validity

what is validity

• Validity = accuracy — whether a test, study, or measurement actually measures what it claims to measure.

• A study can be reliable (consistent) but not valid (not accurate)

face validity

• Simplest form of validity.

• Does the test appear (on the surface) to measure what it’s supposed to?

• Judged subjectively by experts or researchers.

✅ Example:
– A depression questionnaire that asks about sadness, sleep, appetite → has good face validity.

❌ Weakness:
– Only surface-level; doesn’t guarantee actual validity.

Concurrent Validity

• Compares a new test with an established, well-validated test that measures the same thing.

• High correlation between the two = high concurrent validity.

✅ Example:
– New IQ test compared with Stanford–Binet IQ test → strong correlation = valid.

Ecological Validity

• Refers to how well findings generalise to real-life settings.

• Linked to external validity.

• High ecological validity = realistic tasks and environments.

✅ Example:
– Memory study using everyday tasks (e.g., remembering a shopping list) = high ecological validity.
❌ Example:
– Artificial lab tasks (e.g., nonsense syllables) = low ecological validity.

Temporal Validity

• Refers to how well findings generalise over time.

• Are the results still relevant and true today?

✅ Example:
– Asch’s conformity study (1950s USA) → low temporal validity (different social norms now).

assessing validity

. Face Validity

• Ask experts or peers to judge if the test looks like it measures the right thing.

Concurrent Validity

• Compare results with those from an established valid test.

• Calculate correlation between scores (strong positive = valid).

xternal Validity (includes ecological & population & temporal)

• Ecological validity: realistic setting/tasks.

• Population validity: generalises to other people.

• Temporal validity: generalises to other time periods.

improving validity

. Experiments

• Use realistic materials and settings (improves ecological validity).

• Use control groups to rule out confounding variables.

• Ensure participant variables are controlled.

• Use double-blind procedures to reduce researcher bias.

Observations

• Use covert observations → participants behave naturally (reduces demand characteristics).

• Use clear operational definitions of behaviour.

Questionnaires

• Use anonymous responses to reduce social desirability bias.

• Include lie scales to detect dishonesty.

• Ensure questions are clear, neutral, and relevant.

Interviews

• Use non-leading, open questions.

• Use trained interviewers to avoid bias.

• Consider structured interviews to ensure consistency.

General Strategies

• Triangulation → use multiple methods/sources (e.g., questionnaires + observations).

• Standardisation → consistent procedures for all participants.

• Pilot studies → identify unclear tasks or instructions.

• Peer review → check for design flaws.