When the results of a study or test are published, two important aspects to use when interpreting those results are reliability and validity. To accurately judge the usefulness of a study and to determine whether or not the results can be generalized, we must know the difference between reliability and validity.
A study or test's reliability is the measure of whether the same results can be obtained given differing circumstances. For example, having a reliable test means a student will score the same or close to the same score on the test under different circumstances. If five different people graded the test, the scores should be the same or similar. If the student took the test on a Monday morning or a Wednesday evening, it shouldn't affect the scores. In addition, someone scoring the test on Monday and then again on Friday should not arrive at different scores. Reliability controls the results of a test or study so external factors -- such as when or by whom it is conducted -- do not change the results.
A study or test's validity is the measure of whether the results actually relate to what the test or study developer intended to measure or test. If a test claims to evaluate a student's ability to use multiplication, for example, but consists only of subtraction questions, it is not valid. If a study was investigating the mating calls of a certain bird but only investigated male birds, for example, or recorded calls outside of mating season, it would also be invalid. Examining a study's validity allows you to decide whether the results should be accepted and used to generalize or not.