Every four-day pilot eventually produces a slide saying wellbeing improved. The credibility of that slide depends entirely on decisions taken before the pilot: which instrument, asked of whom, how often, and against what baseline.
The good news is that you do not need to invent a questionnaire. The pilot literature converged on a small set of validated short tools, and the strongest recent study — the 2025 Nature Human Behaviour trial analysis covering 2,896 staff with control companies — shows what a defensible battery looks like. This article walks the choices and the traps.
The shortlist that researchers actually use
WHO-5 Well-Being Index. Five questions, two minutes, free to use, scored 0–100. It is the workhorse of pilot evaluation: sensitive enough to detect change over six months, short enough that staff will still complete it at month six. If you adopt one wellbeing measure, adopt this.
Burnout scales. The Copenhagen Burnout Inventory and the burnout subscales of the Maslach tradition are the validated options; short forms exist for survey use. Burnout is the construct the pilots most consistently move — 71% of staff in the UK 2022 pilot reported reduced burnout — so measure it directly rather than inferring it from general satisfaction.
Sleep, fatigue and stress items. Pilots used validated sleep-problem and fatigue items alongside single-item stress questions. Germany’s 2024 pilot surfaced its most quoted people finding here: participants slept an average of 38 minutes more per week. Recovery is plausibly the mechanism behind the wellbeing gains, and the 2025 Nature analysis found exactly that — improvements were mediated by sleep, fatigue and self-rated work ability. Measure the mechanism, not just the mood.
Work ability and work–family conflict. A single-item Work Ability Index question (“how would you rate your current work ability compared with your lifetime best?”) featured in the Nature study and costs fifteen seconds. A short work–family conflict scale captures the care dimension that drives much of the retention story.
When to survey, and how not to spoil it
Survey at baseline (before the pilot, ideally four to eight weeks out), midpoint, end, and at six and twelve months after. Four waves in year one is plenty; monthly wellbeing surveys create fatigue that masquerades as decline. Keep pulse questions — two or three items on intensity and hours — on the shorter cycle, and keep them visibly separate from the formal waves.
Three honesty rules matter as much as the instrument:
- Measure actual hours in the same survey. The Nature study’s dose-response finding — larger real hour cuts, larger gains — means a wellbeing rise with no hours fall is a different phenomenon (novelty, gratitude, or pressure to please). Report the two together or believe neither.
- Guarantee and demonstrate anonymity. If staff suspect managers can see individual answers, you are measuring diplomacy. Small teams need results pooled or suppressed below a minimum group size.
- Publish subgroup cuts carefully. Breaking results down by caring status, contract type and role is how you find the people the average is hiding — a group whose intensity rose while everyone else recovered. But on small numbers, a “carers worsened 40%” line can rest on three people. Report counts with percentages, suppress tiny cells, and never let a dramatic percentage outrun its denominator.
Warning signs your measurement is flattering you
Wellbeing up while after-hours messages climb. Scores improving only in the wave immediately after announcement. Response rates collapsing to the enthusiasts. And the classic: average hours “down” on contracted time while nobody has checked timesheets, overtime claims or Friday log-ins since launch. Hidden overtime is the solvent of pilot evidence; the pilots that caught Friday working are remembered precisely because somebody looked.
One page of method, agreed in advance, is what separates a wellbeing result your board can rely on from a number your sceptics can dismantle in one question. Choose the validated tools, baseline them early, and keep hours on the same page as happiness.
Sources
- Fan, Schor, Kelly & Gu, four-day week trial study (burnout, satisfaction, mental and physical health; sleep and fatigue mediation) — Nature Human Behaviour, 2025.
- UK four-day week pilot people measures and results — Autonomy, University of Cambridge, Boston College and 4 Day Week Global, February 2023.
- Germany four-day week pilot sleep and wellbeing findings — University of Münster and Intraprenör, October 2024.
Day5Group is a consultancy that assists in the transition to a 4 day work week. More about Day5Group · All blog articles