Research confirms that forgiveness is associated with better mental health, lower blood pressure, and reduced anxiety.
What does the science of forgiveness actually show? Four decades of randomised trials and meta-analyses show that structured forgiveness programmes reliably increase forgiveness and hope while reducing anger, anxiety and depression. One imaging-confirmed cardiac trial has shown that they even improved blood flow to the heart muscle itself.
Why this question deserves a proper answer
For a long time, forgiveness was treated as a private matter of conscience. Something for the pulpit, the counsellor’s room, or the quiet hours of the night. It was assumed to be good for the soul, and nobody thought to ask whether it was measurably good for anything else.
That changed. Since the late 1980s, forgiveness has become one of the more thoroughly studied interventions in applied psychology. It has been manualised, randomised, controlled, replicated across cultures, tested against waiting lists and rival therapies, and included in meta-analyses covering dozens of trials and thousands of participants. There are now two well-established intervention models, a substantial body of physical-health findings, and an emerging line of work in trauma and moral injury.
Forgiveness is not just for religious people. It is a practical skill anyone can learn.
What follows is a survey of what that research actually says – the models, the key trials, the meta-analyses, the physical-health findings, and the honest limits of what has been shown so far. I have kept the science plain rather than technical, but I have not softened it. The evidence is strong enough that it does not need help.
And there is a reason to care about this beyond curiosity. Most of us do not forgive because we are told to. We forgive when we can see what it costs us not to. The research, read properly, is a very long and very careful account of exactly that cost.
Most of us do not forgive because we are told to. We forgive when we can see what it costs us not to.
The two models that built the field
Almost all forgiveness research traces back to one of two intervention models. They were developed independently, they look quite different on paper, and – as we will see – they turn out to work for much the same underlying reason.
Enright’s Process Model: depth, over months
Dr Robert Enright of the University of Wisconsin-Madison is generally called the father of forgiveness research, and his Process Model is the field’s deepest intervention. It runs to twenty units across four phases, and it is designed for individual therapy over a long period.
The Uncovering Phase comes first. Before anything is let go of, it has to be looked at squarely. The client examines the anger and resentment they are carrying and how the injury has actually shaped their life – the psychological defences built around it, the shame, the endless comparison of themselves to the person who hurt them, the realisation that their whole view of the world quietly changed after the event.
Before anything is let go of, it has to be looked at squarely.
The Decision Phase follows. This is the point at which the person consciously decides to consider forgiveness as an option and commits to the process. Note that Enright does not treat the decision as the achievement. It is the doorway, not the destination – which is a distinction anyone who has tried to forgive by sheer willpower will recognise immediately.
The Work Phase is the engine of the whole model. Here the client works on reframing: coming to understand the offender in their global context – their upbringing, the pressures they were under, the wounds they themselves were carrying – and developing empathy, and sometimes even compassion, for them. Alongside this comes the idea of absorbing the pain rather than passing it on down the line. This phase is where the actual change happens, and it is worth holding on to that fact, because it becomes the key to the whole comparison later on.
Forgiveness is not just a decision; it is a process.
Finally the Deepening Phase, sometimes called the Outcome Phase. The person finds meaning in what they suffered, realises they are not alone in having suffered, recognises a new sense of purpose, and experiences genuine release from the negative emotion they had been carrying. This is the part that a good many people assume forgiveness starts with. In Enright’s model it is where forgiveness arrives after the work (which of course is the Work Phase) has been done.
Enright’s broader framework is triadic: it treats forgiveness of others, forgiveness of oneself, and forgiveness of the Sacred or of fate as three related domains. That third strand matters more than it might sound, and we will come back to it when we reach the research on moral injury.
Self-forgiveness not only helps you. It contributes to the health of families, workplaces, communities and ultimately to a healthier world.
Worthington’s REACH Model: five steps, built to travel
Dr Everett Worthington of Virginia Commonwealth University took a different route. Where Enright built depth, Worthington built reach – and the acronym is not accidental. REACH is compact, manualised, and designed to be delivered in a workbook or a group format of around twelve hours rather than over many months.
