One interesting position that Ruth Clark takes is that written text probably makes more sense in elearning than audio narration when the audience is composed primarily of non-native speakers. Her argument is that control over pacing is so important here that on-screen text wins out over audio.
I was a little surprised by this, as I would have assumed that with on-screen text it is hard to avoid split-attention effects. Clark cites no sources, and in a quick swing through the literature I can’t find any empirical comparisons of comprehension of reading versus listening in a non-native language. Interesting assertion, and one I’ll have to think about further as we begin to develop elearning for international audiences at McGladrey.