If you read the Persuasion Blog you don’t see much about emotion as a persuasion play. While I can get as moody as a poet – I weep when I listen to Joni Mitchell – I am scared to death of emotion as a practical persuasion play. I’ve been reading and researching human emotion since the beginning of my research career and I still don’t understand it well enough to employ it as a reliable persuasion tactic. As with shame.
States have long used punitive public shaming as a means of curbing bad behavior (Jacquet, 2011), and such institutionally sanctioned shaming practices remain common; examples include the statewide issuance of marked license plates for individuals convicted of DUIs (Nussbaum, 2006)) and online lists of noncompliant taxpayers (Jacquet, Hauert, Pizarro, and Tracy, 2012).
And, I’m sure you remember the cable TV shows on surprise interventions as friends and family members of an addict would draw that hapless Other Guy into a circle then disclose all the pain, woe, and suffering the Other Guy caused everyone with the addiction. Lots of shame and embarrassment. You’ve probably used shame yourself with a child, a bad spouse, and certainly your dog. Shame works!
Except maybe not. Randles and Tracy execute an interesting prospective observational study of recovering alcoholics (during the first six months of self reported sobriety).
Participants were paid $40 for each of two sessions, during which they completed a series of questionnaires. Sessions were conducted approximately 4 months apart (M = 4.24, SD = 1.81, range = 3–11).1 This time frame was chosen to capture a window when relapse was likely (most relapses occur within the first 3 months of sobriety; Foster, Marshall, & Peters, 2000; Hunt, Barnett, & Branch, 1971).
During the first session each participant was asked to recall and report on the last time they drank and felt bad about it while being videotaped. They also completed other survey items on paper and pencil forms. A few months later they returned and self reported about their drinking since the first session. That was the key measure of relapse.
The interesting variable here is the nonverbal indicators of shame during that videotaped description of the last bad experience drinking. Coders unaware of the study purposes used standard protocols to count the number of shame displays showing narrowed chest and slumped shoulders for each participant while they recalled that last bad drinking incident.
If standard persuasion practice applies then you would predict that people who showed greater shame would be less likely to relapse. Since shame is a negative consequence and people want to avoid negative consequences, then people who experienced the most shame from drinking should be more resistant to future drinking. At least that’s the theory and the common sense. Yet, that’s not what happened in this one small prospective observational study.
In all three analyses, nonverbal shame displays predicted relapse, regardless of whether it was operationalized as an increased likelihood of any relapse behavior (LOR = 2.47, SE = 1.20, p < .05), the number of drinks consumed between waves (ln(b) = 2.36, SE = .36, p < .001), or the number of drinks consumed between waves for those participants who relapsed only (n = 24; ln(b) = 1.75, SE = .50, p < .001; see Table 2; Table S2 online).
Those Log Odds Ratio values are Medium Windowpanes, about a 35/65 effect meaning people who showed more shame in those videotaped recollections were obviously more likely to relapse. That’s the exact opposite of common sense and theory prediction.
Now, quickly, there are a ton of caveats, wait a minutes, and hold on, sirs in this study. This is a convenience sample of self reported alcoholics. The researchers reported a lot of statistical whizbangery. They did not replicate the study. They didn’t test anything with a lab analog. And please note the gorgeous jeté in the persuasion train of conceptual logic here. Shame during disclosure in a research interview is correlated with future relapses. That’s not exactly the same thing as a persuasion play that shames Other Guys.
To their credit, the researchers note all these limitations and problems with the study. To address them, they provide a series of alternative analyses which you need to read if you want the details. Addiction research is an extremely difficult area for high quality experimental studies simply due to the nature of the people and the problem. That doesn’t mean you forget the standards of science and this is clearly a shaky study from a methods perspective.
That noted, the findings are consistent with related persuasion concepts. Just consider the typical psychology of addiction. You have an uncontrollable desire to perform a behavior that you love, but tends to bring following negative consequences. Yet, you enjoy the behavior so much that the negative consequences are worth the price of admission. Unfortunately, your addiction behavior also tends to generated negative consequences for people around you and they let you know it. You feel shame and guilt when they express the pain those negative consequences give them (but not you!).
Gee. Wouldn’t it be nice if you could do the behavior without the people around you getting negative consequences? And human nature enters at this point as the addict finds a Wile E. Coyote solution. She rents a hotel room, locks the door, and drinks alone! No one will ever know! She gets to drink her liquor and have it, too! You see the obvious Dissonance Reduction drama here as addicts find ways to justify their drinking behavior without bothering anyone else – at least in the addict’s mind.
And, all the result of that well known emotion persuasion play, Shame The Bastard.
You see why I am so leery of emotion in persuasion. Feelings are more like an explosive that is poorly understood and controlled. Sure, you can produce amazing effects with it – oh, Joni, you break my heart – but you can also blow yourself up, too. Now, if you are a Joni Mitchell persuasion maven, maybe you can do it, but then the trick is not in the play, but in the player if you catch the distinction. That’s persuasion as art, not as persuasion.
Daniel Randles & Jessica L. Tracy. (2013). Nonverbal Displays of Shame Predict Relapse and Declining Health in Recovering Alcoholics Clinical Psychological Science, first published on February 4, 2013.
P.S. Weep with me now! A Woman of Heart and Mind . . . you know the times you impress me the most are the times when you don’t try, when you don’t even try (YouTube).
P.P.S. Joni more lately finding the jazz in Woodstock (YouTube).