Near the end of her ethnography Contesting the Commons, Carolyn Lesorogol discusses some of the limitations of ethnography. She argues that while ethnographic research can obtain fine-grained and highly contextual information, “its very specificity makes it difficult to generalize, within or beyond a community” (Lesorogol 2008: 199). This shortcoming is often seen as a severe limitation for the applicability of ethnographic research to wider social problems and comparisons. How can larger theoretical conclusions be made from ethnography if there is no way to compare different sets of data? “To draw more general conclusions about behavior,” writes Lesorogol, “we require methods that enable us to control for local differences, as far as possible, in order to detect underlying patterns of behavior” (Lesorogol 2008: 199). Lesorogol’s answer to this problem: experimental economics. At least, that's one avenue she travels down.
In order to find a way to compare the behaviors of different groups of people, Lesorogol conducts field experiments using local participants in northern Kenya. She uses money as an incentive to gauge economic behaviors among individuals from the communities of Siambu and Mbarington (Lesorogol 2008: 198). The tests consisted of a series of “games” that were patterned after the MacArthur Foundation research that was conducted in “fifteen small-scale communities across the globe” (Lesorogol 2008: 201). The underlying assumption that Lesorogol (and the MacArthur researchers) make is that the behaviors exhibited in these experimental games are somehow transferable to other social and economic behaviors. This assumption is not clearly explained or justified.
Using similar methodologies and assumptions on another project, Henrich and McElreath (who worked on the MacArthur project and also conducted experimental economic games), write: “we think that the decisions in our model bear some resemblance—in terms of the framing of gains and losses—to the actual cropping decisions that farmers make” (Heinrich and McElreath 2002: 179). Neither Lesorogol nor Heinrich and McElreath explain why it makes sense to assume that context specific behaviors exhibited during these “games” should actually tell us anything about wider economic decision making practices. In a critical essay about Henrich’s use of experimental economics in a similar case, anthropologist Michael Chibnik writes, “The evidence that experimental play often mirrors interaction patterns in daily life is soft by the rigorous standards that the project contributors extol” (Chibnik 2005: 204). Chibnik concludes his essay by stating that the study “tells readers a lot about the reasons for cross-cultural variations” in how people engage in experimental economics games—and not much else (Chibnik 2005: 207).
Despite the fact that experimental economics was originally designed as a means for gaining a better understanding of human universals, in practice they often end up eroding and challenging the very universalist foundations upon which they are built. This was the case in Lesorogol’s research, although she argues that her use of experimental economics was intentionally designed to illustrate “how games can enhance understanding of a localized process of institutional change” (2008: 218). Considering Lesorogol’s earlier arguments about the need for cross-cultural and generalizable data, her conclusions and rationalizations about how she actually implemented the games seem like an afterthought. If she was really just looking to elicit context-specific cultural behavior, why use a generalizing method? Why not, as Chibnik argues, use classic ethnographic methods?
Regardless, Lesorogol’s data ends up illustrating the importance of both culture and context. The participants in Lesorogol’s experimental games clearly filtered their understandings of the games through particular cultural frameworks. In some of the games Lesorogol conducted, participants’ understandings of “ownership” strongly influenced their ensuing conduct in those games. As Lesorogol explains, “Even though the wording of the instructions is clear—the money is given to both players—the fact that Player One gains physical possession of the money and is given the right to decide on the split creates a situation of virtual ownership by Player One” (Lesorogol 2008: 217). This calls into question the cross-cultural value of any experimental games—whether conducted by Lesorogol or any one else. Why? Because the assumption that everyone will understand the situations and meanings of the games equally and comparably is just that—an assumption.
As Chibnik writes, “The isolation of a few variables for analytic purposes in experimental research is diametrically opposed to an entrenched tradition of holistic studies in anthropology in which attempts are made to consider the complex interactions of many variables” (2005: 202). Severe methodological and theoretical problems can arise with such economistic methods. One problem is the fact that the very assumptions of models can ultimately shape the data that they are designed to explain. By limiting research to a small set of pre-selected categories, the realities of the social group under study can become so abstracted that they lose any actual real world relevance. The danger with models is that it is all too easy to mistake them for the phenomena under study, as is illustrated by a passage from Henrich and McElreath:
Are these farmers actually reacting according to a utility curve, or is the curve merely a way of explaining their actions? Is it accurate to describe human behavior in terms of “utility curves”? At what point do these kinds of analytical categories and modes of analysis become completely irrelevant? Should there be any concern for how well these tools actually align with the actual categories that real people use to describe their own behaviors? These questions strike at the heart of the objectivist methodological tendencies that pervade the research of many economists and other mathematically minded social scientists. Of course, the problem goes far beyond the tendency to describe human behavior in these sorts of terms. The deeper issue is when actual human tendencies are conflated with the models that social scientists construct. This is the sort of issue that crops up when economists start talking about the reality of a generalized human "rationality" that adheres to particular cultural and historical ideals.
