We applaud Henrich et al. for their cogent demonstration of the need for more representative samples in psychological research in order to permit generalization to the human species. However, even if participant samples are representative, the psychology elicited by experiments that require participants to make judgments in response to hypothetical situations, answer abstract questions, or behave in response to artificial laboratory tasks may not be representative of – or even very informative about – human psychology across most domains of life. To understand human nature, our methods must explore the psychology of natural human experience. By this, we mean that psychological theory must be grounded in detailed observation and description of everyday life across cultures in order to understand the cognitive mechanisms that operate in the naturally occurring situations to which humans are adapted.
Observation- and description-deprived (ODD) research programs often wander far from real-life psychology because they become increasingly oriented to addressing the precedents and frameworks of previous ODD research and theory. For example, Henrich et al. show how theories about judgments of fairness and cooperation have to take into account the culture of the people making those judgments. But there is a further problem with basing our psychological theory on studies of economic games: Behavior in artificial games does not correlate strongly to social behavior in the community (Gurven & Winking Reference Gurven and Winking2008). If the cooperative behavior and fairness judgments we want to understand are those that occur in everyday behavior in communities, rather than only those that are specific to the particular artificial framework of the Ultimatum Game, then we need real-world validity. That is, we need experiments that make sense to participants because the psychology of the experiment matches the psychology of behavior in the real world. ODD psychological research programs rarely provide evidence regarding the ecological validity of the results based on experiments that typically use extremely impoverished stimuli, severely restrict responses, or are based on hypothetical scenarios and Likert-scale responses to questions about abstract concepts.
Moreover, we can only interpret data if we know how participants have interpreted the research situation, the task, and the stimuli. For example, WEIRD people are used to identifying themselves and stating their interests and values, and typically welcome the opportunity to do so. But even WEIRD Scandinavians find this an uncomfortably unnatural practice; on the first day of seminar, a Scandinavian student whose turn comes to “say a bit about yourself” is embarrassed and confused by this American practice, which feels uncomfortably self-promoting (Lotte Thomsen, personal communication). In Burkina Faso, Moose informants find any personal questions threatening and demeaning: “To ask about my thoughts, desires, or activities is to seek control over me, possibly in order to thwart or harm me.” Given this interpretation of an interview or questionnaire, Moose responses mean something quite different from the responses of Americans, who tend to perceive personal questions as a welcome invitation to assert themselves and make themselves look good.
If we do not know the psychocultural meaning of an experience, we cannot understand the meaning of responses to the experience. Attachment research is based on infants' behavior when their caretakers leave them in a “strange situation.” In most African cultures, infants are on the body of their mother or other close kin much of the time, sleep with them, and are never out of sight of their immediate family members. Families expect children to form inalienable bonds of interdependence with them. In contrast, German infants sleep alone, are often left in daycare for many hours with strangers, and in early childhood are expected to play alone and are often left at home unattended (LeVine & Norman Reference LeVine, Norman, Moore and Mathews2001). German parents foster self-reliance and autonomy. Consequently, being separated from the mother in the strange situation has completely different meanings for African and German infants, so their responses cannot be directly compared.
Psychological theory over the past 40 years has been formulated mostly on the basis of prior theory, data, and intuitions. As researchers are largely from WEIRD populations, the theoretical constructs that inform experimental design tend to be based on WEIRD intuitions and stimulated by ODD data and theory. If our goal is to understand human thinking and behavior in the world, we must leave our desks and begin collecting an extensive and rich body of naturalistic descriptive data based on various kinds of observation. Currently, experimental papers are accorded the highest prestige in psychology and comprise the vast majority of studies published in top journals (Cialdini Reference Cialdini2009). Experimental controls are invaluable, but they are useless if the constructs being tested are invalid or the experiment elicits unnatural psychology. A natural science of psychology should be based on naturalistic study in the real world of diverse situations in diverse cultures. Just as botany, zoology, ecology, geology, astronomy, chemistry, and physics grew out of, constantly return to, and must ultimately be validated by observations of the natural world, so psychology should be. In addition to relying on analytic and functional approaches, psychological constructs should be cultivated inductively from observation and we should grow our theories by contemplating naturally occurring patterns of action. Experiments are one way to test such constructs and theories, but they are not the only way. Often they are insufficient if we want to make inferences about behavior outside the lab, beyond key presses and pencil marks.
