Expression of remorse ... behaviour change ... and a Soution Focused perspective

  • 14 November 2013
  • Michael Durrant

Here in Sydney, in early November 2013, the media is full of the story of an 18-year-old young man, enjoying a night out in Kings Cross, who was "king-hit", apparently randomly, and fell to his death. Evidence in court, which is publicly available, shows that the 19-year-old offender had been drinking significantly before going into the city and that his attack on Thomas Kelly was unprovoked and random. It is clear that the offender also assaulted four other men on the same night. Apparently, the offender was already on probation for gate-crashing a party in 2011 and punching the host, and his prior record also included assaulting a police officer, theft, malicious damage and affray.

The 19-year-old pleaded guilty to manslaughter after prosecutors agreed to withdraw a murder charge. He was sentenced to seven years in prison, with a non-parole period of five years and two months. With time served, he will be eligible for parole in four years. The parents of the the dead boy have publicly expressed their outrage and bewilderment. "We are completely cold, shocked, and just beyond disbelief that the sentence was just so lenient."

In his sentencing remarks, the judge said the offender "for reasons of drunkenness, was unable or unwilling to control his aggressive urges". The judge noted that the offender wept in court when the family read out their victim impact statements, which he said he took as a sincere expression of remorse.

The judge said, "In my judgment the offender is very unlikely to re-offend. I have the very distinct impression that from the tragic consequences his offending has brought about, he has well and truly learnt his lesson ... I find that the combination of the offender's youth, remorse, prospects of rehabilitation and the need to structure sentences for multiple offences constitute special circumstances." There was a clear assumption that the expression of remorse somehow indicated the offender's lesser likelihood of reoffending.

Is there any evidence that "showing remorse" is in any way related to the decreased likelihood of reoffending?

I remember VERY clearly, 26 years ago, I was at some function with my family and another four-year-old, unprovoked, kicked my four-year-old daughter in the shin. The mother of this four-year-old started fussing. "Say you're sorry to Claire ... come on, say you're sorry ... now, say you're sorry ... etc. ... etc. " 

I remember thinking, "I don't care whether or not she says she is sorry, as long as she doesn't do it again!"

Is there any relationship between "saying sorry" and "not doing it again"? Is there any proven relationship between "saying sorry" and behaving differently in the future?

Here in Australia, there is a popular view that "expressing remorse" (or "showing remorse") is an important thing for offenders to do. This taps into what I think is a generally-held belief by the so-called "average person" that accepting responsibility for violent or criminal behaviour and expressing regret, sorrow or remorse for that behaviour is a positive thing. Moreover, there appears to be a general belief that "genuine remorse" means the person is less likely to repeat the behaviour in the future.

Indeed, in our newspapers and on television, we regularly read or hear reports from our courts. It is common to hear that a judge specifically mentioned the fact that "the accused demonstrated remorse for his violent behaviour" as the reason that the defendant is given a lesser sentence. Similarly, when judges impose a heavier or more punitive sentence, it is not unusual to hear that the judge's rationale was that not only was the defendant guilty of his crime, but that he also showed no remorse for his actions.

Judicial sentences are, as I understand it, designed to have two main impacts. First, they are designed to have a punishment effect. Second, they are designed to have a rehabilitative effect. That is, one of the aims of the justice system is to ensure that people are less likely to re-offend.

If one of the aims of the judicial system is to ensure that offenders are less likely to re-offend — and if there is a general view that an offender "showing remorse" is somehow a positive thing — it begs the question of whether there is actually any evidence that "showing remorse" (or "accepting responsibility") is predictive of future non-violent or non-offending behaviour.

Bibas & Bierschbach (2004) observe that,

Remorse and apology are powerful forces in everyday life. Parents make their children apologize for everyday wrongs. "I'm sorry" is a common expression, and confession and forgiveness loom large in both religious rituals and secular reconciliation. ... People value remorse and apology because they heal psychic wounds, teach lessons, and reconcile damaged relationships. (p. 87)

This general valuing of expressions of remorse or sorrow permeates legal and judicial practice. Ward (2006) notes the prevalence of court rulings and judicial statements that find the expression of remorse to be an appropriate mitigating factor in deciding criminal sentences and, conversely, the prevalence of judicial comments that find the absence of remorse to be an appropriate aggravating factor in determining sentences.

