Government · UK · Office for National Statistics
A £10 Incentive Lifts Under-30 Survey Response 11.5 Points
The ONS paid a £10 completion voucher to the groups least likely to answer its flagship social survey. Response among 16 to 29-year-olds rose 11.5 percentage points, from 22.0% to 33.5%.
Source published Summarised
Key findings at a glance
- 11.5 points percentage-point rise in survey response among 16 to 29-year-olds once a £10 conditional incentive was offered
- 22.0% response rate for 16 to 29-year-olds before the incentive was introduced
- 33.5% response rate for the same age group during the incentive trial
- 12.9 points rise among people from ethnic minority groups aged under 30, the largest gain of any subgroup
- 10.1 points rise in response among ethnic minority respondents overall, from 28.7% to 38.8%
- 10.8 points rise across the whole targeted group, from 25.3% to 36.1%
- 3.5 points rise in the survey's overall response rate, from 45.2% to 48.7%, despite only part of the sample being paid
- £10 value of the conditional voucher, paid only on completion, that produced these gains
- 2.8 points increase in the targeted group's share of all achieved responses, from 10.1% to 12.9%
- 1.60 odds ratio for response in the post-trial targeted group against the pre-trial baseline, significant at p below 0.001
- 341 additional respondents aged 16 to 29 gained across the six trial waves
- 186 additional respondents from ethnic minority groups gained across the same six waves
What the report covers
The Opinions and Lifestyle Survey (OPN) is the ONS monthly survey of Great Britain residents aged 16 and over, collected mainly through an online self-completion questionnaire. It is the source behind a large share of the UK public-attitude statistics that marketers quote, including our earlier report on how just 36% of UK adults think AI will benefit them. Like almost every survey, it has a representativeness problem: as of April 2025, people aged 16 to 29 and people from Asian, Black and Mixed ethnic groups were consistently under-represented among respondents, while over-50s and white respondents were over-represented.
Rather than accept that and weight around it, the ONS ran a pilot. From May to July 2025 it offered a £10 conditional voucher, payable only on completion, to a targeted group: anyone aged 16 to 29 and/or from an under-represented ethnic minority group. Everyone else in the sample kept the existing incentive, an unconditional A6 notepad sent with the invitation letter, and acted as the control. Six survey waves during the trial were compared against six waves collected from February to April 2025.
The effect was large and concentrated exactly where it was aimed. Response among the targeted group rose 10.8 percentage points, from 25.3% to 36.1%, while the untargeted control group barely moved. Ethnic minority respondents under 30, the hardest group of all to reach, improved most: up 12.9 percentage points, from 17.7% to 30.6%. The ONS recommendation is to make targeted incentives a standing part of the OPN, subject to ethics and budget sign-off.
What this means for UK marketers
Most UK brands run some form of voluntary research: NPS and CSAT surveys, post-purchase questionnaires, review requests, panel recruitment, user-testing sign-ups. Nearly all of them suffer the same skew the ONS documented, and nearly all of them ignore it. If your customer survey is answered disproportionately by older, settled, already-loyal customers, then every persona, every message test and every product decision downstream inherits that bias. You are not hearing from the customers you most need to win.
The useful finding is not that incentives work, which was already well established. It is that targeting them works, and works cheaply. The ONS did not pay everybody. It paid only the people it was short of, which is why an intervention reaching a minority of the sample still lifted the overall response rate by 3.5 percentage points. The commercial translation is direct: if your under-35 segment responds at half the rate of your over-50s, a modest completion-contingent reward aimed only at that segment will buy you a more representative sample for far less than an across-the-board prize draw. Conditional beats unconditional here too, since the money only leaves the business when a completed response arrives.
Two operational notes. First, the £10 was conditional on completion, so the incentive economics are tied to finished responses rather than clicks, which is the same logic that should govern any measured on-site journey. Getting that journey to convert once someone has agreed to engage is the discipline behind our conversion rate optimisation work, and a survey funnel leaks in exactly the same places a checkout does. Second, differential rewards carry a fairness question. The ONS took it seriously enough to get approval from the National Statistician's Data Ethics Advisory Committee, monitor social media for backlash and brief its enquiry line in advance. It recorded no negative feedback. A brand offering one customer segment £10 and another nothing should expect to explain itself, so decide in advance what the public answer is and whether the differential is disclosed.
The honest caveat: this measured whether people respond, not whether the people brought in by money answer differently from those who would have responded anyway. Paying for participation reduces one bias. Whether it introduces another is still open.
How they got the data
This was a quasi-experimental pilot, not a randomised controlled trial. The ONS considered an RCT and ruled it out because the OPN sample size meant a full trial would have taken longer than the budget and timetable allowed. Instead it used a difference-in-differences design, comparing the change in the targeted group against the change in an untargeted control group of white respondents aged 30 and over, over six waves before the trial (February to April 2025) and six waves during it (May to July 2025). Pre-intervention trends for the two groups were checked and found to be parallel, which is the assumption the method rests on.
Findings were confirmed with a logistic regression controlling for age and ethnicity, modelling whether an individual responded. The post-trial targeted group showed an odds ratio of 1.60 against the pre-trial baseline at p below 0.001, and predicted response rates matched observed rates to within a fraction of a point, with confidence intervals under 0.5%. The ONS flags one real limitation: OPN respondents are drawn from people who already completed the Transformed Labour Force Survey and were paid £10 for it, so some may have arrived expecting payment, which muddies attribution. Fuller detail on the survey design sits in the ONS quality and methodology information for the OPN.
Government · UK
Read the full report
From Office for National Statistics · published 27/07/2026