"Validity" means that something is accurate. In research, validity has to do with the strength of the measurement or how much it does what it says it will. For instance, a psychological test that is meant to measure depression, but that really measures levels of happiness would have low validity. If your washer does not wash your clothes properly but merely spins them, then it is not a valid tool for washing clothes.
"Reliability" refers to the dependability of something. For instance, if your washing machine starts up every time you turn the dial to “wash,” it is reliable. Similarly, in research, the term is interchangeable with consistency. When trying to find out if a measurement is reliable, researchers look to see if the measure is consistent on every repetition. If the measurement is a real or true measurement, then it is accurate each time it is used.
Validity and reliability are different but related terms. For instance, when looking at measurement validity, or how accurately a test measures a specific phenomenon, the test must be reliable for it to be valid. That is, if the test sometimes measures a depression and sometimes measures happiness, it is neither reliable nor valid. At other times, such as measuring groups, a measurement may be valid, in that it accurately measures individuals, but is not reliable, in that scores between group members vary. Ideally, a measurement is both reliable and valid, or, in other words, it accurately and dependably measures what it is designed to measure.
There are different types of reliability and validity. Common types of validity include face validity, which means it measures what it was designed to measure; content validity, when the measurement matches the concept it was designed to measure; and predictive validity, when the measurement makes an accurate predication. Types of reliability vary depending on where the researcher is looking for it. For instance, test-retest reliability shows if the test is consistent when given at different times, while internal consistency reliability looks at whether specific items in a test are consistent in comparison with each other.