Lesson 3 of 8 · 16 min
WEIRD research and international collaboration
Explain why findings from Western, educated, industrialised, rich and democratic samples may not apply elsewhere, and why cross-cultural research needs international collaboration.
What you will be able to do
- Explain what WEIRD stands for and why most psychological research has WEIRD participants.
- Explain why findings from WEIRD samples may not be fully applicable in other contexts, using evidence.
- Identify the design, variables and limitations of a cross-cultural comparison.
- Explain why international collaboration is important when investigating cross-cultural phenomena.
Most of what you have learned in psychology came from studies of a particular kind of person. In 2010, Joseph Henrich, Steven Heine and Ara Norenzayan gave that kind of person a name: WEIRD, from Western, Educated, Industrialised, Rich and Democratic.
| Measure | Value |
|---|---|
| Participants from the United States | 68% |
| Participants from Western industrialised countries (North America, Europe, Australia, Israel) | 96% |
| Share of the world's population living in those countries | 12% |
| American samples in the leading social psychology journal made up only of psychology undergraduates | 67% |
Use the table. What percentage of the world's population lived outside the Western industrialised countries that supplied 96% of participants?
Does it matter? Only if people differ. If basic processes like perception worked identically everywhere, the sample would not matter. Henrich and colleagues reviewed evidence that WEIRD people are often unusual, not typical, on measures from visual perception to fairness and reasoning. A finding from WEIRD participants may not be fully applicable to people in other contexts: its external validity (in particular, its population validity) is limited.
The clearest example is visual. In the Müller-Lyer illusion, a line with inward-pointing arrowheads looks longer than an equal line with outward-pointing arrowheads.
The Müller-Lyer illusion. Line b (arrowheads pointing in, fins spreading out) looks longer to most WEIRD viewers, though the lines are the same length.
Marshall Segall, Donald Campbell and Melville Herskovits (1966) tested how strong the illusion was in 16 societies, 14 of them small-scale societies. They varied the lengths and found the point at which people judged the two lines equal. American undergraduates were the most susceptible: they needed one line about a fifth longer before the two looked equal. The San foragers of the Kalahari were not affected at all. A textbook written from WEIRD data alone would have described a "universal" illusion that is not universal.
Segall and colleagues suggested that growing up surrounded by rectangular buildings and straight edges (a "carpentered world") trains the visual system to read angles as depth. That is a hypothesis about why, and it is still debated; the difference itself is the point here.
Classify the variables in a comparison of the Müller-Lyer illusion in two cultural groups. Notice what kind of variable "cultural group" is.
Key idea
Culture cannot be randomly allocated
A cross-cultural comparison treats cultural group as the independent variable, but researchers cannot assign people to a culture. Groups that differ in culture also differ in schooling, environment, familiarity with tests and much else. So these studies are quasi-experimental (or correlational): they can show a difference between groups, but any of these confounding variables could explain it, and the study cannot show that culture caused it.
A researcher compares the strength of the Müller-Lyer illusion in adults who grew up in a large city and adults who grew up in a remote community with few rectangular buildings. What type of investigation is this?
More traps in cross-cultural research.
- Measures may not mean the same thing. A questionnaire translated word for word can change meaning. Researchers use translation and back-translation (translate, have someone else translate it back, compare) and check that items work the same way in each culture.
- Unfamiliar tasks. People who have never sat a timed test, or seen line drawings on paper, may score lower for reasons unrelated to what is being measured, which threatens validity.
- Researcher bias. Researchers interpret behaviour through their own culture (ethnocentrism) and may misread what they see.
- Ethics and cultural protocols. Research with a community needs that community's consent and involvement. The syllabus names protocols for working with the knowledge of First Nations peoples: research is done with communities, not on them.
Key idea
Why international collaboration matters
Researchers from each culture involved can design measures that make sense locally, translate and check them, recruit participants respectfully, and interpret results through local knowledge. Collaboration across many countries also gives large, varied samples, so a finding can be tested for whether it holds across cultures.
