9.3.6 - Culture and gender in research

Culture and the meaning of a measure

Culture can influence both the behaviour being studied and what a research task means to its participants.

Culture includes shared meanings, practices and expectations learned within communities. It is not interchangeable with nationality: people within one country may use different languages, have different experiences and belong to several communities. A cultural explanation therefore needs evidence about relevant practices rather than a label attached to a passport.

Ethnocentrism is judging other groups through the standards of one's own group. For example, a researcher might treat willingness to challenge a teacher publicly as the universal measure of independent thinking. Where respectful disagreement is usually expressed privately, the same behaviour could have a different meaning. The result may reflect the measure's cultural fit rather than a difference in the underlying ability.

An etic approach seeks constructs or comparisons that work across cultures. An emic approach first investigates meanings within a particular culture. These approaches can complement each other. Local interviews might reveal how respect for authority is expressed; researchers can then develop measures that make a fair cross-cultural comparison. Neither approach guarantees accuracy: a local account may not generalise, while a standardised comparison may standardise a misunderstanding.

Cognitive tasks illustrate the problem. A digit-span task requires a participant to hear and repeat number names, so language and familiarity with the task may affect performance alongside memory. Sebastián and Hernández-Gil's Spanish study tested children aged 5–17 and reported a developmental pattern differing from previously reported Anglo-Saxon data. That challenges a universal age rule based on one population. It does not by itself establish that culture caused the difference: language, schooling and sampling need to be investigated separately.

Samples and the reach of a claim

A finding can provide a useful explanation without establishing that the explanation works equally well for every population.

A representative sample reflects the relevant characteristics of the target population. Representativeness depends on the intended claim. A carefully selected sample may answer a bounded causal question well while remaining a weak basis for a claim about everyone. Increasing the sample size does not repair the systematic absence of relevant groups.

Sherif's Robbers Cave research deliberately selected boys of similar age and social background, including middle-class Protestant families. This helped reduce pre-existing differences while investigating how competition affected group relations. The same choice restricted the evidence about other genders, ages and social settings. Homogeneity can support control and limit generalisability at the same time; calling the study simply 'bad because it used boys' misses that trade-off.

A universal explanation proposes that a mechanism operates across populations. A culturally situated explanation specifies the conditions under which it operates. Competition may generate hostility when groups value the prize and perceive an opposing group as blocking it. Different norms, histories and unequal power may alter how hostility is expressed. Finding variation need not disprove the mechanism, but it may show that the original account needs boundary conditions.

Clinical psychology presents a further challenge. A classification system can help professionals communicate consistently, yet a judgement about an unusual belief or behaviour must consider cultural context, distress and functioning. Unfamiliarity to a clinician is not sufficient evidence of disorder. Conversely, respecting cultural context does not mean ignoring a person's distress or denying them support. Researchers need evidence that diagnostic questions measure comparable experiences in the groups to which the classification is applied.

When applying results, distinguish three claims: 'this pattern occurred in this sample'; 'this mechanism may explain the pattern'; and 'the mechanism works in the proposed new setting'. Each step requires additional justification. Replication across relevant settings and locally validated measures strengthen the final step. A country's average is also not a description of every individual within it.