Search

Mokken’s scalability coefficients take values on the interval (−∞, 1]. The sampling distribution of scalability coefficients is skewed near the boundary, so Wald-based confidence intervals and significance tests may be biased. We introduce a transformation of the scalability coefficients and their standard errors, which can be used to construct range-preserving confidence intervals and significance tests. We demonstrated that for scalability coefficients away from the boundary, the properties of this range-preserving method are similar to the properties of the Wald-based method, but the range-preserving method outperforms the Wald-based method when the coefficient is close to unity. The range-preserving method can be applied to all types of scalability coefficients. Its implementation in software is discussed.

This proceedings volume highlights the latest research and developments in psychometrics and statistics. It represents selected and peer-reviewed presentations given at the 85th Annual International Meeting of the Psychometric Society (IMPS), held virtually on July 13-17, 2020. The IMPS is one of the largest international meetings on quantitative measurement in education, psychology and the social sciences. It draws approximately 500 participants from around the world, featuring paper and poster presentations, symposiums, workshops, keynotes, and invited presentations. Leading experts and promising young researchers have written the included chapters. The chapters address a wide variety of topics including but not limited to item response theory, adaptive testing, Bayesian estimation, propensity scores, and cognitive diagnostic models. This volume is the 9th in a series of recent works to cover research presented at the IMPS.

This proceedings volume highlights the latest research and developments in psychometrics and statistics. It represents selected and peer reviewed presentations given at the 84th Annual International Meeting of the Psychometric Society (IMPS), organized by Pontificia Universidad Católica de Chile and held in Santiago, Chile during July 15th to 19th, 2019. The IMPS is one of the largest international meetings on quantitative measurement in education, psychology and the social sciences. It draws approximately 500 participants from around the world, featuring paper and poster presentations, symposiums, workshops, keynotes, and invited presentations. Leading experts and promising young researchers have written the included chapters. The chapters address a large variety of topics including but not limited to item response theory, multistage adaptive testing, and cognitive diagnostic models. This volume is the 8th in a series of recent volumes to cover research presented at the IMPS.

In survey research, it is often problematic to ask people sensitive questions because they may refuse to answer or they may provide a socially desirable answer that does not reveal their true status on the sensitive question. To solve this problem Warner (1965) proposed randomized response (RR). Here, a chance mechanism hides why respondents say yes or no to the question being asked. Thus far RR has been mainly used in research to estimate the prevalence of sensitive characteristics. It is not uncommon that researchers wrongly believe that the RR procedure has the drawback that it is not possible to relate the sensitive characteristics to explanatory variables. Here, we provide a review of the literature of regression procedures for dichotomous RR data. Univariate RR data can be analyzed with a version of logistic regression that is adapted so that it can handle data collected by RR. Subsequently the manuscript presents extensions towards repeated cross-sectional data that allowed for a change in the design with which the RR data are collected. We also review regression procedures for multivariate dichotomous RR data, such as the model by Glonek and McCullagh (1995), a model for the sum of a set of dichotomous RR data, and a model from item response theory that assumes a latent variable that explains the answers on the RR variables. We end with a discussion of a recent development in the analysis of multivariate RR data, namely models that take into account that there may be respondents that do not follow the instructions of the RR design by answering no whatever the sensitive question asked. These are coined self-protective responses.

In survey research, it is often problematic to ask people sensitive questions because they may refuse to answer or they may provide a socially desirable answer that does not reveal their true status on the sensitive question. To solve this problem Warner (1965) proposed randomized response (RR). Here, a chance mechanism hides why respondents say yes or no to the question being asked. Thus far RR has been mainly used in research to estimate the prevalence of sensitive characteristics. It is not uncommon that researchers wrongly believe that the RR procedure has the drawback that it is not possible to relate the sensitive characteristics to explanatory variables. Here, we provide a review of the literature of regression procedures for dichotomous RR data. Univariate RR data can be analyzed with a version of logistic regression that is adapted so that it can handle data collected by RR. Subsequently the manuscript presents extensions towards repeated cross-sectional data that allowed for a change in the design with which the RR data are collected. We also review regression procedures for multivariate dichotomous RR data, such as the model by Glonek and McCullagh (1995), a model for the sum of a set of dichotomous RR data, and a model from item response theory that assumes a latent variable that explains the answers on the RR variables. We end with a discussion of a recent development in the analysis of multivariate RR data, namely models that take into account that there may be respondents that do not follow the instructions of the RR design by answering no whatever the sensitive question asked. These are coined self-protective responses.

