Meta-regression in metafor - subgroup differences

Question

I have performed a meta-regression following the example found in here. This is the code that I'm using adapted to my dataset:

fit meta-regression model to test for subgroup differences

resMeta <- rma(xi=nphy, ti=ni, mods = ~ pop, data=metaAAS)

'metaAAS' holds the data from a csv file, and 'pop' is the moderator. The problem is that 'pop' has 4 levels, described as 0,1,2 and 3, and the output for the test of moderators (QM) is showing only one degree of freedom, instead of 3 (number of levels - 1):

Test of Moderators (coefficient 2): QM(df = 1) = 8.7150, p-val = 0.0032

What am I doing wrong?

Thank you!