Hi

My question concerns both data structure (and subsequent the exact analysis model to use) and the implementation in the statistical software (SPSS).

Let’s first explain the data structure:
The experiment concerns a forest planted on a site formerly consisting of adjacent parcels of two different land use types. The forest consists of different tree species planted in blocks (several per species) covering both former land use types. Measurements (a value per tree) were taken in small plots (of +/- 5 trees) within the blocks. We have four plots per species-land use combination. We are interested in the effect of species and former land use type, though also we need to check the effect of the plot. My idea was to run a mixed model with ‘species’, ‘land use’ and their interaction as fixed factors, and ‘plot’ as a random factor nested within ‘species’. However, ‘plot’ could also be considered as being nested within ‘land use’? Or both?
Then, what is the best way to implement this in SPSS? Is there also a possibility to perform a post hoc test (e.g. Tukey) when using (linear) mixed models?

Hope my question is clear, many thanks in advance for your welcome suggestions.

mv