Validity

​​​​​VALIDITY

Internal Validity

Means that a researcher may not be recording how people really think and behave.

e.g. investigator effects, demand characteristics, confounding variables, social desirability bias, poorly operationalised behavioural categories.

External Validity

Population Validity - generalising from one sample to all people.

Ecological Validity - generalising a research effect from one setting to another. This is affected by how the dependent variable is measured (which may lack mundane realism) and whether participants are aware they are being studied.

Temporal Validity - generalising a research effect from one historical period to another.

Assessing Validity

Face Validity - whether test/questionnaire items look like they are measuring what they claim to be measuring.

Concurrent Validity - checking if a test correlates well with existing validated measures.

Improving Validity

- Revise questions so they look more relevant.

- Revise and check concurrent validity again.

- Improve design (e.g. use double blind or a more realistic task)