It's been a busy few months for instructor training: along with a two-day class at the University of Manchester, we also wrapped up a pilot of a new kind of multi-week course and a two-day version of the same material. We have also been training four new trainers, and have put the first complete draft of the training materials online. Here's some of what we think we've learned so far.
The multi-week online class met every week instead of every second week (as it had in the past). 54 people started with us in October, and 45 stayed with us until the end, which is the highest completion rate we've ever had for an online class this size. We used Piazza to manage homework and discussion; it's a nice tool, but wasn't noticeably better for our purposes than the shared WordPress blog we used to use. (Both are much easier to manage than the GitHub-based blog we used in the winter, though.) We used BlueJeans for the hour-long meetings, which worked as well as web conferencing ever does.
For the two-day version of the pilot we had people apply in groups. One reason was so that each site would have enough learners to do the video recorded teaching exercise that has become such an important part of the in-person course. Another was so that there'd be critical mass to run workshops after the training: participants have said in the past that this is one of the big challenges. Nine groups took part in total:
Each team used one Google Hangout; there were a few glitches, but it was mostly reliable. (A DDoS attack on UK academic networks during the training didn't help matters, but Team 2 managed nonetheless.)
Around the same time (23-24 November) we also run a face-to-face instructor training at the University of Manchester. This session was supported by ELIXIR UK and the places for trainees were offered to members of the ELIXIR project, which meant that we had a very international group present. This did not make a significant difference to the training itself, but it was great for instructors in training to meet their counterparts from other countries and organisations.
Having the whole group in one physical location means that there is a lot of discussion and interaction, but the trainer still has to facilitate that, especially at the beginning. It also means that for the practical exercises the trainers can ensure that people swap the partners and groups which increases interaction and knowledge exchange.
Overall, we think that the two-day versions worked as well as the multi-week version, and were less effort for both instructors and learners. A few of the learners in the two-day version may have been less enthusiastic than usual (we think they may have been roped in by friends to get numbers up for applications), but on the other hand, all but a couple of the people who started the two-day version finished it. There wasn't much inter-group discussion—sessions tended to be hub-and-spoke with the trainer as the hub—but there seemed to be plenty of conversation within each group. The biggest drawback of the two-day version was that instructors couldn't give the same quality of feedback on exercises as in either the live version or the multi-week version: there wasn't as much time to do so as in the latter, and there wasn't the hands-on aspect of the former (particularly for the live coding exercise). On the other hand, the online live course made a lot more use of the Etherpad, which meant there were good collaboratively-produced notes at the end.
We won't be able to compare the two pilots fully until participants have finished qualifying and started teaching, but until then, we'd be grateful for feedback from people who took part.
This post originally appeared in the Software Carpentry blog.