Read on Twitter
Stop press! We have a new working paper, experimentally comparing In-person with Virtual coaching of South African teachers. https://tinyurl.com/yy2d8x4k . This is important evidence for any org/gov planning to provide virtual professional development for teachers during the pandemic.
Teachers in both programs received the same upfront training and learning aids, and followed the same lesson plans. The focus was on teaching English as a Second Language. The key differences: In-person vs virtual coaching and (ii) paper vs tablet-based lesson plan.
After the 1st year (end of grade 1), both programs were equally effective at increasing students' English oral language skills: https://tinyurl.com/y6bvt8vs . These gains were still seen after 3 years of implementation, although much larger for In-person (0.31 SD) vs Virtual (0.13 SD).
After 3 years of exposure (end of grade 3), In-person coaching improved English *reading* skills (0.12 SD), but virtual coaching had no impact. Moreover, virtual coaching had negative spillover effects on home language literacy, possibly due to crowding out of teaching time.
Why the large difference? The quality of implementation for both programs was high; and almost all teachers in virtual arm accessed the lesson plans on the tablets at least one, but usage was variable. So teacher motivation, not technological barriers, was the binding constraint.
We find evidence that teachers in the In-person arm were held more accountable by the coach, and were more likely to rely on her for support.
Qualitative research suggest two other mechanisms: (i) observing teaching in the classroom, to identify and address specific teaching challenges; (ii) building a trusting relationship, so a teacher is willing to be vulnerable and acknowledge areas that need improvement
What can be done to make virtual coaching effective? (1.) Allow some face-to-face contact to allow a relationship to develop. (2.) Enable virtual classroom observations. (Note: the virtual coach encouraged teachers to send videos, but not all did)
More generally, these results point to strong complementarities between tech interventions and incentives.
This is joint work with @stephengstaylor , Fleisch, Mohohlwane, Tshego, and @JaneliViljoen, who will be discussing it at the @riseprogramme conference on 7/30 4pm GMT. You can watch a presentation here: https://tinyurl.com/y42bztwg . That's all!
