Reliability and validity
Two key words for evaluating any research: reliability (consistency) and validity (accuracy).
Reliability
Studies with standardised procedures and quantitative data (such as lab experiments and closed questionnaires) are easier to replicate, so their reliability can be checked.
In observations, inter-rater reliability means two or more observers record the same behaviour in the same way.
Qualitative methods (unstructured interviews, case studies) are harder to repeat exactly, so they tend to be less reliable.
Validity
Ecological validity: whether results can be applied to real-life settings. Field and natural experiments usually have higher ecological validity than lab experiments.
Population validity: whether results can be generalised to other people. This depends on the sampling method: a representative (stratified or random) sample gives higher population validity than an opportunity or volunteer sample.
Validity is reduced by extraneous variables, demand characteristics and social desirability (people giving answers that make them look good).
Applying them
| Feature | Reliability | Validity |
|---|---|---|
| Sampling: stratified or random | Results more likely to be repeated with a new sample | Higher population validity |
| Design: repeated measures | Order effects need counterbalancing for consistent results | Controls participant variables, so more valid comparison |
| Quantitative methods (lab experiment) | High: easy to replicate | May lack ecological validity |
| Qualitative methods (unstructured interview) | Lower: hard to repeat | Can be high: rich, detailed, real views |
Two observers watch the same playground video and record very different numbers of aggressive acts. What does this suggest, and how could it be improved?
- The observers are not consistent with each other: low inter-rater reliability.
- Clear behaviour categories and observer training would help them record behaviour in the same way.
Answer: Low inter-rater reliability; improve with clearly defined behaviour categories and training.
Reliability means consistency: a reliable study gives the same results when repeated, helped by standardisation and quantitative data. Validity means accuracy: ecological validity concerns real life, population validity concerns other people. Lab experiments tend to be reliable but less ecologically valid; qualitative methods can be valid but less reliable.
- Reliability
- Consistency: a reliable measure or study gives the same results if repeated.
- Validity
- Accuracy: whether a study measures what it claims to and whether results apply beyond the study.
- Quantitative data
- Numerical data, such as scores, counts and times.
- Ecological validity
- How far results apply to real-life settings.
- Demand characteristics
- Clues that let participants guess the aim of a study, so they change their behaviour.
Try two quick questions
See if it has sunk in. Answers are marked straight away.
2 more questions
Try each one, then tap to check.
How can reliability be checked?
- Changing the IV
- Ignoring results
- Repeating the study to see if results are similar
- Using a different DV
Show the answer
Answer: Repeating the study to see if results are similar
Consistent results show reliability.
What is ecological validity?
- How consistent results are
- How well findings apply to real-life settings
- How many people took part
- How quick the study was
Show the answer
Answer: How well findings apply to real-life settings
High ecological validity means realistic.
Check you have got it
Answer 7 quick questions with instant marking. If you get one wrong, GCSE-ready shows you why and gives you another go. It is free, and you do not need an account.
Facts checked against the sources below, 17 October 2026. Last updated 2 October 2026. Spotted a mistake? Tell us and we will fix it.
What do exam command words like explain and evaluate mean? โ