Although few would dispute the usefulness of looking at behavioral change from a stage-sequential perspective, until recently the lack of appropriate modeling techniques has hampered rigorous empirical tests of stage theories. In particular, for behavioral measurements that are ordinal, there is a need for methods that represent the underlying change processes in the form of qualitative and discontinuous shifts. This article introduces a stage-sequential ordinal model by postulating that at any point in time there are a finite number of latent stages. Panel members may shift among these stages over time. The authors show that the stage-sequential model provides a general approach for both the analysis of ordinal time-dependent data and tests of various competing theories and hypotheses about psychological change processes. An analysis of a 5-year study concerning attitudes toward alcohol consumption by teenagers is presented to illustrate the modeling approach. (PsycINFO Database Record (c) 2016 APA, all rights reserved)

This article studies the power of the Lagrange Multiplier Test and the Generalized Lagrange Multiplier Test to detect measurement non-invariance in Item Response Theory (IRT) models for binary data. We study the performance of these two tests under correct model specification and incorrect distribution of the latent variable. The asymptotic distribution of each test under the alternative hypothesis depends on a noncentrality parameter that is used to compute the power. We present two different procedures to compute the noncentrality parameter and consequently the power of the tests. The performance of the two methods is evaluated through a simulation study. They turn out to be very similar to the classic empirical power but less time consuming. Moreover, the results highlight that the Lagrange Multiplier Test is more powerful than the Generalized Lagrange Multiplier Test to detect measurement non-invariance under all simulation conditions.

Latent variable models represent a useful tool in different fields of research in which the constructs of interest are not directly observable. In presence of many latent variables/random effects, problems related to the integration of the likelihood function can arise since analytical solutions do not exist. In literature, different remedies have been proposed to overcome these problems. Among these, the composite likelihoods method and, more recently, the dimension-wise quadrature have been shown to produce estimators with desirable properties. We compare the performance of the two methods in the case of longitudinal ordinal data through a simulation study and an empirical application. Both the methods perform similarly, but the dimension-wise quadrature results less computational demanding. Indeed, for the specific model under investigation, it involves integrals of smaller dimensions than those involved in the computation of the pairwise likelihood, with a better performance than the latter in terms of accuracy of the estimates.

The thesis describes the multivariate analysis of randomized response data. Randomized response was introduced in 1965 as an interview technique to eliminate evasive response behavior (Warner, 1965). The questions are answered based on the outcome of a randomizer (a pair of dice, a deck of cards), so that the privacy of the respondents is protected. Research has shown that randomized response results in more valid answers than direct questions (van der Heijden et al., 2000 and Lensvelt-Mulders et al. 2005). During the past decade randomized response is used regularly by the Dutch administration to assess regulatory noncompliance. A well-known randomized response design is forced response (Boruch, 1971). In this design the respondent tosses two dice and answers the sensitive question based on the outcome of the two dice. If the outcome is 2, 3 or 4 the respondent has to answer is "yes", and if the outcome is 11 or 12, the respondents has to answer "no". If the outcome is 5, 6, 7, 8, 9 or 10, the respondents has answer truthfully. Since only respondent knows the outcome of the dice, confidentiality is ensured. Until recently randomized response was used to obtain a prevalence estimate of the sensitive behavior under study. Lately multivariate techniques have been developed to analyze the dependence of the sensitive behavior on other variables, like age, gender, education, etc. It has also become possible to analyze the associations between different sensitive behavior. The latest development is the estimation of self-protective response behavior, which occurs when respondents consistently give the non-sensitive response, irrespective the outcome of the randomizer. In this thesis four models for the multivariate analysis of randomized response data are presented. The models include the log-linear model, the proportional odds model, and two version of the zero-inflated Poisson model. Te models are applied to data from nationwide social welfare surveys that were conducted by the Dutch administration in the years 200, 200, 2004 and 2006.

The thesis describes the multivariate analysis of randomized response data. Randomized response was introduced in 1965 as an interview technique to eliminate evasive response behavior (Warner, 1965). The questions are answered based on the outcome of a randomizer (a pair of dice, a deck of cards), so that the privacy of the respondents is protected. Research has shown that randomized response results in more valid answers than direct questions (van der Heijden et al., 2000 and Lensvelt-Mulders et al. 2005). During the past decade randomized response is used regularly by the Dutch administration to assess regulatory noncompliance. A well-known randomized response design is forced response (Boruch, 1971). In this design the respondent tosses two dice and answers the sensitive question based on the outcome of the two dice. If the outcome is 2, 3 or 4 the respondent has to answer is "yes", and if the outcome is 11 or 12, the respondents has to answer "no". If the outcome is 5, 6, 7, 8, 9 or 10, the respondents has answer truthfully. Since only respondent knows the outcome of the dice, confidentiality is ensured. Until recently randomized response was used to obtain a prevalence estimate of the sensitive behavior under study. Lately multivariate techniques have been developed to analyze the dependence of the sensitive behavior on other variables, like age, gender, education, etc. It has also become possible to analyze the associations between different sensitive behavior. The latest development is the estimation of self-protective response behavior, which occurs when respondents consistently give the non-sensitive response, irrespective the outcome of the randomizer. In this thesis four models for the multivariate analysis of randomized response data are presented. The models include the log-linear model, the proportional odds model, and two version of the zero-inflated Poisson model. Te models are applied to data from nationwide social welfare surveys that were conducted by the Dutch administration in the years 200, 200, 2004 and 2006.