200 likes | 386 Views
How WEIRD are they? Disentangling Psychological and Pragmatic Processes. Thomas Holtgraves Dept. of Psychological Science Ball State University. Overview.
E N D
How WEIRD are they? Disentangling Psychological and Pragmatic Processes Thomas Holtgraves Dept. of Psychological Science Ball State University
Overview • Overreliance on WEIRD participants distorts behavioral research; wrong to assume one can generalize from research with WEIRD Ps. Old and perennial issue in social psychology. • Three responses • 1. Some psychological differences may reflect differences in pragmatic rules rather than differences in psychological processes. • 2. A focus on phenotypic differences may hinder detection of underlying genotypic principles • 3. Generalizability is an empirical question that is now easier to address
1. Pragmatic not Psychological • Henrich et al. examples are all psychological, many social psychological • Visual perception (Muller-Lyer illusion), cooperation in economic decision making, self-concept, moral reasoning, etc. • Some observed differences in these variables may be due to methodological differences (Schweder; Baumard & Sperber) • Even if identical methods are used, the manner in which those methods are interpreted by Ps may differ; i.e., pragmatic differences
Pragmatics of Experiments • All behavioral science experiments involve communication • E explains to P what the study is about, what they are to do, and so on • Some psychological effects may in fact be pragmatic effects (N. Schwarz; R. Wyer) • Rather than psychological effects, these reflect pragmatic effects (i.e., principles regarding conversation, etc.)
Pragmatics of Experiments • Example: Representativeness heuristic(Kahneman & Tversky) • Ps given individuating (Jack is conservative, apolitical, spends time on his hobbies, etc.) and base-rate (description comes from a panel of 30 engineers and 70 lawyers) information • What is probability Jack is engineer? Ps indicate engineer • Interpreted as a cognitive bias (Ps ignore base-rate info) • However, information provided by E comes with a presumption of relevance; and so Ps use it (even if logically, they shouldn’t).
Pragmatics of Experiments • Fundamental Attribution Error (FAE) = Tendency to make dispositional attributions for behavior and ignore situation factors (e.g., Fred slipped because he is clumsy). • Henrich et al. Western/Individualistic Ps more likely to demonstrate FAE than non-Western Ps • Attribution difference or pragmatic difference?
Pragmatics of Experiments • Classic Fundamental Attribution Error studies (e.g., Jones & Davis): Ps provided with information about another’s (Joe) behavior and asked to make a judgment (what is Joe’s true opinion) • Some of the information is not relevant for an attributional judgment • e.g., Joe was told to write a pro-Castro essay • Ps read Joe’s essay and are asked about Joe’s opinion • Ps indicate that Joe has a favorable attitude toward Castro (even though he was told to write the favorable essay; FAE).
Pragmatics of Experiments • Pragmatic alternative: Ps assume that information (the writer’s essay) is given to them for a reason (i.e., it’s relevant). • And so they use it • Possible pragmatic differences: The presumption of relevance may be less strong in non-Western cultures • OR • What counts as relevant may vary over cultures
Psychological Difference vs. Pragmatic Difference • Economic games • WEIRD Ps understand the “as if” nature of the games; they have a schema that makes it understandable • Non-WEIRD Ps have no schema for this. They try to make sense by…
2. Different Manifestations of Same Underlying Process • Identical underlying psychological process may be expressed differently across cultures • Observed differences (phenotypic expression) may be explicable at more abstract level of analysis (genotypic process) • Overemphasis on differences can be misleading • Example: Is self-enhancement universal? • Henrich et al., argue unique to WEIRD Ps • Gaertner et al. argue self-enhancement is universal but that cultures vary in terms of what traits they enhance on (West = individualistic traits - independence; East = collectivist traits; devotion to family)
Pragmatics: Politeness and Conversation Processing • Brown & Levinson Politeness Theory • Politeness (via facework) is assumed to be universal • But, much cultural variability in terms of: • How politeness is expressed • What is face-threatening • And so on • So, even though much cultural variability in linguistic politeness, that variation reflects an underlying universal motivation (face)
Politeness as Universal and Variable • Empirical evidence: Holtgraves & Yang (1992) • Examined levels of reported politeness as a function of Power, Distance, and Imposition in U.S. and S. Korea • Results: • Culture and gender differences in overall politeness and positive politeness
Underlying Cultural and Gender Similarities and Differences • Similarity: Politeness varied as a function of power and distance for U. S. and S. Korean, and for male and female. • Difference: • Power weighted more heavily by S. Koreans than U.S. • Power weighted more heavily by males than by females • Distance weighted more heavily by females than by males
Request Politeness as a Function of Distance for U.S. and S. Korean Ps
Request Politeness as a Function of Hearer Power for Male and Female Ps
Maxims as Universal and Variable • Grice’s maxims (quantity, quality, etc.) probably not universal • However, violations of maxims (regardless of what they are) prompt deeper processing (i.e., extra-cognitive activity in order to make sense of the violation). • Hence, both universality and variability
3. Generalizability as a Tractable Empirical Question • Whether results using WEIRD samples generalize is an empirical issue • Now easier to undertake these types of analyses • Internet and survey platforms • Amazon Mechanical Turk • Possible to recruit and pay Ps for participation • Possible to select relevant samples (e.g., male S. Koreans between 18 and 35)
Mechanical Turk • Especially useful for pragmatics research • Can create and present scenarios, stimulus sets, etc. • Collect judgments of appropriateness, interpretations, reaction times, etc. • Can specify non-WEIRD samples • Can include validity checks (fake items to leave blank)
Internet (Mturk) Data Quality and Limitations • Internet/Mturk data quality (Buhrmester, Kwang, & Gosling, 2011) • Data are as reliable as data collected via traditional methods • Compensation (e.g., $.01 U.S.) doesn’t effect quality (but does have an effect on speed). • Limitations • IRB • Internet availability
Recap • 1. Some psychological differences are pragmatic differences. • E.g. Fundamental Attribution error. • Others? • 2. Observed differences may reflect universal principles • 3. Tests of generalizability now easier and encouraged