Government physical mail has the benefit that substantial tampering is way harder to do at scale.

It's the same vein as criminals using cash vs Bitcoin; both can hide crime, but one is way easier to scale up.

It's also wild just how cost-effective interventions like this can be. You can pay a thousand here and there, or a few hundred thousand incarcerating these people when they turn to crime out of desperation.

I don't think the parent is even saying that, their point is pretty reasonable: having some objective measure for before and after in any study is more reliable than self-reporting, especially when the subject might be incentivized to lie.

The self reports might be totally true, but the study isn't as good as it might be.

Could you elaborate? The project structure looks extremely normal to me, but I don't know if I'm overlooking red flags all over the place.

This seems to have the positive effect that patching applications on your own device (a la Revanced patching Spotify) appears blessed, since government prosecution would need to demonstrate a public interest case, if I'm reading this correctly.