A direct proof is given for the generalized Spearman-Brown formula for any real multiple of test length. (Author)

The aim of this paper is to explore methods for evaluating the effects of randomized interventions in a longitudinal design. The focus is on methods for modeling the possibly nonlinear relationship between treatment effect and baseline and evaluating the treatment effect taking this nonlinear relationship into account. A control/treatment growth…

A study was conducted to identify local item dependence (LID) in the context-dependent item sets used in an examination prepared for use in an introductory university physics class and to assess the effects of LID on estimates of the reliability and standard error of measurement. Test scores were obtained for 487 students in the physics class. The…

This Digest describes four types of errors in mail surveys and summarizes the ways they can be reduced. Any one of these sources of error can make survey results unacceptable. Sampling error is examined through inferential statistics applied to sample survey results. In general, increasing sample size will decrease sampling error when simple…

The purpose of this study was to examine the potential impact of selected methodological factors on the validity of conclusions from reliability generalization (RG) studies. The study focused on four factors; (1) missing data in the primary studies; (2) transformation of sample reliability estimates; (3) use of sample weights for estimating mean…

An item response theory (IRT) model is used as a measurement error model for the dependent variable of a multilevel model where tests or questionnaires consisting of separate items are used to perform a measurement error analysis. The advantage of using latent scores as dependent variables of a multilevel model is that it offers the possibility of…

Single studies, by themselves, rarely explain the effect of treatments or interventions definitively in the social sciences. Researchers created meta-analysis in the 1970s to address this need. Since then, meta-analytic techniques have been used to support certain treatment modalities and to influence policymakers. Although these techniques…

A test which allows for errors of measurement is derived for the hypothesis that all the members of a population who possess a certain skill are a sub-set of the members who possess another skill. (Author)

Author supports the credibility of the propositions that: (1) the true component of a score is proportional to the number of equivalent elements that contribute to it. And, (2) the error component of a score is proportional to the square root of the number of equivalent elements that contribute to it. (Author/MB)

The standard error of an equipercentile equating is derived for four situations. Some numerical results are checked by Monte Carlo methods. Numerical standard errors are computed for two sets of real data. Standard errors of linear and equipercentile equating are compared. (Author)

Multivariate generalizability deals with the interpretation of a profile or composite of measurements. Using data from the Beginning Teacher Evaluation Study to illustrate the differing interpretations of viewing test data from universe and multivariate perspectives, this chapter concludes with computer programs that the reader can implement to…

A simple, usefully accurate approximation of the standard error of measurement is proposed for use by classroom teachers. An empirical comparison with Lord's approximation indicated that, though not as easy to calculate as Lord's, this approximation is more practical because it is useful at any point in the score distribution. (BW)

For tests used to make pass/fail decisions, the relevant standard error of measurement (SEM) is the SEM at the passing score. If the test is highly stratified, this SEM should be estimated by a split-halves approach. A formula and its derivation are provided. (Author)

A mathematical link between test reliability and test validity is derived, taking into account the correlation between error scores on a test and error scores on a criterion measure. When this correlation is positive, the "paradoxical" nonmonotonic relation between test reliability and test validity occurs universally. (Author/BW)

Glass's estimator of effect size, the sample mean difference divided by the sample standard deviation, is studied in the context of an explicit statistical model. The exact distribution of Glass's estimator is obtained and the estimator is shown to have a small sample bias. Alternatives are proposed and discussed. (Author/JKS)