I finally found some data on the question of the family size of people who have served time in prison. (Reader Mark Wethman e-mailed me about this a few weeks ago). I suspect that the hope is that ex-cons have fewer kids, so that America's current practice of mass incarceration has an unintended eugenic effect.
The MIDUS Study asked people how many children they have, and if they have ever served time in prison. The problem is that researchers wanted to know how the number of biological, adopted, and step-children all added up. These days quite a few people have non-biological children, and I would expect ex-cons to have more because of greater relationship instability. Here are the means:
Mean number of children (sample size = 1,937, ages 45+)
Served time 2.71
Did not 2.62
Served time 2.79
Did not 2.70
Men and women who have been in prison have slightly higher means, but notice how the means are high for all groups. The numbers are inflated by non-biological children. I suspect that the means for ex-cons would drop more if we could pull out the stepkids, but there is little evidence here that prison reduces one's number of offspring.