Порівняння методів

	Conditional Standard Error of Measurement	Теорія реакцій на завдання (IRT)
Галузь	Education	Психометрія
Родина	Latent structure	Latent structure
Рік появи	1980	1952–1968
Автор методу	Test theory (Lord; Feldt; codified in the Standards)	Frederic M. Lord (and Allan Birnbaum for the 2PL/3PL models)
Тип	Measurement-error estimate that varies with the examinee's score or ability level	Probabilistic measurement model
Основоположне джерело	Lord, F. M. (1980). Applications of Item Response Theory to Practical Testing Problems. Lawrence Erlbaum Associates. ISBN: 9780898590067	Lord, F. M. & Novick, M. R. (1968). Statistical Theories of Mental Test Scores. Addison-Wesley.
Інші назви	CSEM, Conditional SEM, Score-Level Measurement Error, Conditional Standard Error	IRT, latent trait theory, item characteristic curve theory, modern test theory
Пов'язані	3	5
Підсумок	The conditional standard error of measurement (CSEM) describes how much measurement error a test score carries at each point along the score scale, rather than as a single average. A test typically measures more precisely in some score ranges than others — often best near the middle and worst at the extremes — and the CSEM captures that variation. Recognized in test theory by Lord and required by the Standards for Educational and Psychological Testing, it is essential for honest score reporting, especially near cut scores where classification decisions are made.	Item response theory models the probability that a respondent answers an item correctly (or endorses it) as a function of the respondent's latent trait level and the item's own statistical properties — difficulty, discrimination, and guessing. Unlike classical test theory, IRT places persons and items on the same scale, yielding measurement that is sample-independent for items and test-independent for persons.