A literature search of articles with the word "remorse" in their title yields many, many articles, many in university law school sponsored journals (The Yale Law Journal, for example). Many of them canvass opinion from criminology and jurisprudence and a great many of them reference judges' actual remarks in their summing up in criminal trials and/or sentencing hearings (I was not always able to access the original court reports and courtroom transcripts and have had to rely upon the accuracy of those legal scholars who have cited them.) Some of these articles argue that remorse is appropriately considered in making sentencing decisions. Others argue that remorse should NOT be considered in sentencing. However, almost none of these articles address the issue of whether or not remorse is actually associated with more or less recidivism.

Those articles that argue AGAINST considering remorse as a mitigating factor tend to argue the difficulties or unreliability of assessing remorse. Ward (2006), for example, seems to accept the general view that remorse ought to be relevant but argues against its use on the grounds that it must always be a subjective decision. He goes on to discuss the difficulty of judging whether a defendant is truly remorseful. Ward identifies "the concern which plagues most prosecutors — the deceptive defendant" and discusses the vagaries of trying to discern a defendant's expression of remorse is genuine or not. "How then can the court tell if the defendant is "truly" remorseful?" (p. 138). These authors seem to say that remorse would be a relevant factor, if only we could assess it more accurately!

Those articles that SUPPORT considering remorse as a valid mitigating factor seem to argue that remorse is important and valuable ... because remorse is important and valuable!

Bagaric & Amarasekara (2001) observe that, in the UK, the presence of remorse is generally seen to be important in judicial decisions, although they comment that, "In some cases, the Court of Appeal appears simply to have assumed that [remorse] is inherently relevant." (p. 366). They note that, in Australia, contrition, repentance and remorse have express statutory foundations as mitigating factors. That is, lawmakers (politicians) have enshrined in legislation the mitigating import of remorse.

[An interesting aside — Indermaur (1996) notes that, in the Australian context, defendants (and defence lawyers) seek to convince the judge that they acknowledge their wrongdoing and are truly sorry for what has happened. Once this statement of remorse is presented, they then embark on an exhaustive attempt to deny any real responsibility for the offence by claiming provocation, the effects of drugs or alcohol, the effects of early family deprivation or abuse, and so on. This makes one wonder about the value — and relevance — of the initial statement of remorse.]

Proeve & Tudor (2010) review research in which experimental subjects (NOT in the judicial system) were given various scenarios regarding violent offences and asked to recommend sentences or punishments for offenders. They showed that expressed remorse on the part of the offender leads to more positive judgments about an offender than no expression of remorse; and expressed remorse leads to recommendations of a less severe punishment of an offender than no expression of remorse. Thus, it seems a general community view that the expression of remorse ought be a mitigating factor.

Common sense says that the justification for expression of remorse mitigating punishment is that it somehow indicates a better prognosis, reduced recidivism or already begun behaviour change.

Ward (2006) suggests that, 

As rehabilitation became the predominant objective of punishment, judges sought indications that defendants were amenable to rehabilitation, and one of the preferred indicators was remorse. Judges concluded that the presence of remorse in a defendant meant he was more likely to respond to attempts at rehabilitation, and the absence of remorse marked a defendant as one unlikely to succeed in any rehabilitation effort. (p. 138)

Bagaric & Amarasekara (2001) comment,

Intuitively it may seem appropriate to treat less harshly offenders who are sorry for their crimes. However, if this is where the appeal of remorse as a sentence mitigating factor starts and ends, its role looks decidedly tenuous. Admittedly we feel that we want to hurt the defiant and boastful offender more than the contrite one, but to justify this we need a stronger basis than the mere impulse to do so — intuition is rarely a good guide for the development of legal rules and principles. (p. 367)

Bibas & Bierschbach (2004) observe,

In the eyes of judges, [expressions of remorse] indicate that an offender is not "lost," that he has some self-transformative capacity that justifies (or requires) a lesser punishment. To be sure, the language of judicial opinions does not always express this sentiment directly. Some judges say that the remorseful offender has a better character than does the unremorseful one. Others say that he is more likely to be rehabilitated, or that he is less hardened, or simply that he deserves less punishment. Whatever the precise language, the judges' point is that the remorseful offender is in some way changed, or likely to change. (p. 95)

Regarding the expression of remorse as an appropriate mitigating factor would make sense if there was evidence that it is predictive of changed future behaviour.

