Sugar, the SAIRR and conflicts of interest

The South African Institute for Race Relations (IRR) identifies itself as a liberal think-tank, focusing on research and ideas that might conduce to investment and economic growth.

As you’d know, the term “liberal” means many different things to different people. On my terms, the IRR has often been empathy-light, in that its idealistic vision of what the data tells us we should do can sometimes appear blind to important social and political contexts.

It is however not those issues that I want to address here, but instead yesterday’s news that the IRR’s research on South Africa’s proposed sugar tax was funded in part by Coca Cola.

The obvious concern here is possible conflict of interest (CoI), and unfortunately, the IRR has done themselves no favours in how they have responded to the news. Furthermore, they’ve done public thinking on the general issue of such conflicts no favours either, choosing to make a good point in exactly the way that guarantees it won’t be heard, in that they beat the accusation away with a logic-stick while ignoring a genuine concern that underlies it.

CoI’s have become a fairly routine accusation, especially on social media, and they generate immediate suspicion and are used to discredit findings even if there is no scientific or other problem with those findings.

They are, in short, sometimes used as lazy shortcuts for people to dismiss some finding they don’t like, rather than to engage with opposing views.

But we forget that public perceptions are usually non-expert views, and that even though appearance matters (in a political sense), it says nothing about the quality of the science or the data. The IRR response (correctly) says that the “expose” doesn’t rebut any of their arguments, but chooses instead to let the claim of CoI alone do the work.

What the IRR get completely wrong, though, is to deny that they ever had any reason to disclose this source of funding (on Twitter), while engaging in a very tenuous bit of tu quoque in the response linked above.

(You’d need to read the piece to follow this, but in short, because there’s no obvious connection between being paid by the private sector and being anti-government, as there is in the case of being paid by Coca Cola and being pro-sugar, whether or not actual bias is manifest.)

There are at least two reasons to be concerned about CoIs: 1) CoIs can compromise the integrity of research; and 2) CoIs can undermine the public’s trust in science or the researchers. It’s the second problem that’s most relevant here.

There is a strong positive correlation between the source of funding and the results of research: for example, studies sponsored by pharmaceutical companies tend to favor the companies’ products. So, this means that we have a prima facie reason to think funding can be a warning flag.

The way you avoid suspicion, or having your research discredited, is a) to make robust arguments, but also b) to disclose things that are going to make people suspicious of you, in order to make it clear you have nothing to hide.

It’s a matter of self-interest. A later revelation of perceived conflict, especially one that is dismissed with a sneer, will have a greater impact on your credibility than initial disclosure would have done. This is the case even if it’s true (as I believe it is) that we shouldn’t automatically dismiss findings because of tainted funding.

A quick note on “perceived conflicts”, to conclude. If you’re in the business of being an “independent think tank”, how things appear to people starts to make a bigger difference, whether or not there is an actual CoI.

When one produces research for public consumption, or even just as a journalist, conflict of interest rules are there precisely to avoid any inquiries into your motives, and to preserve the trust of your readership (who are not in a position to investigate your motives). The CoI rules are a signal of your compact with them.

While I don’t want to reinforce the misconception that tainted funding axiomatically means tainted findings, there is nevertheless a clear ethical failing in not disclosing a source of your funding, where that source would have a clear vested interest in a certain outcome.

This is in part because we know that as much as we might try, we’re not immune to pressure (or even something less sinister, like confirmation bias). We also know that there is a positive relationship between funding sources and sympathetic research outcomes.

You show that you’re aware of these issues, and that you’ve done your best to avoid them, by being transparent about things that might be of concern to the public. When you fail to do so, and respond defensively when that failure is pointed out, your commitment to an ethical relationship with that public is placed in serious doubt.

Also published on Medium.