- R is for Recall the hurt – as objectively as the person can manage, without sliding into self-pity or into the well-worn groove of the victim story.
- E is for Empathise with the offender – actively working to understand their perspective, their pressures, their motivations. This is the empathy-induction core of the model, and it does the same job that Enright’s Work Phase does.
- A is for the Altruistic gift. The person recalls a time they themselves were forgiven for something, and chooses to extend that same gift onward rather than hoarding it.
- C is for Commit to the forgiveness, often in writing – a kind of forgiveness contract with oneself.
- H is for Hold on to the forgiveness when the doubts come back, as they generally do.
Every act of forgiveness you do makes the next act of forgiveness easier.
REACH has been tested in more than thirty randomised controlled trials since the late 1990s. That is an unusual amount of evidence for any psychological intervention, and it exists precisely because the model is compact enough to be handed to a lot of people at once. Enright’s model goes deeper. Worthington’s model goes further. Both facts turn out to matter.
What happened when the models were tested
Freedman and Enright, 1996 – the study that set the standard
The single most cited trial in the field is Freedman and Enright’s 1996 study, published in the Journal of Consulting and Clinical Psychology. It worked with twelve female incest survivors, aged between twenty-four and fifty-four, randomly assigned either to an immediate individual forgiveness intervention or to a waiting-list control group. Sessions were weekly, and they averaged fourteen and a third months. This was not a weekend workshop. It was well over a year of individual therapeutic work with people carrying some of the most severe interpersonal injury there is.
The reported finding was clear:
“the experimental group gained more than the control group in forgiveness and hope and decreased significantly more than the control group in anxiety and depression.”
- Freedman & Enright, Journal of Consulting and Clinical Psychology, 1996
Then the researchers did the thing that makes this study genuinely persuasive. They gave the same treatment to the control group afterwards – and those participants showed comparable gains, including increased Self Esteem. When a waiting-list group catches up after receiving the intervention, it becomes very difficult to explain the results away as the passage of time, or as the natural optimism of people who had volunteered for something.
Long-held feelings of anger, rage, fear and shame are painful. Forgiveness is the way to release them.
I want to sit with that finding for a moment, because it is easy to skim past. These were survivors of one of the deepest betrayals a person can experience. Nobody in that room was suggesting that what happened to them was acceptable, or that the people who harmed them deserved anything. What was on offer was release from carrying it. And it worked – measurably, in hope, in anxiety, in depression, in Self Esteem.
REACH under the microscope
Because REACH is short and portable, it has been tested far more often, and across a much wider range of people.
Forgiveness is available for everyone to learn.
A 2020 trial published in Frontiers in Psychology tested the REACH intervention with 124 Indian college students, delivered as a secular, manualised twelve-hour group programme. The effects on forgiveness and on reduced unforgiveness were large – effect sizes of 1.20 and above, which in psychological research is not a modest nudge but a substantial shift. Participants also showed increased empathy, increased positive affect, and increased Self Esteem. Notably, the programme worked across religious backgrounds without needing to be culturally customised.
The same study offered a useful clue about mechanism:
“Perceiving the offender as having a similar spirituality was related to…more growth in empathy, positive affect, and emotional forgiveness.”
- Frontiers in Psychology, 2020
This might not really be a finding about spirituality, but a finding about similarity. The more a participant could see the person who hurt them as being made of the same stuff as themselves, the more the forgiveness took hold. That is the empathy mechanism showing its hand.
A group randomised trial with 162 middle-aged adults, from Worthington’s line of research, compared the REACH group condition against a waiting-list control. The REACH group showed greater reductions in unforgiveness and in rumination, and greater increases in empathy and benevolence. Crucially, those effects held not only at the end of the programme but at six-month follow-up. Forgiveness work that evaporates the moment the group disbands would be of limited use to anyone. This did not evaporate.
Forgiveness is a skill. The more you practise it, the easier and more natural it becomes.
