On undisclosed conflicts of interest in medicine, science, and skepticism

I’ve written about conflicts of interest (COIs) a lot over the years. COIs are important in medicine and science because, as much as physicians and scientists like to think that they are immune to such things, we are as human as anyone else. We are just as prone to unconsciously (or consciously) being influenced by self-interest related to our COIs. Most of the time, for purposes of science, COIs are considered to be mostly financial in nature: employment or payments from a drug company, a financial interest in a treatment being studied, and the like. Andrew Wakefield is a classic example in that he was developing a separate measles vaccine that would compete with the MMR vaccine that he tried to nail as a cause of autism. However, conflicts of interest are not just financial. They can be ideological. For example, I have come across antivaccine articles published in ostensibly respectable sources in which the author does not divulge that she is affiliated with “vaccine safety advocacy group” (a.k.a. an antivaccine group). These sorts of COIs can be as important, if not sometimes even more so, than financial COIs. After all, money is just money, but an intense ideological belief is unlikely ever to be removed and can warp one’s perspective even more than money. Sometimes, COIs take the form of something about one’s background that is relevant to a topic being investigated or discussed or a personal experience that strongly influences one views. These non-financial COIs are, without a doubt, under appreciated, even in the skeptic movement.

That’s why I’ve become very insistent that we, as skeptics, scientists, and physicians, need to be totally up front about our conflicts of interest, be they financial, ideological, or personal. One reason, of course, is that those who—shall we say?—don’t share our dedication to rationality, science, and critical thinking will be very quick to point them out if we don’t do so first, but that’s not the most important reason. The most important reason is to be better skeptics. We need to honestly admit and recognize anything that might compromise our objectivity or lead us to conclusions that are not the ones best supported by science and the evidence. Once we know our own skeptical weaknesses in the form of COIs, we can work on trying to mitigate them. In many ways, financial COIs are the easiest to deal with, because they’re far more straightforward. When one has a personal experience that informs one’s views on a topic or has a strong ideological commitment to a point of view, it’s often hard to tell where skepticism devolves into motivated reasoning.

You know what’s also bad? False accusations are bad.

Indeed, I think that, without a doubt, we can all agree that false accusations of serious crimes and misdeeds are bad things, horrible things, a terrible things, things that can ruin reputations and lives and even end up with people dying over them. If there’s anyone who disagrees with this contention, he’s one warped person with whom I want no part. I also think that, without a doubt, skeptics can agree that examining false accusations is completely within the purview of skepticism, that no accusation is off-limits to legitimate skepticism that truly uses science, reason, and critical thinking and doesn’t devolve into trying to discredit the presumed victim. In that vein, Ben Radford’s post on his Center for Inquiry (CFI) blog A Skeptic Reads the Newspaper entitled The Anatomy of False Accusations: A Skeptical Case Study, seems, at first glance, like an entirely reasonable deconstruction of false accusations.

At first. It doesn’t take long, however, to notice problems. It starts out with an anecdote (and, as we all know from medicine, the plural of “anecdote” is not “data”), and it’s a damning one. It’s the story of a college student who falsely accused her former boyfriend of abduction and sexual assault. It turns out that she made the false accusation because her grandmother had discovered images of her and the accused engaged in consensual sex acts. The rest of the post is a litany of more of the same. One false accusation recounted by Radford even resulted in the death of the falsely accused because the boyfriend of the woman who made the false accusation shot the man with whom she was having sex. The reason? He thought she was being raped.

The further I read, the more disturbed I became. For one thing, until near the end the article was relentlessly one-sided, its purpose clearly being to give the impression that false accusations of sexual assault are common. Oh, sure, towards the end Radford quotes Alan Dershowitz to concede that “most people who are accused of a crime are in fact guilty.” However, the overall message I got from his blog post was that false accusations of rape and sexual misconduct are common, making his concession that most people don’t lie about such things seem half-hearted, particularly in the context of the lack of high quality evidence to support his view in his post. Again, the plural of “anecdote” is not “data,” and Radford, disappointingly, went for anecdotes instead of data.

Now, here’s where I reveal that I know something that many of you don’t know (although, I daresay, many of you do). What those of you who aren’t into the skeptical movement probably don’t know is that last summer, the author of this piece, Ben Radford, was publicly accused of sexual harassment by Karen Stollznow. Now, let me make one thing very clear. I make no judgment as to whether Radford is actually guilty of sexual harassment. I don’t know. I don’t have enough information to know, because all I know is what Stollznow wrote about it (an article that was later removed) and some of what flew back and forth on atheist blogs for a few weeks. For purposes of this discussion of COIs, it really doesn’t matter. For purposes of my discussion of disclosing COIs, it’s utterly irrelevant to me whether Radford is guilty or not.

Now, how does Radford’s post read? Different, doesn’t it? Knowing this about him, I find it hard to view his post as anything more than an attempt at self-justification and a means of casting doubt on his accuser—even if such was not his intent. How would I have reacted to his post if he had disclosed his COI up front? I don’t know for sure. Probably not as badly as I did with his not having disclosed it. No, definitely not as badly as I did. However, what irritates me is what people who don’t know the back story will see. They will tend to assume that Radford is reasonably disinterested, trying to apply skepticism and critical thinking to the issue of false accusations. He is, after all, a prominent skeptic, writing on his employer’s blog, and his employer is CFI, which is dedicated to promoting skepticism and critical thinking. What Radford denied such readers is a piece of evidence necessary to help them evaluate his arguments, namely the bias of the writer. The closest Radford ever gets to admitting his COI is this paragraph:

It may be hard to sympathize with a man or woman falsely accused of a crime unless you’ve been in that situation yourself. Many people may assume that they would never be in relationship with a person who would falsely accuse them of something as serious as sexual harassment or sexual assault. However the fact is that any of us could be in that position; the man Levitski accused of abduction and assault was a friend and recent sex partner, who presumably had no idea what she was capable of. Think about how you would feel if this happened to your wife, husband, daughter, son, brother, sister, mother or father.

