groupedData, ordered factor and lme

[Karin <twentestat@yahoo.com>]

Hello,

 

I am analyzing some data with a three level structure (persons are measured during a few days). First I neglected the smallest level (day). And I constructed a groupedData object of the data with two levels and did some lme analysis with a random structure for the two levels.

Then I constructed a groupedData object with the smallest level included in it from the original data (not by updating the first GroupedData object but just constructing a new groupedData object).

I did the SAME analysis (in the analysis the lowest level is not modelled) with the GroupedData object with three levels.

 

I thought that I should get exactly the same results but this is not the case; the results are slightly different (okay only in the decimals). Could that be due to the fact that for the GroupedData object with 2 levels the grouping factor is converted in an ordered factor and for the GroupedData object with three levels this is not the case for the level1 grouping factor? Does lme perform the analysis use the ordering of the grouping factor?

 

Does anyone experienced the same or knows what is going on?

 

Thanks in advance,

 

Karin

---

Do you Yahoo!?
Yahoo! Mail Address AutoComplete - You start. We finish.