A LITTLE LIGHTBULB MOMENT
Sociology, research design and CARS reasoning
Validity asks whether the measure really measures the target
Sign in to bookmark this concept and add private notes →
A fake thermometer in a volcano
The idea behind the cue
Validity is the extent to which a measure captures the concept it is intended to measure. A test can be reliable without being valid. Content, construct, and criterion validity are common ways to assess whether a measure truly fits the target concept.
Practice this flashcard →Make it memorable
Short & simple
A fake thermometer in a volcano
Acronym
VCCC — Velvet Cupcakes Cuddle Cucumbers.
Story
A team claims its “happiness scale” is valid, but it only asks about soda purchases. The glittery chart looks impressive until everyone realizes it measures vending-machine behavior, not happiness.
Connect the pieces
- V (Velvet) → Validity: The measure really captures the target concept accurately.
- C (Cupcakes) → Content: The measure covers the full domain of the concept.
- C (Cuddle) → Construct: The measure fits the underlying theoretical trait it claims to measure.
- C (Cucumbers) → Criterion: The measure lines up with a relevant outcome or gold standard.
- Velvet · Validity: The measure really captures the target concept accurately.
- Cupcakes · Content: The measure covers the full domain of the concept.
- Cuddle · Construct: The measure fits the underlying theoretical trait it claims to measure.
- Cucumbers · Criterion: The measure lines up with a relevant outcome or gold standard.
Know the limits
Validity is about accuracy of measurement, not ethics, sample size, or whether the study is causal.
Further reading · OpenStax science textbooks ↗Check your understanding
A survey consistently gives the same result each time, but it clearly measures test-taking anxiety rather than social confidence. The survey is best described as:
A. Both valid and reliable
B. Neither valid nor reliable
C. Reliable but not valid
D. Valid but not reliable
Reveal explanation
Reliable but not valid
Consistent results indicate reliability, but measuring the wrong construct means it lacks validity.