The Impact of Non-Normality on Extraction of Spurious Latent Classes in Mixture IRT Models

Abstract Objectives: Based on a multidimensional perspective, this study aimed to assess the heterogeneous health latent classes of older Chinese, and further examined the effects of health latent classes and associated factors on healthcare utilization. Methods: Data came from the Chinese Longitudinal Healthy Longevity Survey in 2014. Latent class analysis was adopted to identify heterogeneous health latent classes by health indicators of physical, psychological, and social dimensions. Two-part models were used to evaluate the impact of health latent classes and socio-demographic factors on outpatient and inpatient utilization. Results: Among 2,981 participants aged 65 and over without missing health indictors, four health latent classes were identified and labeled as “Lacking Socialization” (10.4%), “High Comorbidity” (16.7%), “Frail Group” (7.7%), and “Relatively Healthy” (65.1%). Among 1,974 participants with complete information, compared with the Relatively Healthy group, those in the Lacking Socialization group costed more inpatient expenditure (p-value =0.02). Those in the High Comorbidity and Frail groups tended to use healthcare services and costed more outpatient expenditure (all p-value <0.01). After controlling for health latent classes, the effects of age, gender, marital status, education, residence area, occupation, and health insurance on healthcare utilization were significant. Conclusions: Four heterogeneous health latent classes were identified by multidimensional health, and had significant effects on healthcare utilization. After controlling for health latent classes, different effects of socio-demographic factors on healthcare utilization were found. It enhances our understanding of heterogeneous health and complex healthcare demands in older Chinese, and is valuable for improving healthcare resource allocation targeted for healthy aging.

Download Full-text

An Exploration of the Heterogeneous Nature of Reading Comprehension Development in First Grade: The Impact of Word and Meaning Skills

Journal of Learning Disabilities ◽

10.1177/00222194211036203 ◽

2021 ◽

pp. 002221942110362

Author(s):

Emily J. Solari ◽

Ryan P. Grimm ◽

Alyssa R. Henry

Keyword(s):

Reading Comprehension ◽

Reading Fluency ◽

Word Reading ◽

First Grade ◽

Discrete Groups ◽

Latent Classes ◽

Linguistic Comprehension ◽

Comprehension Development ◽

The Impact ◽

Reading Decoding

This exploratory study builds upon extant reading development studies by identifying discrete groups based on reading comprehension trajectories across first grade. The main goal of this study was to enhance the field’s understanding of early reading comprehension development and its underlying subcomponent skills, with the intent of better understanding the development of comprehension in students who display risk for reading difficulties and disabilities. A sample of first-grade readers ( N = 314) were assessed at three timepoints across the first-grade year. These data were utilized to derive empirical latent classes based on reading comprehension performance across the first-grade year. Reading subcomponent skill assessments (phonological awareness, word reading, decoding, linguistic comprehension, and reading fluency), measured in the fall of first grade, were compared across latent classes to examine how they related to growth across the first-grade year. Results suggest that there were four distinct latent classes with differential reading comprehension development, each of which could also be distinguished by the subskill assessments. These findings are presented within the context of the broader reading research base and implications for practice are discussed.

Download Full-text

A General Unfolding IRT Model for Multiple Response Styles

Applied Psychological Measurement ◽

10.1177/0146621618762743 ◽

2018 ◽

Vol 43 (3) ◽

pp. 195-210 ◽

Cited By ~ 3

Author(s):

Chen-Wei Liu ◽

Wen-Chung Wang

Keyword(s):

Response Style ◽

Response Styles ◽

Parameter Estimates ◽

Data Set ◽

Parameter Recovery ◽

New Model ◽

Irt Model ◽

Irt Models ◽

Strongly Agree ◽

The Impact

It is commonly known that respondents exhibit different response styles when responding to Likert-type items. For example, some respondents tend to select the extreme categories (e.g., strongly disagree and strongly agree), whereas some tend to select the middle categories (e.g., disagree, neutral, and agree). Furthermore, some respondents tend to disagree with every item (e.g., strongly disagree and disagree), whereas others tend to agree with every item (e.g., agree and strongly agree). In such cases, fitting standard unfolding item response theory (IRT) models that assume no response style will yield a poor fit and biased parameter estimates. Although there have been attempts to develop dominance IRT models to accommodate the various response styles, such models are usually restricted to a specific response style and cannot be used for unfolding data. In this study, a general unfolding IRT model is proposed that can be combined with a softmax function to accommodate various response styles via scoring functions. The parameters of the new model can be estimated using Bayesian Markov chain Monte Carlo algorithms. An empirical data set is used for demonstration purposes, followed by simulation studies to assess the parameter recovery of the new model, as well as the consequences of ignoring the impact of response styles on parameter estimators by fitting standard unfolding IRT models. The results suggest the new model to exhibit good parameter recovery and seriously biased estimates when the response styles are ignored.