More recently, a global study reported by VCU News in May 2024 tested the REACH workbook with a large international sample, comparing those given the workbook against a delayed-access control group. Those who worked through the workbook showed significant reductions in unforgiveness, in depression symptoms, and in anxiety symptoms. A workbook. No therapist in the room. That is a remarkable result for something so cheap to distribute, and it is the strongest single argument for the scalable end of this field.
One more finding from Worthington’s work deserves a mention, mostly because of its title. His 2000 paper “Forgiving Usually Takes Time” established early on that forgiveness interventions need sustained engagement to produce durable change. A single session does not do it. That conclusion was reached from the data, but anyone who has genuinely tried to forgive something serious already knew it in their bones.
Forgiveness is not a single moment. It is a practice that deepens over a lifetime.
Head to head: what the meta-analyses found
Individual trials are suggestive. Meta-analyses – studies that pool many trials and look for what holds up across all of them – are where a field either earns its confidence or loses it. Forgiveness research has three that matter.
Baskin and Enright (2004) analysed intervention studies on forgiveness and found that process-based interventions, meaning the Enright-style long-form models, were more effective than brief, decision-based or cognitive approaches. Depth and duration mattered – and within process-based work, individual delivery outperformed group delivery.
Wade, Hoyt, Kidwell and Worthington (2014), published in the Journal of Consulting and Clinical Psychology, pooled fifty-four studies. The headline result:
“Participants receiving explicit forgiveness treatments reported significantly greater forgiveness than participants not receiving treatment”
- Wade, Hoyt, Kidwell & Worthington, Journal of Consulting and Clinical Psychology, 2014
The effect size was 0.56 against no-treatment controls, and 0.45 against alternative treatments that were not forgiveness-specific. The first number, in percentage terms, means that the average patient in the forgiveness group outperformed 71.2% of individuals who received no treatment at all. That second number is also important. It means forgiveness work did not just beat doing nothing – it outperformed other active therapeutic approaches at producing forgiveness. In other words, if we paired a patient from the forgiveness group with a patient from the alternative treatment group, there is a 62.5% chance the forgiveness-specific therapy would yield a better outcome. The analysis also found gains in depression, anxiety and hope.
And then the finding that reframes the whole comparison: once treatment dosage and modality were controlled for, differences between treatment models disappeared – Wade and colleagues found no significant difference in efficacy between the Enright and Worthington models. But they were candid about a limitation in that comparison, noting that the individual-delivery advantage was:
“most clearly demonstrated for Enright-model interventions, as there have been no studies of individual interventions using the Worthington model”
- Wade, Hoyt, Kidwell & Worthington, Journal of Consulting and Clinical Psychology, 2014
In other words, the individual-versus-group finding may partly be a finding about which model happened to have been tested that way. That is exactly the sort of caveat that gets quietly dropped when research is turned into headlines, and it should not be. Give the two models a comparable dose, and on the evidence available they perform much the same.
Lundahl, Taylor, Stevenson and Roberts (2008), in Research on Social Work Practice, pooled fourteen process-based intervention studies that had a comparison group. The effect sizes were substantial across the board: 0.82 for forgiveness itself, 0.81 for positive affect, 0.60 for Self Esteem, and 0.54 for reduced negative affect. The gains were largely maintained at follow-up.
Forgiveness can help you feel happier, have better relationships, improve your health, and create greater success.
Set those three meta-analyses side by side and a clear picture emerges. Forgiveness interventions work. They work better than nothing, and better than generic alternatives. The gains cover forgiveness, hope, positive affect, Self Esteem, anxiety and depression. And they last.
The shared active ingredient
Here is the most useful thing the comparison research has produced, and it is the part I would most like readers to take away.
Enright’s model and Worthington’s model were built separately, look nothing alike on the page, and are delivered on completely different timescales – months of individual therapy versus a twelve-hour group or a workbook. Yet at the centre of each sits precisely the same operation. Enright calls it the Work Phase: reframing the offender in a wider context until empathy, and perhaps compassion, becomes possible. Worthington calls it E: actively working to understand the offender’s perspective, pressures and motivations. Different vocabulary. Same move.
Forgiving makes it easier to cultivate qualities such as patience, tolerance, compassion and empathy.
