Conversation

Replying to and
Ideally, I'd take a mixed approach, correlating statistical signals with narrative events that seem to be psychologically salient vs. not. There's too much noise in the raw data signal. You need to unconfound for eg. current event effects etc
1
1