1 November 2017 in Behaviour Research and Therapy
1 April 2025 in Psychological Methods
The year 2022 is the 20th anniversary of Joseph Schafer and John Graham's paper titled "Missing data: Our view of the state of the art," currently the most highly cited paper in the history of Psychological Methods. Much has changed since 2002, as missing data methodologies have continually evolved and improved; the range of applications that are possible with modern missing data techniques has increased dramatically, and software options are light years ahead of where they were. This article provides an update on the state of the art that catalogs important innovations from the past two decades of missing data research. The paper addresses topics described in the original paper, including developments related to missing data theory, full information maximum likelihood, Bayesian estimation, multiple imputation, and models for missing not at random processes. The paper also describes newer factored regression specifications and missing data handling for multilevel models, both of which have been a focus of recent research. The paper concludes with a summary of the current software landscape and a discussion of several practical issues. (PsycInfo Database Record (c) 2023 APA, all rights reserved).
3 September 2018 in Multivariate Behavioral Research
Literature addressing missing data handling for random coefficient models is particularly scant, and the few studies to date have focused on the fully conditional specification framework and “reverse random coefficient” imputation. Although it has not received much attention in the literature, a joint modeling strategy that uses random within-cluster covariance matrices to preserve cluster-specific associations is a promising alternative for random coefficient analyses. This study is apparently the first to directly compare these procedures. Analytic results suggest that both imputation procedures can introduce bias-inducing incompatibilities with a random coefficient analysis model. Problems with fully conditional specification result from an incorrect distributional assumption, whereas joint imputation uses an underparameterized model that assumes uncorrelated intercepts and slopes. Monte Carlo simulations suggest that biases from these issues are tolerable if the missing data rate is 10% or lower and the sample is composed of at least 30 clusters with 15 observations per group. Furthermore, fully conditional specification tends to be superior with intraclass correlations that are typical of crosssectional data (e.g., ICC = .10), whereas the joint model is preferable with high values typical of longitudinal designs (e.g., ICC = .50).
24 February 2017 in Journal of Educational and Behavioral Statistics
Multiple imputation methods can generally be divided into two broad frameworks: joint model (JM) imputation and fully conditional specification (FCS) imputation. JM draws missing values simultaneously for all incomplete variables using a multivariate distribution, whereas FCS imputes variables one at a time from a series of univariate conditional distributions. In single-level multivariate normal data, these two approaches have been shown to be equivalent, but less is known about their similarities and differences with multilevel data. This study examined four multilevel multiple imputation approaches: JM approaches proposed by Schafer and Yucel and Asparouhov and Muthén and FCS methods described by van Buuren and Carpenter and Kenward. Analytic work and computer simulations showed that Asparouhov and Muthén and Carpenter and Kenward methods are most flexible, as they produce imputations that preserve distinct within- and between-cluster covariance structures. As such, these approaches are applicable to random intercept models that posit level-specific relations among variables (e.g., contextual effects analyses, multilevel structural equation models). In contrast, methods from Schafer and Yucel and van Buuren are more restrictive and impose implicit equality constraints on functions of the within- and between-cluster covariance matrices. The analytic work and simulations underscore the conclusion that researchers should not expect to obtain the same results from alternative imputation routines. Rather, it is important to choose an imputation method that partitions variation in a manner that is consistent with the analysis model of interest. A real data analysis example illustrates the various approaches.
2 February 2021 in Multivariate Behavioral Research
Missing data are exceedingly common across a variety of disciplines, such as educational, social, and behavioral science areas. Missing not at random (MNAR) mechanism where missingness is related to unobserved data is widespread in real data and has detrimental consequence. However, the existing MNAR-based methods have potential problems such as leaving the data incomplete and failing to accommodate incomplete covariates with interactions, non-linear terms, and random slopes. We propose a Bayesian latent variable imputation approach to impute missing data due to MNAR (and other missingness mechanisms) and estimate the model of substantive interest simultaneously. In addition, even when the incomplete covariates involves interactions, non-linear terms, and random slopes, the proposed method can handle missingness appropriately. Computer simulation results suggested that the proposed Bayesian latent variable selection model (BLVSM) was quite effective when the outcome and/or covariates were MNAR. Except when the sample size was small, estimates from the proposed BLVSM tracked closely with those from the complete data analysis. With a small sample size, when the outcome was less predictable from the covariates, the missingness proportions of the covariates and the outcome were larger, and the missingness selection processes of the covariates and the outcome were more MNAR and MAR, the performance of BLVSM was less satisfactory. When the sample size was large, BLVSM always performed well. In contrast, the method with an MAR assumption provided biased estimates and undercoverage confidence intervals when the missingness was MNAR. The robustness and the implementation of BLVSM in real data were also illustrated. The proposed method is available in the Blimp software application, and the paper includes a data analysis example illustrating its use.
1 October 2025 in Psychological Methods
Estimating power for multilevel models is complex because there are many moving parts, several sources of variation to consider, and unique sample sizes at Level 1 and Level 2. Monte Carlo computer simulation is a flexible tool that has received considerable attention in the literature. However, much of the work to date has focused on very simple models with one predictor at each level and one cross-level interaction effect, and approaches that do not share this limitation require users to specify a large set of population parameters. The goal of this tutorial is to describe a flexible Monte Carlo approach that accommodates a broad class of multilevel regression models with continuous outcomes. Our tutorial makes three important contributions. First, it allows any number of within-cluster effects, between-cluster effects, covariate effects at either level, cross-level interactions, and random coefficients. Moreover, we do not assume orthogonal effects, and predictors can correlate at either level. Second, our approach accommodates models with multiple interaction effects, and it does so with exact expressions for the variances and covariances of product random variables. Finally, our strategy for deriving hypothetical population parameters does not require pilot or comparable data. Instead, we use intuitive variance-explained effect size expressions to reverse-engineer solutions for the regression coefficients and variance components. We describe a new R package mlmpower that computes these solutions and automates the process of generating artificial data sets and summarizing the simulation results. The online supplemental materials provide detailed vignettes that annotate the R scripts and resulting output. (PsycInfo Database Record (c) 2023 APA, all rights reserved).