“Walk in to Work Out”: a randomised controlled trial of a self help intervention to promote active commuting

The researchers tested a single intervention: the “Walk in to Work Out” pack. This was a self-help booklet containing:

Interactive worksheets based on the **transtheoretical model** (stages of change: precontemplation, contemplation, preparation, action, maintenance)

Local maps showing safe walking and cycling routes to the three workplaces

Distance charts from various neighbourhoods to each workplace

Safety information (e.g., how to cross busy roads, visibility tips for cyclists)

Goal-setting and self-monitoring pages (e.g., “Plan your walk for next week”)

Tips for overcoming common barriers (bad weather, lack of time, tiredness)

The **control group** received no materials at baseline — they were promised the pack after six months (a delayed-intervention control). This design tests whether simply giving people a well-designed, theory-based booklet can change commuting behaviour, without any coaching, classes, or financial incentives.

**Primary outcome:** Change in self-reported frequency of walking or cycling to work (days per week) at six months.

**Secondary outcomes:** Proportion of participants who became “regular active commuters” (defined as walking or cycling to work at least three days per week) at six and twelve months; changes in total physical activity; and qualitative feedback from focus groups.

**295 employees** from three workplaces in Glasgow, Scotland (a university, a hospital, and a local council office)

All participants had been **pre-screened** as being in the “contemplation” or “preparation” stage for active commuting — meaning they were thinking about walking/cycling to work, or doing it irregularly (less than three days per week). People who already walked/cycled regularly were excluded.

**Age range:** 18–65 years (mean ~40 years)

**Gender:** Approximately 70% female (typical for these workplace settings)

**Distance to work:** Ranged from <1 mile to >10 miles; median distance ~3–4 miles

**Employment type:** Office-based staff (administrative, academic, healthcare)

**Car ownership:** Not reported, but Glasgow has moderate public transport and many employees lived within walking/cycling distance

**Why this matters for self-experimenters:** This is a motivated-but-not-yet-active population — exactly the kind of person who might try a self-help intervention. They were not athletes, not already committed, and not paid to participate.

**Self-reported commuting behaviour:** Participants completed a short questionnaire at baseline, 6 months, and 12 months. They reported how many days per week they walked, cycled, drove, or took public transport to work. The key metric was “days per week of active commuting” (walking + cycling combined).

**Stage of change:** A validated algorithm classified participants into one of five stages (precontemplation, contemplation, preparation, action, maintenance) for both walking and cycling separately.

**Focus groups:** At six months, a subset of intervention participants (n=~30) attended focus groups to discuss what they liked/disliked about the pack and what helped or hindered them.

**No objective measures:** No pedometers, accelerometers, GPS trackers, or fitness tests. No verification of self-reported walking/cycling (e.g., no employer swipe-card data or direct observation).

**Measurement weakness:** Self-report is notoriously unreliable for physical activity — people tend to overestimate. The authors acknowledge this but argue that any over-reporting should be similar in both groups (non-differential misclassification), which would bias results toward the null (making it harder to find an effect). So the true effect might be even larger than reported.

**Design:** Randomised controlled trial (RCT) with two parallel groups. Participants were individually randomised (not by workplace) using a computer-generated random number sequence. Allocation was concealed (the researcher assigning participants did not know which group the next person would go into).

**Blinding:** This was **single-blind** — the researchers analysing the data did not know group assignments, but participants obviously knew whether they received the pack or not. This is unavoidable for a behavioural intervention (you cannot give someone a booklet and pretend they didn’t get it). However, the lack of participant blinding means **expectation effects** could inflate the results: people who got the pack might report more walking simply because they felt they “should” be doing it.

**Duration:**

Baseline assessment (questionnaire)

Intervention group received pack immediately

6-month follow-up (primary endpoint)

12-month follow-up (maintenance check — control group had received the pack at 6 months, so by 12 months both groups had the intervention)

**Statistical approach:**

Primary analysis: logistic regression to compare the odds of increasing walking (any increase vs. no increase) between groups at 6 months. Adjusted for baseline walking frequency, distance to work, age, and gender.

Secondary analysis: chi-square tests for proportions of “regular active commuters” at 6 and 12 months.

