Psychometrics

Psychometrics is the science of measuring psychological traits and abilities. It develops and evaluates tests and rating scales using statistical models and evidence about what scores mean.

Latent variable: An unobserved quantity inferred from patterns in measured variables. Psychometric models often treat traits such as ability as latent variables inferred from test responses.
Rasch model: An item response model in which the probability of a correct response depends on person ability and item difficulty. It shows how a specific model can place people and test items on a shared measurement scale.
Intelligence quotient: A standardized score summarizing performance on an intelligence test relative to a reference population. IQ illustrates how psychometric scoring converts test performance into a norm-referenced ability estimate.
Ipsative measurement: A scoring approach that expresses a person's standing on one trait relative to that person's standing on others. Unlike norm-referenced scores, ipsative scores emphasize within-person profiles and constrain comparisons between people.
Test bias: Systematic distortion in test scores or interpretations that disadvantages a group for reasons unrelated to the intended construct. Psychometric evaluation asks whether test content or scoring systematically disadvantages particular groups.
Reliability (psychometrics): The consistency of scores across items, occasions, or raters under specified conditions. Reliability estimates how much observed score variation reflects measurement error rather than stable differences.
Differential item functioning: A pattern in which people from different groups with the same measured trait have different probabilities of responding to an item. It helps detect items that behave differently across groups beyond measured ability differences.
Educational assessment: The systematic collection and interpretation of evidence about learning and educational outcomes. Psychometric principles guide the design and interpretation of tests used to evaluate learning.
Norm-referenced test: A test that interprets an individual's score by comparison with a reference group. It contrasts with criterion-based interpretation, which evaluates performance against a defined standard.
Stereotype threat: The risk that awareness of a negative stereotype about one's group impairs performance in a relevant situation. It shows how testing conditions can influence measured performance independently of the ability a test aims to assess.

Linked from 47 pages