Methodological Individualism and the Selfish Gene

23 August 2006 at 1:54 am Nicolai Foss 10 comments

| Nicolai Foss |

As a staunch advocate of methodological individualism in the social sciences, I have often experienced the following comment at seminar presentations and in conversations: “Why take the individual as the explanatory atom? Why not go further in the direction of reductionism and begin analysis with the selfish gene?”

The comment is usually (though not always) intended to suggest that an advocacy of methodological individualism is fundamentally arbitrary and that there is no reason why individuals should have a privileged status in an explanatory sense. However, the comment is based on a fallacy, which Livia Markoczy and Jeff Goldberg (1998) call the “driver-seat fallacy.” To wit:

It is all too common for people to imagine that evolutionary psychologists and others are claiming that our thoughts and emotions are driven by our genes … This fallacy misunderstands the way genes work. Genes build bodies. … Once the body is built, the genes have no control or influence on what those bodies do. It makes no more sense to say that genes drive our thoughts and emotions than it does to say that genes pump our blood. Our heart pumps our blood and our brain drives our thoughts and emotions … Our genes are not in the driver’s seat, we are.

Thus, the selfish gene argument against methodological individualism is a red herring.

Entry filed under: - Foss -, Methods/Methodology/Theory of Science.

Kauffman Entrepreneurship Research Portal Oskar Morgenstern on Economic Data

10 Comments Add your own

1. The Filter^ | 24 August 2006 at 8:25 am

In defense of methodological individualism

One of the underlying reasons why some economist’s disagree with others is a central methodological premise. Austrian’s (and free-marketeers more generally) believe in methodological individualism – i.e. that the unit of analysis should be the indivi…
2. Bo | 24 August 2006 at 8:38 am

Rather than talking about unit of analysis as if one is better (or more relevant) than the other, I believe we should focus on multi-level issues – that is we should identify and model accordingly the multiple levels of analysis necessary to adequately research a phenomena. Of course, this begs the question: how many levels?

Charles Perrow (1986) noted the existence of 12 levels of analysis: individual, group, department, division, organization, inter-organization, organizational set, networks, industry, region, nation, and world.

While I don’t suggest all these levels be included in every study, I DO suggest that one focuses on multi-level issues pertaining to the specific research question under investigation. Aggregation (and disaggregation) runs the risk of ecological or atomistic fallacies. Modelling cross-level effects holds great promise for both traditional (e.g., your Austrian economics stuff) and untraditional research…
3. Nicolai Foss | 27 August 2006 at 1:31 pm

Bo, You are right to the extent that we should let our choice of unit of analysis and level of analysis be dictated by the problem at hand. As you know, I also greatly sympathize with your emphasis on multi-level modelling. That being said, I maintain that the level of the individual is a very special one as only individuals can act. Hence, the requirement that ultimately all aggregate phenomena be reducible to the actions of individuals.
4. Bo | 29 August 2006 at 5:23 am

Nicolai,

I am not sure I entirely agree with you. While you may be right that the individual (level) is a special one (particularly in management studies), one could argue that certain types of questions (e.g., industry analysis or FDI studies) benefit little from disaggregating to the individual level. It is possible (as I illustrated with my soccer team example and the Hoefstede cultural example) that reducing the level of analysis to the individual may lead to incorrect inferences:

Disaggregation may lead to ‘miraculous’ multiplication of the number of units. For instance, it you sample 10 managers in the first stage and 10 managerial decisions by each manager in the second stage, you may end up with in total 10×10=100 decisions. One might then disaggregate the data to the level of decision-making (decisions) and estimate the relation between the experience of the manager and the outcome of the decision, without taking into account that some decisions are made by the same manager – this is like pretending that there are 100 independent observations, when in reality there are only 10 (the managers). Hence, disaggregating the data as if they are independent may lead to a dramatic exaggeration of the sample size (and may lead to serious type I error). If only measures are taken at the micro level, analyzing the data at the micro level is correct, AS LONG AS one takes into account that observations at a macro-unit may be correlated.

My point is that when you deal with a multi-level issue, disaggregating to the lowest (individual) level may not always be beneficial and should be treated with great care. Having read several of your papers, Nicolai, you still seem to owe a convincing argument for WHY it is a REQUIREMENT that all aggregate phenomena can be reduced to the level of the actions of individuals (or maybe this simply escaped me?).
5. Nicolai Foss | 29 August 2006 at 7:47 am

Bo, I think you misunderstand me: According to the doctrine of methodological individualism — to which I happen to subscribe — all aggregate social phenomena must be reducible in principle to the individual level, for the reason that ultimately only individuals act (i.e., not firms, groups, nation states, etc). However, pragmatically we agree that it is not always recommendable to reduce analysis to the individual level. We can use aggregate constructs and neglect the individual level, as long as we trust that it is possible to theorize the relevant individual level mechanisms.
6. Bo | 30 August 2006 at 3:39 am

Thanks Nicolai – this makes sense – however, although the firm may not act, individuals act as a group within its boundaries, influenced by administrative heritage, policies, procedures, culture etc…

Could you provide a reference to this methodological individualism doctrine?
7. Nicolai Foss | 30 August 2006 at 3:57 am

Bo, This is a BIG, BIG issue in the literature on the philosophy and methodology of social science. Classical statements by Popper, Hayek, Kincaid, Elster and countless others. Begin by checking it on Wikipedia and take it from there.
N
PS. I don’t think groups can act. Only individuals can act.
8. Bo | 30 August 2006 at 7:23 am

Thanks,

and of course you are wrong. Groups of fish (called Schools funny enough) are knows to “act” in unison. Some schools of small fish react simultaneously to external stimuli (e.g., threat) by forming a school that resembles a larger fish..how they coordinate this among up to 1000s of individuals remains a mystery.
9. teppof | 30 August 2006 at 7:17 pm

Bo and Nicolai: I see a you have a spirited discussion here.

Bo, you asked about references for the “doctrine” – many of the issues you are bringing up are wrestled with in the following forthcoming paper (sorry[!] for the ‘self-plug’ – but Nicolai alerted me to your comments as they relate to my research):

Felin, T. & Hesterly, W.H. (forthcoming). The knowledge-based view, heterogeneity and new value creation: Philosophical considerations on the locus of knowledge. Academy of Management Review.

Also, some other foundational sources below (from the above paper’s bibliography):

Nagel, E. 1961. The structure of science: Problems in the logic of scientific explanation. New York: Harcourt, Brace & World, Inc.

Popper, K.R. 1959. The logic of scientific discovery. London: Hutchinson & Co.

Rosenberg, A. 1995. The philosophy of social science. Boulder, CO: Westview Press.

Sawyer, R.K. 2001. Emergence in sociology: Contemporary philosophy of mind and some implications for sociological theory. American Journal of Sociology, 107: 551-585.

Scaltsas, T. 1990. Is the whole identical to its parts? Mind, 99: 583-598.

Tuomela, R.1990. Methodological individualism and explanation. Philosophy of Science, 57:133-140.

Udehn, L. 2001. Methodological individualism: Background, history and meaning. London, New York:
Routledge.
10. Bo | 31 August 2006 at 5:56 am

Thank you much!