Intention-to-treat analysis: all participants were analysed in their original groups, even if they didn’t use the pack or dropped out. Dropout was ~15% at 6 months and ~25% at 12 months, which is moderate.

**What this design can prove:**

**Causality:** Because of randomisation, any difference between groups at 6 months can be attributed to the pack (assuming no major confounds). This is the strongest design for proving that an intervention *causes* a behaviour change.

**Real-world effectiveness:** The study was conducted in actual workplaces with real employees, not a lab. Results are likely generalisable to similar office-based populations.

**What this design cannot prove:**

**Why it worked:** The pack contained many components (maps, worksheets, safety info). The design cannot tell you which component was crucial. Was it the goal-setting? The maps? The “stage of change” tailoring? You don’t know.

**Long-term maintenance beyond 12 months:** The 12-month data is confounded because the control group got the pack at 6 months. We don’t know if effects last 2+ years.

**Generalisability to non-motivated people:** Participants were already “contemplating” active commuting. The pack might fail completely in people who have no interest in walking.

**Cycling-specific barriers:** The pack failed for cycling, but the design cannot tell you whether that was because the pack was poorly designed for cyclists, or because environmental barriers (traffic, hills, lack of showers) are simply too strong for a booklet to overcome.

**Major methodological weaknesses:**

1. **Self-report only** — no objective verification of walking/cycling

2. **No blinding of participants** — expectation effects possible

3. **Moderate dropout** (25% at 12 months) — if dropouts differed between groups, results could be biased

4. **Single city (Glasgow)** — weather, infrastructure, and culture may limit generalisability

5. **No cost-effectiveness analysis** — we don’t know if the pack is worth the printing/distribution cost

**Primary outcome (6 months):**

The intervention group was **1.93 times more likely** to have increased their walking to work compared to the control group (odds ratio = 1.93, 95% confidence interval: 1.06 to 3.52, p = 0.03).

In absolute terms: **25% of the intervention group** increased walking (any amount) vs. **15% of the control group** — an absolute difference of 10 percentage points.

**Secondary outcomes:**

**Regular active commuting (≥3 days/week walking or cycling) at 6 months:** 25% of the intervention group (95% CI: 17% to 32%) vs. 15% of controls (95% CI: 9% to 22%). This difference was statistically significant (p = 0.04).

**Regular active commuting at 12 months:** 25% of the intervention group (who got the pack at baseline) were still regularly active commuting. The control group (who got the pack at 6 months) had caught up to ~22% by 12 months — suggesting the effect is reproducible.

**Cycling:** No significant effect. Only ~3% of participants in either group reported cycling regularly at any time point. The pack did not increase cycling.

**Distance to work:** No interaction — the pack worked equally well for people living 1 mile away and people living 5 miles away (though very few people >5 miles walked).

**Gender and age:** No significant differences — the pack worked similarly for men and women, and across ages 20–60.

**Focus group findings (qualitative):**

Participants valued the **local maps** most — they reported not knowing safe routes before.

**Goal-setting worksheets** were used by about half of participants; those who used them reported higher success.

**Barriers to cycling** included: traffic danger, lack of showers at work, hills, and needing to carry work clothes.

**Weather** was the most common excuse, but many participants said the pack’s “bad weather tips” helped them overcome it.

The effect is **modest but meaningful** at a population level:

**Number Needed to Treat (NNT):** To get one additional person to start walking to work regularly, you would need to give the pack to about **10 people** (absolute risk reduction = 10%, so NNT = 1/0.10 = 10). This is comparable to many public health interventions (e.g., smoking cessation leaflets have NNTs of 20–50).

**Effect size in Cohen’s d:** Not reported, but from the odds ratio of ~1.9, the effect is small-to-medium (d ≈ 0.3–0.4). This means the average person in the intervention group walked about **0.5–1 more day per week** than the average control.

**Practical translation:** If you are already thinking about walking to work, a well-designed self-help pack can tip you from “thinking” to “doing” — but it won’t turn a non-walker into a daily walker overnight. The effect is roughly equivalent to having a friend who walks to work and encourages you to join them.