Divorced from all cultural and historical contexts, economistic models are quite “risk-prone” to reification. The conclusion of Henrich and McElreath’s 2002 paper is a wonderful illustration of the problems and dangers of reification. After a lengthy discussion and analysis of the games they conducted, Henrich and McElreath conclude, “it could very well be that humans have some predisposition toward taking risky monetary gambles” (Heinrich and McElreath 2002: 180). In short, after conducting a series of what were little more than tests that involved gambling, the authors wrap up their paper by making assumptions about human behavior based only upon possibilities that were determined by the researchers themselves. The reasoning behind this conclusion is that since most of the non-western participants were more willing to take risks in the games, there must therefore be some deeper human predisposition toward risky gambling. This evolutionary argument about human cognition and behavior is based upon the idea that Westerners, with a long history of interaction in market systems, have “acquired, via social learning, rules and preferences for dealing with risky monetary situations” (Henrich and McElreath 2002: 180). Obviously, the authors conveniently left Las Vegas (and the housing market in California) out of their model of Western society.
By eliminating or ignoring relevant cultural and historical factors that might explain their results, Henrich and McElreath only find answers that stem from the factors they deem (for whatever reason) relevant for consideration. Like a structuralist argument that assumes binary oppositions in human cognition and results in binary-based explanations, economically oriented models conveniently often result (only) in economic conclusions. This is a problem that Pierre Bourdieu criticized when he wrote, “if one fails to recognize any form of action other than rational action or mechanical reaction, it is impossible to understand the logic of all the actions that are reasonable without being the product of a reasoned design, still less of rational calculation” (Bourdieu 1990: 50). As Wilk and Cliggett argue, for Bourdieu human behavior can only be understood empirically through the observation and analysis of real world situations (2007: 186). What does this mean? It means that cultural, historical, and social factors cannot be dismissed in favor of calculatingly efficient, yet semantically barren, economic models. Efficiency is great and all, but what use is an efficient model that has little to do with empirical reality?
*This is part II of a larger project that I am working on called "What the hell do economists do and what does this have to do with anthropology?" Expect more, soon. Caveat: this is a work in progress, so feel free to lambaste me for any reason you see fit.
References
Bourdieu, Pierre. 1990. The Logic of Practice. Stanford: Stanford University Press.
Chibnik, Michael. 2005. “Experimental economics in anthropology: A critical assessment. American Ethnologist, Vol 32(2): 198-209.
Henrich, Joseph, and Richard McElreath. 2002. “Are Peasants Risk-Averse Decision Makers?” Current Anthropology Vol. 43 (1): 172-181.
Lesorogol, Carolyn K. 2008. Contesting the Commons. Ann Arbor: University of Michigan Press.
In order to find a way to compare the behaviors of different groups of people, Lesorogol conducts field experiments using local participants in northern Kenya. She uses money as an incentive to gauge economic behaviors among individuals from the communities of Siambu and Mbarington (Lesorogol 2008: 198). The tests consisted of a series of “games” that were patterned after the MacArthur Foundation research that was conducted in “fifteen small-scale communities across the globe” (Lesorogol 2008: 201). The underlying assumption that Lesorogol (and the MacArthur researchers) make is that the behaviors exhibited in these experimental games are somehow transferable to other social and economic behaviors. This assumption is not clearly explained or justified.
Using similar methodologies and assumptions on another project, Henrich and McElreath (who worked on the MacArthur project and also conducted experimental economic games), write: “we think that the decisions in our model bear some resemblance—in terms of the framing of gains and losses—to the actual cropping decisions that farmers make” (Heinrich and McElreath 2002: 179). Neither Lesorogol nor Heinrich and McElreath explain why it makes sense to assume that context specific behaviors exhibited during these “games” should actually tell us anything about wider economic decision making practices. In a critical essay about Henrich’s use of experimental economics in a similar case, anthropologist Michael Chibnik writes, “The evidence that experimental play often mirrors interaction patterns in daily life is soft by the rigorous standards that the project contributors extol” (Chibnik 2005: 204). Chibnik concludes his essay by stating that the study “tells readers a lot about the reasons for cross-cultural variations” in how people engage in experimental economics games—and not much else (Chibnik 2005: 207).
Despite the fact that experimental economics was originally designed as a means for gaining a better understanding of human universals, in practice they often end up eroding and challenging the very universalist foundations upon which they are built. This was the case in Lesorogol’s research, although she argues that her use of experimental economics was intentionally designed to illustrate “how games can enhance understanding of a localized process of institutional change” (2008: 218). Considering Lesorogol’s earlier arguments about the need for cross-cultural and generalizable data, her conclusions and rationalizations about how she actually implemented the games seem like an afterthought. If she was really just looking to elicit context-specific cultural behavior, why use a generalizing method? Why not, as Chibnik argues, use classic ethnographic methods?
