Multilevel Data
Reading material: Hedeker, D., Gibbons, R.D., & Flay, B.R. (1994). Random-effects regression models for clustered data with an example from smoking prevention research. Journal of Consulting and Clinical Psychology, 62, 757-765. (pdf file)
Slides: Multilevel Analysis: An Applied Introduction (pdf file)
Example using SAS PROC MIXED:
TVSFPMIX.SAS - SAS code for analysis
of TVSFP dataset using a few different MIXED models. Also includes
individual-level and aggregate-level analyses.
TVSFP2B.DAT -
ASCII data file for example above.
Example using SPSS MIXED:
TVSFPC.SPS - SPSS code for analysis
of TVSFP dataset using a few different MIXED models.
TVSFPC.SAV –
SPSS .SAV file for the example above.
Longitudinal Continuous Data
Reading material: Hedeker, D. (2004).
An introduction to growth modeling. In D. Kaplan (Ed.), Quantitative Methodology for the Social
Sciences.
Slides: Introduction to Mixed Models for Longitudinal Data for Longitudinal Continuous Data (pdf file)
Examples using SAS PROC MIXED:
1. Analysis of Riesby dataset. This example has a few different PROC MIXED
specifications, and includes a grouping variable and curvilinear effect of
time. (SAS code and output)
2. This handout shows how empirical Bayes
estimates can be output to a dataset in order to calculate estimated individual
scores at all timepoints. (SAS code and output)
3. This handout has the analysis
considering the time-varying drug plasma levels, separating the within-subjects
from the between-subjects effects for these time-varying covariates. (SAS code and output)
Datasets:
Riesby dataset – for examples 1 and 2, the
variable order and names are indicated in the above syntax files.
Riesby dataset with time-varying covariates – for example 3, the variable order and names are indicated in the above syntax.
Examples using SPSS MIXED:
1. Analysis of Riesby dataset. This example has a few different MIXED
specifications, and includes a grouping variable and curvilinear effect of
time. It also shows how to get plots of
the empirical Bayes estimates. (SPSS code)
2. This handout has the analysis
considering the time-varying drug plasma levels, separating the within-subjects
from the between-subjects effects for these time-varying covariates. (SPSS code)
Datasets:
Riesby dataset – a SPSS .SAV file - for example 1.
Riesby dataset with time-varying covariates – a SPSS .SAV file - for example 2.
Missing Values in Longitudinal Data
Reading material: Hedeker, D., & Gibbons, R.D. (1997). Application of random-effects pattern-mixture models for missing data in longitudinal studies. Psychological Methods, 2, 64-78. (pdf file)
Slides: Mixed Pattern-Mixture and Selection Models for Missing Data (pdf file)
Examples using SAS PROC MIXED:
schizpm.sas - SAS code for pattern-mixture model analysis
of NIMH Schizophrenia dataset.
schizsel.sas - SAS code for shared parameter
(selection) model analysis of NIMH Schizophrenia dataset.
SCHIZREP.DAT
- ASCII datafile for examples above.
Example using SPSS MIXED:
schizpm.sps
- SPSS code for pattern-mixture model analysis of NIMH Schizophrenia dataset.
SCHIZREP.SAV – SPSS .SAV file for example above.
Longitudinal Dichotomous Data
Reading
material: Hedeker, D. (2005). Generalized linear mixed
models. In B. Everitt
& D. Howell (Eds.),
Encyclopedia of Statistics in Behavioral Science. Wiley,
Slides: Mixed Models for Longitudinal Dichotomous Data (pdf file)
Example using SAS:
Analysis of the NIMH Schizophrenia dataset.
This handout provides SAS (PROC LOGISTIC, NLMIXED) code for running
ordinary logistic regression and mixed-effects logistic regression. (SAS code)
Dataset: SCHIZ dataset - the variable order and names are indicated in the example above.
This handout shows how to use PROC
IML to get the marginalized probability estimates from the above NLMIXED
analysis for the random intercept model (SAS code)
This handout shows how to use PROC
IML to get the marginalized probability estimates from the NLMIXED analysis for
the random trend model (SAS code)
GEE analysis of the NIMH Schizophrenia dataset using SAS PROC GENMOD (SAS code)
Longitudinal Ordinal and Nominal Data
Reading
material: Hedeker, D. (in
press). Multilevel models for ordinal and
nominal variables. In
J. de Leeuw & E. Meijer
(Eds.), Handbook of Multilevel Analysis.
Springer,
Slides: Mixed Proportional Odds Models for Longitudinal Ordinal Data (pdf file)
Slides: Mixed Non-Proportional Odds Models for Longitudinal Ordinal and Nominal Data (pdf file)
Examples using SAS PROC NLMIXED:
schzonl.sas - SAS code for mixed-effects proportional
odds regression analysis of NIMH Schizophrenia data. schizx1.dat - ASCII datafile.
sandonl.sas
- SAS code for mixed-effects proportional odds and non-proportional odds
analyses of
sandnnl.sas
- SAS code for mixed-effects multinomial analyses of
sdhouse.dat
- ASCII datafile for examples above.
Sample Size Estimation for Longitudinal Studies
Reading material: Hedeker, D., Gibbons, R.D., & Waternaux, C. (1999). Sample size estimation for longitudinal designs with attrition: comparing time-related contrasts between two groups. Journal of Educational and Behavioral Statistics, 24:70-93. (pdf file)
Slides: (pdf file)
RMASS2.EXE
contains:
- executable program for sample size determination based on this paper.
RMASS2.PDF
contains:
- PDF version of program guide
More information and materials:
Don's short course on Longitudinal Data Analysis
Don's 15-week course on Longitudinal Data Analysis
Any questions or comments to Don: hedeker@uic.edu