9. Multiple Imputation of Incomplete Categorical Data Using Latent Class Analysis

Vermunt, Jeroen K.; Ginkel, Joost R. van; Ark, L. Andries van der; Sijtsma, Klaas

doi:10.1111/j.1467-9531.2008.00202.x

Cited by 93 publications

(130 citation statements)

References 43 publications

(79 reference statements)

Supporting

Mentioning

126

Contrasting

Unclassified

Order By: Relevance

“…For categorical variables, the log-linear model (Schafer, 1997) would be the most appropriate imputation model. However, the log-linear model can be applied only when the number of variables used in the imputation model is small (Vermunt, van Ginkel, van der Ark, & Sijtsma, 2008), whereby it is feasible to set up and process the full multiway cross-tabulation required for the log-linear analysis. As an alternative, imputations can be carried out under a fully conditional specification approach (van Buuren, 2007), where a sequence of regression models for the univariate conditional distributions of the variables with missing values are specified.…”

Section: Multiple Imputation (Mi)mentioning

confidence: 99%

Simple imputation methods versus direct likelihood analysis for missing item scores in multilevel educational data

et al. 2011

View full text Add to dashboard Cite

Missing data, such as item responses in multilevel data, are ubiquitous in educational research settings. Researchers in the item response theory (IRT) context have shown that ignoring such missing data can create problems in the estimation of the IRT model parameters. Consequently, several imputation methods for dealing with missing item data have been proposed and shown to be effective when applied with traditional IRT models. Additionally, a nonimputation direct likelihood analysis has been shown to be an effective tool for handling missing observations in clustered data settings. This study investigates the performance of six simple imputation methods, which have been found to be useful in other IRT contexts, versus a direct likelihood analysis, in multilevel data from educational settings. Multilevel item response data were simulated on the basis of two empirical data sets, and some of the item scores were deleted, such that they were missing either completely at random or simply at random. An explanatory IRT model was used for modeling the complete, incomplete, and imputed data sets. We showed that direct likelihood analysis of the incomplete data sets produced unbiased parameter estimates that were comparable to those from a complete data analysis. Multipleimputation approaches of the two-way mean and corrected item mean substitution methods displayed varying degrees of effectiveness in imputing data that in turn could produce unbiased parameter estimates. The simple random imputation, adjusted random imputation, item means substitution, and regression imputation methods seemed to be less effective in imputing missing item scores in multilevel data settings.Keywords Item response theory . Multilevel data . Missing data . Imputation methods Multilevel data in education settings can contain complex patterns of nested sources of variability. For instance, suppose that exercise items nested in courses or chapters with varying difficulty levels are presented to students nested in schools or classes with varying ability levels. For each student, scores to the items, along with person and item properties, can be recorded. The collected data are clustered or multilevel in nature, consisting of the properties of schools, students, chapters, and items and the students' item scores (e.g., binary, pass/fail) on attempted items. These data can be modeled statistically-say, using item response theory (IRT; van der Linden & Hambleton, 1997)-to explain and understand student characteristics in relation to the item properties.De Boeck and Wilson (2004) described IRT models within the framework of generalized linear mixed models (GLMMs) or nonlinear mixed models (NLMMs), also accounting for more complex multilevel structures than the structure of measurement occasions within subjects,

show abstract

Section: Multiple Imputation (Mi)mentioning

confidence: 99%

Simple imputation methods versus direct likelihood analysis for missing item scores in multilevel educational data

et al. 2011

View full text Add to dashboard Cite

show abstract

“…In this article, a different alternative approach based on LCA was taken which classified students regarding their background characteristics and used these classifications as predictors in the LRM. This approach is similar to applying LCA directly as an imputation model for categorical data which has been shown to yield good recovery of parameters and standard errors when a sufficient number of latent classes is used (Vermunt et al, 2008). Since the parameters of interest in large-scale educational surveys are student proficiencies, we used LCA as a method of obtaining predictors for the LRM.…”

