So I'm new to linear mixed models and know basic R but that's it. I understand the idea of mixed models conceptually but am having trouble with the syntax and interpretation.
I have one outcome measure (blood level) and am trying to see if they change over time (which I have in days, but can also make into a dichotomous baseline/followup variable). However, they were all tested at the end in batches on the testing machine which leads to some inter-batch variation. However, batches are unique to each participant-time combination (e.g. participant 1 may have had things tested on batch 2 at timepoint 1 but batch 3 at timepoint 2, while participant 2 may have had levels tested on batch 4 at both timepoints).
Thus, I'd like to model this using subject id as a random effect and the others as fixed effects.
The syntax might be:
lme.model <- lmer(bloodlevel ~ batch + time + (1 | ID), data = dataset, REML=T)
with time being the time after initial blood draw (at baseline it'd be 0 and followup would be like 90 days for example).
When I do: anova(lme.model) to get p values, the output is as follows:
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
batch 5.8020 5.8020 1 54.080 0.6147 0.4365
Time 4.5306 4.5306 1 52.394 0.4800 0.4915
I'm a little unclear how to interpret this. Does that mean that the p value for the difference in blood levels of my sample between the two timepoints was 0.4915? Or what is the correct way to model this?
I'm still learning this so please forgive my ignorance.
Edit: Please could somebody help me with this?
