Regardless of the field you are in, when you are testing subjects you must already have designed a test to administer. The test you use for such a task can come in many forms. A clear division in the categories of testing and assessment is that of self-report form tests versus skill-based assessments. Self-report tests include surveys and questionnaires, in which individuals state their subjective answers regarding a specific topic or skill or give opinions. Self-reports tend to be subjective and easily influenced by social factors. Skill-based assessments, on the other hand, assess individuals in an objective way. Assessments such as those that test IQ or working memory fall into this category.
After a test is administered and the data collected, researchers must evaluate their findings to make conclusions about the assessments. Evaluations are an important part of testing and assessment, as there are many types of evaluations suitable for different situations. Most evaluations are quantitative in that they use statistical analyses to find differences among groups or investigate how changes on a certain variable affect subjects' answers or performance.
The tests and assessments themselves are a subject of testing and assessment. Validity refers to the specific legitimacy of a given test or assessment. If a test is valid, then researchers can know that the results of the test are likely useful in predicting behavior or indicating certain traits related to the individual who took the test. Validity comes in many forms, including external validity, predictive validity and surface validity.
Reliability is a trait belonging to an established test or assessment that indicates the test's consistency and precision. In testing and assessment, tests with high reliability are desirable, because of their applicability to large groups of people. A common way to describe the idea of reliability is that a test with high reliability will produce the same or exceptionally similar results when the same individual takes the test twice. Like validity, reliability comes in many flavors, including test-retest reliability, internal reliability and cross-rater reliability.