Ward (2006) comments,

Courts routinely fail to provide any substantive evidence that remorse or its absence can be affirmatively correlated with any type of future conduct by the defendant. Similarly, the presence or absence of remorse may or may not be directly related to the "character" of the defendant in any identifiable way. In essence, courts rely on remorse simply because, historically, courts always have. (p. 140)

Adshead (1998) and Proeve & Tudor (2010) both note that lack of remorse is regarded as a predictor of future criminal behaviour, despite the lack of evidence for this connection. Cox (1999, cited in Proeve & Tudor, 2010) expresses some surprise at the lack of research on the relationship between remorse and future re-offending.

Ward (2006) comments that "one gets the sense that scholars are grasping for a theoretical justification for a popular sentencing factor that has become ingrained throughout the history of the criminal sentencing process". (p.139)

So, it seems that there is little or no evidence that the expression of remorse or sorrow is in any way a predictor of future changed behaviour (i.e. reduced recidivism)1. Update.

Lack of remorse and psychopathy

It seems generally accepted within forensic psychology that there is some relationship between the so-called trait of psychopathy and recidivism. Dolan & Doyle (2000) observe that "predicting future risk of violent behaviour has a long and difficult history" (p.303) and suggest that research indicates generally that clinicians have little expertise in predicting future repeated violence. They argue that actuarial measures (so-called objective measures and scales) should be relied upon rather than clinical judgement.

Proeve and Tudor (2010) note that the concept of psychopathy was operationalized with the advent of The Psychopathy Checklist — Revised by Hare in 1991. They note that Lack of remorse and guilt feelings is one of the items that contributes to total scores on the Psychopathy Checklist. They comment that, whilst there is research evidence of the association of high scores on the Psychopathy Checklist with future violent or recidivist behaviour, the specific role of lack of remorse, as distinct from other features of psychopathy, is not clear.

Dolan & Doyle (2000) review a number of meta-analyses of longitudinal studies which explored factors that reliably predict future violent behaviour in offenders and they report that meta-analyses show that actual criminal behaviour (i.e. the extent of previous violence) is a better predictor than clinical factors, including judgements about characteristics such as remorse.

Restorative justice

Recently, there has been an increase in the popularity of various "restorative justice" models of dealing with offenders. Typically, these involve the offender meeting face-to-face with the victim(s) of the crime.

Citing various "leaders" in the restorative justice movement, Bibas & Bierschbach (2004) define restorative justice as "a process whereby all the parties with a stake in a particular offence come together to resolve collectively how to deal with the aftermath of the offence and its implications for the future." They comment that apology and remorse are considered important as part of this process. Maxwell and Morris (1994) say explicitly that one of the benefits of family conferencing for offenders is that "it enables them to understand the consequences of their offence and to express remorse" (p. 16).

There is evidence that both victims and offenders find restorative justice practices (such as Family Group Conferencing) emotionally helpful. Wagland, Blanch, & Moore (2013) show that participants (both victims and offenders) have a high level of satisfaction with restorative justice conferencing. However, Smith & Weatherburn (2012) research outcomes and conclude, "The present study finds no evidence that the NSW juvenile restorative justice scheme (Youth Justice Conferencing) reduces the risk of juvenile re-offending, the seriousness of juvenile re-offending, the time to the next offence or the number of new offences." (p. 17)

So, again, it seems that our view that offenders and victims facing each other and the offender apologising "ought to make a difference", feels good but is not supported by research.

