I heard back from Ian Walker. He didn't address my question about the moth effect directly. Here's what he said:
This question of blinking v static lights is something that my colleague Tim Gamble and I used to talk about frequently, and we kept meaning to do some research on the topic but never quite found the time.
My suspicion (based partly on some knowledge of the physiology of the visual system) is that the two types of light will do different jobs: I suspect stationary lights are best for drivers knowing exactly where a cyclist is and how they are moving, and flashing lights are best for attracting attention to the cyclist in the first place (and, perhaps, for signalling the message "this is a cyclist").
The implication could be that the best would be the type of lights that do both: e.g., the lights that are always on but also pulse or flash. Or having one of each.
As I say, those are my (very) strong suspicions, but we never managed to do the actual testing, even though we did have a very fun experiment planned using flashing tennis balls and an automated dog-ball launcher!