“Honestly, initially I didn’t think it would work,” he says. Anecdotes are considered unscientific; they’re “weak data,” in the lingo. “If one anecdote is unscientific, how can two thousand anecdotes become science? But I accepted the data at face value.
