Experiments on politicians: South Africa edition

1294 South African local politicians were each contacted by a fictitious constituent, whose name signaled either co-ethnicity with the politician or out-group membership. The constituent raised a concern either about a public goods issue that is prioritized equally across South African ethnic groups or about one that is more ethnically divisive.

I found that politicians of all ethnic groups were more likely to respond to co-ethnic constituents than to non-ethnic constituents (even when the non-ethnic constituents were co-partisans) and were much more likely to respond to a unifying issue than to a divisive issue. The paper concludes by discussing whether the findings provide an argument for descriptive representation.

A new paper by Gwyneth McClendon, a PhD candidate at Princeton. Super interesting.

Conversations with a colleague, who has been thinking much on this issue, have gotten me thinking: targeting politicians experimentally is getting increasingly common, and there are clearly important things to learn, but what’s the basis for human subjects approval (and ethical research) in these kinds of experiments? There can’t be consent or it would negate the experiment. These are public figures, so do we get to forego consent?

Maybe people aren’t worried about the ethical treatment of politicians, but the same question applies to experiments with information to voters, or public messaging. My gut tells me that most are okay, but gut feelings are not a very sophisticated basis for the ethical conduct of a profession. Researchers do bad things all the time, and not only is it bad for subjects, but it’s bad for other researchers who might suffer from any spillover of mistrust.

I don’t see much issue with this specific paper, but still, my gut tingles a bit with fictitious constituents. Reader comments?

 

7 Responses

  1. Paul Krugman at about the 0:38 mark addresses this version of racism in the context of why America never embraced socialism. The “us” and “them” problem that was prevalent in the USA but not Europe.
    http://www.youtube.com/watch?v=H9zXdlVDQVc&feature=player_embedded

    If your concerned about contacting politicians just for the sake of response…may I direct you to a series called “Talking To Americans” by Rick Mercer which used to be aired on the CBC (so yes your friendly neighbors to the north used government fund to air/produce this…sorry about that). At least Ms McClendon has loftier goals than a laugh.

  2. I agree with Gwyneth (and with the discussion of this very issue in the Butler and Broockman paper). I think if it is essential for the research question and if the burden placed on the individual is kept as low as possible, it is perfectly ethical. You don’t want to displace real constituent concerns.

    I also would be more hesitant to do this in a resource-constrained place in general (i.e. some underfunded African bureaucracy, versus a Congressman with a large staff).

    I also don’t think you should debrief them. That seems more likely to change future behavior, which I don’t think is a good thing, generally speaking.

  3. Thank you, Chris, for linking to the paper. I had hoped that the paper would generate more discussion around the ethics, logistics, and theoretical value of conducting this kind of research.
    In the case of this particular study, I hope that people will take a look at the section of the paper devoted to ethical considerations. In particular, I used deception in this study only because the hypotheses under scrutiny would be nearly impossible to evaluate without it and because the hypotheses are theoretically, practically, and normatively important. I took care to minimize the burden placed both on the politicians and on the resources at their disposal, as well as to protect the identities of individual politicians. (Using real constituents that live in the actual municipalities represented by these politicians, incidentally, introduces complications for experimental control which might be worth more discussion.) See Butler’s and Brookman’s article in the most recent AJPS for a similar approach in their study of the behavior of U.S. politicians. As I note in my paper, researchers at South African institutions (including at the University of Cape Town and at the South African Institute of Race Relations) also reviewed and advised on the ethics of the experiment, even though this step is not required by current IRB protocol since these researchers were not official collaborators. In the future, IRBs might consider suggesting or requiring this kind of in-country review.
    Another issue that might be worth further discussion is the issue of debriefing. Occasionally, other studies in which subjects were contacted by fictitious people (e.g. fictitious potential employees, fictitious students) have debriefed the subjects afterwards. But I have not seen this done in most field experiments testing for discrimination. In my case, the IRB advised against it, because the subjects are public officials being observed conducting their public duty. As far as I know, there are few guidelines on whether debriefing should be used in field experiments using deception. Any thoughts on this issue would be much appreciated.

  4. Is it any different because it is politicians and not other individuals? The same arguments made by List in his JEP piece presumably apply – i.e. if there is no physical harm and the politicians don’t ever know they are in an experiment, then perhaps no need for consent.
    I don’t see this as much different from the racial discrimination experiments in which fictitious resumes are sent to firms. In both cases there is some minimal cost to the politician or firm (taking time to answer a fake constituent or fake job-seeker), but this cost is quite low.
    But I agree, this is an issue in which standard ethical guidelines are not well-developed, so it would be nice to see more discussion and emerging norms on this – and learn more about the other types of experiments being performed on politicians.

  5. It might be possible to go around the ethical issues by including in the field experiments “real” subjects with legitimate concerns.