×

Model fit criteria curve behaviour in class enumeration – a diagnostic tool for model (mis)specification in longitudinal mixture modelling. (English) Zbl 07546449

Summary: The use of longitudinal finite mixture models (FMMs) to identify latent classes of individuals following similar paths of temporal development is gaining traction in applied research. However, FMM’s users may be unaware of how data features as well as the inappropriate specification of the model’s covariance structure impacts class enumeration. To elucidate this, we investigated model fit-criteria curve behaviour across an array of data conditions and covariance structures. Fit statistic patterns were variable among the fit criteria and across a range of data conditions. This variability was greatly attributable to the level of class separation and the presence/absence of random effects. Our findings support some widely held notions (e.g. BIC outperforms other criteria) while debunking others (adding random effects is not always the solution). Based on the obtained results, we present guidelines on how the behaviour of fit criteria curves can be used as a diagnostic aid during class enumeration.

MSC:

62-XX Statistics
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Muthén, BO; Shedden, K.; Muthén, B., Finite mixture modeling with mixture outcomes using the EM algorithm, Biometrics, 55, 463-469 (1999) · Zbl 1059.62599 · doi:10.1111/j.0006-341X.1999.00463.x
[2] Berlin, KS; Parra, GR; Williams, NA., An introduction to latent variable mixture modeling (part 2): longitudinal latent class growth analysis and growth mixture models, J Pediatr Psychol, 39, 188-203 (2014) · doi:10.1093/jpepsy/jst084
[3] Nagin, DS. Group-based modeling of development. Cambridge (MA): Harvard University Press; 2005. Available from: http://www.degruyter.com/view/books/9780674041318/9780674041318/9780674041318.xml.
[4] Lima Passos, V.; Klijn, S.; van Zandvoort, K., At the heart of the problem - a person-centred, developmental perspective on the link between alcohol consumption and cardio-vascular events, Int J Cardiol, 232, 304-314 (2017) · doi:10.1016/j.ijcard.2016.12.094
[5] Falkenstein, MJ; Nota, JA; Krompinger, JW, Empirically-derived response trajectories of intensive residential treatment in obsessive-compulsive disorder: a growth mixture modeling approach, J Affect Disord, 245, 827-833 (2019)
[6] Grevenstein, D.; Kröninger-Jungaberle, H., Two patterns of cannabis use among adolescents: results of a 10-year prospective study using a growth mixture model, Subst Abus, 36, 85-89 (2015) · doi:10.1080/08897077.2013.879978
[7] Erosheva, EA; Matsueda, RL; Telesca, D., Breaking bad: two decades of life-course data analysis in criminology, developmental psychology, and beyond, Annu Rev Stat Its Appl, 1, 301-332 (2014) · doi:10.1146/annurev-statistics-022513-115701
[8] Enders, CK; Tofighi, D., The impact of misspecifying class-specific residual variances in growth mixture models, Struct Equ Model, 15, 75-95 (2008)
[9] Diallo, TMO; Morin, AJS; Lu, HZ., Impact of misspecifications of the latent variance-covariance and residual matrices on the class enumeration accuracy of growth mixture models, Struct Equ Model, 23, 507-531 (2016) · doi:10.1080/10705511.2016.1169188
[10] Davies, CE; Glonek, GFV; Giles, LC., The impact of covariance misspecification in group-based trajectory models for longitudinal data with non-stationary covariance structure, Stat Methods Med Res, 26, 1982-1991 (2017) · doi:10.1177/0962280215598806
[11] Morin, AJS; Maïano, C.; Nagengast, B., General growth mixture analysis of adolescents’ developmental trajectories of anxiety: the impact of untested invariance assumptions on substantive interpretations, Struct Equ Model A Multidiscip J, 18, 613-648 (2011) · doi:10.1080/10705511.2011.607714
