Jos Twisk1, Frank Rijmen. 1. Department of Clinical Epidemiology and Biostatistics, VU Medical Centre, Amsterdam, The Netherlands. jwr.twisk@vumc.nl
Abstract
BACKGROUND: In many epidemiologic longitudinal studies, the outcome variable has floor or ceiling effects. Although it is not correct, these variables are often treated as normally distributed continuous variables. OBJECTIVES: In this article, the performance of a relatively new statistical technique, longitudinal tobit analysis, is compared with a classical longitudinal data analysis technique (i.e., linear mixed models). STUDY DESIGN AND SETTING: The analyses are performed on an example data set from rehabilitation research in which the outcome variable of interest (the Barthel index measured at on average 16.3 times) has typical floor and ceiling effects. For both the longitudinal tobit analysis and the linear mixed models an analysis with both a random intercept and a random slope were performed. RESULTS: Based on model fit parameters, plots of the residuals and the mean of the squared residuals, the longitudinal tobit analysis with both a random intercept and a random slope performed best. In the tobit models, the estimation of the development over time revealed a steeper development compared with the linear mixed models. CONCLUSION: Although there are some computational difficulties, longitudinal tobit analysis provides a very nice solution for the longitudinal analysis of outcome variables with floor or ceiling effects.
BACKGROUND: In many epidemiologic longitudinal studies, the outcome variable has floor or ceiling effects. Although it is not correct, these variables are often treated as normally distributed continuous variables. OBJECTIVES: In this article, the performance of a relatively new statistical technique, longitudinal tobit analysis, is compared with a classical longitudinal data analysis technique (i.e., linear mixed models). STUDY DESIGN AND SETTING: The analyses are performed on an example data set from rehabilitation research in which the outcome variable of interest (the Barthel index measured at on average 16.3 times) has typical floor and ceiling effects. For both the longitudinal tobit analysis and the linear mixed models an analysis with both a random intercept and a random slope were performed. RESULTS: Based on model fit parameters, plots of the residuals and the mean of the squared residuals, the longitudinal tobit analysis with both a random intercept and a random slope performed best. In the tobit models, the estimation of the development over time revealed a steeper development compared with the linear mixed models. CONCLUSION: Although there are some computational difficulties, longitudinal tobit analysis provides a very nice solution for the longitudinal analysis of outcome variables with floor or ceiling effects.
Authors: Ramit Ravona-Springer; Erin Moshier; James Schmeidler; James Godbold; Jimmy Akrivos; Michael Rapp; Hillel T Grossman; Michael Wysocki; Jeremy M Silverman; Vahram Haroutunian; Michal Schnaider Beeri Journal: J Alzheimers Dis Date: 2012 Impact factor: 4.472
Authors: Thijs T Jansz; Franka E van Reekum; Akin Özyilmaz; Pim A de Jong; Franciscus T J Boereboom; Tiny Hoekstra; Marianne C Verhaar; Brigit C van Jaarsveld Journal: Am J Nephrol Date: 2018-11-13 Impact factor: 3.754
Authors: Peter B Østergren; Caroline Kistorp; Mikkel Fode; James Henderson; Finn N Bennedbæk; Jens Faber; Jens Sønksen Journal: J Urol Date: 2016-12-09 Impact factor: 7.450
Authors: Michael R Dolsen; Adriane M Soehner; Charles M Morin; Lynda Bélanger; Matthew Walker; Allison G Harvey Journal: J Consult Clin Psychol Date: 2017-04-10