What Is the Difference Between Validity and Reliability?
In an experiment, reliability signals how consistently the experiment produces the same results while validity signals whether the experiment measures what it is intended to measure. An experiment’s reliability does not offer any information about its validity.
When designing an experiment, both reliability and validity are important. With an unreliable experiment, results are untrustworthy. If replications produce different results, it is impossible to interpret the results; however, even a highly reliable experiment sometimes lacks validity. A poorly-designed experiment may produce results that are not directly relevant to the hypothesis, despite reliable replications. For example, if there are confounds in the design, the results may arise for reasons other than those hypothesized when designing the study.
There are several ways to measure reliability. Test-retest reliability involves re-running the study multiple times and checking the correlation between results. If the results are consistent, the test is reliable. Split-half reliability is similar; half of the data are selected at random and compared to the other half. If the measure is reliable, the two halves will be quite similar.
There is no quantitative measure of validity. Instead, validity is typically gauged by discussing the measure with experts, considering in close detail how the measure relates to the hypothesis, and comparing the results with other tests designed to measure the same outcome.