There are a couple of factors to consider.
(While I use "I" in the following examples, I do not always do this)
First, in any community there are sub groups within it. When I see a question that has been modified by certain users, I look to see what that person said in the question (assuming they answered it rather than editing the question). If the question has a dozen answers in it, I look to see that person's contribution and vote on that question, leaving other questions unvoted. This will have the effect of boosting a single user's answer score while leaving others the same.
Furthermore, people with high reputations have likely learned how to write for the site. Each target media has a different style of writing that fits it best. One writes for Stack Exchange different than Wikipedia, which is different than Everything2, which is different than a blog post, which is different than a news paper article, which is different than a glossy magazine article, etc... Certain styles of writing will be more effective in attracting people to read it, and vote on it. It is a learned skill.
Personally, I am sometimes amazed at the more opinionated nearly content free answers that I have that have received more votes than the boring content rich answers that languish - but that's because they are better written (and more interesting) as something to read. That doesn't mean write content free things (unless its a grammar).
Lastly, as Jimmy mentioned in the comment, there is the 'mob mentality' of following the general trend of the votes. If something is upvoted, people tend to upvote it. If something is downvoted, people tend to down vote it (though that is less so because of the associated cost for a downvote). Not everyone spends the time reading all the answers, especially when there are lots of them it becomes more intimidating to read past the second or third one. Thus, the top one gets an upvote, the second one may get one if it is interesting and useful and then it starts petering out. Having a 'boring' higher voted question may be even more impacting on questions that are scored lower than it as its boring to read.
(Related to this - a techcrunch article The Science Of Reddit: Why Some Ideas Dominate The Net which looks into how initial votes on a question or answer may influence the following votes - the actual paper and abstract can be seen at Social Influence Bias: A Randomized Experiment)
Sometimes, there is a subtle difference between two answers that is the key to one being the 'right' answer. That one answer picks up a specific point that another does not include. This may not always be the something the person who answers it picks up on (the '-ers' being an indication of procedural code being asked about, or one of my own where I pointed out the dry being used as a justification for clever code - I was expecting that question to sit at the bottom with a few votes and it got accepted and now has the most).