Hi Glen,
I've used the method described in the reference below (it's shorter than a
book!), which basically involves regressing the measure over time, to
produce a coefficient that describes the trajectory of the data for each
individual. You then do the stats on one outcome measure instead of
differences between discrete timepoints, which can sometimes tell a
different story. It's useful where assessments may be missed or don't
necessarily fall nicely into their allocated annual assessment slots (not
suggesting that other techniques can't also handle this, like multilevel
modelling for instance). It can also highlight subgroups that you might not
necessarily see in the visit data, such as those who decline at a faster
rate than most.
Hope that helps!
Brian Saxby
Matthews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial
measurements in medical research. British Medical Journal 1990; 300:230-235.
-----Original Message-----
From: Research of postgraduate psychologists.
[mailto:[log in to unmask]] On Behalf Of Glen Pennington
Sent: Tuesday, November 16, 2004 2:53 PM
To: [log in to unmask]
Subject: Longitudinal Data Analysis
Hello,
Can anyone recommend a good, easy to grasp book on longitudinal data
analysis?
Or alternatively offer any advice?
As a short overview I have 104 participants being tested on 8 working memory
measures and a single maths measure (then including a measure of IQ in the
next phase and subsequent one), they are being tested annually for three
years.
If you have direct/specific advice then please email me direct at
[log in to unmask] but book recommendations reply to the list in case
anyone else is having the same panic that I am right now.
Best wishes
glen
|