Social psychology

What apology research measures

A 2016 study separated apologies into six components and asked participants to rate their effectiveness. The finding was widely summarised. The definition of effectiveness inside the study was not.

Published 29 July 2026 · Sources: 4 · Record 03

Roy Lewicki, Beth Polin and Robert Lount published a paper in Negotiation and Conflict Management Research that treated an apology as a composite object rather than a single act. Drawing on earlier work in organisational behaviour, they identified six components: an expression of regret, an explanation of what went wrong, an acknowledgement of responsibility, a declaration of repentance, an offer of repair, and a request for forgiveness.

Across two experiments, participants read descriptions of a transgression followed by an apology containing some subset of the six components, and rated the apology on scales measuring how credible and how adequate it seemed. Apologies containing more components were rated higher. Among individual components, acknowledgement of responsibility drew the highest ratings and an offer of repair was close behind. A request for forgiveness contributed least.

The gap between the design and the summary

The design is a vignette study. Participants read about a transgression that did not happen to them, involving people they had never met, and rated a written apology they did not receive. The transgressions were largely workplace and trust scenarios rather than personal ones, and participants were adults recruited online in the United States.

Effectiveness in this study is an operational term. It refers to a rating on a scale, given by a reader in a laboratory task minutes after reading a paragraph. It does not refer to whether a relationship recovered, whether the wronged party's behaviour changed, or whether the apology was accepted by anyone. The study does not claim otherwise, and the paper is explicit about the limitation. Secondary coverage frequently was not, and the six components have since circulated as a template for producing an effect the research did not measure.

There is also a difference between rating an apology and receiving one. A participant judging a written apology is evaluating a text. A person receiving an apology has a history with the person delivering it, a view of whether the delivery is sincere, and an interest in the outcome. Vignette designs deliberately strip those things out to isolate the wording, which is a legitimate method and a real limit on what the result can be extended to.

The wider literature

A meta-analysis by Ryan Fehr, Michele Gelfand and Monisha Nag, published in Psychological Bulletin in 2010, synthesised a large number of studies on forgiveness and its correlates. Apology emerged as one of the more consistent situational correlates of forgiveness, alongside factors such as the perceived intent behind the offence and its severity. Dispositional factors, including trait agreeableness, were also associated.

The authors note the character of the evidence base they were pooling. Much of it is cross-sectional and self-report, which measures association at a single moment rather than change over time. Many samples are undergraduates. Correlational findings of this kind are consistent with several causal stories, including one in which people who are already inclined to forgive also recall having received better apologies.

The defensible summary is limited. When people are asked to evaluate apologies, an explicit acknowledgement of responsibility is rated more favourably than the components that do not admit fault, and this pattern is stable across several studies and samples. What happens between two particular people after a particular apology is not what these studies were built to observe, and no finding here should be read as describing it.

Back to the record