Dear All
I have to estimate a missing value in a crossover experiment (and even in swichback designs). I haven't found anything but the estimation of a missing value in a Latin square of at least 3x3 (and disparture of the formulas for bias in the SS for treatement from one book to another !!!). The formulas for the missing value (of course) don't work with 2x2 squares. How can I then estimate such a value that comes from a crossover or swichback design ?
Any hints will be much appreciated
Merry Christmas to all
Steve Méthot
Statisticien
Agriculture et agroalimentaire Canada
tel : (819) 565-9174 poste 262
fax : (819) 564-5507
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|