Treatments of Missing Data: A Monte Carlo Comparison of RBHDI, Iterative Stochastic Regression Imputation, and Expectation-Maximization

By Michael S. Gold, Peter M. Bentler. 2000

354 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/d0482b5ee82be3d8cbf18b147ee76bbb59dd1bfb

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Structural Equation Modeling in the Communication Sciences, 1995–2000 (2002, background): “It is important for researchers to define how they dealt with missing values (Gold & Bentler, 2000).”
Missing Data in Educational Research: A Review of Reporting Practices and Suggestions for Improvement (2004, neutral): “…of these methods over traditional methods such as listwise and pairwise deletion (e.g., Arbuckle, 1996; Enders, 2001a, 2001b, 2003; Enders & Bandalos, 2001; Gold & Bentler, 2000; Graham, Hofer, & MacKinnon, 1996; Graham & Schafer, 1999; Muthén, Kaplan, & Hollis, 1987; Kaplan, 1995; Wothke, 2000).”
Multilevel Causal Ordering of Academic Self-Concept and Achievement: Influence of Language of Instruction (English Compared With Chinese) for Hong Kong Students (2002, background): “A growing body of research emphasizes potential problems with traditional pairwise, listwise, and mean substitution approaches to missing data (e.g., Brown, 1994; Gold & Bentler, 2000; Graham & Hoffer, 2000; Little & Rubin, 1987).”
Presidential IQ, Openness, Intellectual Brilliance, and Leadership: Estimates and Correlations for 42 U.S. Chief Executives (2006, methodology): “Because it takes into consideration the entire data structure, EM has been shown to be superior to alternative imputation procedures, such as regression (see, e.g., Gold & Bentler, 2000).”
The Influence of Childhood (2012, methodology): “The EM–ML approach, and use of the minimum N per variance term, ensured that we accounted for missing data and mild nonnormality (Gold & Bentler, 2000).”
Security-related factors in extended UTAUT model for NFC based mobile payment in the restaurant industry (2017, methodology): “Missing values checked to be missing completely at random and imputed by Expectation Maximization (EM) method (Gold & Bentler, 2000; Hair Jr., et al., 2010).”
Fully conditional specification in multivariate imputation (2006, methodology): “Variations of iterated regression imputation have been studied later by Finkbeiner [9], Raymond and Roberts [10], Jinn and Sedransk [11], and Gold and Bentler [12].”
A Statistically Justified Pairwise ML Method for Incomplete Nonnormal Data: A Comparison With Direct ML and Pairwise ADF (2005, methodology): “Other authors (e.g., Gold & Bentler, 2000) refer to the two-stage method as unstructured EM and to the direct ML method as structured EM.”
A Comparison of Maximum-Likelihood and Asymptotically Distribution-Free Methods of Treating Incomplete Nonnormal Data (2003, methodology): “Graham et al. (1996) then replicated their simulation with two changes: (a) the simulated data were univariately nonnormal, with skew ranging from –.”
How does school climate impact academic achievement? An examination of social identity processes (2017, methodology): “To secure better estimates and include as many cases as possible for SEM analyses the school climate and identification missing items were imputed with the Expectation Maximisation (EM, Gold & Bentler, 2000) method with SPSS 20.”
In Search of Underlying Dimensions: The Use (and Abuse) of Factor Analysis in Personality and Social Psychology Bulletin (2002, methodology): “A recent Monte Carlo analysis by Gold and Bentler (2000) has indicated that this approach provides a more accurate estimate of model parameters than restricting the analysis to cases with complete data.”
Longitudinal Modeling with Randomly and Systematically Missing Data: A Simulation of Ad Hoc, Maximum Likelihood, and Multiple Imputation Techniques (2003, neutral): “Initial research by Gold and Bentler (2000) indicates that parameter bias under the EM algorithm is not increased by univariate skewness and kurtosis.”
Psychological capital bolsters motivation, engagement, and achievement: Cross-sectional and longitudinal studies (2018, neutral): “Expectation maximization is considered one of the best ways to deal with missing data (Gold & Bentler, 2000).”
The Development and Validation of the Game User Experience Satisfaction Scale (GUESS) (2016, neutral): “…techniques (Enders, 2003; Fox-Wasylyshyn & El-Masri, 2005; Graham, 2009; Musil, Warner, Yobas, & Jones, 2002) and that this method outperformed other methods (e.g., iterative stochastic regression imputation) under nonideal conditions (e.g., non-normally distributed data; Gold & Bentler, 2000).”