And when Wade and colleagues controlled for duration and modality, the difference between the two branded models disappeared. That is a strong hint that neither brand is the active ingredient. What is doing the work is the empathy induction, plus enough sustained engagement for it to take hold.
This matters practically, and it matters a great deal. It means the question is not “which model should I use?” It means the question is “am I actually doing the empathy work, and am I giving it enough time?” A person who genuinely engages with understanding the human being who hurt them – properly, not as a box-ticking exercise – is doing the thing the research says produces the outcome, regardless of whose manual is open on the table.
It also explains the pattern in where each model shines. Depth and duration – Enright-style, individual, running over months – carry an edge for severe and complex injury: incest survivors, veterans carrying moral injury. Brief, scalable, group-based delivery – REACH-style – performs comparably for the everyday interpersonal offences that make up most of human hurt, and it can be handed to thousands of people at a time. Neither is the winner. They are answers to different questions.
The decision to forgive starts the process. The willingness to see it through completes it.
Does it have to be religious?
A fair question, given that both flagship models grew partly out of pastoral and religious counselling traditions. The research has an unusually clean answer.
Rye, Pargament, Pan, Yingling, Shogren and Ito (2005), also in the Journal of Consulting and Clinical Psychology, randomised 192 divorced individuals into three conditions: a secular group forgiveness programme, a religiously-framed group forgiveness programme, or no treatment. Eight sessions, measured before, after, and at six-week follow-up.
Both intervention arms produced significantly greater gains in forgiveness than the control group. The secular condition also showed notably reduced depression symptoms relative to control. And religious background did not predict which format worked better for whom – religious participants did not need the religious version to benefit.
Forgiveness can be approached as a psychological process, a spiritual process or a religious process. Just make the approach that works for you.
The mechanism, in other words, is psychological rather than necessarily spiritual. If forgiveness is part of your faith, the research gives you no reason to abandon that framing and every reason to keep it. If it is not, you lose nothing. The empathy work does the work either way.
What forgiveness does to the body
The mental-health findings are substantial, but they rest largely on questionnaires – people reporting how they feel. That is legitimate, and it is how most psychological research is done. It is also the point at which a sceptic reasonably starts asking whether we are measuring anything more than mood and expectation.
Which is why one particular trial deserves far more attention than it gets.
The cardiac trial
Waltman, Russell, Coyle, Enright, Holter and Swoboda (2009), published in Psychology & Health, worked with coronary artery disease patients. Thirty-two patients were screened using anger-recall stress imaging – they were asked to recall an anger-provoking event while their heart was imaged. Seventeen of them showed anger-induced myocardial perfusion defects: measurable reductions in blood flow to the heart muscle, triggered by remembering being wronged.
Those seventeen were randomised either to ten weekly sessions of Enright-model forgiveness therapy or to a control therapy. At the ten-week follow-up, the forgiveness group showed:
“significantly fewer anger-recall induced myocardial perfusion defects from pre-test to the 10-week follow-up”
- Waltman et al., Psychology & Health, 2009
They also showed, in the authors’ words, “significantly greater gains in forgiveness.” The paper’s conclusion was appropriately measured:
“Forgiveness intervention may be an effective means of reducing anger-induced myocardial ischemia in patients with coronary artery disease.”
- Waltman et al., Psychology & Health, 2009
Notice what makes this different from every self-report study in the field. The endpoint here was imaging. A machine looked at blood flowing through the heart muscle of people with diagnosed coronary artery disease, before and after forgiveness therapy, and the flow had improved. You cannot produce that result by feeling more positive about a questionnaire.
Resentment is like cycling with the brakes partly on all the time.
That is a metaphor I have used for years, long before I read this trial. I had meant it as a description of how unforgiveness feels – the drag, the effort, the sense of pushing uphill on level ground. It is a little sobering to find the cardiologists producing images of something so close to it.
Blood pressure and stress recovery
A related line of research has looked at trait forgiveness – not a programme someone completed, but their general disposition towards forgiving. Higher trait forgiveness predicted lower diastolic blood pressure at baseline, and faster blood pressure recovery after a stress task.
