Explainer · Plain-language

What Is Validity in Research? Types, Examples & How to Ensure It | CASRAI

Validity in research refers to the extent to which a study measures what it claims to measure and produces findings that accurately reflect the phenomena under investigation. It encompasses both the quality of measurement instruments and the soundness of inferences drawn from data.

CASRAI plain-language explainers — clear answers to recurring research-administration questions

The step most authors miss

Doing CRediT right? Don’t stop at the statement.

A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.

Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.

Internal validity

Internal validity concerns the confidence with which a causal inference can be drawn from a study: does the observed change in the dependent variable genuinely result from the independent variable, or could it be explained by confounding factors, biases, or methodological artefacts? Threats to internal validity include selection bias (groups differ at baseline), history (external events affect outcomes), maturation (participants change over time), testing effects (repeated measurement alters responses), regression to the mean, and attrition (differential dropout). Randomised controlled trials maximise internal validity by randomly assigning participants to conditions, controlling confounding variables, and blinding participants and assessors where possible.

External validity

External validity concerns the generalisability of findings — can the results observed in this study, with this sample, in this setting, be applied to other populations, settings, and time points? Threats include sample unrepresentativeness (the study sample differs systematically from the target population), artificial laboratory conditions that do not reflect real-world contexts, and the Hawthorne effect (participants behave differently because they know they are being studied). External validity is often in tension with internal validity: highly controlled experimental designs improve internal validity but may reduce generalisability. Large, representative surveys often have high external validity but make causal inference more difficult.

Construct, content, and criterion validity

Construct validity is the degree to which a measure captures the theoretical construct it is intended to assess — whether a questionnaire measuring "resilience" actually measures resilience rather than social desirability or optimism. Evidence includes convergent validity (correlating with measures of related constructs) and discriminant validity (not correlating with measures of unrelated constructs), as well as confirmatory factor analysis. Content validity assesses whether the items in a measure cover the full domain of the construct (commonly evaluated by expert panels). Criterion validity assesses whether the measure predicts or correlates with an external criterion: concurrent validity (at the same time) and predictive validity (at a future point).

Validity in qualitative research

The term "validity" originates in quantitative methodology, and some qualitative researchers prefer alternative language. Lincoln and Guba (1985) proposed four parallel trustworthiness criteria. Credibility (analogous to internal validity) refers to confidence in the truth of the findings, supported by member checking, prolonged engagement, and peer debriefing. Transferability (analogous to external validity) refers to the applicability of findings to other contexts, supported by thick description. Dependability (analogous to reliability) refers to the consistency of the research process, supported by an audit trail. Confirmability (analogous to objectivity) refers to whether findings reflect participants' meanings rather than researcher bias, supported by reflexivity and documentation.

Key facts

At a glance

Definition: Whether a study measures what it claims to measure
Internal: Confidence in causal inference; threats include confounding and bias
External: Generalisability to other populations, settings, and time points
Construct: Whether the measure captures the intended theoretical concept
Content: Whether items cover the full domain of the construct
Qualitative: Lincoln & Guba's credibility, transferability, dependability, confirmability

Common misconceptions

What people often get wrong

Often heard: A highly reliable measure is automatically a valid measure.

Actually: No — reliability and validity are independent. A measuring instrument can produce consistent results (reliable) but consistently measure the wrong thing (invalid). A broken thermometer that always reads 37 °C is reliable but not valid.

Often heard: Validity only applies to quantitative research.

Actually: No — validity concerns apply to all research. In qualitative research, parallel criteria (Lincoln and Guba's trustworthiness criteria) assess the equivalent qualities of credibility, transferability, dependability, and confirmability.

Often heard: A large sample size guarantees validity.

Actually: No — sample size affects statistical power, not validity. A large biased sample (such as a convenience sample) may still produce systematically invalid results, while a small carefully selected sample may be more valid.

Going deeper