Part of the problem with +/- user moderation is that it is often dependent sequence. The digg system hides negatively ranked comments after a certain number (like -5 or something) which can cause a user glaze over them. Comments can even be glazed over if they have a visual indication that they’re considered “bad” but aren’t hidden (for example, the title is grayed-out or turned red).
If the initial 5 people who see a comment don’t like it they can unfairly influence the subsequent possible votes. The same is true for positive votes but on a lesser scale (since they usually aren’t visually differentiated beyond the display of their score). I’ve talked to a handful of friends who frequent digg and they agree that when they read comments they scan the digg ranking of the comment as they scroll down. Now I know that isn’t exactly scientific but I think its relevant.
What might work is a silent, timed variation of digg/bury. Users could initially mod answers up or down but the results are never shown. After a given period of time, each ranking takes effect. The only noticeable difference would be that answers below the negative threshold would simply not appear (or maybe that would be optional).
Part of the reason I like fluther is that there isn’t this black and white control mechanism throughout, like digg. Marking an answer as “great” doesn’t necessarily imply that you agree with it. While it would be valid to argue that only being able “great” is akin to only being able to digg a comment up, it has a different contextual meaning. Maybe then it would be an idea to assign different attributes to an answer beyond just “great”. Or possibly breaking down the components of what makes an answer great and using them (eg. “informative”, “funny”, “poetic”, “logical” or whatever).