You have
Access
We applaud Henrich et al. for their cogent demonstration of the need for more representative samples in psychological research in order to permit generalization to the human species. However, even if participant samples are representative, the psychology elicited by experiments that require participants to make judgments in response to hypothetical situations, answer abstract questions, or behave in response to artificial laboratory tasks may not be representative of – or even very informative about – human psychology across most domains of life. To understand human nature, our methods must explore the psychology of natural human experience. By this, we mean that psychological theory must be grounded in detailed observation and description of everyday life across cultures in order to understand the cognitive mechanisms that operate in the naturally occurring situations to which humans are adapted.
Observation- and description-deprived (ODD) research programs often wander far from real-life psychology because they become increasingly oriented to addressing the precedents and frameworks of previous ODD research and theory. For example, Henrich et al. show how theories about judgments of fairness and cooperation have to take into account the culture of the people making those judgments. But there is a further problem with basing our psychological theory on studies of economic games: Behavior in artificial games does not correlate strongly to social behavior in the community (Gurven & Winking Reference Gurven and Winking2008). If the cooperative behavior and fairness judgments we want to understand are those that occur in everyday behavior in communities, rather than only those that are specific to the particular artificial framework of the Ultimatum Game, then we need real-world validity. That is, we need experiments that make sense to participants because the psychology of the experiment matches the psychology of behavior in the real world. ODD psychological research programs rarely provide evidence regarding the ecological validity of the results based on experiments that typically use extremely impoverished stimuli, severely restrict responses, or are based on hypothetical scenarios and Likert-scale responses to questions about abstract concepts.
Moreover, we can only interpret data if we know how participants have interpreted the research situation, the task, and the stimuli. For example, WEIRD people are used to identifying themselves and stating their interests and values, and typically welcome the opportunity to do so. But even WEIRD Scandinavians find this an uncomfortably unnatural practice; on the first day of seminar, a Scandinavian student whose turn comes to “say a bit about yourself” is embarrassed and confused by this American practice, which feels uncomfortably self-promoting (Lotte Thomsen, personal communication). In Burkina Faso, Moose informants find any personal questions threatening and demeaning: “To ask about my thoughts, desires, or activities is to seek control over me, possibly in order to thwart or harm me.” Given this interpretation of an interview or questionnaire, Moose responses mean something quite different from the responses of Americans, who tend to perceive personal questions as a welcome invitation to assert themselves and make themselves look good.
If we do not know the psychocultural meaning of an experience, we cannot understand the meaning of responses to the experience. Attachment research is based on infants' behavior when their caretakers leave them in a “strange situation.” In most African cultures, infants are on the body of their mother or other close kin much of the time, sleep with them, and are never out of sight of their immediate family members. Families expect children to form inalienable bonds of interdependence with them. In contrast, German infants sleep alone, are often left in daycare for many hours with strangers, and in early childhood are expected to play alone and are often left at home unattended (LeVine & Norman Reference LeVine, Norman, Moore and Mathews2001). German parents foster self-reliance and autonomy. Consequently, being separated from the mother in the strange situation has completely different meanings for African and German infants, so their responses cannot be directly compared.
Psychological theory over the past 40 years has been formulated mostly on the basis of prior theory, data, and intuitions. As researchers are largely from WEIRD populations, the theoretical constructs that inform experimental design tend to be based on WEIRD intuitions and stimulated by ODD data and theory. If our goal is to understand human thinking and behavior in the world, we must leave our desks and begin collecting an extensive and rich body of naturalistic descriptive data based on various kinds of observation. Currently, experimental papers are accorded the highest prestige in psychology and comprise the vast majority of studies published in top journals (Cialdini Reference Cialdini2009). Experimental controls are invaluable, but they are useless if the constructs being tested are invalid or the experiment elicits unnatural psychology. A natural science of psychology should be based on naturalistic study in the real world of diverse situations in diverse cultures. Just as botany, zoology, ecology, geology, astronomy, chemistry, and physics grew out of, constantly return to, and must ultimately be validated by observations of the natural world, so psychology should be. In addition to relying on analytic and functional approaches, psychological constructs should be cultivated inductively from observation and we should grow our theories by contemplating naturally occurring patterns of action. Experiments are one way to test such constructs and theories, but they are not the only way. Often they are insufficient if we want to make inferences about behavior outside the lab, beyond key presses and pencil marks.