A scoring system in which a response will receive the same score, no matter who does the scoring. No judgment is required to apply the scoring rule. Compare with subjective scoring. Also see analytic scoring and holistic scoring.

In item response theory (IRT), the a parameter is a number that indicates the discrimination of a test item — how sharply the item differentiates between generally strong and generally weak test takers. If the a parameter for an item is large, the probability that the test taker will answer the item correctly increases sharply within a fairly narro [..]

The knowledge, skills or other characteristics of a test taker measured by the test.

A type of testing in which the questions presented to the test taker are selected on the basis of the test taker's previous responses. Good performance by the test taker leads to harder questions; poor performance leads to easier questions. The purpose of adaptive testing is to use testing time more efficiently, by not giving test takers any q [..]

A statistic that is used to estimate the reliability of scores on a test. What alpha actually measures is internal consistency — the extent to which the test takers performed similarly on all the items. Under some assumptions that are usually reasonable, alpha also indicates the extent to which those test takers would perform similarly on two diffe [..]

A procedure for scoring responses on a constructed-response test, in which the scorer awards points separately for specific features of the response. (Compare with holistic scoring.)

For equating the scores on two forms of a test that are taken by different groups of test takers, it is necessary to know how those groups differ in the ability measured by the test. An anchor test is a test given to both groups to obtain this information. The anchor test can be a set of test questions appearing in both forms (called "common i [..]

These terms all refer to devices or procedures for getting information about the knowledge, skills or other characteristics of the people being assessed, tested or examined. The three terms are often used interchangeably, but there are some differences between them. "Assessment" is the broadest of the three terms; "examination" [..]

In item response theory (IRT), the b parameter is a number that indicates the difficulty of a test question. In general, a higher b parameter indicates a more difficult test question.

A statistic used at ETS to describe the relationship between performance on a single test item and on the full test. It is an estimate of the correlation between the test score and an unobservable variable assumed to determine performance on the item and assumed to have a normal distribution (the familiar "bell curve"). Compare to correla [..]