It is quite a big difference, but look at this one.
Avg (all users): 5.8732
Avg (commenters): 6.2500
Avg (participants): 5.3846 Avg (non-participants): 6.1556 0.771 difference
I think these are much more interesting stats than the "old" ones.
The sample sizes can be the cause of this difference. Based on your numbers, 34 voters were non participants and 114 were participants. With a sample size of only 34, the average can be greatly skewed by a small number of impressed voters. In other words, had just one of the 34 voters switched from a 10 to a 1, your non-participants average would drop to 6.4412.
Also, looking at the histogram, you have a nice normal distribution of votes which indicates fairness. Had there been a nefarious attempt to vote lower, you would see a bimodal distribution.
I believe it is very hard to not be extra critical while running in the same challenge. It is why I try to avoid voting in the challenges I enter. Even if I want to be 100% partial I don't think I can be because I'm human. I think those scores are showing the same thing.
