Just to make sure, is this talking about something different from people committing less crimes when they are themselves practicing TM or in daily contact with someone who does?

I'm not sure if personal contact with practitioners has an effect, but the studies I'm thinking of were on the level of cities -- put a group of meditators in Chicago, the Chicago crime rate goes down.

What arm TM-Sidhis, are they something distinct from regular TM (are these different types of practicioners).

TM-Sidhis is a separate/additional practice that has TM as a dependency in the sense of package management. If you know TM, you can learn TM-Sidhis.

And what's with the sqrt(1%)? One in ten people in the total population need to be TM-Sidhis for the crime rate reduction effect to kick in?

Sorry, I meant sqrt(.01p) where p is the population of the group to be affected. For example, a city of one million people would require ten thousand TM meditators or 100 TM-Sidhis meditators.