[12] Sijbrandij, JJ; Hoekstra, T.; Almansa, J., Identification of developmental trajectory classes: comparing three latent class methods using simulated and real data, Adv Life Course Res, 42, 100288 (2019)
[13] McNeish, D.; Harring, JR., The effect of model misspecification on Growth Mixture Model class enumeration, J Classif, 34, 223-248 (2017) · Zbl 1373.62313 · doi:10.1007/s00357-017-9233-y
[14] Sijbrandij, JJ; Hoekstra, T.; Almansa, J., Variance constraints strongly influenced model performance in growth mixture modeling: a simulation and empirical study, BMC Med Res Methodol, 20, 276 (2020) · doi:10.1186/s12874-020-01154-0
[15] Nagin, DS; Land, KC., Age, criminal careers, and population heterogeneity: specification and estimation of a nonparametric, mixed poisson model, Criminology, 31, 327-362 (1993)
[16] van der Nest, G.; Lima Passos, V.; Candel, MJJM, An overview of mixture modelling for latent evolutions in longitudinal data: modelling approaches, fit statistics and software, Adv Life Course Res, 43, 100323 (2020)
[17] Brame, R.; Nagin, DS; Wasserman, L., Exploring some analytical characteristics of finite mixture models, J Quant Criminol, 22, 31-59 (2006) · doi:10.1007/s10940-005-9001-8
[18] Kim, ES; Wang, Y., Class enumeration and parameter recovery of growth mixture modeling and second-order growth mixture modeling in the presence of measurement noninvariance between latent classes, Front Psychol, 8 (2017) · doi:10.3389/fpsyg.2017.01499/full
[19] Lo, Y.; Mendell, NR; Rubin, DB., Testing the number of components in a normal mixture, Biometrika, 88, 767-778 (2001) · Zbl 0985.62019
[20] Muthén, BO., Statistical and substantive checking in growth mixture modeling: comment on Bauer and Curran (2003), Psychol Methods, 8, 369-377 (2003) · doi:10.1037/1082-989X.8.3.369
[21] Nylund, KL; Asparouhov, T.; Muthén, BO., Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study, Struct Equ Model A Multidiscip J, 14, 535-569 (2007) · doi:10.1080/10705510701575396
[22] Tofighi, D.; Enders, CK.; Hancock, GR; Samuelsen, KM, Adv latent var mix model, Identifying the correct number of classes in growth mixture models, 317-341 (2008), Greenwich (CT: Information Age, Greenwich (CT
[23] Grimm, KJ; Mazza, GL; Davoudzadeh, P., Model selection in finite mixture models: a k -fold cross-validation approach, Struct Equ Model A Multidiscip J, 24, 246-256 (2017) · doi:10.1080/10705511.2016.1250638
[24] van de Schoot, R.; Sijbrandij, M.; Winter, SD, The GRoLTS-checklist: guidelines for reporting on latent trajectory studies, Struct Equ Model, 24, 451-467 (2017) · doi:10.1080/10705511.2016.1247646
[25] Celeux, G.; Soromenho, G., An entropy criterion for assessing the number of clusters in a mixture model, J Classif, 13, 195-212 (1996) · Zbl 0861.62051 · doi:10.1007/BF01246098
[26] Depaoli, S., Mixture class recovery in GMM under varying degrees of class separation: frequentist versus Bayesian estimation, Psychol Methods, 18, 186-219 (2013) · doi:10.1037/a0031609
[27] Tolvanen, A. hideLatent growth mixture modeling: a simulation study [Internet]. Unpublishe. University of Jyvaskyla, Finland; 2007. Available from: http://urn.fi/URN:ISBN:951-39-2971-8. · Zbl 1153.65012
[28] Li, M.; Harring, JR., Investigating approaches to estimating covariate effects in growth mixture modeling: a simulation study, Educ Psychol Meas, 77, 766-791 (2017) · doi:10.1177/0013164416653789
[29] Nowakowska, E, Koronacki, J, Lipovetsky, S. Tractable Measure of Component Overlap for Gaussian Mixture Models. 2014;1-24. Available from: http://arxiv.org/abs/1407.7172.
