Why do I keep doing things I do not want to do?

Daniel Gavalec · Updated October 23, 2026

Repeating something you regret can reflect a familiar cue, an immediate benefit and a response you have practised many times. Start by mapping one recent situation and changing the setup before it happens again. You do not need a new identity or a perfect streak to run a useful experiment.

Why your intention and your next action can disagree

You decide on Sunday to stop checking messages while studying. On Tuesday, a difficult paragraph appears and your hand reaches for the phone before you finish deciding. The resolution was real, and so was the learned response. Wood and Rünger's review connects habitual action to recurring contexts. A familiar setting can matter even when your long-term preference has changed.

This is not proof of “dopamine addiction”, a damaged attention span or a hidden desire to fail. Repetition alone does not tell you whether something is a habit, a response to distress or part of a clinical problem. Begin with an observable sequence: what happened, what you did, what happened immediately afterwards. Leave personality labels out of the first draft. They tend to close the investigation before it becomes useful.

Use the habit loop, then notice its limits

The cue–routine–reward model can organise a notebook page. “Saw a difficult email; opened a game; felt a brief escape” is a workable sketch. But the sketch is not a brain scan. You might be wrong about the cue or reward. The behaviour may continue without much pleasure, and several causes can operate at once. Treat your explanation as a hypothesis you can update.

Ask two different questions. What made the behaviour easy to start? What made the alternative hard? If the phone is beside your keyboard and the next work step is unclear, both deserve attention. Move the phone and specify the task. You do not have to discover one deep emotional reason before making a practical change. Equally, a tidier desk will not resolve every source of distress.

Your first ten minutes: build a small experiment

For two minutes, describe one recent event as if a camera recorded it. Use a place, an action and a rough time: “After lunch at the kitchen table, I opened the feed before returning to the essay.” For the next two minutes, identify the immediate benefit you were seeking: a break, stimulation, reassurance or avoidance of uncertainty. A guess is enough.

Spend three minutes changing the environment. Charge the phone elsewhere, close an unnecessary tab or put the first worksheet on the desk. Then spend two minutes choosing an alternative that is easy enough to do in that exact situation. A proper break may fit better than another productivity rule. Use the final minute to write when you will review the experiment. You are testing a setup, not deciding your worth.

Choose a replacement that serves the same moment

Consider Sam, a freelancer who opens entertainment tabs after sending a difficult proposal. The original response marks the end of an uncomfortable task. Replacing it with another difficult task ignores the need for a transition. Sam tries standing up, getting water and taking a short break away from the desk. The workday plan includes the break rather than treating it as stolen time.

A different example is a student who checks messages because they do not know how to begin revision. Their replacement may be opening one practice question and attempting a rough answer. Similar behaviour can need different alternatives. Write the smallest action that meets your own situation. If your experiment fails repeatedly, ask whether you chose a response that sounds virtuous but does not provide anything you actually needed.

Turn the decision into an if–then plan

Gollwitzer and Sheeran's meta-analysis supports specifying a situation and a response in advance. Write “When I finish lunch, I will leave my phone charging and open one question.” Avoid “When I lack discipline, I will be disciplined.” The cue should be recognisable and the response observable. Keep the plan available where the decision happens, rather than in a notebook you never open.

Prepare a fallback as well: “If I need the phone for study materials, I will open that document before I begin and close the feed.” Real life contains exceptions. A useful plan accommodates them without silently dropping the goal. After several relevant opportunities, review whether the cue actually occurred, whether the response was available and whether the alternative was tolerable. Change one part so you can tell what helped.

How long should change take?

Lally and colleagues studied the development of automaticity in new everyday behaviours. Their findings do not supply a personal deadline for quitting an old behaviour or recovering from addiction. A widely repeated number of days can become another way to judge yourself. Instead, ask whether starting the alternative is becoming easier and whether the unwanted behaviour is interfering less with what matters to you.

Count opportunities as well as outcomes. If the situation arose twice this week, a daily grid may tell a misleading story. A practical weekly review could include what you tried, whether it was possible and what you want to change. You can stop an experiment that has become an elaborate bookkeeping task. The point is a more workable life, not maintaining a perfect record of attempts.

Common mistakes, and the day after a lapse

Changing five behaviours at once makes it difficult to see what works. Waiting for the urge before preparing the alternative adds another decision at the hardest moment. Publicly declaring a permanent transformation can make an ordinary setback feel like humiliation. Keep the experiment small, specific and revisable. You can ask a trusted person for a check-in without turning them into an inspector.

After a lapse, record the sequence and repair any practical consequences. Breines and Chen link self-compassion with motivation to improve after mistakes. Translate that into a sentence and an action: “I had a rough evening; tonight I will prepare a real break.” With substances or serious harm, NIDA describes treatment as a continuing process. Do not use a habit worksheet to postpone contacting a professional.

How CORTEX helps you run the experiment

Create one replacement in Habits and choose a weekly target that matches the available opportunities. Daily habits have rest-day handling; an explicit slip is not silently changed into rest. Numeric habits can keep partial entries. Treat these records as practical information, not a recovery score or evidence that your body is safe.

Use your plan to schedule the preparation task, for example moving the charger. Use Focus for a bounded attempt at the work you were avoiding, and the journal's evening reflection for a short review. CORTEX does not detect your triggers, restrict other apps or treat compulsions. You enter the information yourself. Continue through habit building, urge surfing and the guide hub as your next question changes.

Print your behaviour experiment

Keep the original hypothesis when you review the sheet. Add what actually happened beside it. A mismatch is useful: perhaps the supposed evening habit really began when you skipped the afternoon break. Prepare the next version around that earlier moment rather than adding a harsher rule at bedtime.

This is general information, not medical advice. Persistent distress, lost control or harm from substance use deserves professional support, even when you can explain part of the pattern yourself.