Whatever Will Bee, Will Bee

November 8, 2015

I warned you last week that I was heading to a meeting about the difficulties of industry-generated science, so you shouldn’t be too surprised that I am following up on that this morning. Here’s the context: we just don’t trust claims made by representatives of the food industry, even when those claims are putatively backed up by science. And even if the claims are made by nominally disinterested parties (like university scientists), we still don’t trust them if the putatively disinterested party was paid to do the research by the food industry. Most people who have been looking at this have come to the conclusion that the problem is actually quite subtle. On the one hand, there are obvious cases where for-profit firms have perpetrated frauds: they have effectively been lying to us. Volkswagen came up in this connection, as did the very recent allegations that Exxon perpetrated a disinformation campaign on climate change that was directly contrary to what their own scientists’ were reporting in internal documents.

On the other hand, these incidents don’t speak to the heart of the problem. It would be rare for university scientists or other disinterested parties to be enrolled in this kind of deceit. The problem resides more in the way that certain research methods can be predictably associated with certain kinds of results, and that though it might be scientifically valid to do research of that sort, there may be unasked questions (to which other methods might have been amenable). As I intimated a couple of weeks back, it’s easy enough to identify the university scientists who are going to do stuff you like, and then to give the money to them, rather than someone who is doing stuff you might not like. We don’t trust the industry or the scientists who work for them not because we think they are lying to us, but because we suspect that they are just not asking the questions that really matter.

As it happens, there were a couple of bee papers at this conference. You may have read that our bees are in trouble. The technical name for this is colony collapse disorder, and the leading hypothesis is that it is caused by a relatively new class of pesticides called neonicotinoids. In this connection, the discussion took up some industry funded research that did not involve collaboration with university scientists or even disinterested parties. It was a collaboration between a company that makes neonicotinoid pesticides and does not want them banned (and hence has an interest in disproving this hypothesis) and beekeepers who have an interest in regulating neonicotinoids, if the hypothesis is shown to be true.

One paper by Sainath Suryanarayanan from the University of Wisconsin claimed that industry research in cooperation with beekeepers had used methodologies that skewed results by limiting the timeframe in which observations would be collected and by discounting beekeeper observations that were out of the mainstream. They got away with this, Suryanarayanan claimed, because they were able to assert an authority for Science (with a capital S) that the claims of working beekeepers lacked. He then sketched an alternative method that accords more epistemic authority for the knowledge claims of beekeepers.

Now here would be an excellent opportunity for the obligatory Thornapple Blog tangent to discuss “epistemic authority”, but instead I’ll just say that it’s a term that is intended to convey whether or not some person or group’s claim to know something should be taken seriously. If, for example, you think that I am a trusted source of knowledge about neonicotinoids just because I can spell the word, you have seriously overestimated my epistemic authority.

There was also a paper by Iain Kelly who works for Bayer CropScience, a company that is a major supplier of neonicotinoids. Kelly claimed that the evidence for “the leading hypothesis” is not so clear and that one of the difficulties that his company had in conducting research with beekeepers is that they would say one thing the first week and something entirely different the next. Frankly, I’m not really sure that there was a deep logical contradiction between the way that Kelly described this research with beekeepers and the way that Suryanarayanan described it. These two papers did not get us to the point where we could ask ourselves, “Should the EPA ban neonicotinoids just because beekeepers don’t like them?” or even to “How should the evidence provided by beekeepers be weighed in making regulatory decisions about neonicotinoids?” Kelly did say that he (and we can presume Bayer CropScience) did learn some things that will influence future research from their collaboration with beekeepers. He didn’t say exactly what it was, but it might have had something to do with epistemic authority.

I can’t entirely bee sure. Que sera, sera.

Paul B. Thompson is the W.K. Kellogg Professor of Agricultural, Food and Community Ethics at Michigan State University



2 thoughts on “Whatever Will Bee, Will Bee

  1. Professor,

    He y, research is research. Assumptions a re assumptions.Conclusions are conclusions. Dictate the right ass umptions and buy me a c uppa and you’re entitled to the right conclusions.

    Or, put it this wa y — in one meeting of a hospital corp board I was on we were presented two financial reports. In themorning, things looked bleak. Wemighjt go bankrupt. This was the report submitted to maximize Medicare reimb ursement.

    In the afernoon we were presented a glowing report.Very profitable. We were assured as the Board that this is thejob the CEO is really doing.

    Can we bee pretty sure that this kind of honey pot is wherever you want to take this metaphor?

    Regular Reader #2


  2. Congratulations on your favorite son getting into medical school.

    He will not allow his conclusions to be bought. I bee sure.


Comments are closed.