[30] Cohen, J., Statistical power analysis for the behavioral sciences (1988), Hillsdale (NJ: Lawrence Erlbaum Associates, Hillsdale (NJ · Zbl 0747.62110
[31] McLachlan, GJ., Mahalanobis distance, Resonance, 4, 20-26 (1999) · doi:10.1007/BF02834632.pdf
[32] Lubke, G.; Neale, MC., Distinguishing between latent classes and continuous factors: resolution by maximum likelihood?, Multivariate Behav Res, 41, 499-532 (2006) · doi:10.1207/s15327906mbr4104_4
[33] Diallo, TMO; Morin, AJS; Lu, HZ., Performance of growth mixture models in the presence of time-varying covariates, Behav Res Methods, 49, 1951-1965 (2017) · doi:10.3758/s13428-016-0823-0
[34] Martin, DP; von Oertzen, T., Growth mixture models outperform simpler clustering algorithms when Detecting longitudinal heterogeneity, even with small sample sizes, Struct Equ Model A Multidiscip J, 22, 264-275 (2015) · doi:10.1080/10705511.2014.936340
[35] Peugh, J.; Fan, X., How well does growth mixture modeling identify heterogeneous growth trajectories? A simulation study examining GMM’s performance characteristics, Struct Equ Model A Multidiscip J, 19, 204-226 (2012) · doi:10.1080/10705511.2012.659618
[36] McLachlan, G, Peel, D. Finite mixture models [Internet]. Wiley Ser. Probab. Stat. Appl. Probab. Stat. Sect. Hoboken, NJ, USA: John Wiley & Sons, Inc.; 2000. Available from: http://doi.wiley.com/. · Zbl 0963.62061
[37] Tueller, S.; Lubke, G., Evaluation of structural equation mixture models: parameter estimates and correct class assignment, Struct Equ Model A Multidiscip J, 17, 165-192 (2010) · doi:10.1080/10705511003659318
[38] He, J.; Fan, X., Evaluating the performance of the K-fold cross-validation approach for model selection in growth mixture modeling, Struct Equ Model, 1-14 (2018) · doi:10.1080/10705511.2018.1500140
[39] Liu, Y.; Luo, F.; Liu, H., Factors of piecewise growth mixture model: distance and pattern, Acta Psychol Sin, 46, 1400 (2014) · doi:10.3724/SP.J.1041.2014.01400
[40] Mattsson, M.; Maher, GM; Boland, F., Group-based trajectory modelling for BMI trajectories in childhood: a systematic review, Obes Rev, 20, 998-1015 (2019) · doi:10.1111/obr.12842
[41] Watson, L.; Belcher, J.; Nicholls, E., Latent class growth analysis of gout flare trajectories: a three-year prospective cohort study in primary care, Arthritis Rheumatol, 72, 1928-1935 (2020) · doi:10.1002/art.41476
[42] Coyne, SM; Padilla-Walker, LM; Holmgren, HG, Instagrowth: a longitudinal growth mixture model of social media time Use across adolescence, J Res Adolesc, 29, 897-907 (2019)
[43] Hu, J.; Leite, WL; Gao, M., An evaluation of the use of covariates to assist in class enumeration in linear growth mixture modeling, Behav Res Methods, 49, 1179-1190 (2017) · doi:10.3758/s13428-016-0778-1
[44] Morgan, GB; Hodge, KJ; Baggett, AR., Latent profile analysis with nonnormal mixtures: a Monte Carlo examination of model selection using fit indices, Comput Stat Data Anal, 93, 146-161 (2016) · Zbl 1468.62143
[45] Kim, SY., Sample size requirements in single- and multiphase growth mixture models: a Monte Carlo Simulation study, Struct Equ Model, 19, 457-476 (2012)
[46] Sher, KJ; Jackson, KM; Steinley, D., Alcohol Use trajectories and the ubiquitous Cat’s cradle: cause for concern?, J Abnorm Psychol, 120, 322-335 (2011)
[47] Bonanno, GA., Loss, trauma, and human resilience: have we underestimated the human capacity to thrive after extremely aversive events?, Am Psychol, 59, 20-28 (2004) · doi:10.1037/0003-066X.59.1.20
[48] McNeish, D.; Harring, J., Covariance pattern mixture models: eliminating random effects to improve convergence and performance, Behav Res Methods (2019) · doi:10.3758/s13428-019-01292-4
[49] Bauer, DJ; Curran, PJ., Distributional assumptions of growth mixture models: implications for overextraction of latent trajectory classes, Psychol Methods, 8, 338-363 (2003) · doi:10.1037/1082-989X.8.3.338
[50] Hallquist, MN; Wiley, JF., Mplusautomation : an R Package for Facilitating Large-Scale LateNT VARIable analyses in M plus, Struct Equ Model A Multidiscip J, 25, 621-638 (2018) · doi:10.1080/10705511.2017.1402334
[51] Muthén, LK; Muthén, BO., MPlus user’s guide. eighth Edi (2017), Los Angeles (CA: Muthén & Muthén, Los Angeles (CA
[52] Wickham, H. ggplot2: elegant graphics for data analysis. New York: Springer-Verlag; 2016. Available from: https://ggplot2.tidyverse.org. · Zbl 1397.62006
[53] Li, L.; Hser, Y-I., On inclusion of covariates for class enumeration of growth mixture models, Multivariate Behav Res, 46, 266-302 (2011) · doi:10.1080/00273171.2011.556549
