The study, conducted by researchers from the University of British Columbia, shows that behavioral displays of shame strongly predicted whether recovering alcoholics would relapse in the future.

Public shaming has long been viewed as a way to encourage people to amend their ways and research suggests that experiences of shame can motivate people to improve their self-image and contribute to a common good.

But it's not clear that improvement on a general level extends to specific behaviors. Researchers don't know whether experiencing shame about a DUI, for example, actually deters drinking and driving. In fact, some studies suggest that shame might do more harm than good, as it can motivate hiding, escape, and general avoidance of the problem.

Psychological scientists Jessica Tracy and Daniel Randles of the University of British Columbia wondered whether the distinction between shame and guilt might play an important role in determining future behavior.

People who feel shame may blame themselves for negative events and view their "bad" behavior as an unchangeable part of who they are. Thus, shame may actually be a risk factor for certain behaviors rather than a deterrent. But this doesn't seem to be the case for guilt.

"One reason that certain sobriety programs may be effective," the researchers say, "is because they encourage people to see their behaviors as something they should feel guilty, but not necessarily shameful, about."

Feeling guilt about previous behavior, as opposed to shame about being a "bad" person, may be an important component of recovery.

To investigate the influence of shame and guilt on recovery from addiction, Tracy and Randles looked at drinking and health outcomes in a sample of newly sober recovering alcoholics.

Shame is difficult to measure because people often avoid acknowledging feelings of shame. To account for this, the researchers used measures that assessed both self-reported shame and shame-related behaviors, such as a narrowed chest and slumped shoulders. The researchers hypothesized that participants would be less able to voluntarily control their behavioral displays of shame.

In the first session, participants were asked to "describe the last time you drank and felt badly about it," and their responses were video-recorded by the researchers. In a second session about four months later, participants were asked to report their drinking behaviors. The participants completed questionnaires about their physical and mental health at both of the sessions.

The results were revealing.

People who displayed more shame-related behavior were likely to be in poorer physical health at the time of the first session.

More surprising, though, was the finding that behavioral displays of shame predicted whether participants would relapse after the first session.

"How much shame participants displayed strongly predicted not only whether they relapsed, but how bad that relapse was -- that is, how many drinks they had if they did relapse," say Tracy and Randles.

Shame behaviors at the first session also predicted distressing psychiatric symptoms at the second session. And the data indicate a possible association between shame and worsening health over time.

In contrast, self-reported shame did not predict likelihood of relapse, number of drinks consumed, or health outcomes, providing further evidence that self-report may not be an accurate way of measuring shame.

Consistent with the researchers' hypotheses, self-reported guilt was not associated with any of the outcomes measured.

This study provides the first evidence that feeling shame about one's addiction can directly promote relapses.

"Treatment providers have long suspected that shame is a barrier to recovery, but this is the first time we've seen this link evidenced so robustly," note Tracy and Randles.

The results have clear implications for anyone who struggles with addiction or who has loved ones struggling with addiction, and it also has implications for researchers and clinicians who study emotion and addiction.

The findings are also important in light of the fact that some policymakers and judges have argued for the use of public shaming as a punitive measure against crime.

"Our research suggests that shaming people for difficult-to-curb behaviors may be exactly the wrong approach to take," Tracy and Randles argue. "Rather than prevent future occurrences of such behaviors, shaming may lead to an increase in these behaviors."

The researchers are conducting additional studies to test whether the effect of shame on behavior change generalizes to problematic behaviors beyond addiction.

This research was supported by Social Science and Humanities Research Council of Canada, a Michael Smith Foundation for Health Research Scholar Award and establishment grant, and a Canadian Institute for Health Research New Investigator Award.