Motivational interviewing: the complete practical guide

I have run motivational interviewing workshops for a long time now, and they almost always begin the same way. I ask the room to pair up. One person talks about something they’ve been meaning to change but haven’t: more exercise, less phone, the shed that needs clearing. The other person’s job is to convince them to do it. Give good advice. Explain the benefits. Be as persuasive as you can.

Then I ask the talkers how it felt. The answers barely vary. “Annoying.” “I started defending myself.” “I came up with reasons I didn’t even know I had.” And then, every time, someone says: “I actually want to do it less now.”

That exercise is the whole case for motivational interviewing in five minutes. When we argue for change, people argue back. When we help them hear their own reasons, they tend to move. This guide is about how to do the second thing, reliably, with real clients.

What motivational interviewing is

Motivational interviewing, MI for short, was first described by the American psychologist William Miller in 1983, and developed with the British psychologist Stephen Rollnick in a series of books, the latest edition published in 2023. They define it, in its simplest form, as a collaborative conversation style for strengthening a person’s own motivation and commitment to change.

Every word in that definition is doing work:

• Collaborative: it’s done with people, not to them.
• Conversation style: it’s a way of talking, not a technique you apply or a set of forms.
• The person’s own motivation: you aren’t installing motivation. You’re finding and strengthening what’s already there.
• Commitment to change: it has a direction. MI isn’t just pleasant listening; it’s listening with a purpose.

MI began in the treatment of problem drinking, but it’s now used across healthcare, social work, probation, education and psychotherapy: anywhere a person is weighing up a change that would be good for them and finding it hard to make.

Ambivalence, and the righting reflex

Most people who need to make a change are not unmotivated. They’re ambivalent. They want to change and they don’t. They can see the reasons to stop smoking and also the reasons it’s the best five minutes of their day. Ambivalence is normal. It’s the place most change starts from, and people can sit in it for a long time.

When a helper meets ambivalence, something kicks in that Miller and Rollnick call the righting reflex: the urge to fix things, to set the person straight, to argue for the healthy side. It comes from a good place. It also backfires, for a simple reason. If someone feels two ways about something and you take one side, they will naturally voice the other. You argue for change; they argue for staying the same. And people tend to believe what they hear themselves say.

The spirit of MI

Before any technique, MI is a way of being with people. Miller and Rollnick describe its spirit as having four parts. Without them, the skills become manipulation dressed up as listening, and clients can usually tell.

Autonomy deserves a special mention. In MI you accept, genuinely, that the decision belongs to the client, including the decision not to change. Paradoxically, saying this out loud often makes change more likely. People are much freer to consider something when they don’t feel cornered into it.

The four processes

MI conversations move through four overlapping processes. They aren’t rigid stages; you’ll move back and forth. But each builds on the one before, and skipping ahead is one of the most common mistakes.

1. Engaging: building a working relationship. Is this person willing to talk with you? Do they feel understood? Without engagement, nothing else works.
2. Focusing: agreeing what the conversation is about. People often bring several concerns; focusing is the shared choice of direction.
3. Evoking: drawing out the person’s own reasons for change. This is the heart of MI, and what makes it different from simply being client-centred.
4. Planning: when the person is ready, working out together how and when to change.

A useful check when a conversation feels stuck: which process are you in? Often the answer is that you’ve jumped to planning before the person has finished evoking, or started evoking before you’ve agreed a focus.

The core skills: OARS

MI is built on four everyday communication skills, remembered as OARS. None of them is unique to MI. What’s distinctive is how deliberately they’re used, and the balance between them.

Open questions

Questions that can’t be answered with a word or two. “What brings you here today?” rather than “Are you here about your drinking?” Open questions invite the person to think and to tell their story. But beware: even open questions, asked one after another, turn a conversation into an interview. A good rule of thumb is to ask a question, then follow it with at least one reflection, ideally two.

Affirmations

Genuine recognition of the person’s strengths, efforts and values. Not praise (“well done!”), which positions you as judge, but noticing: “You’ve kept going with this, even when it felt pointless.” Affirmations build confidence and make the conversation feel safer. They must be true. People can smell a manufactured compliment.

