By Alex Fradera
To punch holes in a scientific claim, it’s legitimate to critique the supporting evidence, or query the way the evidence has been interpreted. More questionable is to throw dirt on the character or capability of the researcher making the claim. New research in PLOS One explores the effectiveness of ad hominem attacks against scientists and shows that some are more damaging than others.
Ralph Barnes of Montana State University and his team asked around 500 undergraduate students to review a number of scientific claims, each attributed to a named researcher, such as “Dr Martinez”. The claims were either obscure or invented, to avoid hot-button issues where the participants might already have strong immovable views (e.g. “According to Dr. Martinez at the University of Oklahoma, dibutylphthalate, a chemical used in Gold Bond foot powder, decreases the risk of some kinds of cancer”).
All the claims came with additional information that was either neutral (it didn’t challenge the claim) or that critiqued the researcher (ad hominem attacks) or the science behind his or her claim (e.g. pointing out issues with the supporting study itself, like the lack of a control group).
The ad hominem attacks either referred to past misconduct, like fabrication of previous data, conflict of interest (e.g. financed by a drug company with interests in the finding), education (e.g. low standards at the university from which he got a degree), or reports of sloppiness from his or her peers. One final type of criticism pointed at both the researcher and the study, for example an investigation that suggested he or she had fabricated this particular dataset.
Participants rated how much faith they held in each scientific claim. Ad hominems involving education and sloppiness did not sway participants from the default level of support that they showed when the claim was presented without criticism. But a conflict of interest or past misconduct lowered faith in the claim just as much as hearing about faults in the research supporting the claim.
What’s more, personal allegations of vested interest or past misconduct were just as damning as hearing that the researcher was implicated specifically in falsifying the very research that supported the claim they were making. It seems that participants could forgive suggestions that a researcher was not top-notch, but suspicions about their past motives and integrity were sufficient to pour cold water on their claims. This pattern was replicated in a sample of around 200 non-students with a wider age range (23-83, mean 49 years).
One criticism of the new findings is the abstract measure of participants’ faith in the claims (rather asking whether they would act on the claims, for instance). But previous use of the same paradigm has shown that this measure predicts more concrete intentions, like which of two prescription drugs a participant says they would personally use.
Although an ad hominem attack is considered a logical fallacy, the participants’ reactions in the new research are not strictly speaking illogical: if someone has cheated in the past, it may be reasonable for that to colour your view of their current work. Barnes and his team expected that this would especially be the case when the allegations of bad behaviour pertained specifically to the evidence underlying a researcher’s claims – but in fact the findings suggest that any allegations of past bad behavior, whether directly relevant or not, made a researcher’s claims appear equally suspect.
The strong effect of the conflict of interest ad hominem is also worth noting. It’s reasonably easy for critics to allege or imply a conflict of interest, especially for those willing to entertain a conspiratorial perspective. This may explain why groups hostile to scientific developments, such as vaccination, put such weight on these issues – because it works.