Weighting for Item Non-Response in Attitude Scales by Using Latent Variable Models with Covariates

Moustaki, Irini; Knott, M.

doi:10.1111/1467-985x.00177

Cited by 59 publications

(50 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…to establish a measurement model for a so-called latent response propensity θ. The person estimators of θ can be used to compute weights for each observation (Moustaki & Knott, 2000). Alternatively, the measurement model for the response propensity θ has been added to the measurement model for the ability variable of interest, ξ.…”

mentioning

confidence: 99%

“…MIRT models with either a between-item-multidimensional structure or a within-item-multidimensional structure have been introduced to account for nonignorable missing data. Another possibility is to avoid modeling a second latent variable by regarding the missing value code as an additional response category in a nominal response model (Moustaki & Knott, 2000). The basic idea of these models involves the specification of a selection model (Heckman, 1979) that accounts for the stochastic dependency between the latent proficiency ξ and the occurrence of missing data.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Modeling Nonignorable Missing Data With Item Response Theory (Irt)

Rose

Davier

2010

ETS Research Report Series

198

View full text Add to dashboard Cite

Large-scale educational surveys are low-stakes assessments of educational outcomes conducted using nationally representative samples. In these surveys, students do not receive individual scores, and the outcome of the assessment is inconsequential for respondents. The low-stakes nature of these surveys, as well as variations in average performance across countries and other factors such as different testing traditions, are contributing factors to the amount of omitted responses in these assessments. While underlying reasons for omissions are not completely understood, common practice in international assessments is to employ a deterministic treatment These two model-based approaches were compared on the basis of simulated data and data from about 250,000 students from 30 Organisation for Economic Co-operation and Development (OECD) Member countries participating in an international large-scale assessment.

show abstract

mentioning

confidence: 99%

mentioning

confidence: 99%

Modeling Nonignorable Missing Data With Item Response Theory (Irt)

Rose

Davier

2010

ETS Research Report Series

198

View full text Add to dashboard Cite

show abstract

“…Rapid responders (DeMars and Wise 2010) who may not provide reliable response data can potentially be identified using this data. Nonresponse models (Glas and Pimentel 2008;Moustaki and Knott 2000;Rose et al 2010) can be used to gain a deeper understanding of situations in which certain types of respondents tend not to provide any data on at least some of the items. Elaborate response-time models that integrate latency and accuracy (Klein Entink et al 2009;Lee 2008) can be integrated with current large-scale assessment methodologies.…”

Section: Forward-looking Design For Using Context Information In Compmentioning

confidence: 99%

Large-Scale Assessments of Adult Literacy

Kirsch

Lennon

Yamamoto

et al. 2017

Methodology of Educational Measurement and Assessment

View full text Add to dashboard Cite

Educational Testing Service's (ETS's) work in large-scale adult literacy assessments has been an ongoing and evolving effort, beginning in 1984 with the Young Adult Literacy Survey in the United States. This work has been designed to meet policy needs, both in the United States and internationally, based on the growing awareness of literacy as human capital. The impact of these assessments has grown as policy makers and other stakeholders have increasingly come to understand the critical role that foundational skills play in allowing individuals to maintain and enhance their ability to meet changing work conditions and societal demands. For example, findings from these surveys have provided a wealth of information about how the distribution of skills is related to social and economic outcomes. Of equal importance, the surveys and associated research activities have contributed to large-scale assessment methodology, the development of innovative item types and delivery systems, and methods for reporting survey data in ways that ensure its utility to a range of stakeholders and audiences.The chronology of ETS's large-scale literacy assessments, as shown in Fig. 9.1, spans more than 30 years. ETS served as the lead contractor in the development of these innovative assessments, while the prime clients and users of the assessment outcomes were representatives of either governmental organizations such as the National Center for Education Statistics (NCES) and Statistics Canada, or transgovernmental entities such as the Organisation for Economic Co-operation and Development (OECD). These instruments have evolved from a single-language, paper-based assessment focusing on a U.S. population of 16-to 25-year-olds to an adaptive, computer-based assessment administered in almost 40 countries and close to 50 languages to adults through the age of 65. By design, the assessments have been linked at the item level, with sets of questions from previous assessments

show abstract

“…Holman and Glas (2005) use reformulations of the models of O'Muircheartaigh and Moustaki (1999) to assess the extent to which the missing data are non-ignorable. Within the same framework, Moustaki and Knott (2000) present a latent variable model for binary and nominal observed items which includes covariate effects on attitudinal and response propensity items. In our study, we extend this approach to the longitudinal case with nonrandom dropout.…”

Section: Introductionmentioning

confidence: 99%

Analysis of Multivariate Longitudinal Data Subject to Nonrandom Dropout

Hafez

Moustaki

Kuha

2014

Structural Equation Modeling: A Multidisciplinary Journal

Self Cite

View full text Add to dashboard Cite

Longitudinal data are collected for studying changes across time. We consider multivariate longitudinal data where multiple observed variables, measured at each time point, are used as indicators for theoretical constructs (latent variables) of interest. A common problem in longitudinal studies is dropout, where subjects exit the study prematurely. Ignoring the dropout mechanism can lead to biased estimates, especially when the dropout is nonrandom. Our proposed approach uses latent variable models to capture the evolution of the latent phenomenon over time while also accounting for possibly nonrandom dropout. The dropout mechanism is modeled with a hazard function that depends on the latent variables and observed covariates. Different relationships among these variables and the dropout mechanism are studied via two model specifications. The proposed models are used to study people's perceptions on women's work using three questions from five waves from the British Household Panel Survey.

show abstract

Weighting for Item Non-Response in Attitude Scales by Using Latent Variable Models with Covariates

Cited by 59 publications

References 7 publications

Modeling Nonignorable Missing Data With Item Response Theory (Irt)

Modeling Nonignorable Missing Data With Item Response Theory (Irt)

Large-Scale Assessments of Adult Literacy

Analysis of Multivariate Longitudinal Data Subject to Nonrandom Dropout

Contact Info

Product

Resources

About