Consider the class of two parameter marginal logistic (Rasch) models, for a test of m True-False items, where the latent ability is assumed to be bounded. Using results of Karlin and Studen, we show that this class of nonparametric marginal logistic (NML) models is equivalent to the class of marginal logistic models where the latent ability assumes at most (m + 2)/2 values. This equivalence has two implications. First, estimation for the NML model is accomplished by estimating the parameters of a discrete marginal logistic model. Second, consistency for the maximum likelihood estimates of the NML model can be shown (when m is odd) using the results of Kiefer and Wolfowitz. An example is presented which demonstrates the estimation strategy and contrasts the NML model with a normal marginal logistic model.

A plausible s-factor solution for many types of psychological and educational tests is one that exhibits a general factor and s − 1 group or method related factors. The bi-factor solution results from the constraint that each item has a nonzero loading on the primary dimension and at most one of the s − 1 group factors. This paper derives a bi-factor item-response model for binary response data. In marginal maximum likelihood estimation of item parameters, the bi-factor restriction leads to a major simplification of likelihood equations and (a) permits analysis of models with large numbers of group factors; (b) permits conditional dependence within identified subsets of items; and (c) provides more parsimonious factor solutions than an unrestricted full-information item factor analysis in some cases.

Maximum likelihood estimation of item parameters in the marginal distribution, integrating over the distribution of ability, becomes practical when computing procedures based on an EM algorithm are used. By characterizing the ability distribution empirically, arbitrary assumptions about its form are avoided. The Em procedure is shown to apply to general item-response models lacking simple sufficient statistics for ability. This includes models with more than one latent dimension.

A two-step weighted least squares estimator for multiple factor analysis of dichotomized variables is discussed. The estimator is based on the first and second order joint probabilities. Asymptotic standard errors and a model test are obtained by applying the Jackknife procedure.

In tailored testing, it is important to determine the optimal difficulty of the next item to present to the examinee. This paper shows that the difference that maximizes information for the three-parameter normal ogive response model is approximately 1.7 times the optimal difference θ − b for the three-parameter logistic model. Under the normal model, calculation of the optimal difficulty for minimizing the Bayes risk is equivalent to maximizing an associated information function.

Additional information contained in incorrect responses calls for a multicategorical rather than a binary analysis of multiple choice data. A nonparametric divided-by-total model for joint maximum likelihood estimation of probability-of-choice functions (for particular responses) and of latent ability is proposed. The model approximates probability functions by rational splines. Some illustrative examples of real test data analysis and the results of a Monte Carlo study are presented.

Relations are examined between latent trait and latent class models for item response data. Conditions are given for the two-latent class and two-parameter normal ogive models to agree, and relations between their item parameters are presented. Generalizations are then made to continuous models with more than one latent trait and discrete models with more than two latent classes, and methods are presented for relating latent class models to factor models for dichotomized variables. Results are illustrated using data from the Law School Admission Test, previously analyzed by several authors.

In social science research and clinical practice, and in educational settings, psychological tests/scales are being increasingly used because of their reliability and the ease and speed of gathering, comparing and differentiating data. A new scale is usually created when instruments or questionnaires to measure the construct of interest are not be readily available or if existing questionnaires do not fully satisfy requirements. Scales are also translated and revalidated if they are not in the language required. This article takes the reader through steps in developing, validating and translating tests in the field of social sciences.

This study is about the validity and item analysis for the Positive and Negative Affect Schedule (PANAS), respectively through the Exploratory Factor Analysis (principal components method) and the Partial Credit Model (PCM). The scale has been largely used in areas ranging from clinical to social psychology since its release in 1988 by Watson, Clark, and Tellegen. In order to assess validity and item properties (Item Response Theory paradigm), this is study administered PANAS to 354 respondents, 115 male and 239 female subjects, with an average age of 29.5 (SD = 10,18). The results show PANAS’s excellent psychometric properties, with consistent dimensions and reliable item functioning, considering the Rasch measurement paradigm expressed in the PCM as an Item Response Theory model for polytomous data. The study considers important cultural issues and the results support more cautious translations for scales as well as further studies concerned with cross-cultural differences on the perception of affect states.

This research deals with assessing humor appreciation and highlights some of the strategies that can be used in two necessary stages in the construction of a test: the content validity study and the item analysis. First, we analyzed the content validity of a battery of 200 items developed to assess humor appreciation. Second, we analyzed the metric properties of the selected items by means of two studies. The first study was a pre-pilot analysis of the items in a sample of 212 participants, and the second study was a new item analysis in a sample of 344 Spanish people aged between 18 and 71 years. To determine content validity, we calculated interjudge agreement on item-facet theoretical match. Each item was assessed by seven judges, and the selection criterion used was a minimum agreement of 70%. This procedure led to eliminating 27 items.

Subsequent item analyses led to a preliminary proposal for a 40-item scale (Escala de Apreciación del Humor, EAHU [Humor Appreciation Scale]) with appropriate descriptive statistics as well as discrimination and homogeneity values. The internal empirical structure of the scale matched the operative definition of humor appreciation, and the Cronbach’s alpha of the EAHU scores ranged from .72 to .89.

A twin item factor analytic model was developed to test for the presence of noninvariant age, sex, and age by sex interaction effects on the individual DSM-III-R criteria for major depression (MD). Based on 1-year reports, six of the nine MD criteria and duration requirement were found to have covariate factor loading and/or threshold effects that significantly deviated from their corresponding factor level expectations. A significant age effect was found for the binary duration variable factor loading. The ‘loss of interest’, ‘weight problems’ and ‘psychomotor problems’ criteria all displayed forms of threshold only effects. ‘Depressed mood’, ‘fatigue’, and ‘feeling worthless’ had more complex patterns that included both factor loading and threshold effects. A significant factor age by sex interaction effect indicating an increasing female mean difference with age was found to be largely associated with the presence of differential threshold covariate effects. Disagreement between estimated factor scores and DSM-derived affected vs. unaffected classification was ∼ 1.3%. Status on the duration requirement was found to be the one feature common to all discrepancies. The MD criteria set provided maximum information for calibrating MD factor scores in the scale region where discrepancies occurred. The dimensional modeling results are discussed in the broader context of epidemiological research and clinical assessment of major depression.

Background: The present study aims to develop and validate a Chinese version of the Dementia Rating Scale (DRS) for use with Chinese populations in psychogeriatric settings.

Methods: The DRS was translated into Chinese and its content validity was evaluated by an 11-member expert panel. To assess reliability and concurrent validity, 52 subjects with dementia were recruited from medical and psycho-geriatric settings using purposive sampling.

Results: With percentage of agreement as an indicator, 28 out of 36 items (78%) had satisfactory content validity. Items with a percentage of agreement below 70% were reviewed and modified, based on the comments of the experts. The CDRS had excellent test-retest and inter-rater reliability, with intraclass correlation coefficient (ICC) at 0.94 and 0.93 respectively. Intraclass correlation coefficients ranged between 0.75 and 0.89 for the subscales. The internal consistency of the CDRS subscale, as measured by Cronbach's α, ranged from 0.57 to 0.82. The CDRS had high correlations with the Chinese Mini-mental State Examination (r=0.80 for total score, r=0.58 to 0.84 for subscales.

Conclusions: The CDRS is a valid instrument for the assessment of dementia in Chinese–speaking subjects.