Currently, moderator elections on Stack Exchange sites have three phases: nomination, primary and election. The primary phase is skipped if there are 10 or less candidates.
The purpose of a primary is to whittle down the field of suitable candidates to a manageable number. But I'd argue that the primary phase is not necessary for elections on SE sites for several reasons:
- We already have some reasonable ways to limit the field to serious candidates. The reputation requirement alone was not sufficient, but combined with the badge requirements it is currently sufficient to limit the number of nominations. The last election had 18 candidates, while the first 2011 election (whithout badge requirements) had 55.
- The STV voting system means that voting for a candidate that has no real chance of being elected won't waste my vote in most cases, it'll just be transferred to my second or third choice. The primary helps to sort out candidates that have no chance of being elected, but with the STV system this probably makes no difference after all
The disadvantages of the primary phase are:
- It prolongs the election considerably, and might lead to some election fatigue. I have no hard data to prove it, but I'd guess that there is a small but significant number of users that take notice of the early election phases, but forget or don't care to vote in the end.
- It might confuse a few users, they might think that by voting in the primary they already cast their real vote. Primaries are very familiar to US citizens, they are uncommon in some other countries and non-native English speakers might not automatically understand what a primary actually is.
It would be interesting to know how many users voted in the primary, but didn't vote in the main election. If that is a significant number it would be a strong argument for eliminating the primary.
But apart from that, even if the confusion the primary causes is minimal, I don't see the big advantage of having a primary phase. I might certainly miss something here, so I'd appreciate any additional arguments in favor of the primary.
Thanks to Shog9 there is now some evidence that the primary indeed causes some confusion. In the latest Math election, 437 users voted in the election, but an additional 251 voted only in the primary.