Bias

A tendency to see or report things in a way that is not fully objective — a tilt away from the truth.

Judging based on facts, not on personal feelings or expectations.

When the people studied don't represent the wider group you want to describe.

When a researcher's expectations shape how they run or interpret a study.

When people change their behaviour because they know they are being studied.

Noticing evidence that fits what you expect and ignoring evidence that doesn't.

When mostly positive results get published, hiding the studies where nothing happened.

It threatens objectivity, so the findings become less trustworthy.

Use a double-blind design so neither the participant nor the tester knows the condition.

Bias — alongside causality, measurement and responsibility.

Judging or measuring one culture by the standards of another.

Seeing your own culture as the normal or correct one, and judging others by it.

Studying a culture from the inside, on its own terms and in its own concepts.

Comparing cultures from the outside using shared, general measures.

Using a method built in one culture on another as if it were neutral for everyone.

It makes conclusions unfair and can cause real harm, like wrong diagnoses or unfair testing.

Use emic methods, local researchers, and translate then back-translate the measure.

Translating a measure into a language and back again to check the meaning survived.

Western, Educated, Industrialised, Rich, Democratic — over-used and wrongly treated as universal.

Bias (mainly) and perspective (one culture's viewpoint treated as the truth for all).

Research treating one gender unfairly — as better, worse, or the default.

Exaggerating the differences between genders, often locking in stereotypes.

Ignoring or minimising real gender differences, often by studying one gender and applying it to all.

Treating male behaviour as the normal standard for everyone.

Testing a heart drug only on men, then giving it to everyone at the same dose.

It damages validity and can cause real harm — unfair theories, wrong doses, stereotypes as science.

Use a balanced sample of all genders and report results for each group.

A researcher reflecting on how their own views might shape the study.

'Treating everyone the same' ignores real differences, which can quietly disadvantage the untested group.

Alpha exaggerates gender differences; beta ignores them.

Protecting objectivity, so findings reflect the truth and can be trusted.

It makes the sample mirror the wider group, so results generalise — fixing sampling bias.

Giving every participant the same instructions and conditions, so differences don't creep in.

Neither the participant nor the tester knows the condition — reduces participant and researcher bias.

A researcher reflecting on how their own views might shape the study.

Others repeat the study; a one-off, biased result usually fails to repeat.

Emic methods, local researchers, and translating then back-translating the measure.

A balanced sample of all genders, with results reported for each group.

Posting the plan before data collection — reduces publication bias, so failures can't vanish.

Match the fix to the bias, and explain why it restores objectivity.