GUIDE · CBT

Downward arrow technique: follow one thought down to the belief under it

You text a friend at noon, and by six there’s still no reply. The thought “She’s annoyed with me” brings a sinking feeling too big for one unanswered text, because it rests on something older. The downward arrow is a technique from cognitive behavioral therapy (CBT) for finding what’s underneath: you keep asking what the thought would mean if it were true.

What it is and where it comes from

You start with an automatic thought, the quick sentence your mind produced in an upsetting moment. Write it at the top of a page, draw an arrow under it, and ask: if this were true, what would it mean to me, and why would that upset me? The answer goes under the arrow, and you ask the same question about it. Keep going until a line sounds less like a reaction to today and more like a rule you live by or a verdict on who you are.

David Burns introduced the method in his 1980 book Feeling Good as the vertical-arrow technique, a way to expose the silent assumptions behind automatic thoughts. In a 2017 memo for therapists, he calls it the downward arrow and writes that he developed it. Several clinical papers that use it to find core beliefs cite Judith Beck’s textbook Cognitive Therapy: Basics and Beyond, first published in 1995.

The arrow finds a belief and leaves changing it for later. Burns contrasts it with the usual thought record, which moves sideways from a negative thought to a balanced answer. The arrow goes down instead, so don’t argue with each line on the way.

Thoughts, rules, and core beliefs

Cognitive therapy describes thinking in layers. On top are automatic thoughts, quick and tied to one situation. Below them sit intermediate beliefs: attitudes (“It’s awful to be disliked”), if-then assumptions (“If I keep everyone happy, they’ll stay”), and rules, often phrased as shoulds (“I should never need too much”). At the bottom are core beliefs, broad conclusions like “I’m unlovable” or “I’m not good enough.”

The arrow usually passes through the middle layer. Burns noticed that the lines in a chain don’t follow logically: “I should have seen it coming” doesn’t prove “I’m a fraud.” What connects them is an unwritten rule, here something like “If I miss anything, I’m a failure.” For more on rigid rules, see “should” statements.

People tend to land in a few familiar places. In a 2015 study, 1,813 people in an online CBT program each recorded a core belief they’d reached with the arrow. The most common themes were a verdict on the whole self (“I am useless”), being loved or left, and competence. A 2022 study that ran the arrow through a scripted chatbot with 308 adults who had mild depression symptoms found the same top three.

Illustrative examples: how the arrow might go

Both people here are made up, to show the shape of the exercise. They aren’t case studies or evidence of anything.

Maya texted a friend at noon and hasn’t heard back by evening. Under each line, she asks what it would mean about her if it were true.

1. “She’s annoyed with me.”
2. “I came on too strong again.”
3. “I’m too much for people.”
4. “Sooner or later, people leave me.”

The top line is a guess about her friend’s mind, the trap called mind reading. By the third line she isn’t talking about this friend at all. The jump between lines two and three points to a rule she lives by: “If people see how much I need them, they’ll pull away.”

Jordan starts with a should. Shoulds are hard to follow down, because the honest answer to “why does it matter?” is “because I should have.” Burns’s fix is to grant it: suppose you should have and didn’t, so what does that mean?

1. “I should have finished the report last night.”
2. “My manager will see I’m behind.”
3. “She’ll realize I’m not as capable as she thinks.”
4. “I’ve been faking it this whole time.”
5. “I’m not good enough.”

The rule in the gaps sounds like “If I’m not ahead on everything, I’m a fraud.” Unlike the core belief, it makes a prediction Jordan can check this week. If his chain sounds like a voice you know, see the guide to the inner critic.

How to do the downward arrow on paper

Use paper or a notes app, and pick a recent moment that upset you a moderate amount. Save the most painful material for work with a therapist.

1. Write one thought as a statement. Use the sentence you told yourself in that moment. Turn questions into statements: “Why do I always ruin things?” becomes “I always ruin things.” If you only find a feeling, ask what you were telling yourself when it hit.
2. Draw an arrow and ask. Under the thought, ask: “If this were true, what would it mean about me? Why would that upset me?” Write the answer in one sentence under the arrow.
3. Repeat on each new line. Ask the same question of every answer. Don’t argue with the lines yet; if a rebuttal pops up, jot it in the margin.
4. Turn feelings and wishes into thoughts. If you write “Then I’d feel ashamed,” ask what you’d be telling yourself, and write that. If a wish shows up, like “I want them to be proud of me,” assume it didn’t come true and ask what that would mean.
5. Grant the shoulds. When a line says you should have done something, suppose that’s right and ask what it means that you didn’t, as Jordan does above.
6. Stop at the bottom, or sooner. Stop when you start repeating yourself, when answers sound silly, or when a line reads like a verdict on who you are or how life will go. Stop sooner if your distress climbs fast.
7. Name the rule in the gaps. Find the jumps where one line doesn’t follow from the last. Write the rule that would bridge each one, as an if-then (“If I make a mistake, people will lose respect for me”) or a should (“I must never let anyone down”).
8. Rate the bottom line twice. Score it from 0 to 100 for how true it feels and how true you think it is. In a 2007 study, the felt ratings ran higher than the reasoned ones, so don’t be surprised by a gap.
9. Pick one next move. Log the next time this belief shows up in a thought record, or test the rule with a behavioral experiment.