A perceptive reader, even one who knows nothing about the back story here, might suspect from this paragraph that Radford’s interest in the topic is more than just academic, but he would not have any way of doing more than suspecting this.

Unfortunately, Radford’s post is also badly reasoned and lacking in evidence. I was going to provide some examples and pick it apart a bit in my own inimitable way, other than pointing out its near-total reliance on anecdotes as I’ve already done, but it turns out that I don’t have to. Here’s what I mean. When I first saw Radford’s post and decided to write about it, I was also annoyed at CFI. Why, I thought, did CFI allow Radford to use its blog as a platform to grind his his own personal axes? Believe it or not, given how happy and pleased I was that my very first major article had just seen print in CFI’s flagship publication, Skeptical Inquirer (it’s a primer on Stanislaw Burzynski coupled with an article about how skeptics have become active again opposing him), I even felt a little trepidation as I wrote this. I wondered whether I would ever be invited to give a talk at a national CSI conference again, the way I was in 2012, or whether I’d ever see any of my articles in print again in the pages of Skeptical Inquirer. It was almost enough to make me stay my typing hands and look to another topic I had had in mind for today before I became aware of Radford’s post. Radford is, after all, very influential in CFI. If I were to piss him off, it wouldn’t result in a profane rant directed at me at TAM this year in which a certain large magician took umbrage about something I wrote about him, but it could have negative effects on my aspirations to be more influential. I don’t know if those fears are unreasonable, but I’m less worried now that I’ve seen another post on a CFI blog.

It turns out that Ron Lindsay, president of CFI, has actually written a response in which he noticed the same sorts of problems that I did. His post is reasoned and balanced, and he basically eviscerates Radford’s arguments right from the very title of his post, Evidence-Based Reasoning: Comments on a Blog Post. Now, I’ve had my issues with Lindsay in the past, in particular over an incident three years ago, but in this case Lindsay is spot on. For example:

In the first paragraph, Ben notes, in referencing the Iowa case, that “The relative obscurity of this case suggests its prevalence.” No, it doesn’t. Obscurity does not imply prevalence. This is fallacious reasoning. Right now, someone in obscure, rural Latvia could be falsely accusing someone else of being a philosopher. The obscurity of this event does not imply that false attributions of philosophizing are prevalent.

Exactly. Radford’s argument is a non sequitur. It does not follow from the obscurity of a case of a false accusation that false accusations are prevalent. It could just as easily imply the opposite. Evidence is needed to make the connection, and Radford didn’t provide any.

Then there’s this:

That false reports happen is not disputed. Nor does anyone dispute that for the individual falsely accused, it’s a very unfortunate, sometimes tragic, situation. But is this a widespread problem? That’s the key question. One might think so from the attention Ben has given to it and his use of the adverb “often,” but, actually, the evidence seems to indicate it is not a widespread problem. For example, a British study last year indicated that there were 35 prosecutions for false accusations of rape during a 17-month period while there were 5,681 prosecutions for rape in the same period of time. The suggestion that false accusations of rape are commonplace does not appear to be supported by the evidence. Moreover, this suggestion can be very harmful if it persuades people that reports of rape should be treated with special suspicion.

Exactly (again). Radford provided no context and no evidence to support his implication that such false accusations happen “often.” Most evidence, as Lindsay points out, actually tends to point in the opposite direction, namely that false accusations of rape are uncommon and that, if anything, rape is underreported, although he doesn’t mention that the issue is highly politicized and you can find outlier studies with very high numbers. In any case, Radford didn’t make even the most superficial attempt to look at the evidence. He just slung anecdotes. That’s the point. That’s where another major skeptical fail was, in addition to Radford’s glaring failure to disclose his personal COI regarding false accusations of sexual misconduct. We don’t let quacks, cranks, pseudoscientists, and antivaccinationists get away with making assertions using only anecdotes to support their conclusions. We should hold the luminaries of the skeptical movement to the same standards.

Think of it this way. No one disputes that in scientific and medical research it’s important to disclose one’s financial COIs. If discussed the way I discussed above, few would argue that it’s not also important to disclose COIs that might imply a strong ideological COI, such as antivaccinationists who publish review articles and research purporting to find a link between vaccines and autism who don’t mention that, oh, by the way, they are on the board of directors of an antivaccine group, although such COIs tend to be treated much less seriously than financial COIs. Fewer people would insist that disclosing COIs like those of Ben Radford, life events that have the potential to massively impact one’s objectivity, is critical, but I would. If you want to claim to be a skeptic and to persuade an audience of skeptics, you need to be completely open about such a potent personal COI. More importantly, if you want to be honest with yourself, it’s even more imperative to do so. The same is true of science. Ruthless self-examination and openness about sources of our potential biases can only help us develop as skeptics. We all have biases, and we all have potential COIs. Acknowledging them and being honest about them, are the first step in overcoming them, because you can’t overcome them if you fail to admit that they exist.