**Acknowledged by authors:**

Self-report bias (over-reporting of walking)

Lack of blinding

Only three workplaces in one city (Glasgow)

Low cycling rates made it impossible to detect cycling effects

The pack was tailored to Glasgow — maps and routes would need localisation

**Additional critical notes:**

**No measure of total physical activity:** Did people who walked to work simply walk less elsewhere? The study did not measure whether total daily steps increased or whether participants compensated by being more sedentary at other times.

**No dose-response analysis:** The pack was given as a single dose. Would a booster pack at 3 months have helped? Unknown.

**Selection bias:** Participants volunteered for the study after seeing recruitment posters. They were likely more motivated than the average employee. The pack might fail in a less motivated population.

**No cost data:** The pack cost money to print and distribute. Is it cost-effective compared to, say, a free smartphone app or a workplace walking group? Unknown.

**Publication date (2002):** This study is over 20 years old. Commuting patterns, smartphone use, and cycling infrastructure have changed dramatically. The results may not replicate today, especially for cycling (which has become more popular in many cities).

For someone running their own n=1 experiment to increase active commuting:

### What to test

**The intervention:** Create your own “Walk in to Work Out” pack. This means:

1. **Map your route** — find the safest, most pleasant walking or cycling path from home to work. Use Google Maps, local cycling apps, or walk the route yourself. Note distance and estimated time.

2. **Set a specific goal** — e.g., “I will walk to work at least 3 days this week” (not “I’ll try to walk more”).

3. **Identify barriers** — write down what stops you (weather, tiredness, time pressure) and plan solutions (e.g., “If it rains, I’ll use a waterproof jacket and pack my work shoes”).

4. **Self-monitor** — keep a simple log of days you walked/cycled vs. drove/took transit.

**Dose:** One pack (or self-created equivalent) at baseline. No booster needed for at least 6 months based on this study.

### Minimum meaningful duration

**Run the experiment for at least 6 weeks** to see if the habit sticks. The study found effects at 6 months, but behaviour change often happens within the first 2–4 weeks. If you haven’t increased walking by week 4, the pack alone may not be enough for you.

**For maintenance:** Track for 12 months. The study showed that 25% of people who started were still going at 12 months — so if you’re still walking at 3 months, you have a good chance of making it a long-term habit.

### What to measure

**Primary metric:** Number of days per week you walk or cycle to work (not total steps, not minutes — just commute days). This is what the study used.

**Secondary metrics:**

- Total weekly physical activity (use a pedometer or smartphone step counter to see if you’re compensating)

- Subjective energy/mood (rate 1–10 each morning after commuting)

- Time saved (compare total commute time walking vs. driving/transit — walking often takes longer but combines exercise with travel)

**Track for at least 2 weeks at baseline** before starting the intervention, then continuously for 6–12 weeks.

### Key confounds to control for

**Weather:** Glasgow is rainy; your city may be different. If you start in summer, you might overestimate your ability to walk in winter. Run the experiment across at least one season change.

**Work schedule:** If you have early meetings or late shifts, walking may be harder. Note your work schedule each day.

**Car access:** If you have a car, the temptation to drive on cold/rainy days is high. Consider leaving your car at home or parking it far away.

**Social support:** The study didn’t test this, but having a walking buddy or telling a colleague about your goal may boost adherence.

**Injury/illness:** If you get injured, pause the experiment and restart when healthy. Don’t push through pain.

### What a positive result would look like

**Small win:** You increase from 0 days/week walking to 1–2 days/week. This matches the study’s effect size (OR ~1.9).

**Big win:** You reach 3+ days/week (the study’s definition of “regular active commuting”). At this level, you’re likely getting significant health benefits (reduced cardiovascular risk, weight maintenance, improved mood).

**Negative result:** No change after 6 weeks. This doesn’t mean the approach is useless — it may mean you need a stronger intervention (e.g., a walking group, financial incentive, or workplace shower facilities). The study’s pack worked for 25% of people, not 100%.

**Bottom line for self-experimenters:** This is a low-cost, low-risk intervention to try. Spend 30 minutes creating your own “pack” (map, goal, barrier plan, log sheet). Commit to 6 weeks. If you see even 1 extra walking day per week, that’s a meaningful win — and you’ll have replicated a published RCT on yourself.