Are economic experiments performative?22.01.2018
Last week, a workshop should have been held at Witten, however, it had to be cancelled because of the impact of ‘Friederike’ on German railways and air traffic. That was a pity, because we planned to discuss interesting topics. One is the PhD thesis by Johanna Jauernig at Technical University of Munich which applies economic experiments on ethical issues.
I am deeply sceptical about economic experiments, especially in ethics. Therefore, I was keen to debate this at the workshop. Psychological and economic experiments differ in two fundamental respects: First, economists always tell the truth about design and content of an experiment to the participants, and secondly, they always use monetary incentives. These rules are most basic, if you submit an experimental paper violating these rules to an economics journal, it will receive a desk rejection on the spot.
I think that this is a quandary in research about human behaviour, because psychologists and economists often deal with the same issues, although shaped in a different language (for an excellent discussion, see Tyler and Amodio 2015). At a conference in Japan about the methodology of economics last year, we discussed these problems in depth. I presented a paper that argues that economic experiments are performative. What does that mean?
First, psychologists would not tell the truth about an experiment because they assume that knowing the true content of an experiment would bias the responses of the subjects. This is tantamount to recognizing that the experiment is performative: Subjects perform the experiment based on the information that they have, and there is no guarantee that they do it as the experimenter conceives. However, economists give strong reasons why they apply full transparency (what Johanna does in her work, too): For example, if subjects know that they are not truthfully informed about the experiment, they might guess experimenters’ intent, thus introducing another bias, which, however, is not transparent to the experimenter, in turn, because s/heshe cannot access internal mental states of subjects. Personally, I think that these are good reasons, but the psychologists’ reasons are good reasons, too. Thus, taking both together, we may turn rather pessimistic about the value of experiments of any kind, at least in the lab (this kind of critique is also familiar from economics, where some researchers think that only field experiments, especially ‘natural’ ones, are valuable). Anyway, the economists factually also recognize the performative nature of experiments, they only presume, as economists always do, that full information will trigger rational responses, and therefore, subjects will perform the experiment as the designer imagined, since this design has been done having rational subjects in mind.
Second, I think that the use of money as an incentive creates real troubles, especially when dealing with ethical issues. Following Vernon Smith’s classical work on methodology, monetary incentives are used because they allegedly trigger the necessary effort and attention of the subjects, who might otherwise introduce too much random noise in their responses. But, and I think this is a big ‘but’, there is rich and strong empirical evidence in psychology that money primes (or, in economic parlance, ‘frames’) change behaviour in a substantial way (for an exemplary overview, see Vohs et al. 2006). Now, the entire set-up of economic experiments is a very strong priming procedure, beginning with advertisements that emphasize the easy money that you can earn with participating, and with all the exhortations that make people aware of the pecuniary losses that they might incur when they do not follow the rules. In other words, it is a reasonable assumption that participants in economic experiments are money-primed.
The behavioural effects include, for example, individualization of effort, less cooperation, or increasing personal distance from others. In other words, people primed with money act as the proverbial rational economic agents. Turning to ethics, this is exactly what philosophers such as Michael Sandel argue when they suggest that the scope of markets in regulating social relationships should be subject to deep ethical scrutiny. That means, however, that the use of economic experiments in ethics creates a methodological quandary: If priming with money causes participants to perform the experiment as rational economic agents, we cannot learn anything about how they behave as ethical subjects in other contexts. Of course, that does not mean that the experiment has no empirical value: We only need to look at the context in the field and show that people are primed with money there, too. This argument is too familiar from intellectual history: Simmel’s entire ‘philosophy of money’ assumes that the modern personality is primed by money, so to say. But we also need to recognize that people are ethically wise: As Zelizer has shown in a classical study, people design rules which govern the reach of money in social life, creating space of ethical behaviour.
Therefore, I think that experiments are performative indeed, and if you want to treat them as empirically valid, you need to find the same performative structures in the field.
Herrmann-Pillath, Carsten, Creating Social Ontology: On the Performative Nature of Economic Experiments (August 27, 2016). Available at SSRN: https://ssrn.com/abstract=2830769 or http://dx.doi.org/10.2139/ssrn.2830769
Tyler, Tom R. and David M. Amodio (2015): Psychology and Economics: Areas of Convergence and Difference, in: Fréchette, Guillaume R., and A. Schotter, eds. Handbook of Experimental Economic Methodology. New York, NY: Oxford University Press, pp. 181-196.
Vohs, K. D., Mead, N. L. & Goode, M. R. (2006). The Psychological Consequences of Money. Science, 314, 1154-1156.
Zelizer, Viviana A. (1997): The Social Meaning of Money, Princeton: Princeton University Press.