When tests contain few items, observed score may not be an accurate reflection of true score, and the Mantel Haenszel (MH) statistic may perform poorly in detecting differential item functioning. Applications of the MH procedure in such situations require an alternate strategy; one such strategy is to include background variables in the matching…

This study, based on the Rasch model, used R. M. Smith's (1986) classification of measurement disturbances to assess the Rasch model approach to error control and statistical prediction. Partitioning the error component into a person component, an item-person interaction component, and a random unexplained error component has the net effect of…

The Kuder Richardson-20 Formula is shown to be a special case, where each examinee is given sufficient time to answer each item, of a more general formula where each examinee may not be allowed the necessary time. The formula is extended to allow two scores, knowledge and speed, to be extracted from each examinees test score. Using a sample of 82…

The effectiveness of alternative procedures to correct for rater leniency/stringency effects was studied when true scores were known. Ordinary least squares, weighted least squares, and imputation of the missing data consistently outperformed averaging the observed ratings; and the imputation technique was superior to the least squares methods.…

Equivalence in reliability and validity across 4-point and 6-point scales was assessed by fitting different measurement models through confirmatory factor analysis of a multitrait-multimethod covariance matrix. Responses to nine Likert-type items designed to measure perceived quantitative ability, self-perceived usefulness of quantitative…

Statistical prediction problems often involve both a direct estimate of a true score and covariates of this true score. Given the criterion of mean squared error, this study determines the best linear predictor of the true score given the direct estimate and the covariates. Results yield an extension of Kelley's formula for estimation of the true…

This report builds on and extends existent research on population invariance to new tests and issues. The authors lay the foundation for a deeper understanding of the use of population invariance measures in a wide variety of practical contexts. The invariance of linear, equipercentile and IRT equating methods are examined using data from five…

Methods of computerized adaptive testing using conventional scoring methods in order to develop a computerized placement test for the Test of English as a Foreign Language (TOEFL) were studied. As a consequence of simulation studies during the first phase of the study, the multilevel testing paradigm was adopted to produce three test levels…

The four main approaches to measuring treatment effects in schools; raw gain, residual gain, covariance, and true scores; were compared. A simulation study showed true score analysis produced a large number of Type-I errors. When corrected for this error, this method showed the least power of the four. This outcome was clearly the result of the…

The first four chapters of this report primarily provide an extensive, critical review of the literature with regard to selected aspects of the criterion-referenced and mastery testing fields. Major topics treated include: (a) definitions, distinctions, and background, (b) the relevance of classical test theory, (c) validity and procedures for…

Racial differences in the reporting accuracy of parental status characteristics by White and Black high school seniors were investigated using Joreskog's general framework for simultaneous covariance structure analyses of multiple populations. Reliability estimates for Whites were significantly higher than for Blacks due to differences in true…

The validity of the assumptions under which the ordinal true score test theory was derived was examined using (1) simulation based on classical test theory; (2) a long empirical test with data from 321 sixth graders; and (3) an extensive simulation with 480 datasets based on the 3-parameter model. (SLD)

This study compared 4,594 student responses from three different surveys of incoming students at the University of South Florida (USF) with data from Florida's State University System (SUS) admissions files to determine what proportion of error occurs in the survey responses. Specifically, the study investigated the amount of measurement error in…

The Rasch model for test analysis is described and compared with two-parameter and three-parameter latent-trait models. Conditional maximum likelihood equations for estimating item parameters are derived, and estimates of person parameters are described together with their confidence intervals. Goodness of fit tests are discussed, including a…

Action research often necessitates the use of intact groups for the comparison of educational treatments or programs. This paper considers several analytical methods that might be used for such situations when pretest scores indicate that these intact groups differ significantly initially. The methods considered include gain score analysis of…