I'm involved in analysing data from a toxicological experiment where
there are four groups, i.e. no-dose control and three log-spaced doses.
Standard procedure at the lab that does the testing is to do an ANOVA
for differences between the means, followed by Dunnett's procedure for
comparing (three) treatment means with the control. Group sizes are
decent (~70) and variances homogeneous.
The use of Dunnett's here worries me, since its alternative hypothesis
has no ordering. Since the scientists and policy regulators are
interested in how big a difference is significantly different from
control, the use of Dunnett to compare the lowest dose with control
seems to me greatly over-conservative.
Can anyone point me in the direction of documented procedures or
references e.g. similar to Duncan's multiple-range test, but for a
priori ordering of treatments?
As always, replies to me please and I'll summarise to the list
Brian Miller
Brian G Miller, BSc, PhD, CStat
Director of Research Operations
Direct Line +44 (0) 131 449 8044
e-mail [log in to unmask]
Institute of Occupational Medicine
Research Avenue North
Riccarton
EDINBURGH EH14 4AP
Scotland, UK
Tel +44 (0) 870 850 5131
Fax +44 (0) 870 850 5132
visit the IOM at www.iom-world.org <http://www.iom-world.org/>
|