Analysis of data with pseudoreplicates

[Mutua]

Dear SAS community,

I performed two repeated experiments with potted plants in two climate chambers set at different day and night temperatures (20°C day and 12°C night temp for climate chamber 1) and (26°C day temp and 19°C temp for night temp in chamber two).  There were three treatments in each chamber which included one level of fungi infestation, one level of bacteria infestation and control plants (not infested), each replicated 20 times.

To assess the levels of damage on each plants, 5 plants from each treatment (making a total of 15 plants per climate chamber were harvest per week. Also above fresh plant weight was measured per pot.

The objective of the experiment was to evaluate the influence of temperature on disease severity and development on plants in two climate chambers.

Statistical questions:

1: should i treat the two climate chambers as pseudoreplicates? (i.e n=2)?

2. If yes to question 1, what is the best SAS procedure to handle this type of data? Proc GLIMMX?

Your support with this issue will be truly appreciated.

Best regards,

Peter

[SteveDenham]

I would not consider the climate chambers as pseudoreplicates.  This looks like a classic split-plot design--whole plot is chamber, sub-plot is treatment, with replicate measures within each subplot.  In any case, GLIMMIX or MIXED are indicated, and which is used all depends on the nature of the dependent variable.

Steve Denham

[Mutua]

Thanks Steve for your informative answer. I will run the analysis and if i have further question i will post it to the SAS support communities again.

Best regards,

Peter