Modeling Unobserved Heterogeneity Using Latent Profile Analysis: A Monte Carlo Simulation

By J. Peugh, Xitao Fan. 2013

321 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/e36ee3763a4c9fb1638a8b662eb30918b3b95703

How researchers see this paper

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

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Dual commitment to organization and supervisor: A person-centered approach (2015, background): “Alternative models with the variances and/or covariances of the indicators freely estimated in all profiles (e.g., Morin, Maïano, et al., 2011; Peugh & Fan, 2013) generally tended to converge on improper solutions (negative variance estimates, non-positive definite Fisher Information matrix, etc.)…”
Multiple-Group Analysis of Similarity in Latent Profile Solutions (2016, unknown): “…best recovers the sample’s true parameters, whereas the AIC is not and presents a marked tendency for overextraction (e.g., Henson, Reise, & Kim, 2007; McLachlan & Peel, 2000; Nylund, Asparouhov, & Muthén, 2007; Peugh & Fan, 2013; Tein, Coxe, & Cham, 2013; Tofighi & Enders, 2008; Tolvanen, 2007).”
Performance of the entropy as an index of classification accuracy in latent profile analysis: A Monte Carlo simulation study (2017, unknown): “…< 0.60 时相当于超过 20%的个体存在分类错误; Entropy≥0.80 表明分类准确率超过 90%。这一结果是否能推广到研究设定因素之外的情况呢？另外, 在他们的研究中还存在如下几个方面的不足。首先, 该研究考查的样本量范围有限, 仅考查了 N = 300 时的情况。在本研究之前的预实验中 , 我们发现样本量与 Entropy 呈负相关, 即样本量越大 Entropy 值越小, 所以扩大样本量范围对全面了解 Entropy 的表现具有重要意义。另外, 她们的研究只考查了存在 2 个潜在类别组即 k = 2 的情况, 然而在实际研究中, 潜类别数量通常多于 2…”
Complementary Variable- and Person-Centered Approaches to the Dimensionality of Psychometric Constructs: Application to Psychological Wellbeing at Work (2016, background): “Simulation studies indicate that four of these indicators (CAIC, BIC, ABIC, and BLRT) are particularly effective and that when the indicators fail to retain the optimal model, the ABIC and BLRT tend to overestimate the number of classes, whereas the BIC, CAIC, and aLMR tend to underestimate it (Nylu”
A person‐centered approach to commitment research: Theory, research, and methodology (2016, background): “Relaxing this default has been shown to result in less biased parameter estimates (Peugh & Fan, 2013) and a more realistic representation of the complexity of human nature (Morin, Maïano, et al., 2011).”
Latent profile analysis: A review and “how to” guide of its application within vocational behavior research (2020, methodology): “…review about LPA, for the sake of completeness and clarity, we introduce a common LPA model equation as shown in Lazarsfeld and Henry (1968) or Peugh and Fan (2013): ( footnote continued) analysis model is therefore the specified distribution of the underlying latent variables (discrete…”
Fuzzy Profiles (2018, methodology): “We allowed profile indicators’ intercepts and variances to be freely estimated across profiles (e.g., Morin et al., 2011; Morin & Marsh, 2015; Peugh & Fan, 2013).4 Second, we obtained the mostlikely class membership from the posterior distribution from the profile enumeration process in the prior…”
Stability, change, and implications of students’ motivation profiles: A latent transition analysis (2017, methodology): “…that four of these indicators (CAIC, BIC, ABIC, and BLRT) are particularly effective (Henson, Reise, & Kim, 2007; Nylund, Asparouhov, & Muthén 2007; Peugh & Fan, 2013; Tein, Coxe, & Cham, 2013; Tofighi & Enders, 2008; Yang, 2006), while the AIC and LMR/ALMR should not be used in the class…”
Impact of Misspecifications of the Latent Variance–Covariance and Residual Matrices on the Class Enumeration Accuracy of Growth Mixture Models (2016, background): “…this conditional independence assumption (i.e., assuming the orthogonality of the residuals conditional on the model) could also lead to potentially severe biases in terms of class enumeration accuracy (Peugh & Fan, 2013), as well as parameter or standard error estimates (Heggeseth & Jewell, 2013).”
Variable-Centered, Person-Centered, and Person-Specific Approaches (2018, methodology): “Based on prior recommendations and empirical works (Meyer, Morin, & Vandenberghe, 2015; Morin et al., 2011; Peugh & Fan, 2013), we originally allowed all profile means and variances to be freely estimated.”
Ten Frequently Asked Questions About Latent Class Analysis (2018, unknown): “…the sensitivity of the within-class variance assumptions and class separation (Bauer & Curran, 2003; B. O. Muthén, 2003), where its implications for the class enumeration process in growth mixture models (Tofighi & Enders, 2008) and LPA models (Peugh & Fan, 2013) have been subject to investigation.”
Profiles of dual commitment to the occupation and organization: Relations to well-being and turnover intentions (2015, methodology): “In all models, the means and variances of the indicators (i.e., the commitment PVs) were freely estimated in all profiles (e.g., Morin et al., 2011a; Peugh & Fan, 2013).”
Motivation profiles at work: A self-determination theory approach (2016, methodology): “However, the entropy should not be used in itself to determine the optimal number of profiles (Lubke & Muthén, 2007; Peugh & Fan, 2012, 2013, 2015; Tein, Coxe, & Cham, 2013).”
Mixture Modeling for Lifespan Developmental Research (2019, background): “…and BLRT) are particularly effective, while the AIC and LMR/ALMR should not be used (Diallo, Morin, & Lu, 2016, 2017; Henson et al., 2007; McLachlan & Peel, 2000; Nylund, Asparouhov, & Muthén, 2007; Peugh & Fan, 2013; Tein, Coxe, & Cham, 2013; Tofighi & Enders, 2008; Tolvanen, 2007; Yang, 2006).”
Further Reflections on Disentangling Shape and Level Effects in Person-Centered Analyses: An Illustration Exploring the Dimensionality of Psychological Health (2016, methodology): “For each parameterization, models with 1 to 8 latent profiles were estimated with the indicators’ (psychological health scores) intercepts and residuals freely estimated in all profiles (Morin, Maïano, et al., 2011; Peugh & Fan, 2013).”