This post is based on a paper, “Cross Cultural Differences in Decisions from Experience: Evidence from Denmark, Israel, and Taiwan,” authored by Sibilla Di Guida, Ido Erev, and David Marchiori. It is a 2015 working paper of ECARES. It immediately reminded me Richard Nisbett’s Geography of Thought.
Richard Nisbett in Geography of Thought provides interesting insights into such differences. He divides the world into Easterners and Westerners. Easterners have difficulty in recognizing changes in objects, while Westerners cannot recognize changes in backgrounds. Easterners believe that the world is complicated and inscrutable. Westerners believe that they can understand the world. Westerners create simple and useful models that can be tested, but tend to focus on the object and slight the possible role of context. Westerners are particularly susceptible to the fundamental attribution error–thinking other people’s actions are explained by what they are, while my actions are explained by circumstances. The table below sets out some more distinctions.
The paper examines the effects of different cultural backgrounds (participants are students in Denmark, Israel, and Taiwan) on decisions from experience. I wanted to like this paper, but the samples are so small and the chances of confounding so big that I could not make much of anything from it. Admittedly, Nisbett expanded on similar studies to create the table above, but at least he made it interesting. For all I know there were more students native to Taiwan in the Denmark sample than in the Taiwan sample. It really is my fault for thinking that the title represented something more definitive than a hundred or so students pushing some buttons. A hundred studies like this and you might have something. Nevertheless, let me give the gist.
Previous research shows that individuals from East Asian cultures tend to expect more
changes in the environment than individuals from Western cultures. For example, when presented with a graph that summarizes a decreasing trend in economic growth rates, subjects from the Peoples’ Republic of China (PRC) were twice more likely than American to predict a change of the trend in the next period.
The main goal of the current paper was to clarify the implications of these cultural differences in the context of repeated decisions from experience. They examined behavior in repeated choice tasks in which people had to rely on their own personal experience. Specifically, they focused on behavior in the basic clicking paradigm. Instructions were something like this: “The current experiment includes many trials. Your task, in each trial, is to click on one of the two keys presented on the screen. Each click will be followed by the presentation of the keys’ payoffs. Your payoff for the trial is the payoff of the selected key.”
For instance, In Problem 1, the action corresponds to the gamble (-10 with p = .1; +1 otherwise); this choice has negative expected return (EV = -0.1), but it yields the best payoff in 90% of the trials. In Problem 2, the action, which corresponds to (+10 with p = .1; -1 otherwise), has positive expected return (EV = +0.1), but yields the worst payoff in 90% of the trials. The results in both problems reflects deviation from maximization. That is, the typical participant behaves “as if” he/she does not pay enough attention to the rare (10%) outcomes. The authors predicted that since people from East Asian cultures expect more changes in the environment experience in East Asia would be characterized by lower best reply rates than in the West. According to the researchers the belief that the environment is likely to change should reduce the attractiveness of the action that has led to the best payoff in the previous trial. In actuality, analysis of only sequential dependencies revealed significant cultural differences.
The explanation provided by the authors assumes that people use two types of similarity rules, i.e., temporal similarity, and sequence-based similarity. The use of temporal similarity implies that the current trial is more similar to the most recent one. The use of sequence-based similarity implies that a trial that occurs after a particular outcome pair (e.g., 0 from S, +10 from R) is more similar to the previous trials that have occurred after that specific outcome pair. The authors hypothesize that the subjects from the West are more likely to use sequence-based similarity rules when the sequences of recent outcomes are relatively simple (i.e., with only three possible outcomes), but are less likely to use sequence-based similarity rules when the number of possible outcomes is large (i.e., at least seven possible outcomes). So more complex trials lead Westerners to follow simple temporal rules, while the opposite is true for Easterners.
To summarize, the study found similar indications of underweighting of rare events in Denmark, Israel, and Taiwan. but found that subjects from the West are less likely to expect a change in a multi-outcome world, and more likely to expect a change when the set of possible outcomes is small.