[R-sig-eco] Bivalve study

[Dixon, Philip M [STAT]]

Chris,

This is an example of a study with 2 factors that is much easier to analyze as a 1 way ANOVA.  You  have three treatments.   The structure of those treatments implies two interesting treatment contrasts:
the effect of raking, estimated by the difference between control and procedural control, and
the effect of fishing in disturbed plots, estimated by the difference between fished and procedural control.

Because these follow from the treatment structure, I would do inference about those differences (tests, confidence intervals) without any adjustment for multiple comparisons.

You could also set up the problem as a regression problem with two indicator variables (the raking and fishing variables in your treatment table).  I think it is more straight forward to think about means of treatments.

A 2 way factorial with interactions will run into serious problems because you do not have all 4 cells of the raking x fishing table.

Best wishes,
Philip Dixon