Reflections

Statements that say back what you understand the person to mean. Reflection is the most important skill in MI, and the one people find hardest to learn, because it feels strange at first to answer someone with a statement rather than a question. There’s much more on reflections below.

Summaries

Longer reflections that gather several things the person has said. Summaries show you’ve been listening, link ideas together and let you choose what to emphasise. They’re also useful for changing direction: “Let me check I’ve got this so far…” gives you a natural moment to move on.

Reflections, in depth

A reflection is a guess at what the person means, offered as a statement. It ends with the voice going down, not up. “You’re worried about your health” rather than “You’re worried about your health?” The statement form invites the person to continue; the question form invites them to confirm or deny.

There are several kinds, and good MI uses a mix:

• Simple reflections stay close to the words: “You’re tired of people telling you what to do.”
• Complex reflections add meaning, feeling or something unsaid: “You want to be treated like an adult who can make your own decisions.”
• Double-sided reflections hold both sides of ambivalence: “Part of you enjoys it, and part of you is worried about where it’s heading.” Put the change side last; people tend to pick up the last thing you said.
• Continuing the paragraph: guessing the next thing the person would say. “…and you’re not sure how much longer you can keep it up.”
• Amplified reflections slightly overstate sustain talk, so the person corrects you towards change: “So there’s really nothing about your drinking that concerns you.” Use these sparingly and without sarcasm, or they sound like a trap.

Don’t worry about getting reflections exactly right. A wrong reflection isn’t a failure. The person will correct you, and in correcting you, they’ll tell you something more accurate. The only real failure is not reflecting at all.

Practise

Here are five things clients commonly say early in a conversation about change. For each, pick the response you’d give. All the options are things real practitioners say; some just work better than others.

Change talk and sustain talk

The single most useful thing MI teaches is how to listen to what kind of thing a person is saying. When people talk about a possible change, some of what they say favours change, and some favours staying the same.

• Change talk is any language in favour of change: “I want to feel healthier”, “I could cut down at weekends”.
• Sustain talk is language in favour of the status quo: “It helps me relax”, “I’ve tried before and it never works”.

Research on MI has repeatedly found that the balance between them matters: when clients voice more change talk, and especially stronger commitment language, change is more likely. So the practitioner’s job is, quite simply, to evoke and strengthen change talk, while responding respectfully to sustain talk without arguing with it.

The kinds of change talk: DARN-CAT

Miller and Rollnick sort change talk into types, remembered as DARN-CAT. The first four are preparatory change talk: the person is thinking about change. The last three are mobilising change talk: they’re moving towards it.

Once you start hearing change talk, you’ll hear it everywhere, often tucked inside a sentence that’s mostly sustain talk. “I know I should stop, but…” contains a small piece of need. Your job is to notice it and pick it up. Try it on these ten.

Evoking: how to draw out change talk

Change talk doesn’t only arrive by chance. There are reliable ways to invite it.

Evocative questions

Ask questions whose answers are change talk. “What worries you about how things are now?” “If you did decide to make a change, what would be different?” “What makes you think you could do it, if you decided to?” Each question invites one of the DARN types.

Importance and confidence rulers

One of the most practical MI tools. Ask the person to rate, on a scale of 0 to 10, how important the change is to them, and then how confident they are that they could do it. The magic is in the follow-up question. Always ask why they’re not lower: “Why a 6 and not a 3?” The answer is, by definition, change talk. (Never ask why they’re not higher. The answer to that is sustain talk.)

The two numbers also tell you where to put your effort. Someone who rates importance at 9 and confidence at 2 doesn’t need persuading; they need help believing they can do it. Try different combinations below.

