There is a tradition (of which I do not approve)of analysing repeated
measures as if the data had been obtained from a split-plot experiment
(which they have not). It seems to be very popular in the psychometric
literature and less so with medical statisticians.
Does anybody know who first proposed this approach? I have heard it
associated with the name of Wiener but can this be the Norbert Wiener
(1894-1964) of cybernetics fame? It seems implausible. Does anybody know?
Stephen Senn
Department of Statistics
Glasgow
|