An apology can begin repairing trust when it clearly names the mistake, accepts responsibility, and recognizes its impact. Research finds stronger average repair after competence failures than after integrity breaches such as concealment. Follow-through matters when colleagues need the harm corrected or need evidence that a risky pattern has changed. Apologize without demanding forgiveness, fix what you can, and make one relevant change visible over time.
What kind of trust did the mistake put at risk?
Before deciding what to say, identify what the mistake changed for the person relying on your work. A missed handoff may make them question whether you can complete a task. Concealing the missed handoff gives them reason to question whether your account of the work is dependable. Those are different problems, even when one incident contains both. The question is not whether you meant to cause harm; it is what the other person now has to risk by accepting your next update or handoff.
A 2017 systematic review and meta-analysis by Yuan, Dong, and Li combined 18 papers, 36 effect sizes, and 4,731 participants. It found a positive average effect of apology on trust repair (d = 0.44, p < .001), with larger effects for competence violations than integrity violations. Effects were also larger when apology was compared with silence than denial. The accessible publisher abstract does not report study-level methods or a risk-of-bias assessment, so the pooled result describes an average across varied studies rather than the likely response of a particular coworker.
That evidence answers a narrower question than whether a workplace relationship feels repaired: an apology can improve trust on average, especially when the breach concerns execution. It does not put the deliverable right or make an inaccurate account reliable. If the mistake left a practical problem, the first useful test is whether the person can safely continue the work—not whether the conversation ended warmly. The meta-analysis supports taking an apology seriously as one repair response, while leaving open whether a colleague has enough evidence to delegate the same work again.
Sources: The trust repair effect of apology: A systematic review and meta-analysis
What should the apology make clear?
Lewicki, Polin, and Lount’s two 2016 studies asked people to judge apology statements after an accounting job candidate’s tax-filing violation. Study 1 used 333 US adults recruited through Amazon Mechanical Turk (mean age 33.7; average 14 years of work experience); Study 2 used 422 undergraduate business students at a Midwestern university (mean age 21.8). Participants rated hypothetical statements; they had not experienced the violation as harmed coworkers. More components generally received higher effectiveness ratings, and competence violations were rated as more repairable than integrity violations. The authors also note that studies have not used consistent definitions of apology components, so their component labels do not add up to a settled recipe. Study 1 presented component definitions; Study 2 used written apology statements in a hiring scenario. Both measured judgments in that scenario, not restored trust at work. Text-only presentation also removed tone, and US samples may not represent other cultural settings. Apply the ratings as clues about what people may value, not a formula guaranteed to repair a real relationship.
A useful apology names the act, the impact, and your responsibility, then offers a repair you can actually make. “I missed the handoff, so you had to redo the final check. I’m sorry. I’ll correct the record and send the updated file by noon.” The promised correction gives the other person something concrete to expect. It does not set a deadline for trust. Keep the explanation brief unless the cause changes the repair: if a system outage interrupted the handoff, naming that fact may help prevent a repeat; it should not obscure your part in notifying the person who depended on it.
The researchers also found that a request for forgiveness, presented as a component of a hypothetical apology, received the lowest effectiveness ratings. That was not a finding that recipients forgave less, or that asking always harms repair. Still, there is no need to press the person for reassurance: let them decide whether to answer, ask for more information, or wait. Their response is not a score of how well you delivered the apology.
Sources: An Exploration of the Structure of Effective Apologies
What action would make reliance safer?
After the apology, match the next action to the remaining problem. Correct a wrong file and notify the people who used it. If a repeated handoff failure created doubt about completion, agree on a temporary check that catches the omission. If information was withheld, correct the record and disclose what others need to make sound decisions. These actions are practical recommendations drawn from the nature of the harm; they are not a tested universal sequence. The correction should reach the people affected, not only the person you apologized to. Otherwise the work can remain wrong even if the conversation between two colleagues feels resolved.
A 2019 leader–follower study examined violation severity and intentionality through a critical-incident study and a laboratory study. Its abstract reports that apology quality related to forgiveness and later trust at moderate combinations of severity and intentionality; at minor or extreme combinations, apology affected trust directly without forgiveness. The accessible university record gives no sample sizes or participant profiles, so these results need that qualification and cannot establish what happens between peers at work. The findings describe a conditional pattern, not a guarantee that an apology or forgiveness restores reliance.
This evidence does not support a simple “bigger mistake, bigger apology” rule. It suggests that severity and intent together can shape what an apology is associated with, while leaving the right response to a particular case unresolved. Start with the risk the other person still carries if they rely on you today.