Remorse and behaviour change within Solution-Focused Brief Therapy

Michael Clark, a clinical social worker who has been both a Probation Officer, a court officer and a magistrate, comments about the role of "admission of guilt" and accountability. Here, he is specifically referring to young offenders in the juvenile justice system; however, I think his comments are more widely applicable.

Compare how both approaches regard accountability. The problem-solving approach requires hard work to understand the problem, to ascertain who's responsible, how the problem originated, and how it's maintained. Accountability is realized when an adolescent owns up to the wrong. Admission is paramount for the assumption of responsibility. Strength-based practice, on the other hand, does not assume the ownership of guilt is somehow automatically curative. (Clark, 1998, p. 47)

Rather, he argues, a Solution-Focused approach actually holds the client accountable for future changed behaviour.

I have a case example I use often in training courses of a therapy session with a man who was referred by the police after a number of domestic violence incidents. His violence goes back many years. In the style of most domestic violence offenders, he minimises his violent behaviour ... "I don't do it all the time ... it's only once every two or three weeks!". While I suspect it is more frequent than that (!), I spend a great deal of time exploring what he does the times he gets angry but doesn't hit her. I do this carefully, I think ensuring that I do not give the message that his violent behaviour is in any way acceptable or can be ignored. Nonetheless, as we discover more and more examples of times (some seemingly trivial) where he has handled his anger in some other way, he gradually begins to see himself and the whole issue of self-control differently.

In my training, I usually add that we interviewed him and his wife (who had not been part of the therapy) at 12-month follow-up, and she said, "Not only has he not hit me since he saw you, but I don't feel scared of him anymore!", which suggests to me a fundamental change in the dynamics of their relationship.

At this point in training, people often say, "Yes, but he didn't actually express remorse ... so how can you know this change is real?"

Can we really continue to assert that "expressing remorse" is actually a predictor of behaviour change? This man's behaviour changed for at least twelve months (after maybe six weeks of therapy) without his having to "show" or "express" remorse. Rather, he had to "own" examples of different, non-violent behaviour.

This man's expressions of regret or remorse (whether or not they are genuine) mean nothing in and of themselves. CHANGED behaviour in the future actually comes from finding and building on examples of changed (or different) behaviour in the past ... that old Solution-Focused idea about exceptions! I'm reminded of what my colleagues in the HR business say when they are discussing what you ask about in job interviews ... they say that the best predictor of future successful behaviour is past successful behaviour (even if only small or embryonic!).

Dolan & Doyle (2000) suggest that previous violent behaviour is the best predictor of future violence. Perhaps the corollary is that PAST non-violent behaviour is the best predictor of future non-violent behaviour? Lee et al (2007) describe "a goal-directed, solution-based, domestic violence group treatment program ... which primarily utilizes goals to create a context for participants to identify, notice, rediscover, and reconnect with their strengths and resources in addressing problems with domestic violence". They observe that the approach "holds domestic violence offenders accountable for solutions rather than responsible for problems".

Adapting Clark (1997), my training handout says:

Offender responsibility is increased by a recognition of success and an expectation of change. Traditionally, "admission of guilt", or "remorse", is paramount for the assumption of responsibility. However, Strength-based practice does not assume the ownership of guilt is somehow automatically curative. This approach holds that accountability is realized through behaviour change, not passive admission; through building an expectation that the offender will DO something and by exploring the offender's responsibility for existing NON-violent, NON-criminal, NON-offending actions.

SO ... my belief is that, by focussing on past successes or differences, SF is focussing on something MUCH more "real" and useful than a focus on statements of regret or remorse.



Adshead, G. (1998). The heart and its reasons: Constructing explanations for offending behavior. Journal of Forensic Psychiatry9, 231-236.

Bagaric, M. & Amarasekara, K. (2001). Feeling sorry? Tell someone who cares: The irrelevance of remorse in sentencing, The Howard Journal40(4), 364-376.

Bibas, S. & Bierschbach, R. A. (2004). Integrating remorse and apology into criminal procedure, Yale Law Journal114, 85-148.

