Behavioural Science Section
Gerontology 2002;48:13–17
Correct Identification and Explanation
of Age-Related Changes Are Always
Possible
Peter C.M. Molenaar
University of Amsterdam, The Netherlands
Peter C.M. Molenaar
University of Amsterdam
Roetersstraat 15, Room 502
NL–1018 WB Amsterdam (The Netherlands)
E-Mail op_molenaar@macmail.psy.uva.nl
ABC
Fax + 41 61 306 12 34
E-Mail karger@karger.ch
www.karger.com
© 2002 S. Karger AG, Basel
0304–324X/02/0481–0013$18.50/0
Accessible online at:
www.karger.com/journals/ger
Key Words
Age-related changes, identification W Research design W
Longitudinal design
Abstract
In this commentary, it is argued that the definition of the
time or the age variable in the models presented by Ho-
fer and Sliwinski is ambiguous, and a standard interpre-
tation of this time of age variable is proposed instead.
Some simulation studies are presented, the results of
which indicate that, even with age-heterogeneous cross-
sectional data, correct identification of the true sources
of age-dependent variation is possible. An important
qualification, based on new methodological work in pro-
gress, is given of the scope of longitudinal designs. It is
concluded that all kinds of age-related changes are po-
tentially interesting and can be correctly identified and
explained.
Copyright © 2002 S. Karger AG, Basel
Introduction
I have read with great interest the paper by Hofer and
Sliwinski [1] (to which I will henceforth refer to as ‘the
authors’). It deals with some of the methodological issues
when using cross-sectional designs in order to investigate
age-dependent changes. In what follows I will present a
critical discussion of various aspects of the paper. Analyt-
ic discourse necessarily has a critical undertone. There-
fore, I would like to stress at the outset that the authors
deserve praise because they make a fundamental state-
ment and raise important questions about the informa-
tion which can be obtained from age-heterogeneous cross-
sectional data.
This commentary is organized as follows. First I will
shortly discuss the basic equation 1 given in the paper.
From that discussion I will draw some conclusions con-
cerning the correctness of the remaining equations. Sec-
ond, I will present some results from simulation studies
based on the basic equation 1 given in the paper. The
results of these simulation studies indicate that, even with
age-heterogeneous cross-sectional data, correct identifica-
tion can be made of the true sources of age-dependent
variation. I also will consider the remarks about partiall-
ing out of age made by the authors. Third, I will comment
on the statements made by the authors about longitudinal
analysis. These comments will include new methodologi-
cal developments that qualify the type of information
which can be obtained even from what is regarded as the
most powerful design to study within-subject variation,
namely the longitudinal design. In closing this commenta-
ry, some general remarks will be made concerning designs
with several homogeneous age groups and the relation
thereof with the age-homogeneous cross-sectional and
sequential designs favored by the authors. I end with a
general evaluation of the paper.