That second finding is the more interesting of the two. Everybody’s blood pressure rises under stress; that is what it is for. The question is how quickly it comes back down afterwards. More forgiving people recovered faster. Whatever forgiveness is, it is tied to how efficiently the body stands down from alarm – not merely to how cheerful someone feels.
We cannot have mental and emotional health if we are clogged up with unforgiven – and therefore undigested – life experiences.
Cortisol, immunity and pain
Toussaint and colleagues (2015), along with the broader review work collected in Forgiveness and Health: Scientific Evidence and Theories Relating Forgiveness to Better Health, found inverse correlations between forgiveness and markers across the cardiovascular, endocrine and immune systems in healthy adult populations. Unforgiveness, in that body of work, is associated with poorer physical health across the board. Reduced cortisol levels and enhanced immune response have both been linked to participation in forgiveness therapy.
There is also a systematic review in the Irish Journal of Medical Science looking at forgiveness and chronic pain. Of seven papers reviewed, five found a relationship between lower forgiveness and either increased pain or decreased pain tolerance. Qualitative work with fibromyalgia and chronic fatigue syndrome patients found that the patients themselves perceived forgiveness as healthy and as pain-reducing – which is worth noting, because these are people with a great deal of hard-won expertise about what does and does not help them.
Unforgiveness blocks our chances for happiness, health, and success.
None of this says forgiveness is a medical treatment. It says that carrying unforgiveness is not a neutral act. The body is keeping the account whether we are or not.
The frontier: PTSD and moral injury
This is the newest and least settled area of the field, and it needs to be described carefully rather than enthusiastically.
Research on forgiveness and post-traumatic stress has found that forgiveness’s protective effect on PTSD symptoms is mediated by reductions in anger and negative affect. That is an important qualification. It suggests forgiveness does not act directly on trauma symptoms at all. What it does is lower the anger and rumination load that keeps the nervous system switched on – and reducing that load happens to reduce the symptoms downstream.
Forgiveness nourishes the nervous system with positive, life-enhancing feelings.
The moral injury work is where Enright’s triadic model comes back into play. Moral injury is not quite the same thing as trauma. It is the wound carried by someone who did, or failed to prevent, something that violated their own deepest sense of right – a burden of shame and guilt rather than of fear. Work with combat veterans has applied the triadic framework – forgiveness of others, of oneself, and of the Sacred or of fate – as a way of addressing all three strands of that burden at once. For a veteran, anger at others may be the smallest part of the problem. Self-forgiveness, and some reckoning with fate, may be the larger part.
It is genuinely promising work. But it must be said plainly: forgiveness therapy is not currently a first-line or preferred treatment for PTSD or complex trauma. It is an active area of research and a plausible adjunct, not a replacement for trauma-focused treatments such as cognitive processing therapy or EMDR. If you are living with post-traumatic stress, please work with someone qualified. Forgiveness may well have a part to play in your recovery. It should not be asked to carry the whole thing.
Forgiveness does not mean that what happened was okay. It means you are choosing to be free of it.
What the science does not say
An honest survey has to include the boundaries, so here they are.
The research does not say that forgiveness makes what happened acceptable. Not one of these trials asked participants to conclude that the harm done to them was fine. Enright’s Uncovering Phase begins by facing the injury squarely – the opposite of minimising it.
The research does not say forgiveness means reconciliation. Forgiveness, in my terms is” letting go of the desire to punish”. Reconciliation is re-establishing a relationship. The Rye study’s participants were divorced; nobody was suggesting they go back. You can forgive someone completely and still, quite reasonably, decide never to see them again.
Reconciliation is a process of re-establishing our relationship with someone. Forgiveness is letting go of the desire to punish. These are different things.