Clark, M. D. (1998). Strength-based practice: The ABC's of working with adolescents who don't want to work with you, Federal Probation62(1), 46-53

Clark, M.D. (1997). Strength-based practice: A new paradigm, Corrections Today59(2), 201-202.

Dolan, M. & Doyle, M. (2000). Violence risk prediction: Clinical and actuarial measures and the role of the Psychopathy Checklist, British Journal of Psychiatry177, 303-311

Indermaur, D. (1996). Offender psychology and sentencing, Australian Psychologist31(1), 15-19.

Maxwell, G. M. & Morris, A. (1994). The New Zealand model of family group conferences. In C. Alder & J. Wundersitz, Family conferencing and juvenile justice: The way forward or misplaced optimism? Canberra: Australian Institute of Criminology. 15-43.

Proeve, M. & Tudor, S. (2010). Remorse: psychological and jurisprudential perspectives. Farnham, UK: Asgate Publishing.

Smith, N. & Weatherburn, D. (2012). Youth justice conferences versus Children's Court: A comparison of reoffending. Crime & Justice Bulletin160. Sydney: NSW Bureau of Crime Statistics and Research.

Wagland, P., Blanch, B. & Moore, E. (2013). Participant satisfaction with Youth Justice Conferencing, Crime & Justice Bulletin170. Sydney: NSW Bureau of Crime Statistics and Research.

Ward, B. H. (2006). Sentencing without remorse. Loyola University Chicago Law Journal38, 131-167.


1. Update — January 2022.

In the eight years since this blog post was originally published, not much seems to have changed. Taylor, Eastman, Latham and Holloway (2021) review the recent literature and conclude, "It seems logical to expect that someone who becomes genuinely remorseful would be less likely to reoffend and more likely to become a good citizen but evidence is slight that remorse or ‘self-culpability’ is associated with good outcomes". Dandawate, Kalebic, Padfield, Craissati and Taylor (2019) comment "There is weak evidence that lack of remorse is associated with reoffending generally". While both of these papers are about offenders with psychiatric conditions, the authors are clear that their conclusions about the paucity of research evidence are conclusions about offenders in general.

Rocksheng, Baranoski, Feigenson, Davidson, Buchanan and Zonana (2014) conclude that “Remorse is held to be an appropriate consideration, particularly during the sentencing phase of criminal proceedings. However, it remains a poorly formulated concept, lacking clarity and uniformity in both its definition and the characteristics that signal its presence or absence.” They conclude that their interviews with judges confirm that there are no objective or agreed-upon criteria for assessing remorse. Thus, even if it were to be a valid variable, its measurement is so unreliable as to make it meaningless.

Bandes (2015) observes:

There is currently no evidence that remorse can be accurately evaluated in a courtroom. Conversely there is evidence that race and other impermissible factors create hurdles to evaluating remorse. There is thus an urgent need for studies about whether and how remorse can be accurately evaluated. Moreover, there is little evidence that remorse is correlated with future law-abiding behavior or other legitimate penal purposes, and, moreover, there is evidence that remorse is often conflated with shame, which is correlated with increased future criminality.


Bandes, S. (2015). Remorse and criminal justice. Emotion Review8, 14-19.

Dandawate, A., Kalebic, N., Padfield, N., Craissati, J. & Taylor, PJ. (2019). Remorse in psychotic violent offenders: an overvalued idea? Behav Sci Law37, 579–88.

Rocksheng, Z,, Baranoski, M., Feigenson, N., Davidson, L., Buchanan, A. & Zonana, H. (2014). So you're sorry? The role of remorse in criminal lawJournal of the American Academy of Psychiatry and the Law Online, 42(1) 39-48

Taylor, P.J., Eastman, N., Latham, R. & Holloway, J. (2021). Sentencing offenders with mental disorders, developmental disorders or neurological impairments: what does the new Sentencing Council Guideline mean for psychiatrists? The British Journal of Psychiatry218(6), 299-301 [Return to previous place]


Share this post

Leave a comment


Email address

This is never shown to the public.