Presuasion

Rigorous and useful about attention, leaning hard on priming research the replication crisis has since made smaller.

Presuasion, Robert Cialdini, 2016.

A rigorous, useful book about attention that leans hard on priming research from before the replication crisis reshaped how much weight the field puts on subtle unconscious cues, and it never once revisits a study in that light. Assume some of these effects are smaller in the wild than in the chapter that reports them.

Cialdini argues that what happens in the moment just before a message, not the message itself, decides much of whether it persuades, because attention is a limited resource and whatever occupies it right before a choice gets treated as relevant to that choice. A large number, a warm cup, a photo of a runner winning a race, any of these can prime a decision without ever entering the argument for it. He is less careful distinguishing primes that work because they are genuinely informative, a two sided satisfaction question, from ones that work through pure unconscious contamination, and the book mostly treats both as the same craft.

Two things kept. Attention is not neutral; whatever is made prominent right before a decision gets treated as important to that decision, whether or not it actually is. And a two sided question, how satisfied or dissatisfied were you, gets more honest data than a one sided one, because the one sided version primes only half of the possible answer.

I write the intake questions for a client feedback form at work, four questions, sent after every project closes. Six months ago I switched every yes or no satisfaction line to a two sided version. Response detail went up; the free text field averages twenty two words now against nine before the change, measured across the last thirty one responses. It has not moved the actual satisfaction scores, which sit almost exactly where they did before, so whatever the framing changed, in this case it changed how much people were willing to say, not what they actually felt.