Regardless, Lesorogol’s data ends up illustrating the importance of both culture and context. The participants in Lesorogol’s experimental games clearly filtered their understandings of the games through particular cultural frameworks. In some of the games Lesorogol conducted, participants’ understandings of “ownership” strongly influenced their ensuing conduct in those games. As Lesorogol explains, “Even though the wording of the instructions is clear—the money is given to both players—the fact that Player One gains physical possession of the money and is given the right to decide on the split creates a situation of virtual ownership by Player One” (Lesorogol 2008: 217). This calls into question the cross-cultural value of any experimental games—whether conducted by Lesorogol or any one else. Why? Because the assumption that everyone will understand the situations and meanings of the games equally and comparably is just that—an assumption.
As Chibnik writes, “The isolation of a few variables for analytic purposes in experimental research is diametrically opposed to an entrenched tradition of holistic studies in anthropology in which attempts are made to consider the complex interactions of many variables” (2005: 202). Severe methodological and theoretical problems can arise with such economistic methods. One problem is the fact that the very assumptions of models can ultimately shape the data that they are designed to explain. By limiting research to a small set of pre-selected categories, the realities of the social group under study can become so abstracted that they lose any actual real world relevance. The danger with models is that it is all too easy to mistake them for the phenomena under study, as is illustrated by a passage from Henrich and McElreath:
From this perspective, farmers are risk-averse because of the concave shape of their utility curves. In contrast, risk neutral farmers would have straight-line utility curves and always prefer the option with higher expected income/wealth. Risk-prone farmers would have convex utility curves (accelerating upward instead of decelerating downward) and prefer options with more variation, even when the expected income from a high-variance option is less from than a low-variance option (Henrich and McElreath 2002: 172-173).
Are these farmers actually reacting according to a utility curve, or is the curve merely a way of explaining their actions? Is it accurate to describe human behavior in terms of “utility curves”? At what point do these kinds of analytical categories and modes of analysis become completely irrelevant? Should there be any concern for how well these tools actually align with the actual categories that real people use to describe their own behaviors? These questions strike at the heart of the objectivist methodological tendencies that pervade the research of many economists and other mathematically minded social scientists. Of course, the problem goes far beyond the tendency to describe human behavior in these sorts of terms. The deeper issue is when actual human tendencies are conflated with the models that social scientists construct. This is the sort of issue that crops up when economists start talking about the reality of a generalized human "rationality" that adheres to particular cultural and historical ideals.
Divorced from all cultural and historical contexts, economistic models are quite “risk-prone” to reification. The conclusion of Henrich and McElreath’s 2002 paper is a wonderful illustration of the problems and dangers of reification. After a lengthy discussion and analysis of the games they conducted, Henrich and McElreath conclude, “it could very well be that humans have some predisposition toward taking risky monetary gambles” (Heinrich and McElreath 2002: 180). In short, after conducting a series of what were little more than tests that involved gambling, the authors wrap up their paper by making assumptions about human behavior based only upon possibilities that were determined by the researchers themselves. The reasoning behind this conclusion is that since most of the non-western participants were more willing to take risks in the games, there must therefore be some deeper human predisposition toward risky gambling. This evolutionary argument about human cognition and behavior is based upon the idea that Westerners, with a long history of interaction in market systems, have “acquired, via social learning, rules and preferences for dealing with risky monetary situations” (Henrich and McElreath 2002: 180). Obviously, the authors conveniently left Las Vegas (and the housing market in California) out of their model of Western society.
By eliminating or ignoring relevant cultural and historical factors that might explain their results, Henrich and McElreath only find answers that stem from the factors they deem (for whatever reason) relevant for consideration. Like a structuralist argument that assumes binary oppositions in human cognition and results in binary-based explanations, economically oriented models conveniently often result (only) in economic conclusions. This is a problem that Pierre Bourdieu criticized when he wrote, “if one fails to recognize any form of action other than rational action or mechanical reaction, it is impossible to understand the logic of all the actions that are reasonable without being the product of a reasoned design, still less of rational calculation” (Bourdieu 1990: 50). As Wilk and Cliggett argue, for Bourdieu human behavior can only be understood empirically through the observation and analysis of real world situations (2007: 186). What does this mean? It means that cultural, historical, and social factors cannot be dismissed in favor of calculatingly efficient, yet semantically barren, economic models. Efficiency is great and all, but what use is an efficient model that has little to do with empirical reality?
*This is part II of a larger project that I am working on called "What the hell do economists do and what does this have to do with anthropology?" Expect more, soon. Caveat: this is a work in progress, so feel free to lambaste me for any reason you see fit.
References
Bourdieu, Pierre. 1990. The Logic of Practice. Stanford: Stanford University Press.
Chibnik, Michael. 2005. “Experimental economics in anthropology: A critical assessment. American Ethnologist, Vol 32(2): 198-209.
Henrich, Joseph, and Richard McElreath. 2002. “Are Peasants Risk-Averse Decision Makers?” Current Anthropology Vol. 43 (1): 172-181.
Lesorogol, Carolyn K. 2008. Contesting the Commons. Ann Arbor: University of Michigan Press.
No comments:
Post a Comment