Section: Discussionmentioning

confidence: 99%

“…Vermunt et al (2008) note that in the context of using LCA for the purposes of density estimation, fitting a larger number of classes than necessary (overfitting) is advantageous since sample specific variability can be captured whereas underfitting is problematic because important associations or interactions between variables are ignored. In this study, the LCA approaches were evaluated with respect to how well they recovered the results from the operational approach since this is the method currently used in most large-scale educational surveys and it has been shown to provide unbiased estimates (e.g., Oranje & Ye, 2014).…”

Section: Discussionmentioning

confidence: 99%

“…We used 27 classes simply because the set of background variables contains about twice as many observed variables, so we halved the number of observed variables. Vermunt, van Ginkel, van der Ark, and Sijtsma (2008) explain why identification and local maxima are less of a concern if an LCA classification is used as an auxiliary variable to explain dependencies in an imputation model, which is the case here for the latent regression, since it is essentially an imputation model for plausible values.…”

Section: Evaluation Of the Operational Approach And Alternative Appromentioning

confidence: 99%

See 1 more Smart Citation

An Alternative Way to Model Population Ability Distributions in Large-Scale Educational Surveys

Wetzel

Davier

2014

Educational and Psychological Measurement

View full text Add to dashboard Cite

In large-scale educational surveys, a latent regression model is used to compensate for the shortage of cognitive information. Conventionally, the covariates in the latent regression model are principal components extracted from background data. This operational method has several important disadvantages, such as the handling of missing data and the high model complexity. The approach introduced here to identify multiple groups that can account for the variation among students is to conduct a latent class analysis (LCA). In the LCA, one or more latent nominal variables are identified that can be used to classify respondents with respect to their background characteristics. These classifications are then introduced as predictors in the latent regression. The primary goal of this study was to explore whether this approach yields similar estimates of group means and standard deviations compared with the operational procedure. The alternative approaches based on LCA differed regarding the number of classes, the items used for the LCA, and whether manifest class membership information or class membership probabilities were used as independent variables in the latent regression. Overall, recovery of the operational approach's group means and standard deviations was very satisfactory for all LCA approaches. Furthermore, the posterior means and standard deviations used to generate plausible values derived from the operational approach and the LCA approaches correlated highly. Thus, incorporating independent variables based on an LCA of background data into the latent regression model appears to be a viable alternative to the operational approach.

show abstract

“…In order to minimise any loss of information, multiple imputation is performed (Rubin 1987;Vermunt, Van Ginkel, Van der Ark and Sijtsma 2008 ). A highly important methodological challenge of this kind of complex model is the sparseness of data.…”

Section: 1mentioning

confidence: 99%

The Global Omnivore: Identifying Musical Taste Groups in Austria, England, Israel and Serbia

Leguina¹,

Widdop²,

Tampubolon³

2016

Sociological Research Online

View full text Add to dashboard Cite

This research offers a unique opportunity to revisit the omnivore hypothesis under a unified method of cross-national analysis. To accomplish this, we interpret omnivourism as a special case of cultural eclecticism (Ollivier, 2008;Ollivier, Gauthier and Truong, 2009).Our methodological approach incorporates the simultaneous analysis of locally produced and globally known musical genres. Its objective is to verify whether cultural omnivourism is a widespread phenomenon, and to determine to what extent any conclusions can be generalised across countries with different social structures and different levels of cultural openness. To truly understand the scope of the omnivourism hypothesis, we argue that it is essential to perform a cross-national comparison to test the hypothesis within a range of social, political and cultural contexts, and a reflection of different historical and cultural repertoires (Lamont, 1992).

show abstract

9. Multiple Imputation of Incomplete Categorical Data Using Latent Class Analysis

Cited by 93 publications

References 43 publications

Simple imputation methods versus direct likelihood analysis for missing item scores in multilevel educational data

Simple imputation methods versus direct likelihood analysis for missing item scores in multilevel educational data

An Alternative Way to Model Population Ability Distributions in Large-Scale Educational Surveys

The Global Omnivore: Identifying Musical Taste Groups in Austria, England, Israel and Serbia

Contact Info

Product

Resources

About