Lewicki and Brinsfield’s review distinguishes short-term verbal strategies, including apologies, from compensatory arrangements and longer-term structural changes. It describes categories of trust repair rather than a fixed workplace protocol. Applying that distinction, restitution fits a loss that can be made good; a check fits a process that can fail again. Keep any safeguard connected to the actual risk and revisit it when the evidence changes.
A safeguard should address the risk without making the coworker your permanent monitor. Own the check and report when it has worked; where honesty was the issue, verification may matter longer than a corrected file. Trust is not something the harmed person must restore on your timetable.
Sources: Trust Repair; Repairing broken trust between leaders and followers: how violation characteristics temper apologies
When does an apology need follow-through?
The practical decision turns on what remains after you have acknowledged the mistake. If the harm is only that your action went unrecognized, a direct apology may be the needed first step. If a file is still wrong, correct it and tell everyone relying on it. If the breach leaves a reasonable question about whether the same failure will recur—or whether the account is honest—add a safeguard tied to that risk. These are recommendations drawn from the harm and the trust-repair categories in the research, not a tested sequence that guarantees renewed trust.
The evidence does not tell you to make every apology elaborate. In Lewicki and colleagues’ hypothetical ratings, more components tended to look more effective, but inconsistent definitions across studies and a tax-filing scenario limit what that result can prescribe. Yuan and colleagues’ pooled result favors apology on average, especially after competence violations; the leader–follower study also suggests severity and intent can change how apology quality relates to trust. Neither establishes the exact response of a coworker. A proportionate correction is enough when it resolves the remaining problem; use a continuing check only while a real recurrence or honesty risk remains, and take responsibility for running it.
The Emotional Skills Profile can help you reflect on habits that shape this exchange, such as noticing impact and regulating your response before you apologize. Its recent-behavior questions and scenario judgments can suggest a behavior to practice; they cannot decide whether this breach needs a correction, whether a safeguard is adequate, or whether a colleague should rely on you again. Make that decision from the harm still present and the evidence your follow-through can provide.
Sources: The trust repair effect of apology: A systematic review and meta-analysis; An Exploration of the Structure of Effective Apologies; Repairing broken trust between leaders and followers: how violation characteristics temper apologies
Questions readers ask
Should I apologize if the mistake was accidental?
Yes, when your action affected someone else. State what happened, recognize the impact, and accept responsibility without presenting the apology as a demand for forgiveness. If the error created an ongoing problem, pair the apology with a correction or a relevant prevention step.
Sources
- The trust repair effect of apology: A systematic review and meta-analysis
The publisher record's abstract reports 18 papers, 36 effect sizes, and 4,731 participants; a positive pooled apology effect on trust repair (d = 0.44, p < .001); and moderation by violation type and comparison group. It reports a larger effect for competence-based than integrity-based violations and for silence rather than denial controls. The record does not provide study-level methods or risk-of-bias details.
- An Exploration of the Structure of Effective Apologies
The full paper reports two scenario experiments evaluating hypothetical apology components in trust-violation scenarios. Study 1 recruited 333 US Amazon Mechanical Turk adults (mean age 33.7; mean work experience 14 years); Study 2 included 422 undergraduate business students at a Midwestern university. More components were generally rated more effective, competence-based violations were rated more favorably than integrity-based ones, and a request for forgiveness received the lowest effectiveness ratings. These are scenario judgments, not observed forgiveness or restored trust in real workplace relationships.
- Trust Repair
The review classifies short-term strategies as including verbal statements (such as apologies) and compensatory arrangements, and longer-term strategies as including structural changes such as contracts and monitoring. This is a taxonomy, not a tested fixed workplace protocol.
- Repairing broken trust between leaders and followers: how violation characteristics temper apologies
The university abstract describes a critical-incident study and a laboratory study of leader–follower trust violations. It says severity and intentionality jointly condition the relationship between apology quality, forgiveness, and subsequent trust; it supplies no sample sizes or participant profiles, and concerns leader–follower rather than coworker relationships.
Apply it to the real situation
Notice what you do after a difficult moment
From this guide: Use the article's repair sequence to reflect on how you recognize impact, regulate your response, and follow through after a mistake.
An apology is one moment in a longer interaction. If you want to reflect on how you notice emotion, take responsibility, and respond when work gets tense, the private 32-item Emotional Skills Profile offers a structured view of recent habits and scenario judgments. Use it to identify a behavior to practice, then observe that behavior in real conversations. It does not predict whether another person will restore trust.