Download Full-text

The Impact of Multidimensionality on Extraction of Latent Classes in Mixture Rasch Models

Journal of Educational Measurement ◽

10.1111/jedm.12185 ◽

2018 ◽

Vol 55 (3) ◽

pp. 403-420 ◽

Cited By ~ 1

Author(s):

Yoonsun Jang ◽

Seock-Ho Kim ◽

Allan S. Cohen

Keyword(s):

Rasch Models ◽

Latent Classes ◽

The Impact

Download Full-text

The Impact of Test and Sample Characteristics on Model Selection and Classification Accuracy in the Multilevel Mixture IRT Model

Frontiers in Psychology ◽

10.3389/fpsyg.2020.00197 ◽

2020 ◽

Vol 11 ◽

Author(s):

Sedat Sen ◽

Allan S. Cohen

Keyword(s):

Model Selection ◽

Classification Accuracy ◽

Irt Model ◽

Mixture Irt ◽

Sample Characteristics ◽

The Impact

Download Full-text

Challenges in modelling the random structure correctly in growth mixture models and the impact this has on model mixtures

Journal of Developmental Origins of Health and Disease ◽

10.1017/s2040174414000130 ◽

2014 ◽

Vol 5 (3) ◽

pp. 197-205 ◽

Cited By ~ 4

Author(s):

M. S. Gilthorpe ◽

D. L. Dahly ◽

Y.-K. Tu ◽

L. D. Kubzansky ◽

E. Goodman

Keyword(s):

Mixture Models ◽

Later Life ◽

Random Structure ◽

Latent Classes ◽

Growth Mixture Models ◽

Autocorrelation Structure ◽

Growth Mixture ◽

Ideal Number ◽

The Impact ◽

The Ideal

Lifecourse trajectories of clinical or anthropological attributes are useful for identifying how our early-life experiences influence later-life morbidity and mortality. Researchers often use growth mixture models (GMMs) to estimate such phenomena. It is common to place constrains on the random part of the GMM to improve parsimony or to aid convergence, but this can lead to an autoregressive structure that distorts the nature of the mixtures and subsequent model interpretation. This is especially true if changes in the outcome within individuals are gradual compared with the magnitude of differences between individuals. This is not widely appreciated, nor is its impact well understood. Using repeat measures of body mass index (BMI) for 1528 US adolescents, we estimated GMMs that required variance–covariance constraints to attain convergence. We contrasted constrained models with and without an autocorrelation structure to assess the impact this had on the ideal number of latent classes, their size and composition. We also contrasted model options using simulations. When the GMM variance–covariance structure was constrained, a within-class autocorrelation structure emerged. When not modelled explicitly, this led to poorer model fit and models that differed substantially in the ideal number of latent classes, as well as class size and composition. Failure to carefully consider the random structure of data within a GMM framework may lead to erroneous model inferences, especially for outcomes with greater within-person than between-person homogeneity, such as BMI. It is crucial to reflect on the underlying data generation processes when building such models.

Download Full-text

Comparison of Scale Identification Methods in Mixture IRT Models

Journal of Modern Applied Statistical Methods ◽

10.22237/jmasm/1556669700 ◽

2020 ◽

Vol 18 (1) ◽

Author(s):

Youn-Jeng Choi ◽

Allan S. Cohen

Keyword(s):

Simulation Study ◽

Constraint Effect ◽

Identification Methods ◽

Irt Models ◽

Correct Model ◽

Mixture Irt ◽

3Pl Model

The effects of three scale identification constraints in mixture IRT models were studied. A simulation study found no constraint effect on the mixture Rasch and mixture 2PL models, but the item anchoring constraint was the only one that worked well on selecting correct model with the mixture 3PL model.

Download Full-text