Case deletion diagnostics in multilevel models

Shi, Lei; Chen, Gemai

doi:10.1016/j.jmva.2008.01.023

Cited by 14 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Among the functions outlined in Table 4 it is important to note that HLMdiag provides two implementations of cooks.distance, mdffits, covratio, and covtrace: one based on the full model refit, and the other based on a one-step approximation. The example discussed in the above sections illustrated the "fast" implementation based on one-step approximations for the fixed effects and associated covariance matrices (for further details we refer the reader to Christensen et al 1992;Shi and Chen 2008;Zewotir 2008). The implementation of these approximations utilizes smaller, dense submatrices resulting in more efficient computation.…”

Section: Package Descriptionmentioning

confidence: 99%

HLMdiag: A Suite of Diagnostics for Hierarchical Linear Models inR

Loy¹,

Hofmann²

2014

J. Stat. Soft.

View full text Add to dashboard Cite

Over the last twenty years there have been numerous developments in diagnostic procedures for hierarchical linear models; however, these procedures are not widely implemented in statistical software packages, and those packages that do contain a complete framework for model assessment are not open source. The lack of availability of diagnostic procedures for hierarchical linear models has limited their adoption in statistical practice. The R package HLMdiag provides diagnostic tools targeting all aspects and levels of continuous response hierarchical linear models with strictly nested dependence structures fit using the lmer() function in the lme4 package. In this paper we discuss the tools implemented in HLMdiag for both residual and influence analysis.

show abstract

Section: Package Descriptionmentioning

confidence: 99%

HLMdiag: A Suite of Diagnostics for Hierarchical Linear Models inR

Loy¹,

Hofmann²

2014

J. Stat. Soft.

View full text Add to dashboard Cite

show abstract

“…Local influence—perturbing various aspects of the model and examining the sensitivity of the model through the normal curvature of the likelihood displacement—is one method for simultaneous assessment and has been applied to the mixed ANOVA model,39 the linear mixed model,40 and the HLM 41. Another method, based on a Taylor series approximation of V ( i ) , has been proposed for the linear mixed model20,24 and the HLM 42. In this approach, calculations of deletion diagnostics are based on a covariance structure for fixed effects, with the i th observation or group deleted, i.e.…”

Section: Influence Analysismentioning

confidence: 99%

Diagnostic tools for hierarchical linear models

Loy

Hofmann

2012

WIREs Computational Stats

View full text Add to dashboard Cite

Hierarchical structures are omnipresent in today's society—this is reflected in the data that we collect on all aspects of this society. Hierarchical linear models allow a representation of structural levels in a statistical modeling framework. Diagnostic tools are used to assess the quality of model estimation and explore features of the data not well described by the model. Residual and influence diagnostics are familiar tools for the classical regression model with independent observations. For hierarchical linear models, these diagnostic tools must be adjusted to reflect the dependence introduced by the nested data structure. Residual analysis now includes the assessment of distributional assumptions at each level of the model. This requires the use of level‐dependent residual quantities. Similarly, the parameter estimates may be influenced at each level of the model, requiring influence diagnostics that can pinpoint specific levels of the model, as well as specific aspects of the model. We present an overview of the diagnostic tools available for hierarchical linear models that are familiar from linear models. Additionally, we discuss the utility of the lineup protocol for residual analysis with complex models. WIREs Comput Stat 2013, 5:48–61. doi: 10.1002/wics.1238 This article is categorized under: Statistical Models > Bayesian Models Statistical and Graphical Methods of Data Analysis > Bayesian Methods and Theory Statistical and Graphical Methods of Data Analysis > Modeling Methods and Algorithms Statistical and Graphical Methods of Data Analysis > Statistical Graphics and Visualization

show abstract

“…We can find the method in Ronald and Larry (1992) and Shi and Chen (2008). On the other hand, for local influence considered in Cook (1986), instead of removing the case completely, giving each case a weight, the influence is assessed by perturbing these weights, and measured by normal curvature of an influence graph based on likelihood displacement.…”

Section: Introductionmentioning

confidence: 99%