I don't think different behaviours between the two groups can be discounted. If unvaccinated has to be tested to participate in social activities, they'll get

I don't think different behaviours between the two groups can be discounted. If unvaccinated has to be tested to participate in social activities, they'll get tested more often and isolate if they get a positive test preventing spread to others in their group (this assumes that unvaccinated group together somewhat, which I don't have any evidence for besides anecdotally) Being vaccinated might also lead to risky behaviour disproportionate to the risk mitigation provided by the vaccine - it might be a consequence of risk homeostasis, but the risk has now increased, which skews the bias towards more infections in the vaccinated group.