Scope and Approach
About the survey
About the survey
Additional documents
A PDF version of this report is available to download here.
The data tables provide the full results for each question asked in the PAS Survey 2025. They also provide a breakdown of results by key subgroups.
This Technical Report provides details of the PAS Survey 2025
conducted by Ipsos on behalf of UKRI. The report will be available soon.
Background
Over the last 25 years, the Public Attitudes to Science (PAS) studies have provided data on what people in the UK think about science, scientists and science policy, and how informed and engaged they feel when it comes to science issues. This is a rich and varied dataset, of use and interest to scientists, science communicators and others working in public engagement, academics, advocacy groups and policymakers across government.
This edition, the seventh under the PAS banner, was undertaken by Ipsos, in partnership with the British Science Association, on behalf of UK Research and Innovation (UKRI), the UK’s national funding agency for science and research. It is the first of three surveys that will be carried out biennially for UKRI, providing more regular data than was previously available, to inform relevant discussions across government and wider society. It is also notably the first edition to take place since the COVID-19 pandemic and since the widespread integration of AI tools in public life.
PAS 2025 covers new ground, while maintaining the same core questions that have been asked previously. Three major new areas of exploration include representation in science, how people seek and receive information on science, and comparing “science” to “research and innovation”, to see if the latter term evokes different considerations among the public. Several questions were asked to half the sample as being about “science” and to the other half as “research and innovation”. Ultimately these two framings received very similar responses.
This report covers the headline findings from PAS 2025. It focuses on a subset of the questions that best reflect the emerging narrative from the wider data, as well as the questions that were wholly new for this year. The data tables, providing access to the full question set, are downloadable for readers to interrogate and quote. The full questionnaire is also included in the Technical Report. Other outputs are also available for download, complementing this report.
This edition shifts from a face-to-face survey to a predominantly online one, known as a “push-to-web” methodology. This means participants were chosen at random and sent multiple postal invites with details of how to take part in the survey online. Those that did not complete an online survey were sent a paper version in the final mailing. The change in the methodology futureproofs the survey. It means that later editions can be undertaken and reported more quickly, with a much bigger sample size than before.
In total, 5,281 UK adults aged 16 or over completed a survey between 6 February and 4 July 2025. Of these, the majority, 4,647, took part online and the remainder used the paper version. The data has been weighted to be representative.
Previous editions have typically included qualitative insights alongside the quantitative survey data, whereas this and upcoming ones will focus on the survey. This will ensure more consistent data and complement UKRI’s wider programme of qualitative research.
Full details of the sampling, fieldwork and weighting can be found in the Technical Report.
How to interpret this report
Here are the key things to bear in mind when reading through this report, with more comprehensive details in the Technical Report:
How to interpret this report
Here are the key things to bear in mind when reading through this report, with more comprehensive details in the Technical Report:
Margins of error – the survey findings are based on a sample of UK adults rather than the whole population, so are subject to margins of error. For questions based on the full sample, the margin of error is approximately 2 percentage points. Some questions were asked of a random half of the sample, and the margin of error for these is approximately 2-3 percentage points, depending on the finding. Margins of error for subgroup findings (e.g., for specific age groups or by gender) will be higher.
Rounding of percentages in text and charts – we only report percentages rounded to the nearest whole number. In addition, some charts exclude certain response categories (e.g., those saying “don’t know”) for simplicity. Therefore, the percentages shown will not necessarily add to 100%.
Trends over time – the PAS series has undergone substantial changes in the methodology across years. Since 2011, it has used a random probability sampling methodology, for greater representativeness. This edition moves from face-to-face to predominantly online data collection. This means we must be cautious about one-off, unexplained changes in findings over time. In this report, we aim to contextualise all changes from 2019 to 2025 in terms of the longer-term trend where possible.
Subgroup differences – when we compare a subgroup (e.g., 16 to 24 year-olds) to the average, we only report differences that are statistically significant, i.e., unlikely to have occurred by chance. We have focused on differences that are meaningful in size, thematically consistent across multiple questions, or important in the context of a single question. Differences by country or region are rarely mentioned, because they did not typically meet these criteria. The differences by gender focus on women and men, rather than other gender categories (e.g., non-binary) that had relatively small sample sizes. The full data tables and raw data will be available to find specific subgroup percentages.
The science capital index – this index was developed for the 2019 study and has been maintained for PAS 2025. It brings together data across multiple questions in the survey, to be able to segment people into three groups – high, medium and low “science capital”. The more science capital someone has, the more they have interacted with science or scientists in their day-to-day lives, either through studying, careers, leisure activities, or friends and family. This is more nuanced than science literacy, i.e., simply knowing about scientific processes and facts. Across this report, we look at how those with high science capital were different to those with low science capital, and the final chapter brings together this content. This helps to demonstrate how a more rounded experience of science can deepen engagement.