I'm trying to perform a mixed model with repeated measures, and I'm not sure if I'm doing it correctly, as the time periods are not included within my formula. I have 2 treatments occurring in 2 blocks, replicated over 2 years. I've taken measurements (eg pre and post) during the year. Concentration are the readings to be analyzed. My database is in long format, eg.
| Treatment | Year | Block | Time | Concentration |
|---|---|---|---|---|
| 1 | 2021 | 1 | Pre | |
| 1 | 2021 | 1 | Post | |
| 1 | 2021 | 2 | Pre | |
| 1 | 2021 | 2 | Post | |
| 1 | 2022 | 1 | Pre | |
| 1 | 2022 | 1 | Post | |
| 1 | 2022 | 2 | Pre | |
| 1 | 2022 | 2 | Post | |
| 2 | 2021 | 1 | Pre | |
| 2 | 2021 | 1 | Post | |
| 2 | 2021 | 2 | Pre | |
| 2 | 2021 | 2 | Post | |
| 2 | 2022 | 1 | Pre | |
| 2 | 2022 | 1 | Post | |
| 2 | 2022 | 2 | Pre | |
| 2 | 2022 | 2 | Post |
M3<-lme(Concentration~Treatment*Year, random=~1|Block, data=Interlong, method="REML")
I've put my treatment x year to determine if there are any interactions between the two. Within this formula I don't have a factor for time. Does it need to be included for a repeated measure? Or does the formula automatically put it in order without having to specify? Is that what "REML" is specifying already? I've also run into problems as to whether I should include CorAR1, but can't figure out the variables to include. I don't know if it's needed or not.
