You know, it's wrong when any group makes crap up... or at the very least smudges the numbers.
Okay, the numbers are actually there but there are strong objections on two points, prisons and marriage.
The prison statistic is a self-declared note. Atheists make up 8-16% of US population and only 0.209% of the prison population. Well, those stats aren't completely wrong. In fact, a 2006 Harris poll could be read to clock atheist at 27% of the population. If 73% of the population believes in God, then the remaining 27% don't and are thusly atheist (even if only 11% say they don't believe in God). Now, this is how the study screws up, and how many studies screw up. Taking a look at the recent Pew study. We see that those unaffiliated people claiming atheism was 1.6% with agnosticism at 2.4% and nothing in particular at 12.1%. Largely almost everybody in that unaffiliated grouping is an atheist, in the true meaning of the word... however, about 10% of atheists actually admit to atheism. There is a massive gap between declared atheists and atheists. One of the side effects of being simply a lack of belief is that a person can be an atheist even if they don't know what it is or want to be one.
Now look at the prison stats again and consider for a moment that parole boards like religion, your fellow inmates are largely religious, religions provide protections in prison, and atheists are the most hated minority in the US. Now, take into account that many atheists are better educated and your average criminal is more poorly educated. Further, consider that blacks make up a larger percent of the prison population than the general population (and crime rate) and that blacks are largely more religious. On top of this, realize the entire fucking thing is asking you to assume that correlation implies causality! Maybe atheists are just better criminals and don't get caught? There are too many problems with this stat to give it a second thought. I would be very interested in a real set of statistics on this point. In the mean time, this stat should be disregarded with prejudice.
Marriage is another stat I wish people would stop tossing around. That stat is fully explained by the age of atheists. It turns out that the 18-24 range of population is quite atheistic. Looking at the statistics over time they aren't getting any more religious.
As the atheist population ages they don't seem to be finding religion (in fact, no religion is one of the largest religious groups and has some of the best retention of any of the groups). The 18 year old atheists of today will be the 25 year old atheists of 7 years from now (while padding the numbers as previous believers who fall away). This provides a rather bottom heavy group of atheists. So, when you ask a bunch of people their religion and whether or not they are divorced you find that the atheists haven't been divorced nearly as much as the other religious groups... because they haven't been married as long or lived as much. They are a bunch of kids! Adjusting for such seems to make that stat go away. Further, consider the increased educational attainment and income. Not to say it isn't true, just that the statistics don't seem bear this out.
It turns out that people who care to learn what is and isn't true do better in life in a plethora of ways and end up finding out that religion is bunk. Pretty much all the stats flow from this. There's nothing about atheism that makes people more moral (or less moral).
If I had to venture a guess, I daresay that there something about good moral people and the truth.