But you haven’t told me how to persuade!

People in my field have been gaining clients (or students) for several millennia now based on explicit or implicit promises to help them persuade their audiences. For example: science communication scholars are sought out for advice about how to persuade audiences that climate change is real, bad, imminent and anthropogenic. “Just tell me what to say–effective messaging strategies!” is a common request. I’ve always wanted to reply: “Well, it’s complicated, and maybe persuasion isn’t what you want to seek.” Now Dan O’Keefe, who is much wiser than me, has confirmed this response in a just-out paper informatively titled “Message Design Choices Don’t Make Much Difference to Persuasiveness and Can’t Be Counted On.”

This essay extends O’Keefe’s longstanding project of meta-analysis of the empirical research on persuasive effects. Here he and his colleague are doing a re-analysis and synthesis of the meta-analyses on specific message variations, e.g. humor v. non-humor, high intensity v. low intensity language, deep cultural tailoring v. no cultural tailoring, lots of evidence v. limited evidence. His finding: for all message design features, effect sizes are small, and effects are variable–all message variations produced some backfire effects.

Helpfully, O’Keefe gives me some advice about what I should be saying to those badgering me for persuasion tips:

Those advising message designers should be modest, cautious, humble. If message-variation effects were substantial and entirely consistent, one could be unreservedly confident in one’s recommendations about persuasive message design: “Always choose message form A rather than message form B. Not only will A always be more persuasive, it will be a lot more persuasive.”

But given the results reported here, advisers will want to be rather restrained, even if there is a statistically significant meta-analytic mean difference in persuasiveness between the message kinds: “You should probably choose message form A rather than message form B, because on average A is more persuasive. However, A is likely to be only a little more persuasive than B, not enormously so. And A will not always be more effective than B—sometimes B will turn out to have been the better choice. So my advice is that you choose A, because you should play the odds. But it’s not a sure thing, and it probably won’t make a huge difference to persuasiveness.”

I imagine a follow-up question from the disappointed would-be climate persuader might be: “Well, why should I bother?” To this O’Keefe answers that there are times when even small persuasive effects can make a big difference (looking at you, Georgia voters)–and further, small effects can accumulate over time into bigger ones. To this I’d also add: there is more to communication than trying to change others. Sometimes, we all need to focus just on doing our part and living up to our responsibilities–focus on ethics, not effectiveness–and trust others to do their parts as well.