Other ways in

• Looking back: “What were things like before this became a problem?”
• Looking forward: “If you carry on as you are, what might things be like in five years? And if you changed?”
• Querying extremes: “What’s the worst that could happen if you don’t change? The best if you do?”
• Exploring values: “What matters most to you in life?” Then: “How does the drinking fit with that?” The gap between values and behaviour is one of the strongest sources of motivation there is.
• Elaborating: when change talk appears, ask for more. “In what way?” “Tell me more about that.” “What else?”

Responding to sustain talk and discord

Earlier versions of MI talked about “resistance”. Miller and Rollnick dropped the word, because it located the problem inside the client. They now distinguish two things.

• Sustain talk is about the behaviour: the person’s own reasons for not changing. It’s a normal part of ambivalence.
• Discord is about the relationship: arguing, interrupting, disengaging, “you don’t understand”. It’s a signal that something in the conversation, often the practitioner, has pushed too hard.

In both cases, the answer is never to push harder. Instead:

• Reflect it. Simple, complex, double-sided. Being understood takes the heat out of most sustain talk.
• Emphasise autonomy. “It’s your decision. Nobody can make it for you.”
• Reframe. “I’ve failed five times” becomes “You’ve kept coming back to this. It really matters to you.”
• Come alongside. “Maybe now isn’t the right time to change this.” Siding with the status quo often prompts the person to argue for change.
• Shift focus. Step away from the sticking point to something the person is ready to talk about.
• Apologise if discord is your doing. “I think I got ahead of you there. Sorry.” It repairs things remarkably quickly.

A word on decisional balance

Many people associate MI with the “pros and cons” grid: listing the benefits and costs of changing and of staying the same. It appears in lots of training materials. It’s worth knowing that Miller and Rollnick themselves now advise caution with it.

Writing with Gary Rose in 2015, Miller reviewed the evidence and concluded that a neutral, evenly balanced exploration of pros and cons can actually reduce commitment in people who are ambivalent, because it evokes as much sustain talk as change talk. Their advice: if your aim is to help someone move towards a change, evoke change talk. Keep decisional balance for situations where you genuinely want to stay neutral, such as helping someone decide between two options when you have no view on which is better.

Planning: when, and how

At some point, the conversation turns. The person stops talking about whether and starts talking about how. Signs of readiness include:

• More change talk, especially commitment and activation language
• Less sustain talk
• Questions about change: “How do people usually do this?”
• Imagining the future after change
• Small experiments already happening: “I didn’t drink on Tuesday, actually”

When you hear these, test the water with a summary of the person’s reasons for change followed by a key question: “So, where does that leave you?” or “What do you think you’ll do?” Then let them answer. If they’re ready, plan together. If not, go back to evoking. There’s no penalty for testing too early, as long as you don’t push.

Making a plan that sticks

1. Start with their ideas. “What do you think might work for you?” Offer options only after.
2. Keep the first step small and specific: what, when, where, how often.
3. Anticipate obstacles: “What might get in the way? What would you do then?”
4. Identify support: who and what could help.
5. Strengthen commitment: summarise the plan and ask how ready they feel to do it.
6. Agree how you’ll review it, without making it feel like a test.

Sharing information and advice

MI doesn’t forbid giving information or advice. It changes how you do it. The most widely taught method is Elicit, Provide, Elicit:

1. Elicit: ask permission and find out what they already know. “Would it be OK if I shared some information about that? What do you already know about how alcohol affects sleep?”
2. Provide: give the information clearly, briefly and neutrally, in small chunks. “What we know is that alcohol helps people fall asleep but breaks up sleep later in the night.”
3. Elicit: ask what they make of it. “What do you make of that?” “How does that fit with what you’ve noticed?”

The final “elicit” is the bit people forget, and it’s the bit that matters. It turns information into the person’s own conclusion, which is far more powerful than yours.

Putting it together: an annotated transcript

Here’s a short, fictional exchange with a man whose wife has asked him to come and talk about his drinking. Read it through once without notes and see what you notice. Then switch on the annotations to see the skills in action and how the client’s language shifts.

Notice what the therapist doesn’t do. No information about units, no warnings about health, no agreement with the wife, no plan. And yet by the end, the client has given four reasons for change and voiced his own intention to cut down. That’s evoking.

