Phase 1: Prepare
Step 1 Warm-up recall
From M7.L4: write the no-effect value for d and for an odds ratio.
Private: nothing you type here is sent or stored.
Step 2 Learning goal
By the end of this lesson you can:
- report the design and findings of Lally and colleagues (2010) (Foundation);
- explain why a range and a median say more than one number (Core);
- convert days into weeks (Core).
Phase 2: Teach
Step 3 A familiar example
A friend says “it takes exactly a month to make a habit”. Another says it took them half a year. A single number hides a lot of variety.
Step 4 Clear explanation
Lally, van Jaarsveld, Potts and Wardle (2010) asked volunteers to choose a new eating, drinking or activity behaviour and to do it daily in the same context for 12 weeks (84 days). They recruited 96 volunteers and 82 provided enough data for analysis. Each day participants reported how automatic the behaviour felt. The authors modelled how automaticity rose and levelled off, and reported that the time to reach a plateau ranged from 18 to 254 days across participants. The results section reports a median of 66 days for the 39 participants whose data fitted the model well. The abstract also states that missing one opportunity to perform the behaviour did not materially affect the habit-formation process.
What can you take from this? Habits did not form on a fixed schedule: some people reached a plateau in about two and a half weeks, others took much longer than the 12 weeks observed, and the 254-day end of the range comes from fitting a curve beyond the observed period. A median of 66 days is about 9.4 weeks. Your own four weeks may be too short to see a habit form, which is why this course tests whether the habit helps, not whether it has become automatic.
Limits: volunteers chose their own behaviours, the sample is small, and the median rests on 39 people with a good model fit.
Step 5 Worked example (4 marks)
A blog says “science shows habits take 66 days”.
- 66 is the median among 39 participants with a good model fit. (1 mark)
- The range across participants was 18 to 254 days. (1 mark)
- 66 days ÷ 7 = 9.4 weeks. (1 mark)
- Fair statement: in one study, the median time to plateau was 66 days, with wide variation; missing one day did not materially affect the process. (1 mark)
Phase 3: Practise
Step 6 Learner practice (instant-marked)
2. How many volunteers did Lally and colleagues (2010) recruit and how many provided sufficient data?
Ninety-six were recruited and 82 provided sufficient data.
3. What range of days to reach a plateau did the abstract report?
The abstract gives a range of 18 to 254 days.
4. What did the abstract say about missing one opportunity to perform the behaviour?
Missing one opportunity did not materially affect the process.
Phase 4: Close
Step 7 Recap
- Lally and colleagues (2010): 96 recruited, 82 analysed, 12 weeks; time to plateau ranged from 18 to 254 days.
- Median 66 days among 39 participants with a good model fit.
- A missed opportunity did not materially affect habit formation.
Step 8 Exit ticket (ungraded)
Write one sentence reporting the time habits take without using a single number.
Private: nothing you type here is sent or stored.
Step 9 Go further (Extension)
Read the Lally and colleagues (2010) abstract in the European Journal of Social Psychology. Find what the authors say about the shape of the curve.
Common mistakes
| Mistake | Why it fails |
|---|---|
| Quoting 66 days as a rule | It is a median in one study, with a wide range. |
| Believing a missed day ruins a habit | The study found one missed opportunity did not materially affect it. |
| Treating a short experiment as proof a habit has formed | Four weeks may be too short. |
Limits. Figures come from the paper’s abstract and results section. The median depends on the model fit and the 39 participants who fitted it well.