject to IRB review, until regs change, they apply to all.

(6) Real issue neither free speech nor hum subj rsch but: does his statement reflect bias that affects performance of his academic duties?

(7) Lesson to academics trying to disown controversial statement: claiming “research” will only move you from the frying pan into the fire.

(8) Oh yeah. Even if he thot his rsch was exempt, like most, NYU's IRB gets to decide when research is exempt or not [link omitted]

I stand by this analysis -- which isn't to say that I disagree with you that Miller has "a case." As we know, IRBs come out differently on “studies” like this. If I wanted to argue that he obtained "data through intervention or interaction with" living human beings, I’d focus on interaction. He didn't merely passively post the tweet, sit back, and collect data about reactions, as you suggest; rather, he responded to at least one person who questioned what she understood to be his claim that obesity and IQ are inversely correlated (

He also "unfollowed" at least one person who disputed his tweet (yours truly). This interaction was, presumably, also part of his study: Why react with sincere anger to someone who was provoked by a tweet which you order to provoke? (I await his refollow, now that the study is over.)

Alternatively, consider that his research involved "manipulations of the subject's environment." He claims that he issued "provocative" tweets over a few months in order to gauge responses. You might say that Twitter is already rife with trolling, such that this couldn't possibly have altered that environment. There's something to that. We'd need to know more about his hypothesis and methods. But it's certainly plausible that he sought to ratchet up the level of provocativeness in his tweets until they sparked a reaction that was measurable even against Twitter’s baseline of trolling and outrage. Not every asinine tweet goes viral, prompts Tumblrs, and makes national news, but after months of hard work, he managed to create one -- for science! -- that did.

Consider, also, the expectations of those following a tenured academic on Twitter rather than, say, a teenage boy. Arguably, such a "provocative" tweet manipulated the environment of *his* tweeps in a way that it might not have for a teenage boy’s tweeps. Although Miller had been provocatively tweeting for months, note that he claims that all of those tweets were part of his study. So at some point he deliberately altered the environment of his tweeps by shifting from non-provocative to provocative tweets.

Finally, consider the ripple effects in that manipulated environment. I spent an hour going back and forth with one of his followers who insisted that I caused the tweet to go viral, “tattled" to his employer (I did not), and cost Miller his career (for the record, I don't think he should be fired and I fully support his right to make claims that are patently falsifiable with a single counter-example; I'd just like some oversight to ensure that his views about obese academics don't ruin *their* careers). If this was all good (deceptive) research fun, then I wish Miller would debrief his follower-subjects so that they can stop blaming others for reacting in exactly the ways Miller presumably hypothesized that they would react.

Some Responses from Schrag

Thanks for these comments, Michelle.

We agree the most about your point 6: "[The] real issue [is] neither free speech nor [human subjects research] but: does his statement reflect bias that affects performance of his academic duties?"

Assuming it does, what should his department do with a faculty member who may have violated the American Psychological Association's standard against unfair discrimination? (The standard prohibits discrimination on the basis of disability, and the EEOC has maintained that "severe obesity" is a disability. Moreover, I doubt that the authors of the ethical principles would be happy about any discrimination based on body weight.) Dare we apply the standard in Dixon v. University of Toledo to a professor?

We also agree about the first part of your point 1. I too am "skeptical he intended this as research," and the University of New Mexico seems skeptical as well, noting, "We are looking into the validity of this assertion." It strikes me as more likely that when Miller claimed to be conducting "research," he was using the term to mean, "I am a troll."

I think you are mistaken to write, "intent is (alas) key under regs." Intent matters under 45 CFR 46.119, but the relevant section here is §46.102, which defines research by design, not intent. Thanks to Robert Levine, we know exactly why this is so. So if Miller intended to conduct research, but failed to design a project "to develop or contribute to generalizable knowledge," he's clear.

You ably lay out the grounds for considering Miller's tweet and subsequent actions to constitute intervention or interaction, but I find them unpersuasive. I would be very unhappy were any IRB to find that any speech in a public forum constitutes "manipulations of the subject or the subject's environment"; if so, what speech is not?

And I would like to know if you can distinguish between Miller's provocation and that of Daniel Callahan. Callahan came in for his fair share of outraged attention in the national press and on Tumblr. He, too, has engaged with his critics, albeit more professionally than Miller. Is this, too, sufficient interaction to trigger IRB jurisdiction?


Michelle Meyer said...

Hi Zach,

On the design/intent point, I acknowledge Bob Levine’s history of the Commission, in which he says the staff eventually came to conclude that “intent should not be used because you could never measure intent. We compromised on the term, ‘design.’ You can measure the design.” Although I’m sympathetic to his point about the difficulty of measuring intent, I am not inclined to rely on such “legislative history” of a bioethics commission in interpreting agency regulations that seem, to me, to say something a bit different, even if unwittingly.

I read § 102's definition of research--"a systematic investigation . . . designed to develop or contribute to generalizable knowledge"--to require both intent and design: both are necessary, and neither is sufficient, for an activity to constitute “research.” To say that an investigation is “designed to” do X implies an agent who did the designing with that purpose (X) in mind. (See, e.g., intelligent design theory.) If § 102 had instead referred to “a systematic investigation . . . *capable of* developing or contributing to generalizable knowledge,” that would be different. But, alas, it does not.

So, as you say, someone could intend to conduct research but be so inept at research design that her activity is insufficiently systematic to count as “research” for purposes of the Common Rule. But, conversely, I think that an actor who engages in a systematic activity whose fruits *could* contribute to generalizable knowledge but who lacks that intent (e.g., she takes a poll to satisfy her own curiosity) isn’t conducting “research” for purposes of the regs. I don’t think IRBs are often faced with a scenario in which design is divorced from intent in either of these ways. And in any case, Miller said he intended this to be research, and as I said originally, IRBs tend to have a low bar for what constitutes a sufficiently systematic investigation, so I doubt that he would escape IRB jurisdiction that way.

Instead, I’d think the more natural way for him to try to wriggle out of IRB jurisdiction would be to argue that no “human subjects” were involved in his research. On this point, you rightly press me to distinguish Miller (research) from Callahan (speech). This takes us back a bit to the first point, about intent.

Callahan didn’t publish his essay *in order to* elicit, measure, learn from, and report others’ reactions; his essay itself was the *result* of his (non-human subjects) research. Similarly, Internet trolls do what they do, I gather, to entertain themselves or to hurt others, but not to contribute to generalizable knowledge. Institutions engage in quality improvement or assurance activities in order to benefit themselves and/or their clients. And so on. Intent became legally relevant when Congress asked the Commission to distinguish between research and practice, and the Commission decided that the latter “refers to interventions that are designed solely to enhance the well-being of an individual patient or client and that have a reasonable expectation of success” (Belmont Report). You might subject someone to the exact same experimental procedure, with the same risk-benefit profile, but your intent in doing so – whether to try to help them or to produce generalizable knowledge -- makes all the difference for how your activity is regulated: as research or as "innovative treatment." That’s what I meant when I lamented that intent was central to which activities get regulated and which don’t; we ought to focus more on activities' foreseeably harmful consequences, not actors’ intent.

Zachary M. Schrag said...

In your comment on my earlier post on Miller, you pointed me to evidence that Miller said he sought to "to measure people’s reactions," which strikes me as a statement of both design and intent. I agree that this could distinguish Miller from Callahan.

Since design and intent seem to be aligned in this case, I would say it's not such a good opportunity after all to split these hairs. And if we are very lucky, before such a case comes along, the Common Rule will be revised with a clearer definition of what needs to be overseen.