Hello,
I agree with what Sage said about limited success in recruitment in any event targeting new users. My experience is that the fastest way to recruit new regular contributors is for them to enter a one on one mentoring and collaborative relationship with an experienced Wikipedian. They can enter this relationship either by being found and invited after they are identified doing something on Wikimedia projects, or by proposing a partnership with them when they themselves come to Wikimedia projects looking for a partnership. There are a lot of potential new collaborators to be found and developed in this way, but processing them takes either a tremendous amount of work from a single Wikipedian acting alone or a communications manager to create a list of contacts and a plan for staying in touch with them over time. In my opinion, anyone who took a communications role to manage contacts in this way should be paid staff, because the responsibility and workload are great. By default, most people seeking to become more engaged are never invited to do so and become lost contacts. I regret this loss because it is mostly a failure to connect community volunteers with people seeking mentoring, and not a lack of mentors. Any Wikimedia community group which could begin to track and manage requests for support would be building a valuable contact list.
Editathons and public outreach events can deepen the engagement of established Wikipedians but have a low rate of conversation from new users to engaged users. The usual figure passed around is 2-5% conversion, and I think those numbers are generous because those figures typically mean something like "makes 5 edits three months after the event" which is not worth noting. Editathons can be good for getting content onto Wiki (rarely - it usually does not work),, converting the organizer of the editathon from a newer user to a more engaged Wikipedian, or soliciting geographically local contacts for the one on one mentorship which I just described above.
I am a fan of community outreach models which minimize the work burden on the volunteers who organize them. Something manageable for a small group of people, in my opinion, is a regular monthly meetup which can be replaced with a special event any time someone steps up to request to organize one. So for example, if some institution has a staffperson who wants to engage more deeply, they can be primed to host an editathon. That editathon becomes the event for the month, and that staffperson gets more engaged to join the meetup team, but it is hard to expect any new user who joins the editathon to convert to become an engaged user, and the chances of actually having stable useful content contributed at a typical editathon by new users is rather low too without thoughtful experienced planning.
I like Pine's idea of having events to train multiple new editors at once but I do not recall seeing that work in practice anywhere ever. I have seen a lot of people try. Institutional partnerships work when someone inside the institution becomes a Wikipedian and converts their colleagues, and I still think that is a fruitful outreach path.
Everything that Peaceray says is a viable agenda for grooming potential new contributors assuming that they have already committed to learn, and are a step beyond casual exploration. He is especially correct about the breakdown of interests - some people will only care about some topical field and will contribute excellent information in that field, but not want to learn more than necessary about Wikimedia infrastructure, while a different demographic will be more interested in the infrastructure and technical execution than any topical focus.
yours,