Articles

A thought record is six boxes and a pen. It is also the single most abandoned exercise in self-help CBT, usually within four days.

Part of that is because it gets taught badly — as a worksheet you fill in correctly, rather than an experiment you run on yourself. Here is how it actually goes, including the bits that go wrong.

The six columns

Beck’s original version had five substantive columns [1]. Greenberger and Padesky later split the evidence step in two and introduced the idea of the hot thought — the one thought in the pile that is actually carrying the feeling [2]. That refinement is the useful one, so this version keeps it.

1. Situation

Where were you, who with, what was happening? Facts only, one line. “Sunday, 9pm, reading work email in bed.”

2. Emotions

One word each, and a number out of 100. “Anxious 80. Ashamed 60.” The number matters more than it looks — it is the before-reading, and without it you cannot tell whether anything moved.

3. Thoughts — then find the hot one

Write everything that went through your head. Then read back and circle the one that makes your stomach drop. That is the hot thought and it is the only one you work on. Most failed thought records are people diligently arguing with a lukewarm thought while the hot one sits there untouched.

4. What supports it

Evidence, not feelings. Things a camera would have recorded. “He replied in four words” counts. “He was clearly annoyed” does not — that is already an interpretation.

5. What does not support it

Same standard. This is the hard column and the one people skip. Useful prompts: what would I tell a friend who said this? Have I been wrong about this kind of prediction before? What am I leaving out because it does not fit?

6. Where that leaves you

Write a sentence you can actually believe, not a nicer one. Then re-rate the emotions. If anxious went from 80 to 65, that is the exercise working. It is not supposed to go to zero.

A worked example

Situation: Sunday 9pm, reading a short reply from my manager. Emotions: anxious 80, ashamed 60. Hot thought: “She thinks I’m not up to this.” Supports it: the reply was four words with no greeting. Does not support it: she writes like that to everyone; she asked me to run the client call on Thursday; I have had no critical feedback in eleven months; it was 9pm on a Sunday and she was probably on her phone. Where that leaves me: “A short email on a Sunday night is weak evidence about what she thinks of my work.” Re-rated: anxious 45, ashamed 30.

Nothing was solved. The point is that 80 became 45 on the strength of evidence you already had.

What the evidence says about this exercise

Thought records are standard, widely taught and belong to treatment packages that clearly work. Whether the thought record is the ingredient doing the work is a different question, and the answer is uncomfortable.

A 2021 component analysis of 76 randomised trials of internet CBT, covering 17,521 people, found cognitive restructuring contributed no detectable independent benefit, while behavioural activation clearly did [3]. In a head-to-head study, behavioural experiments — going and finding out — produced more belief change than thought records [4]. Across anxiety disorders, cognitive therapy does not outperform exposure except possibly in social anxiety [5]. What does track with outcomes is simply doing the homework at all, consistently [6].

Read that as a steer rather than a reason to stop: a thought record that ends in a small experiment is worth more than one that ends in a tidier paragraph.

When it does not work

You picked the wrong thought. Nothing shifts because the hot one is still unexamined.

You wrote a question. “What if I get fired?” cannot be tested. Turn it into a statement first: “I am going to get fired.”

You are a worrier and the worries multiply. If every worry you examine spawns two more, thought records can feed the machine rather than starve it. That pattern needs a different approach — and probably a person.

It is a belief, not a thought. “I am fundamentally unlovable” will not move on four lines of evidence, and failing to shift it can confirm the very thing you are afraid of.

You know it but do not feel it. Extremely common, not a failure. Belief change tends to lag behind the logic by some distance.

You are doing it in a cruel voice. A technically correct balanced thought delivered with contempt does not land.

Doing this alone

An honest note about self-help. In a network meta-analysis of 36 trials with 8,107 people, guided internet CBT beat unguided — and the gap depended on severity. At mild symptom levels the two were roughly equivalent; above moderate, guided was clearly better [7]. Standalone apps produce real but modest effects on depression and anxiety [9], and in the wild most people stop using them within a month [10].

The reassuring finding: an analysis of 13 trials found people using self-guided internet CBT were less likely to deteriorate than controls, not more [8]. Doing this on your own does not appear to make things worse. It just works better with someone alongside you.

Where MoodFire fits

Reframe is a thought record with the friction removed — it prompts for the hot thought rather than letting you circle a safe one, and it keeps the before-and-after ratings so you can see whether anything actually moved. MoodFire is not therapy and is not a substitute for it.

If the thoughts you are recording are about harming yourself, stop and talk to someone instead — a doctor, a crisis line, someone you trust. A worksheet is the wrong tool for that, and you deserve a person.

Otherwise: one record, six boxes, five minutes. Then go and test the prediction. That last part is the one the evidence keeps pointing at.