Psychological testing | |
---|---|
Medical diagnostics | |
ICD-10-PCS | [1] |
ICD-9-CM | 94.02 |
MeSH | D011581 |
Psychological testing refers to the administration of psychological tests. A psychological test is "an objective and standardized measure of a sample of behavior" (p. 4). The term sample of behavior refers to an individual's performance on tasks that have usually been prescribed beforehand. The samples of behavior that make up a paper-and-pencil test, the most common type of test, are a series of items. Performance on these items produce a test score. A score on a well-constructed test is believed to reflect a psychological construct such as achievement in a school subject, cognitive ability, aptitude, emotional functioning, personality, etc. Differences in test scores are thought to reflect individual differences in the construct the test is supposed to measure. The technical term for the science behind psychological testing is psychometrics.
A psychological test is an instrument designed to measure unobserved constructs, also known as latent variables. Psychological tests are typically, but not necessarily, a series of tasks or problems that the respondent has to solve. Psychological tests can strongly resemble questionnaires, which are also designed to measure unobserved constructs, but differ in that psychological tests ask for a respondent's maximum performance whereas a questionnaire asks for the respondent's typical performance. A useful psychological test must be both valid (i.e., there is evidence to support the specified interpretation of the test results) and reliable (i.e., internally consistent or give consistent results over time, across raters, etc.).
It is important that people who are equal on the measured construct also have an equal probability of answering the test items correctly. For example, an item on a mathematics test could be "In a soccer match two players get a red card; how many players are left in the end?"; however, this item also requires knowledge of soccer to be answered correctly, not just mathematical ability. Group membership can also influence the chance of correctly answering items (differential item functioning). Often tests are constructed for a specific population, and this should be taken into account when administering tests. If a test is invariant to some group difference (e.g. gender) in one population (e.g. England) it does not automatically mean that it is also invariant in another population (e.g. Japan).