Shame Versus Guilt

What it is

The distinction organises a large body of work on what the psychologists call the moral emotions, and June Tangney, a Distinguished University Professor of psychology at George Mason University, is one of its main empirical sources. Her research interests include shame and guilt, and her books include Shame and Guilt, written with Ronda Dearing, and Shame in the Therapy Hour, which she co-edited with Dearing. A 2007 review of hers in the Annual Review of Psychology is titled Moral emotions and moral behavior.

The vocabulary is precise. Proneness means a person's general tendency to respond to wrongdoing in one way or the other, measured by questionnaire, not an emotion felt on a particular day. Guilt-proneness is a tendency to feel bad about an act and to want to repair it. Shame-proneness is a tendency to feel bad about the self. The distinction is attributional in that it depends on where the person aims the feeling, at the act or at the self.

The idea of a shame culture and a guilt culture, associated with Ruth Benedict in 1946, is a different use of the same two words. It classifies whole societies and is contested as a historical artefact. This page concerns the narrower claim about two emotions within a person.

The practical reading of the distinction is easy to overstate. It does not say that people should be made to feel guilty instead of ashamed. It says that, among the people studied, the tendency to aim bad feeling at an act went with better outcomes than the tendency to aim it at the self and then blame others. Whether any treatment, rebuke or programme can move a person from one tendency to the other is a separate question, and the studies cited on this page did not try. That is why the concept sits beside the shaming debate rather than inside it. It describes who tends to do well after wrongdoing and leaves open what, if anything, a third party can do about it.

It is also a distinction about measurement. Because proneness is assessed by questionnaire, the findings depend on people describing how they usually react, and a person who is defensive about the self may answer differently from one who is not. The jail study addressed part of that problem by checking self-reported reoffending against official arrest records, but the questionnaire side of the design still rests on self-description.

In effect

The jail study

Tangney, Stuewig and Martinez assessed 476 jail inmates soon after incarceration. They interviewed 332 of them a year after release and obtained official arrest records for 446. Participants were asked whether they had been re-arrested or had committed a crime that went undetected, and the answers were checked against official records. Guilt-proneness directly and negatively predicted reoffending in that first year. Shame-proneness did not predict reoffending directly. It predicted more, indirectly, through its strong link to externalising blame. When shame was not offset by defensive blame-shifting, it inhibited reoffending.

Why the shame result splits

The split is the finding the authors titled Two Faces of Shame. A person who feels bad about the self and turns the feeling outward, towards other people and circumstances, is the person in the data who reoffended more. A person who feels bad about the self and does not shift blame is in the data as someone who reoffended less. The study does not say what produces the difference.

Where it meets the shaming debate

Debates about public shaming often assume that more shame produces better conduct, or that shame is always harmful. This research cuts across both views. It measures dispositions, and nobody in the jail study was publicly rebuked. It therefore says little about whether a public penalty produces guilt, shame or blame-shifting. The closest policy-relevant evidence is the survey of tax offenders by Kristina Murphy and Nathan Harris, in which a felt reintegrative experience was linked with less later evasion, and the exposure experiment on tax delinquents by Ricardo Perez-Truglia and Ugo Troiano, which found an effect on payment of small debts and none on large ones.

What it does not say

It does not say that guilt can be induced by a public rebuke. Proneness was measured, not assigned.

It does not say that shame is harmless. Shame-proneness predicted more reoffending through blame-shifting, and the same study found a constructive side only when that route was absent.

It does not say how much of the difference between people is stable. The study followed inmates for about a year after release.

It does not generalise by itself from jail inmates to other settings, and it makes no clinical claim. This site uses it only for what the study measured.

Evidence status: mixed. The distinction is the standard one in the field, and the single study used here is a longitudinal sample of jail inmates with a one-year follow-up and arrest records for most of them.

---