[54] Hipp, JR; Bauer, DJ., Local solutions in the estimation of growth mixture models, Psychol Methods, 11, 36-53 (2006) · doi:10.1037/1082-989X.11.1.36
[55] Engle, S.; Whalen, S.; Joshi, A., Unboxing cluster heatmaps, BMC Bioinformatics, 18, 63 (2017) · doi:10.1186/s12859-016-1442-6
[56] Henson, JM; Reise, SP; Kim, KH., Detecting mixtures from structural model differences using latent variable mixture modeling: a comparison of relative model fit statistics, Struct Equ Model, 14, 202-226 (2007)
[57] Lennon, H.; Kelly, S.; Sperrin, M., Framework to construct and interpret latent class trajectory modelling, BMJ Open, 8, e020683 (2018)
[58] Ram, N.; Grimm, KJ., Methods and measures: growth mixture modeling: a method for identifying differences in longitudinal change among unobserved groups, Int J Behav Dev, 33, 565-576 (2009)
[59] Kreuter, F.; Muthén, BO., Analyzing criminal trajectory profiles: bridging multilevel and group-based approaches using growth mixture modeling, J. Quant. Criminol (2008)
[60] Lubke, G.; Muthén, BO., Performance of factor mixture models as a function of model size, covariate effects, and class-specific parameters, Struct Equ Model A Multidiscip J, 14, 26-47 (2007) · doi:10.1207/s15328007sem1401_2
[61] Masyn, KE.; Little, TD, Oxford handb quant methods, 2, Latent class analysis and finite mixture modeling, 784 (2013)
[62] Dodge, HH; Shen, C.; Ganguli, M., Application of the pattern-mixture latent trajectory model in an epidemiological study with non-ignorable missingness, J Data Sci, 6, 247-259 (2008)
[63] Little, RJA., Pattern-mixture models for multivariate incomplete data, J Am Stat Assoc, 88, 125 (1993) · Zbl 0775.62134
[64] Bouguessa, M., A mixture model-based combination approach for outlier detection, Int J Artif Intell Tools, 23, 1460021 (2014) · doi:10.1142/S0218213014600215
[65] Guerra-Peña, K.; Steinley, D., Extracting spurious latent classes in growth mixture modeling with nonnormal errors, Educ Psychol Meas, 76, 933-953 (2016) · doi:10.1177/0013164416633735
[66] Depaoli, S.; Winter, SD; Lai, K., Implementing continuous non-normal skewed distributions in latent growth mixture modeling: an assessment of specification errors and class enumeration, Multivariate Behav Res, 54, 795-821 (2019) · doi:10.1080/00273171.2019.1593813
[67] Nam, Y.; Hong, S., Growth mixture modeling with nonnormal distributions: implications for data transformation, Educ Psychol Meas, 81, 698-727 (2021) · doi:10.1177/0013164420976773
[68] Elmer, J.; Jones, BL; Nagin, DS., Using the beta distribution in group-based trajectory models, BMC Med Res Methodol, 18, 152 (2018) · doi:10.1186/s12874-018-0620-9
[69] Nagin, DS; Tremblay, RE., Developmental trajectory groups: fact or fiction?, Criminology, 43, 873-904 (2005)
[70] Lubke, GH; Miller, PJ; Verhulst, B., A powerful phenotype for gene-finding studies derived from trajectory analyses of symptoms of anxiety and depression between age seven and 18, Am J Med Genet Part B Neuropsychiatr Genet, 171, 948-957 (2016) · doi:10.1002/ajmg.b.32375
[71] Hall, TO; Stanaway, IB; Carrell, DS, Unfolding of hidden white blood cell count phenotypes for gene discovery using latent class mixed modeling, Genes Immun, 20, 555-565 (2019)
[72] De Rubeis, V.; Andreacchi, AT; Sharpe, I., Group-based trajectory modeling of body mass index and body size over the life course: a scoping review, Obes Sci Pract, 7, 100-128 (2021) · doi:10.1002/osp4.456
[73] Kosmidis, I. brglm2: bias reduction in generalized linear models [Internet]. 2020. Available from: https://cran.r-project.org/package=brglm2.
[74] Wickham, H, François, R, Henry, L, et al. dplyr: a grammar of data manipulation [Internet]. 2020. Available from: https://cran.r-project.org/package=dplyr.
[75] Lüdecke, D., Ggeffects: tidy data frames of marginal effects from regression models, J Open Source Softw, 3, 772 (2018) · doi:10.21105/joss.00772
[76] Bates, D.; Mächler, M.; Bolker, B., Fitting linear mixed-effects models using lme4, J Stat Softw, 67 (2015)
[77] Hlavac, M. stargazer: well-formatted regression and summary statistics tables [Internet]. Bratislava, Slovakia: Central European Labour Studies Institute (CELSI); 2018. Available from: https://cran.r-project.org/package=stargazer.
[78] Garnier, S. viridis: default color maps from “matplotlib” [Internet]. 2018. Available from: https://cran.r-project.org/package=viridis.
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.