I read that the default setting of lme4 is listwise deletion. My data is in long format (repeated measures with two time points) and it doesn’t appear that it really deletes listwise (as I understand it, this would mean that subjects with missing values on the dependent variable at either T0 or T3 are not taken into account in the analysis). I get this impression when I look at the ‘number of observations’ on the output. This number corresponds exactly to the number of my data points, but only if I take into account the observations of the subjects who only have data for either T0 or T3. Any help?

Here is the function:

model <- lmer(Dependent_variable ~ Group*Timepoint + Age + Centre + (1|Subject), Data_longformat)

The result when printing the model:

Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
Formula: BE_S ~ Gruppe * Zeitpunkt + Alter + Zentrum + (1 | Patient)
   Data: BES_alle_long_change

REML criterion at convergence: 622.7

Scaled residuals: 
     Min       1Q   Median       3Q      Max 
-1.82915 -0.57270  0.05619  0.59276  2.02843 

Random effects:
 Groups   Name        Variance Std.Dev.
 Patient  (Intercept) 0.4582   0.6769  
 Residual             0.6039   0.7771  
Number of obs: 216, groups:  Patient, 131

Counting the Subjects/Patients with Data for T0:

> result <- BES_alle_long_change %>%
+   filter(Zeitpunkt == "T0" & !is.na(BE_S)) %>%
+   summarize(count = n_distinct(Patient))
> print(result)
# A tibble: 1 × 1
  count
  <int>
1   119

Counting the Subjects/Patients with Data for T3:

result <- BES_alle_long_change %>%
+   filter(Zeitpunkt == "T3" & !is.na(BE_S)) %>%
+   summarize(count = n_distinct(Patient))
> 
> print(result)
# A tibble: 1 × 1
  count
  <int>
1    97