Transitional modeling of experimental longitudinal data with missing values. (English) Zbl 1414.62013

de Rooij, Mark

Adv. Data Anal. Classif., ADAC 12, No. 1, 107-130 (2018).

Summary: Longitudinal categorical data are often collected using an experimental design where the interest is in the differential development of the treatment group compared to the control group. Such differential development is often assessed based on average growth curves but can also be based on transitions. For longitudinal multinomial data we describe a transitional methodology for the statistical analysis based on a distance model. Such a distance approach has two advantages compared to a multinomial regression model: (1) sparse data can be handled more efficiently; (2) a graphical representation of the model can be made to enhance interpretation. Within this approach it is possible to jointly model the observations and missing values by adding a new category to the response variable representing the missingness condition. This approach is investigated in a Monte Carlo simulation study. The results show this is a promising way to deal with missing data, although the mechanism is not yet completely understood in all cases. Finally, an empirical example is presented where the advantages of the modeling procedure are highlighted.

MSC:

62-07	Data analysis (statistics) (MSC2010)
62P25	Applications of statistics to social sciences
62H30	Classification and discrimination; cluster analysis (statistical aspects)

Keywords:

longitudinal data; missing values; multinomial data; multidimensional scaling; multinomial regression

References:

[1]	Agresti A (2002) Categorical data analysis. Wiley, New York · Zbl 1018.62002
[2]	Akaike H (1973) Information Theory as an extension of the maximum likelihood principle. In: Petrov BN, Csake F (eds) Second international symposium of information theory. Akademia Kiado, Budapest, pp 267-281
[3]	Albert, PS, A transitional model for longitudinal binary data subject to nonignorable missing data, Biometrics, 56, 602-608, (2000) · Zbl 1060.62572
[4]	Anderson DR (2008) Model based inference in the life sciences. Springer, New York · Zbl 1277.62002
[5]	Bartolucci F, Farcomeni A, Pennoni F (2013) Latent Markov models for longitudinal data. Chapman and Hall/CRC, Boca Raton · Zbl 1341.62002
[6]	Bonney, GE, Logistic regression for dependent binary observations, Biometrics, 43, 951-973, (1987) · Zbl 0707.62153
[7]	Cheng, G.; Yu, Z.; Huang, JZ, The cluster bootstrap consistency in generalized estimating equations, J Multivar Anal, 115, 33-47, (2013) · Zbl 1258.62057
[8]	Rooij, M., Ideal point discriminant analysis revisited with a special emphasis on visualization, Psychometrika, 74, 317-330, (2009) · Zbl 1243.62097
[9]	Rooij, M., Trend vector models for the analysis of change in continuous time for multiple groups, Comput Stat Data Anal, 53, 3209-3216, (2009) · Zbl 1453.62080
[10]	Rooij, M., Transitional ideal point models for longitudinal multinomial outcomes, Stat Model, 11, 115-135, (2011)
[11]	Rooij, M., An application of the mixed effects trend vector model to asymmetric change data with auxiliary variables, Behaviormetrika, 39, 75-90, (2012)
[12]	Rooij, M.; Schouteden, M., The mixed effects trend vector model, Multivar Behav Res, 47, 635-664, (2012)
[13]	Rooij, M.; Worku, HM, A warning concerning the estimation of multinomial logistic models with correlated responses in SAS, Comput Methods Programs Biomed, 107, 341-346, (2012)
[14]	Diggle PJ, Heagerty P, Liang K-Y, Zeger SL (2002) Analysis of longitudinal data. Oxford University Press, Oxford · Zbl 1268.62001
[15]	Donders, ART; Heijden, GJMG; Stijnen, T.; Moons, KGM, Review: a gentle introduction to imputation of missing values, J Clin Epidemiol, 59, 1087-1091, (2006)
[16]	Eliason SR (1993) Maximum likelihood estimation: theory and logic. Sage Publications, Newbury Park
[17]	Enders CG (2010) Applied missing data analysis. The Guilford Press, York
[18]	Fox, J.; Anderson, R., Effect displays for multinomial and proportional odds logit models, Sociol Methodol, 36, 225-255, (2006)
[19]	Greenland, S.; Finkle, WD, A critical look at methods for handling missing covariates in epidemiologic regression analysis, Am J Epidemiol, 142, 1255-1263, (1995)
[20]	Hedeker D, Gibbons RD (2006) Longitudinal Data Analysis. Wiley, Hoboken · Zbl 1136.62075
[21]	Jeličić, H.; Phelps, E.; Lerner, RM, Use of missing data methods in longitudinal studies: the persistence of bad practices in developmental psychology, Dev Psychol, 45, 1195-1199, (2009)
[22]	Liang, K-Y; Zeger, SL, Longitudinal data analysis using generalized linear models, Biometrika, 73, 13-22, (1986) · Zbl 0595.62110
[23]	Little RJ, Rubin DB (2002) Statistical analysis with missing data. Wiley, New York · Zbl 1011.62004
[24]	Meulman J (1982) Homogeneity analysis of incomplete data. DSWO Press, Leiden
[25]	Molenbergs G, Verbeke G (2005) Models for discrete longitudinal data. Springer, Berlin
[26]	Paik, MC, Repeated measurement analysis for nonnormal data in small samples, Commun Stat Simul Comput, 17, 1155-1171, (1988) · Zbl 0695.62172
[27]	Pan, W., Akaike’s information criterion in generalized estimating equations, Biometrics, 57, 120-125, (2001) · Zbl 1210.62099
[28]	Rubin, DB, Inference and missing data, Biometrika, 63, 581-592, (1976) · Zbl 0344.62034
[29]	Sherman, M.; Cessie, S., A comparison between bootstrap methods and generalized estimating equations for correlated outcomes in generalized linear models, Commun Stat Simul Comput, 26, 901-925, (1997) · Zbl 0901.62088
[30]	Takane, Y.; Bozdogan, H.; Shibayama, T., Ideal point discriminant analysis, Psychometrika, 52, 371-392, (1987) · Zbl 0624.62053
[31]	Van Buuren S (2012) Flexible imputation of missing data. Chapman & Hall/CRC Press, Boca Raton · Zbl 1256.62005
[32]	Yu, HT; Rooij, M., Model selection for the trend vector model, J Classif, 30, 338-369, (2013) · Zbl 1360.62408