I ran across an article from Ann Memmott recently about conflicts of interest in autism research. The article proper is worth your read, and so is the research paper she was responding to, but I also wanted to give a bit of background in what she’s talking about.
Her main point is that a lot of autism research has been done by people that have no right to be doing it. Most notably, a lot of Applied Behavioral Analysis professionals have produced most of the research on the subject. Obviously they’re going to want to say, “Yes, ABA works!” If they don’t, they lose their jobs. (Alas, this is not the only problem with ABA…)
A more typical example of this kind of thing would be a teacher being asked to grade her daughter’s homework. Or a judge presiding over the trial of a family member or close friend. A doctor who owns stock in a pharmaceutical company taking a job to test their products. In each case, someone with an obvious bias accepts the role of an impartial judge.
Poisoning the Well
In science, we call this a conflict of interest. Because it produces inaccurate, unfair, and untrustworthy data, educators always teach avoidance of conflicts of interest. Authors should be honest about their conflicts of interest. As a failsafe, publishers should keep them honest by doing their homework on each author. Presently, my research suggests publishers don’t do that homework much, if at all.
Left unchecked, the compromised researchers may publish data they made up. The funders of the research can push for types of research that’s likely to show the results they want. Or, more subtly, companies can demand a certain type of analysis and interpretation of the research results. Finally, companies can try to suppress research results they don’t like. In all cases, greed and ego twists or hides the truth.
It’s worth noting that even if the researcher chooses not to act on their biases, it still counts as a conflict of interest. The issue is that they have that chance, and may unconsciously act on it. Even as they do their best not to in service of the truth.
A single false study can influence future research. The way science works, researchers pick topics to study based on the results of previous research. So if product A is effective according to a false study, future research teams (who may or may not be corrupt) may choose to study variations of Product A. We hope that the future teams’ data will show the truth, but that doesn’t always happen. The future teams may simply conclude that Variation 1 of Product A was not as effective as the original Product A.
This is how whole streams of research theory can be based in lies, and turn out to be pointless. This is, Ann Memmott says, exactly what has been happening in autism research.
Hypothetically
When I served on the DOD’s Autism Research Program as a community reviewer, each application had a section to disclose conflicts of interest. We had every possible option to avoid committing the offense, right down to, “I suddenly realized I have a conflict of interest, weeks after I’ve read, rated, and reviewed this paper.” The organizers were very studious about sending identified conflicts of interest out of the room before each application.
The thing is, I don’t know whether any research was done into the reviewers to be sure they were honest. I never had any conflicts, because I’m not a member of any institution that does research. Beyond my typical existing biases, like “please don’t commit eugenics and erase autistic people from humanity,” I don’t have the connections and relationships with such entities. I’d certainly like to think the government did its due diligence and made certain no conflicts of interest affected the reviewers, but I truly don’t know.
I strongly suspect the Department of Defense wouldn’t publicize a conflict of interest in their scientific reviewers. It would reflect badly on the organization. They would probably blacklist the lying scientist and not hire them again, but not publicly denounce the offender. If they did publicly denounce the scientist, that scientist’s reputation would be ruined, and their career would be over. If the conflict of interest was accidental, that’s a good scientist’s life and career ruined forever. Instead, the organization would likely hire a replacement and move on.
Generalizing
It’s not much of a stretch to assume prestigious publications might take similar courses of action. In the current culture, people seem to distrust organizations that admit they made a mistake. Rather than risk their own reputations, a publication might decide to simply let the conflict of interest slide. Assuming, of course, they even noticed it in the first place.
Also, the number of research articles and proposals submitted for publication has risen over time. It costs money and time to hire people to look into every potential author. In the industry, the norm appears to be simply taking the potential authors at their word. In high profile cases, a publisher might retract a published study and condemn the author. But this doesn’t happen often.
Moving Forward
Honestly, I’d like to see jobs specifically hunting for conflicts of interest in the publishing industry. Clearly, we cannot rely solely on the honesty of researchers. It would be wise to create an additional check on the publication process.
As I understand it, this problem is not specific to the autism research community. Pharmaceutical research suffers from a plague of corruption exactly like this. Which means many possible jobs. These jobs could be remote, requiring only an Internet connection and a knowledge of how to find people’s publishing history, resumes/CVs, etc.
Honestly, this is the sort of job I might like doing. It’s a job detail-oriented autistic people overall might thrive in.
realisticautistic 