Validity and Reliability
If you mastered this topic, Click here to try our time based online test with graphical result analysis
Click here to Read more interesting case study based article on each TLS sub topic in my second website
Click here to Read more interesting case study based article on each TLS sub topic in my second website
| Validity | Reliability |
| Validity implies the extent to which the research instrument measures, what it is intended to measure. | Reliability refers to the degree to which assessment tool produces consistent results, when repeated measurements are made. |
| It refers to the ability of the instrument/test to measure what it is supposed to measure | It refers to the reproducible of the results when repeated measurements are done |
| It relates to the correct applicability of the instrument/test/procedure in a needed situation | It relates to the extent to which an experiment, test or any procedure gives the same result on repeated trials. |
| Can relate to question of ‘Does it measure what it is supposed to measure’? | Can relate to question, ‘How representative is the measurement’? |
| Answers, ‘Is it the right instrument/test for what I need to measure?’ | Answers, ‘Can the results obtained be replicated if the test is repeated?’ |
| Validity looks at accuracy | Reliability looks at repeatability/consistency |
| Validity mainly focuses on the outcome | Reliability mainly focuses on maintaining consistent result |
| Influencing factors for validity are: process, purpose, theory matters, logical implications, etc. | Influencing factors for reliability are: test length, test score variability, carcinogenicity, etc. |
| Validity has more analysis and is harder to achieve. | Reliability is comparatively easier and yields faster results. |
| There cannot be validity without reliability | There can be reliability without validity. |
| Even if validity of an instrument is poor (for certain test), it can have high reliability (for other tests) | When reliability/repeatability is poor, validity may also be poor. Thus, usefulness of a test/experiments are negligible. |
| If the results are not valid, the test is of no use at all |
If the results cannot be replicated, the test is of little use |