The research does not say forgiveness is quick. Worthington’s own title says forgiving usually takes time; Freedman and Enright’s participants averaged over fourteen months; and Baskin and Enright’s meta-analysis found duration and depth mattered. Anyone selling instant forgiveness is not selling what the studies tested. However, in my own experience working on myself and with others (in groups and as individuals) using the GFI Four Steps to Forgiveness, time and depth do not need to be particularly extensive. Also, although I completely concur with the benefits of empathy, oddly enough, the GFI Four Steps to Forgiveness bypasses empathy and takes an entirely different approach. These four steps focus on acknowledging the pain and the benefits the person will receive from releasing it. I have found this effective cross-culturally. However, there has been no scientific study of this method, so there is currently no research to support these findings.
And the research does not say that forgiveness replaces medical or psychological care. It is an adjunct with good evidence behind it, not a substitute for treatment.
Forgiveness is not an act of will; it is an act of willingness.
From the evidence to your own life
So what does all of this mean for a person who is not running a trial, and who simply has someone they cannot stop thinking about?
I think it means this. The thing that stops most of us from forgiving is not that we lack the technique. It is that some part of us believes holding on is doing a job – keeping us safe, keeping the score straight, making sure the wrong is not forgotten. The research is, in effect, a very large and very patient body of evidence that holding on is not doing that job. It is not protecting us. It is running our nervous system at a level it was never designed to sustain, restricting blood flow, slowing our recovery from stress, dulling our immunity, and quietly draining energy that could be going into a life we would actually enjoy.
Resentment is an attempt to hurt someone by harming yourself.
The studies also tell us something encouraging about how the change happens. Nobody in those trials was required to feel warmly towards the person who hurt them, and nobody was asked to start from a place of generosity. They started from where they were, with the anger fully acknowledged, and worked – slowly, deliberately – towards seeing the other person as a whole human being with a history. That is the empathy step, and it is the thing that moved the numbers in both models.
And it is learnable. That is perhaps the most encouraging finding in the whole field. A twelve-hour group programme moved forgiveness by effect sizes above 1.20. A workbook, handed to people internationally with no therapist attached, reduced depression and anxiety symptoms. This is not a gift some people are born with and others are not. It is a skill, and skills respond to practise.
Even a little bit of willingness to explore forgiveness, even if very tentatively, can bring big results.
Which brings us to the practical point, and to the reason I have spent thirty years on this. The benefits in the research papers are real, but they are also narrower than what people actually report. Forgiveness does not only lower anxiety scores. It gives you your attention back. It stops a part of your mind lurking permanently in an unhappy past. It makes you easier to be around, and it makes other people’s goodness easier to accept. It clears your thinking, because a mind busy prosecuting an old case has less of itself available for the present one.
Every step in forgiveness is a step toward the life you truly want.
You do not have to take my word for that, and you do not have to take the researchers’ word for it either. Both are second-hand. The one piece of evidence that will actually persuade you is the piece you generate yourself, by forgiving one thing – preferably something small – and noticing what changes.
Where to start
If the research has made you curious enough to try it, start somewhere manageable. Take one small, ordinary grievance rather than the great injury of your life. Practising on the small things makes the big things possible later, and it also lets you see the effect clearly, without the noise.
Practising on the small things makes it easier to forgive the big things.
The method I teach is The GFI Four Steps to Forgiveness, and the ebook is completely free – no email address required, no sign-up, nothing to unsubscribe from later. You can download it in PDF, Kindle or EPUB format and be using it in the next few minutes:
If you would like to go further – into a fuller picture of what forgiveness is – my book Forgiveness is Power (ISBN 9781844096282) is available from Amazon and from Inner Traditions. It has also been published in China (ISBN: 9787559606006), India (ISBN: 9789325975170) and Taiwan (ISBN: 9789863591184).
And if you would rather be walked through it than read it, we have a collection of videos on the Global Forgiveness Initiative YouTube channel covering the same ground in plain language, including the Four Steps themselves.
The Global Forgiveness Initiative exists so that forgiveness is available to everyone, everywhere, for free.
The science took forty years to arrive at what many people have suspected all along: that carrying unforgiveness costs us far more than it ever costs the person we are carrying it about. The good news is that you do not have to wait another forty years to benefit from what it found. You can begin with one small thing, today.
–
William Fergus Martin
English website: w-f-m.com
Chinese website: wfm-cn.com