MI in therapy

MI started as a brief intervention, but it’s now widely used alongside other therapies. It fits naturally with CBT, where it can be used:

• Before treatment, to build commitment to starting something difficult, such as exposure work.
• When homework isn’t happening, to explore the ambivalence behind it rather than repeating the rationale.
• When therapy stalls, to step back from techniques and revisit what the person wants.
• In health behaviours alongside mental health work: sleep, exercise, alcohol, smoking, medication.

Henny Westra and colleagues, for example, have studied MI as a lead-in to CBT for generalised anxiety, with encouraging results, particularly for clients with more severe worry. More broadly, MI tends to work best as part of a wider plan rather than on its own.

When MI isn’t the right tool

• When there’s no ambivalence. If someone is already ready and asking how, help them plan. Evoking reasons they already have can feel patronising.
• In an emergency, where safety comes first and a directive approach is needed.
• When it becomes a technique for getting people to do what you want. MI’s spirit is about the person’s own goals. Used to push an agenda, it stops being MI.

The evidence

MI has a large research base: hundreds of controlled trials across many settings. A meta-analysis by Brad Lundahl and colleagues in 2010 found MI produced small but meaningful effects across a range of problems, comparable to other active treatments, and often in fewer sessions. Effects are strongest and best established for alcohol and other substance use, with good evidence in some health behaviours too.

Two caveats. First, effects vary a lot between practitioners and studies, which is partly why MI fidelity, how well it’s actually delivered, gets so much attention. Second, “brief MI” in a trial can mean anything from a single conversation to several sessions, so results don’t always transfer neatly to a given setting.

Learning MI

Here is the uncomfortable finding that every MI trainer quotes. In 2004, Miller and colleagues tested how well clinicians learned MI. Those who only attended a workshop improved a little, then drifted back towards their old habits. Those who also received feedback on recordings of their own sessions, and individual coaching, became much more skilled and stayed that way.

In other words, MI is learned the way any skill is learned: through practice and feedback. Reading this guide is a start. Recording yourself is where the learning really begins.

Measure your own practice

The standard tool for rating MI practice is the Motivational Interviewing Treatment Integrity code, or MITI, developed by Theresa Moyers and colleagues. Full coding takes training, but two of its measures are easy to track yourself: the ratio of reflections to questions, and the proportion of your reflections that are complex. Listen back to 20 minutes of a recorded session, with your client’s consent, and tap as you hear each one.

Common traps

Miller and Rollnick describe several traps that pull conversations away from MI. Most of us fall into all of them regularly.

• The question-and-answer trap: question after question, answer after answer. The person becomes passive, and you end up doing the work.
• The expert trap: believing you know best what the person should do, and telling them.
• The premature focus trap: zeroing in on what you think the problem is before the person is ready.
• The labelling trap: arguing about whether someone is “an alcoholic” or “depressed”. The label rarely matters; the arguing does harm.
• The blaming trap: getting drawn into whose fault it is.
• The chat trap: pleasant, aimless conversation that never gets to anything that matters.

Summary

• Ambivalence is normal. Arguing for change usually strengthens the argument against it.
• The spirit comes first: partnership, acceptance, compassion and empowerment.
• Move through engaging, focusing, evoking and planning, and don’t skip ahead.
• Use OARS, and reflect more than you ask: aim for at least two reflections per question.
• Listen for change talk, reflect it, and ask for more.
• Respond to sustain talk and discord with reflection and respect for autonomy, never with pressure.
• Ask permission before giving information, and always ask what the person makes of it.
• Learn it the way you’d learn an instrument: practice, recording, feedback.

I’ll end where my workshops end. After the persuasion exercise, I ask the pairs to swap and try again, but this time the listener can only ask open questions and reflect. No advice, no persuasion. The room gets quieter. Conversations slow down. And at the end, when I ask the talkers how it felt, someone always says the same thing: “I actually talked myself into it.”

That’s the point. It was always theirs to talk themselves into.