Upgrade your account to a Community Supporter account and remove most of the site ads.

Reply to thread

Message: <blockquote data-quote="fusangite" data-source="post: 2635781" data-attributes="member: 7240">Some observations about how the results are correlating to different things:PostcountThe five candidates winning more than 200 votes all have post counts over 4000, whereas only two candidates behind the pack has a post count of over 4000. It should be noted, however, that these two candidates both have higher post counts than two of the five leading candidates. It does not appear that post counts higher than 4000 strongly correlate to placement within the top five candidates. The fifth place candidate, who currently holds about 35% of the vote has a post count of more than double that of the first place candidate who holds about 70%. Similarly, the second place candidate, who holds about 65% of the vote accounts for about 40% of all posts made by all candidates. I would argue, based on this data, that there is probably a floor post count that a candidate would likely have to achieve in order to be viable. I would set that number around 4000. However, once this floor is achieved, it is unclear that post count is a useful predictor of candidate placement. ReviewsOf the seventeen candidates, eight have published reviews. Of these eight, four occupy the top four positions, three occupy places 9-11 and one is currently dead last. Of the candidates winning more than 300 votes, every single one has published at least one review. There does, then, seem to be a strong correlation between having published at least one review and placing in the top five candidates. However, the number of reviews published does not seem to be a good predictor of placement once this number exceeds zero. For instance, the tenth place candidate has published the third largest number of reviews of all candidates. I would argue, based on the data, that reviews function much as post count does: exceeding a certain floor number is all but essential for winning a judging post. But whereas this number seems to be around 4000 for posts, it is 1 or 2 when it comes to reviews. And, as with posts, some highly prolific writers do not appear to benefit from producing these reviews.Posts Per MonthOf the leading five candidates, all but one exceed a posting rate of 150 per month and all exceed the posting rate of 100 per month. The fifth and sixth place candidates have the third and second highest posting rates, respectively exceeding 250 per month. Of the candidates not in the leading five, only two exceed a posting rate of 150 (or 100 for that matter) per month. I would suggest, then, that 150 posts per month is functioning as another "floor" for determining candidate viability. It is interesting to note, at this point, that two of the candidates who have the highest ratio of votes to posts per month are the two that are publicly identified as female. It is also noteworthy that posts per month, much more than reviews or total posts, appears to have a fairly direct correlation to number of votes in addition to constituting a floor.SeniorityExcepting the fourth place candidate, all of the leading five candidates were active on ENWorld prior to January 2002 while only a third of those outside of this group have this much seniority. There do not seem to be any strong correlations between join date and placement within the group who are currently losing.Nevertheless, I am tempted to see seniority as operating much like total postcount and reviews as establishing a mimimum criterion that either 80% or 100% of the leading candidates fulfill.ModerationAlthough last year's first place finisher was a moderator, the one moderator running this year is just a few votes above dead last place. I would suggest, therefore, that moderation, contrary to what we might have suspected based on last year's results, does not seem to be much of a factor.What does this tell us?In my view, I think we can learn a few things from this list from two perspectives. For those wishing to win a position as an ENNies judge next year, some things that are certain to help you: 1. Make 3-5 posts per day.2. Publish at least one or two reviews.3. Push your post count past 4000.4. Stick around.More importantly, from these results I think we can learn a bit about what subset of ENWorlders are voting. I would suggest that our average voter, not mobilized by the candidates has the following characteristics:1. Is interested in the forums and forms an impression of people who post consistently. 1a. Reads posts critically, voting both based on name recognition and post content.2. Is interested in the capacity of candidates to review products but can make this decision based on fairly few reviews.2a. Reads reviews critically, voting both based on name recognition and review content BUT may not be a consistent reader of the reviews section of the site.3. Has been active on ENWorld for some time.4. Likes female candidates better than males, all other things being equal.EDIT: I'm attaching the data I used in an Excel spreadsheet for those who want to play with it themselves, especially those (that means you Umbran) who can actually calculate correlation coefficients and the like.</blockquote>

Verification