I have no strict criteria and will generally !vote based on my overall feeling of a candidate. However, the criteria listed in the collapsed box are often used by myself to determine my overall feeling. It is not exhaustive or complete, and I may ignore my criteria if I so want to. I use these criteria to work out if I support a RfA or RfB candidate.

My criteria boil down to:

If I can trust that a candidate will use the tools to the benefit of the community, then I will support.
criteria
  • If the candidate can show / or has been shown to be trusted by the community, then I will generally support.[a]
  • If the candidate has worked to improve the encyclopedia, then I will generally support.[b]
  • If the user has a need for the tools, I will generally support.[c]
  • If the nominee has recent XfD nominations which were closed as keep and the reasons for the keep could have been reasonably seen by the candidate at the time of nomination, then I am more likely to oppose the candidate.[d]
  • Repeated issues with civility is a reason for me to oppose.[e]
  • Any socking,[f] or non-self requested blocks which have not been overturned (except the odd 3RR block)[g] will lead me to oppose.
  • If the nominee has had recently imposed community / ArbCom sanctions, or they have violated their active community / ArbCom sanctions recently, I will likely oppose. If the community / ArbCom sanctions have been adhered to, they will become less of a red flag. If the nominee seems to have moved on from the issues that lead to the community / ArbCom sanctions, this may lead me to discount said sanctions.[h]
  • If the candidate has been desysoped recently (unless the desysop was because of inactivity or because they resigned the tools not "under a cloud"),[i] or has applied unsuccessfully for administrative rights recently,[j] then this may make me oppose.
  1. ^ This can mean that if a user is receiving a large majority of supports (around above 90% usually), then I may use this to justify a support, as the candidate is being shown through the RfA to be trusted by the community. This won't be a direct reason for support, but is a useful starting ground for me. However, this does not mean that if I see a big issue I will ignore it and support the candidate; if the issue is big enough, I may still oppose.
  2. ^ This does not mean "articles created" et. al., but instead is if this user improved Wikipedia through doing things like writing articles, maintaining templates, helping new users, patrolling new changes, reviewing drafts, dealing with endless backlogs etc. Pretty much, as long as they are not WP:NOTHERE, they meet this criteria.
  3. ^ By this I mean, if the candidate wants to work administratively in their area of work / an area which is backlogged or can say why they need the tools, then as long as I find them trustworthy enough, this is another reason to support. Any user who cannot at least what area of Wikipedia they intend to work in is troubling for me, as it may indicate they are hat collecting or have no actual use for the tools. Any reason for needing the tools, as long as it is valid, meets this criteria for me; this includes viewing deleted content
  4. ^ Recent is not a fixed value; the more issues with a XfD nomination is directly proportional to the cut of point for what is recent. However, recent will usually not exceed 6 months and is usually only 1-2 months.
    I assume good faith until I feel I can't, so I won't count innocent mistakes.
    Some kind of acknowledgement and good faith efforts to not make the same mistake again, would mean I would not count that particular XfD.
  5. ^ Repeated civility issues are a problem because admins need to be civil. They are role models for the community, so being uncivil will only encourage others to be uncivil or lead newbies to believe being uncivil here is the "right thing". The odd minor comment or issue which has been apologised for I will probably ignore. If the incivility is from a long time ago, I will probably ignore it unless there are recent incivility issues.
  6. ^ This is a big red flag to me, but I understand that many sites don't require that you have one user account, so in one off incidents where I can assume good faith I will likely overlook this. Any more than once or when good faith cannot be assumed, use of illegitimate accounts is an almost instant reason for me to oppose.
  7. ^ As usual I would try to assume good faith, although longer blocks or blocks for reasons of copyright infringement, promotion, editing while not disclosing being WP:PAID would be for me enough grounds for opposing a candidate. Blocks that were a long time ago, and for which the user has shown they will not make the same mistakes / cause the same issues which lead to the block I will likely ignore.
  8. ^ I cannot trust a candidate who needs to have sanctions applied to themselves. However, even if the sanctions are still in place, it does not mean that they need the sanctions applied. Furthermore old, successfully appealed or small sanctions I may exclude. Any site bans, even if expired / successfully appealed, may still make me oppose for a while after the expiration / appeal date. This is because, unlike a topic ban or editing restriction, it has been shown that the issues leading to the site ban were bad enough / outweighed their good contributions. If the user can show that they have learnt from the mistakes they made which lead to the sanction, this will make me more likely to ignore the sanction
  9. ^ Adminstrators who have been desysoped, for reasons which are not inactivity or having resigned not "under a cloud", I cannot trust until they have shown they are now trustworthy. They need to deal with the issues leading to the desysop before reapplying and regain trust of the community.
  10. ^ This is due how the user needs to first address the issues presented by the opposers before reapplying. This should take time, and so going for RfA again too soon after a unsuccessful RfA does not give the candidate long enough to address the issues and so show they are now trustworthy enough.