- The International Comparative Study of Ethnocultural Youth brought researchers from 13 societies together to study 5,366 immigrant youth aged 13 to 18, plus 2,631 national youth, with the same measures (Berry, Phinney, Sam & Vedder, 2006).
- Many Labs 2 (Klein et al., 2018) repeated 28 published findings in 125 samples, 15,305 participants from 36 countries and territories. For most findings, results varied little between samples, which no single-country study could have shown.
Worked example
Evaluate a claim from a WEIRD sample
A news article reports: "Psychologists have shown that people are happiest when they make decisions on their own." The study it describes asked 200 psychology undergraduates at a university in the United States to rate their happiness after choosing a prize themselves or having it chosen for them.
Evaluate the article's claim with reference to the sample.
Step 1: Describe the sample
Psychology undergraduates at one US university: young, educated, from a Western, industrialised, rich, democratic society, a WEIRD sample.
Step 2: Identify the problem
The claim is about "people" in general, but the finding may reflect the culture of the sample. The US is strongly individualist, where personal choice is valued; in cultures that tend to value group decisions, having a choice made by family or the group may not reduce happiness.
Step 3: Name the validity issue
The study may have good internal validity, but its external (population) validity is limited: the results may not be fully applicable to people in other cultural contexts, or even to older or less educated Americans.
Step 4: Suggest how to test it
Repeat the study in several cultures, with researchers from each culture adapting and translating the materials (international collaboration).
Step 5: Judge the claim
The claim overgeneralises. The evidence supports the conclusion only for people like the participants.
Try it yourself first, then check the working one step at a time.
Answer
The claim overgeneralises: the sample is WEIRD (US psychology undergraduates), and valuing personal choice may be a feature of an individualist culture. The finding may not apply in cultures that tend to value group decisions. Testing it across cultures through international collaboration would show whether it generalises.
Watch out for
Saying the study is "wrong". It may be valid for its sample; the problem is generalising beyond it.
Worked example
Analyse a cross-cultural comparison
A researcher measures the strength of the Müller-Lyer illusion in 30 adults who grew up in a city and 30 adults who grew up in a rural area with few straight-edged buildings. The city group's illusion is significantly stronger ().
(a) Identify the type of investigation, the independent variable and the dependent variable. (b) Explain why the researcher cannot conclude that growing up among straight-edged buildings causes a stronger illusion.
Step 1: Identify the investigation type
Participants cannot be randomly allocated to where they grew up, so this is a quasi-experiment comparing pre-existing groups.
Step 2: Identify the variables
IV: where the participants grew up (city or rural). DV: the strength of the illusion, for example how much longer one line must be before the two look equal.
Step 3: Identify confounding variables
The groups may also differ in schooling, familiarity with drawings and tests, age, eyesight or time spent on screens. Any of these could explain the difference.
Step 4: Conclude
Without random allocation, the difference may be due to these confounds, so the study shows a difference between the groups, not its cause.
Try it yourself first, then check the working one step at a time.
Answer
(a) Quasi-experiment; IV: where participants grew up (city or rural); DV: strength of the illusion. (b) Participants could not be randomly allocated to their environment, so the groups may differ in other ways (schooling, experience with drawings, age) that could explain the difference.
Watch out for
Calling it an independent groups experiment. That needs random allocation to conditions.
Common mistake
WEIRD does not mean wrong
The problem with WEIRD samples is not that the findings are false, but that we cannot assume they generalise. Some findings hold widely; others do not. The way to find out is to test them in other contexts, ideally with researchers from those contexts.
Summary
WEIRD: Western, educated, industrialised, rich and democratic. In 2003–2007, 96% of participants in top psychology journals came from countries holding 12% of the world's population.
WEIRD participants are often unusual, as the Müller-Lyer illusion shows: strongest in American undergraduates, absent in San foragers (Segall et al., 1966).
Findings from WEIRD samples may not be fully applicable in other contexts: their external (population) validity is limited.
Cultural comparisons are quasi-experimental: culture cannot be randomly allocated, so confounds are likely.
International collaboration gives local knowledge, valid translated measures, respectful recruitment and large, varied samples.
Finished the lesson?
Try the 2 quick checks above and the lesson completes itself, or mark it complete now.