Hello all,
I'm currently involved in a project, where we are comparing two reef fish monitoring methods, but we've had some problems with some of the data analysis.
The effect of different methodologies and researcher experience on the census results, were examined by a Mixed Model Analysis of Variance with 4 factors in the design. A random factor "SITE" (with 3 levels: Site 1, Site 2), and 3 fixed factors: Method, fish size estimates, and Observers (each with 2 levels). Satterthwaites method of calculation was used to determine the synthesis of the denominator errors for the degrees of freedom and for the mean squares (MS).
We ran this design through 3 statistical programs (Statistica 8, Minitab and SPSS) and with all of them we consistently got no answers (blank cases) for several of the interactions. Apparently the problem is that the denominator used to calculate the F value is cero (in those cases). We have have no clue what could be causing this behavior of the results, or how could we solve this problem !!???? More over, sometimes when we ran multiple comparisons where significant differences were detected, the tests (for example Tukey) showed no differences what so ever.
If anyone has any idea what the mistake might be, or how could we interpret these results, please don't hesitate to write me back at, [log in to unmask]
Best regards,
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|