Methods for determining equivalence of measures.

Gulliksen, Harold

doi:10.1037/h0026721

Cited by 31 publications

(17 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, the true scores are not necessarily measured equally accurately for each person by different tau-equivalent tests. In fact this is the first liberalization from the stringent parallelism requirements proposed by Gulliksen (1968). In terms of the model for congeneric tests tau-equivalent tests satisfy the requirement:…”

Section: Parallel and Tau-equivalent Testsmentioning

confidence: 96%

“…In a sense the second approach proposed by Gulliksen (1968) to move away from the strict parallelism requirements is implicit in the model for congeneric tests; tests are equivalent if they measure the same general ability except for errors of measurement.…”

Section: Congeneric Testsmentioning

confidence: 99%

“…For the specific problem of rater agreement as a form of equivalence Gulliksen (1950) suggested an approach based on the Wilks-Votaw tests for compound symmetry. This procedure provides a statistical criterion upon which to base a statement whether the sets of ratings are random samples from the same population having equal means, equal variances and equal covariances, and also for each specific criterion, that the set of ratings have the same validity coefficients.…”

Section: Equivalence Of Ratingsmentioning

confidence: 99%

“…This procedure provides a statistical criterion upon which to base a statement whether the sets of ratings are random samples from the same population having equal means, equal variances and equal covariances, and also for each specific criterion, that the set of ratings have the same validity coefficients. If ratings satisfy the Wilks-Votaw criterion the sets of ratings are said to be interchangeable (Gulliksen, 1968). In this situation the most stringent requirements of equivalence are met.…”

Section: Equivalence Of Ratingsmentioning

confidence: 99%

See 3 more Smart Citations

Agreement Between Raters

Kamp

Mellenbergh

1976

Educational and Psychological Measurement

View full text Add to dashboard Cite

In this paper Jöreskog's model of congeneric tests is used to analyze agreement between raters. Raters are treated as measuring instruments. The model of congeneric tests, of which classical prallelism and tau-equivalence are shown to be special cases, is applied to teachers' ratings of students' responses on open-end questions. The findings suggest that the ratings are tau-equivalent and that the ratings given by the teachers are reliable.EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT 1976, 36, 311-317. It is well known that the reliability of a test is usually estimated by correlating two sets of scores from equivalent forms, split halves, two administrations of the test, or from analysis of variance computations. When dealing with ratings, a special class of test scores, it may occur that more than two comparable sets are available. In a typical rating study different raters rate a number of subjects on a number of attributes or items. This situation frequently arises in educational and psychological research. In order to estimate the reliability of the ratings several procedures have been proposed, estimation procedures based upon analysis of variance (e.g., Ebel, 1951; Stanley, 1961;Maxwell and Pilliner, 1968) and the ones based upon a multivariate analysis of variance (e.g., Abelson, 1960;Fleiss, 1966) or a factoranalytic approach (e.g., La Forge, 1965). Apart from the problem of reliability of ratings, the problem of rater agreement is a pertinent one in view of the frequent use of the average of ratings as e.g. a criterion for validating a selection test. If we take the theoretical position that each rater may be regarded as a test instrument, then the problem of rater

show abstract

Section: Parallel and Tau-equivalent Testsmentioning

confidence: 96%

Section: Congeneric Testsmentioning

confidence: 99%

Section: Equivalence Of Ratingsmentioning

confidence: 99%

Section: Equivalence Of Ratingsmentioning

confidence: 99%

See 2 more Smart Citations

Agreement Between Raters

Kamp

Mellenbergh

1976

Educational and Psychological Measurement

View full text Add to dashboard Cite

show abstract

“…Koutsopoulos (1961) presented a linear practice effect solution for a counterbalanced case of equating, in which two equally random groups (alpha and beta) take two forms, X and Y, of a test, alpha in the order X, Y and beta in the order Y, X. Gulliksen (1968) presented a variety of solutions for determining the equivalence of two measures, ranging from a criterion for strict interchangeability of scores to factor methods for comparing multifactor batteries of measures and multidimensional scaling. Boldt (1972) laid out an alternative approach to linking scores that involved a principle for choosing objective functions whose optimization would lead to a selection of conversion constants for equating.…”

Section: Early Equating Proceduresmentioning

confidence: 99%