Guest blog by Paul Allin, RSS Honorary Officer for Public Statistics
“You have to be counted to count”.
Survey response rates have been declining, in a trend accelerated by the Covid-19 pandemic and particularly acute in the UK. This poses a threat to data representativeness and quality.
The ONS has a critical role to play in producing high-quality economic statistics, underpinned by data from social and business surveys. The issue of declining response rates and representativeness has been highlighted in recent reviews of the UK statistical system, prompting the ONS to launch a new Survey Improvement and Enhancement Plan.
In the summer of 2025, the Royal Statistical Society (RSS), the Office for National Statistics (ONS), and the Office for Statistics Regulation (OSR) joined forces to convene a roundtable to consider this pressing issue facing our statistical system. This was the first in a planned series of collaborative roundtables aimed at strengthening the UK statistical system for the public good.
The roundtable brought together leading voices from across the sector – commercial organisations and academia as well as ONS and OSR, to share insights and explore innovative solutions to falling response rates. Today we are publishing a write-up of the roundtable, and in this blog we highlight key messages from the discussions.
Understanding non-responders is crucial
One of the most urgent issues is the difficulty in reaching certain groups who are typically underrepresented in surveys, eg young people and young males specifically. The roundtable discussion emphasised the need for both qualitative and quantitative research to understand how best to overcome these challenges. This includes a need to trial different methods to motivate people to respond to surveys, for example by highlighting the benefits of responding, as well as exploring how best to build trust and awareness, and tailor communication methods and style to increase responses.
A key area for improvement was working to reduce response burden and fatigue – this could include a role for the ONS to work collaboratively with other organisations to coordinate surveys, including considering how to shorten questionnaires. Building an evidence base of ‘what works’ to increase response rates would be valuable.
Innovative statistical techniques
To help improve survey representativeness, roundtable discussions highlighted an important role for using ONS information to aid sampling, for example to identify areas where under-represented groups are over-represented. Discussions also focussed on looking at differences between responders and non-responders using paradata (data that tells you about how the data was collected).
The roundtable discussions highlighted that statistical methods alone cannot fix a broken data collection system. However, they can help make the most of imperfect data. Key recommendations included continuing work on exploring alternative data sources (eg administrative data) and focussing efforts on incorporating additional data sources, for more refined weighting.
Reflections on the ONS Survey Improvement Plan
The ONS’s recently published Survey Improvement and Enhancement Plan was welcomed for its ambition and transparency. However, participants noted that successful implementation will require investment in technology, citizen engagement, and workforce development. Methods to increase the effectiveness of the workforce were discussed, including career pathway development for interviewers.
We also discussed the importance of coordination across surveys to ensure that interventions that boost response rates for some surveys do not negatively impact others.
Collaboration is key
We are pleased to be working with the ONS and the OSR on this – the challenges we face with survey representativeness are system-wide, and solving them will require coordinated action across the statistical community. We look forward to working with ONS, other government statisticians and experts across the wider statistical system, including those who attended this roundtable.
We also look forward to continuing this partnership to address other critical statistical issues at future roundtables. Watch this space for future roundtable topics to be announced, to be held over Autumn and Winter.
Take a look at the full report, including ONS and OSR reflections.
If you’d like to get in touch with us about this work, please email policy@rss.org.uk.