Keep the page where you can find it. If different starting thoughts keep reaching the same bottom line, that’s the belief to work on first.

For fears: “What’s the worst part about that?”

Anxious thoughts are usually predictions, like “I’ll freeze in the meeting,” and asking what a prediction means about you can stall. For fears, ask “What’s the worst part about that?” or “If that happened, then what?” One research team defines the downward arrow as exactly these two moves: asking why a line would be upsetting, or what the worst outcome would be if it were true.

Sam, another made-up example, has a presentation on Thursday.

1. “I’ll freeze in the meeting.”
2. “Everyone will see my hands shake.”
3. “They’ll decide I can’t handle this job.”
4. “I’ll get pushed out.”
5. “I’ll have failed at the one thing I was supposed to be good at.”

The chain starts with a scene and ends with a verdict about competence. Worry researchers see the same drift in the catastrophizing interview, which asks “What is it that worries you about that?” about each answer in turn. Across six studies, Davey and Levy found that chronic worriers kept the chain going longer than other people, and personal inadequacy crept into their steps whatever the topic. A 2007 study found that worriers’ mood got worse as the chain went on.

Keep the fear version short. Stop at the first line that names the outcome you dread, then test it. Decatastrophizing sets the worst case beside a best and a most realistic one, and a behavioral experiment checks the prediction directly. If what-ifs run on their own most days, read about catastrophizing.

What to do with the belief you find

Treat the bottom line as a hypothesis, not a verdict. It’s a conclusion your mind reached, often long ago and from hard experiences, and it may be partly right. Judith Beck, president of the Beck Institute, writes that CBT therapists don’t tell clients their thoughts are wrong. They check together whether a belief is 100 percent true, 0 percent true, or somewhere in between, using evidence and real-life experiments.

Beck names three habits that keep a negative core belief going: watching for proof, coping habits like avoidance or overcompensating, and discounting good experiences that don’t fit. Log evidence in a thought record, and include what you’d normally wave away, like the friend who did text back the next morning.

A rule is a good place to start, because it makes a prediction you can test. In a 2012 study with 91 volunteers, one session of either a thought record or a behavioral experiment shifted a target belief more than a control session did, and the experiment shifted it sooner. The volunteers weren’t a clinical group, so take that as a hint.

Expect the belief to loosen rather than vanish. Beck notes that even people without a mental health condition can have a negative belief flare after a setback, like being passed over at work, and then settle again. If yours comes back, you’ll have the page and the evidence.

Common snags, and when to stop

If you go in circles between two lines, like “I’d be a failure” and “No one would love me,” join them. Burns suggests asking: if both were true, what would that mean? If the loop holds, you’ve likely reached the bottom.

If you write a feeling, like “Then I’d feel terrible,” ask what you’d be telling yourself in that moment and write that sentence instead. Burns insists that the chain start from a thought and stay made of thoughts.

If you hit “I don’t know,” offer yourself a few meanings and see which one stings: “I can’t cope alone,” “I’m unlovable,” “I’ll never be happy.” Burns also suggests explaining why it would be bad to someone from another planet.

If the answers turn vague or silly, like “It would mean the universe hates me,” you’ve passed the useful part. One study told people to stop right there, or when they began repeating themselves.

If you land on something painful, slow down. Burns tells therapists that when a painful memory surfaces mid-chain, they should set the technique aside and attend to the person. Do the same for yourself: put the page down, take a few slow breaths, and come back another day or bring it to a therapist.

Stop and get support if the chain pulls up memories of abuse or other trauma, especially ones that feel like they’re happening again. Stop too if the bottom line is hopeless, like “Nothing will ever change” or “Everyone would be better off without me.” In the 2015 online-program data, people who’d had thoughts of suicide in the past week landed on whole-self verdicts like “I am useless” more often than chance would predict. A belief like that is a reason to talk with a doctor or therapist soon. If you’re thinking about ending your life, reach out right away; the crisis numbers are below.

What’s been tested, and what hasn’t

The downward arrow is taught in CBT trainings and written into treatment protocols, but it has barely been studied on its own. A PubMed search for the exact phrase returns four records. Full-text searches add about 50 papers that mention it, mostly as one technique among many in case reports and trial protocols. We found no trial that tested the downward arrow by itself.

Studies that focus on the arrow mostly use it as a way to find beliefs. Besides the 2015 and 2022 studies above, a 2007 interview study used it to reach core beliefs behind images about eating and body shape. The interview gave consistent results when it was repeated or scored by a second rater. None of these studies measured whether finding the belief helped anyone feel better.

The arrow also sits inside tested programs. Preventive cognitive therapy, an eight-session program for people in remission from recurrent depression, uses it among other tools. In a 2019 trial with 214 people who had recovered with CBT, adding the program delayed relapse over 15 months compared with usual care, though the number and severity of relapses didn’t differ. The trial can’t isolate what the arrow added, since it came bundled with other techniques.

For CBT as a whole, a 2012 review of 106 meta-analyses called the evidence very strong, with some of the strongest support for anxiety disorders. A 2007 review adds a caution: component studies have found little evidence that specific cognitive techniques add to what CBT already does. Judge the arrow by whether the belief it finds gives you something useful to test.

Follow one thought down with Helpy

The chat below runs with Helpy, the site’s AI assistant. Give it one thought from a recent upset, and it can ask the arrow question with you, help you spot the rule in the gaps, and suggest a way to test it.

Evidence and sources

• Burns (1980), Feeling Good: The New Mood Therapy (Morrow): Introduces the vertical-arrow technique for exposing silent assumptions. We checked the 1980 wording through a full-text search of a scanned first edition, not a full read.
• Burns (2017), memo for therapists on the downward arrow: Says he developed the technique; source of the tips on feelings, shoulds, loops and painful memories. Expert advice, not data.
• Judith Beck, Cognitive Behavior Therapy: Basics and Beyond, 3rd ed. (Guilford Press, 2020): Publisher’s page. Clinical papers such as Griffioen and colleagues (2017) cite its 1995 first edition for the arrow; that 30-person trial used it to pick target beliefs before CBT or eye movement desensitization and reprocessing (EMDR), and both groups improved. We didn’t read the book.
• Picardi and Gaetano (2014), Clinical Practice and Epidemiology in Mental Health: Describes core beliefs, intermediate beliefs and automatic thoughts as layers. A review of the model, not a test of it.
• Judith Beck (2024) and (2026), Beck Institute blog: Testing beliefs with evidence instead of calling them wrong, and three habits that keep negative core beliefs going. A clinician’s blog posts, not studies.
• Millings and Carnelley (2015), Journal of Affective Disorders: 1,813 online CBT users recorded a core belief reached with the arrow; global self-evaluation came up most. Program completers only, with no symptom scores from the same time.
• Somerville and Cooper (2007), Eating Behaviors: The arrow reached core beliefs reliably in women with and without bulimia, and emotional belief ratings ran higher than rational ones. An assessment study.
• Burger and colleagues (2021), PLOS ONE and (2022), Frontiers in Digital Health: Online thought records with arrow steps from 320 and 308 people, with stop rules and, in 2022, a scripted chatbot. Not Helpy, and no symptom outcomes.
• Davey and Levy (1998), Journal of Abnormal Psychology: Worriers’ catastrophizing chains ran longer, and personal inadequacy crept in whatever the topic. Davey and colleagues (2007) found their mood worsened along the chain, and Talbot and colleagues (2010) describe the interview’s wording. Lab tasks, not treatment.
• McManus and colleagues (2012), Journal of Behavior Therapy and Experimental Psychiatry: One session of a thought record or a behavioral experiment shifted a belief more than a control session in 91 volunteers. Nonclinical, single session.
• de Jonge and colleagues (2019), Journal of Consulting and Clinical Psychology: Preventive cognitive therapy, whose 2015 protocol includes the arrow, delayed depression relapse over 15 months in 214 people. It tests a bundle of techniques.
• Hofmann and colleagues (2012), Cognitive Therapy and Research: Review of 106 meta-analyses of CBT, with the strongest support for anxiety disorders among other problems. Whole CBT, not single techniques.
• Longmore and Worrell (2007), Clinical Psychology Review: Little component evidence that specific cognitive techniques add to CBT. It covers those techniques as a group.
• PubMed search: “downward arrow technique”: Four records on October 1, 2026. Full texts add about 50 papers that mention it, mostly as one technique among many, plus training reports such as Gupta and Aman (2012). A phrase search misses other wording.

Read next

How to Fill Out a Thought RecordWhere the first thought gets written down"Should" Statements: The Tyranny of "I Have To"Rules that often sit at the bottom of the arrowYour inner critic: where the harsh voice in your head comes from and how to tame itWhen the belief underneath is about youCore beliefs: the old conclusions under your automatic thoughtsThe deep beliefs behind recurring thoughts

← All CBT guides