Readme file

SERIES C
Applied Statistics

E. Lesaffre and B. Spiessens: On the effect of the number of quadrature points in a logistic random-effects model: an example - Journal of the Royal Statistical Society, Applied Statistics, Series C, Volume 50 (2001) part 3, pages 325 - 335

The data have kindly been made available by Novoartis, Belgium. The source of the data must be acknowledged in any publication which uses them.

Data structure of C4827r.txt

Column Variable

1 identification number of the patient (idnr)
2 binary response (y)
3 treatment group (trt)
4 time of measurement (time)
5 visit number (visit)

Programs used :

GEE :
proc genmod data=C4827r;
class visit idnr;
model y = trt time time*trt / dist = b;
repeated subject=idnr/ withinsubject=visit type=unstr corrw;
run;

NLMIXED :
proc nlmixed data=onycho;
bounds sigma2 > 0;
parms beta0=-1.63 beta1=0.19 beta2=-0.40 beta3=-0.14 sigma2=16.5;
eta = beta0 + beta1*treat + beta2*time + beta3*timetr + b;
expeta = exp(eta);
p = expeta/(1+expeta);
model y ~ binary(p);
random b ~ normal(0,sigma2) subject=idnr;
run;

E. Lesaffre
Biostatistical Centre
Katholieke Universiteit Leuven
Universitaire Ziekenhuizen Sint-Rafaël
Capucijnenvoer 33
B-3000 Leuven
Belgium

E-mail: emmanuel.lesaffre@med.kuleuven.ac.be

 

Dataset (C4827r.txt - size, 34KB)

 Journals

SERIES A
Statistics in Society

SERIES B
Statistical Methodology

SERIES C
Applied Statistics

SERIES D
The Statistician