On Thu, 19 Aug 2004, Hem Nalini Morzaria-Luna wrote:
> Hi,
> I am running a ANOVA on a factorial design, and using TukeyHSD
> for post hoc comparisons. I have 2 factors with three levels each:
> Factor B
> Factor A 1 2 3
> 1
> 2
> 3
>
>
> When I look at the Tukey output (on the interaction of the factors) the
> comparisons come out numbered 1-36.
Rather, they come out as unlabelled and this is a print of a matrix
without rownames.
> e.g.
> $"A:B"
> diff lwr upr
> [1,] 49.1666667 -160.041022 258.3744
> [2,] 50.0000000 -159.207689 259.2077
> ....
>
> [35,] 84.3333333 -124.874355 293.5410
> [36,] 57.6666667 -151.541022 266.8744
>
> Does anyone know how the function numbers the comparisons
It doesn't. However, this is standard R order, that is down columns
(1:1, 2:1, 3:1 etc).
> or an easy way of figuring out the order without having to go back and
> calculate the means between each pair and then going back to compare
> them with the output?
It should be easy to add labels for the next release.
--
Brian D. Ripley, ripley at stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595