How easy or otherwise would it be to calculate how many of those selfscratches are only included by that DT1 user in question, for the full list of DT1s, the same as you have done for Dabs?
The negative effects of Selfscratching could be mitigated by requiring 2 DT1 inclusions to be on DT2.
Seeing these data, I agree with this even more strongly than I did before. It's simply too much a conflict of interest for DT1 members to be able to artificially inflate their own trust scores so easily.