Thursday, March 27, 2008

A Study in Mistrust

Should research on lung cancer be trusted if it is funded by a tobacco company?

This is the question raised by a recent article in the New York Times (Cigarette Company Paid for Lung Cancer Study, March 26, 2008). The article notes a study which concluded that CT (computer tomography) scans could significantly decrease lung cancer deaths. The revelation the study was largely funded by money from tobacco companies has resulted in widespread skepticism about its results.

But is that a fair conclusion? The skeptics are arguing something like this: "The tobacco companies funded it, which creates a conflict of interests. The company funding the study is going to want a result that's in their favor. Therefore, the results are tainted, and cannot be trusted."

But isn't it ad hominem reasoning to say that, because of somebody's bad or selfish intentions, what they say cannot be trusted?

In defense of this reasoning, the article does note that:

Corporate financing can have subtle effects on research and lead to unconscious bias. Studies have shown that sponsored research tends to reach conclusions that favor the sponsor, which is why disclosure is encouraged.

But that's different from saying corporate-financed research is never accurate, or even that it's more often false than accurate. The article doesn't say what percentage of corporate-financed is flawed to the point of being worthless, or HOW MUCH more likely corporate-financed research is to be flawed than research funded by other means.

We and the media spend a lot of time focusing on conflicts of interest, full disclosure, and "following the money". But how much of this focus is truly valuable, and how much of it is just ad hominem?

Should we ignore any corporate-funded research that makes corporations look good?

Should we ignore any government-funded research that makes government look good?

Should we ignore any scientific research that makes scientists look good?

Is this kind of mistrust always reasonable? Or just sometimes? If so, how do we determine the difference between justified suspicion and ad hominem argument?

No comments: