What does a medical records database in England have to do with genetically modified crops or climate change? Gideon Skinner and Jayesh Navin Shah of Ipsos MORI summarise a comprehensive new study on people's attitudes to science and scientists in the UK, and the lessons for policy-makers and researchers.
In January of this year, the National Health Service in England announced plans to upload the medical data held by GPs and hospitals to a single repository called Care.data, intended to improve information sharing across the service. One month later, after some vocal opposition and public confusion about the scheme, the project was delayed for six months to allow more time for debate and discussion with patients and GPs.
At a time when revelations about mass surveillance programmes and phone hacking are hot news, this response might have been expected; that it was not is perhaps an indication that we have failed to learn the lessons of seemingly unrelated but similarly “hot button” issues like genetically modified (GM) crops or climate change – topics where the weight of scientific opinion is not always reflected in the public discourse.
Why do these disconnects occur between scientific knowledge and public understanding? Sometimes it is to do with levels of trust (or the lack of it); other times it may come down to the inability of scientists to communicate in ways that resonate with people.
There is a big distinction between knowing scientific facts and understanding how science is achieved
Since 1988, the UK government has sought to keep tabs on this relationship through the Public Attitudes to Science (PAS) studies. The 2014 study1, carried out by Ipsos MORI, comprised a nationwide survey of adults aged 16 and over, between July and November 2013, as well as qualitative research and the monitoring of open conversations about science on social media, blogs and other websites. Several qualitative “public dialogue” workshops were also carried out for the UK's Economic and Social Research Council and Office for National Statistics, looking specifically at attitudes towards the linking together of large government administrative datasets to supplement the survey's findings in this area2 (see ‘Attitudes to big data', page 16). The findings from these pieces of research provide not only insights into issues like government data linking, but also more general lessons on how complex science and social science topics can be better communicated to a lay public.
But why should attitudes matter? Isn't science about right and wrong answers, not degrees of opinion? Shouldn't we be measuring knowledge of scientific facts – the “scientific literacy” of a population – as the primary measure of where a country stands on science? Perhaps, you might say – but this approach has important limitations.
Firstly, it is difficult to come up with objective measures. Scientific literacy is culturally specific and likely to reflect a country's historic science base, so what might be considered basic knowledge in the US would not necessarily be the same in India3. There is also a big distinction between knowing scientific facts and understanding how science is achieved, through experiments, the scientific method and statistical significance.
Attitudes to big data
On the surface, it may seem like people are relaxed about where their personal data ends up, as long as certain conditions are met. While other Ipsos MORI research1 has shown that privacy around data is important to people, especially in the current climate, the PAS 2014 survey suggests that six in ten people (61%) do not mind how their personal information is used, “provided that it is anonymised and cannot be linked back to them”. There is also strong public support for big data being collected and used in contexts where there are tangible public service benefits – for example, nine in ten (88%) support developing a DNA database of cancer patients to help develop more effective treatments for cancer, while three-quarters (73%) support using data from electronic travel cards to improve the scheduling of buses and trains for passengers (Figure 1).
However, this surface reading of the data again hides a more complex story. In the qualitative workshops, there was no consensus view from participants on what they considered to be “personal” data. While researchers may consider any type of data that cannot be linked back to individuals as “anonymised”, some participants still felt that this was their data, since they were still a row on a database. It is possible that they were conflating anonymised data with aggregated data, not broken down to the individual level.
There are also indications that people can simply overlook the anonymised nature of big data sets, or treat anonymisation with scepticism. In the qualitative workshops, the linking of two government data sets and the anonymisation process around this was described very explicitly using a visual role-play, with cards representing different pieces of data, such as data on earnings held by the tax office, or data on benefit claims held by job centres. In spite of this, participants regularly forgot that the data was anonymised and needed reminding from the workshop facilitators. Some participants also did not trust the anonymisation process and worried that somehow those using the data might still find a way to reveal individuals’ identities.
Framing is also important. The survey finds that while people largely support big data being collected for public benefit, they are less keen on it being used commercially – seven in ten (70%) say they oppose companies offering discounted mobile phone calls and texts funded by personalised ads based on the content of people's text messages. In addition, relatively less tangible public benefits – the kind that can arise from much social science or “blue sky” research – can often be treated with scepticism, as was often the case in the workshops. Both of these may have been factors in the Care.data case, with media coverage often focusing on the potential commercial applications of the data, but less so on the potential benefits for medical and public health research.
Beyond this, many workshop participants also had several nuanced views on the use of big data. While they did not oppose it being collected in principle, they were concerned about what would actually come of the analysis. There were strong suspicions that governments do not put research findings into practice, so that efforts would be wasted. Some participants also thought there could be some negative consequences from big data mining, for instance if it meant that individuals lost out by being part of a “pigeonholed” group, or if it meant that the qualitative context of certain social issues were overlooked, leading researchers to draw incorrect conclusions from the data.
Moreover, it could be argued that better informing people about a science topic or new technology should lead to greater public support – but this is not always borne out by the data. PAS 2014 looked at the perceived risks and benefits of specific science topics, and found a broad relationship between feeling informed about a topic and being more favourable to it (Figure 2). However, taking GM crops as an example, among those who feel very well informed, perceptions of risks and benefits are more polarised – more think the benefits outweigh the risks (47%, versus 36% overall) and more also think the risks outweigh the benefits (40%, versus 28% on average). A similar polarisation of views among the more scientifically literate has also been found in other research with regard to climate change4. Simply informing people of the scientific facts around a topic will have limited impact.
The state of UK public attitudes
From a starting point that looks at public attitudes, rather than at knowledge, we find on the surface that the UK is in rude health. According to PAS 2014, four-fifths (81%) agree that “science will make people's lives easier”, and over half (55%) think that “the benefits of science outweigh any harmful effects” – only 16% disagree with the latter point of view.
Where it is possible to look at changes over time, it is evident that attitudes have come a long way in the last 25 years. As Figure 3 shows, substantially more people now agree that it is important to know about science in their daily lives (72%, versus 57% in 1988), and fewer think that “science makes our way of life change too fast” (34%, versus 49%). On several of these indicators, the balance of opinion has changed over time, with more now seeing science in a positive light than in a negative one.
People are overwhelmingly upbeat about the contribution science makes to the UK economy, in terms of growth, international competitiveness and future prosperity. For example, three-quarters (76%) think “scientific research makes a direct contribution to economic growth in the UK”, and nine in ten (91%) agree that “young people's interest in science is essential for our future prosperity”.
Many also value the contribution it has made to them personally. Half (51%) think the science they learnt at school has been useful in their everyday lives, while three-quarters (76%) think this of the maths they learnt at school.
Perhaps recognising these impacts, the UK public continue to support government funding of science. Eight in ten (79%) agree that, “even if it brings no immediate benefits, scientific research which advances knowledge should be funded by the government”, while two-thirds (65%) disagree with the statement that funding “should be cut because the money can be better spent elsewhere”.
Overall perceptions of those who work in scientific and academic fields are also strongly positive. Nine in ten (90%) think that “scientists make a valuable contribution to society”, and eight in ten (83%) agree that “scientists want to make life better for the average person”. Nine in ten (90%) trust scientists working for universities “to follow any rules and regulations”, while the same proportion trust “researchers” working for universities (90%) and “university lecturers” (89%).
Does this mean we are entering a new age of reason? Perhaps not. Digging deeper into the results quoted above shows some important complications.
Younger generations have been and continue to be less likely than their grandparents to think that the benefits of science outweigh any harmful effects
First of all, overall time series data hide a generational divide. As can be seen in Figure 4, younger generations have been and continue to be less likely than their grandparents to think that the benefits of science outweigh any harmful effects. And, looking at the 2014 results in isolation, younger people are somewhat less positive about the economic contribution or government funding of science – for example, while four in ten UK adults overall (38%) strongly agree that “the UK needs to develop its science and technology sector in order to enhance its international competitiveness”, only a quarter (27%) of 16–24-year-olds say the same. Altogether, these differences indicate that the upward trend in positivity towards science and scientists seen since 1988 is not guaranteed to continue indefinitely, as today's younger (and more sceptical) generation become a proportionally larger part of tomorrow's population.
There are also apparent contradictions in people's opinions when it comes to trust in scientists and researchers. While seven in ten (71%) think scientists are “honest”, half (50%) also think they are “secretive”. Indeed, looking just at the seven in ten people who say scientists are honest, over two-fifths (44%) also believe scientists are secretive. Moreover, while people generally trust scientists and researchers to follow rules and regulations, the survey findings and qualitative research both suggest that many people still have concerns about hidden data, or about the intentions of scientists and researchers – with these concerns influencing what information they trust and what they reject. The following quote from one of the qualitative workshop participants shows that the “highly regulated industry” argument will only go so far in allaying people's concerns: “There are standards and regulations. But a moral education, the moral attitudes of the researcher – these are more important.”
While people generally trust scientists and researchers to follow rules and regulations, many still have concerns about hidden data
Contradictions and concerns may well stem from a lack of awareness of how scientists and researchers go about their work. Although two-thirds (68%) feel they know what scientists do, one-third (35%) nonetheless think that scientists adjust their findings to get the answers they want. Only one-third (34%) disagree with that statement, with almost as many (31%) undecided or neutral about it. The concept of peer review does not seem to be widely understood, or is often treated with scepticism, with three in ten (29%) believing that scientific research is never or only occasionally checked by other scientists before being published.
The UK public also lack awareness of how research is funded, particularly outside of government funding. When asked unprompted who funds scientific research in the UK, just over a third (36%) mention private companies and just under two in ten (17%) say they do not know, compared to seven in ten (70%) who mention the government.
Alongside this lack of understanding, there are indications that overall positive attitudes to science and scientists are not underpinned by any deep thought process. Even when people report that they do not feel informed about scientific research and developments, 70% still agree that “scientific research makes a direct contribution to economic growth in the UK”, versus 76% of all adults. A similar proportion (67%) of those who do not feel informed about economics and the way the economy works also agree with the statement. These attitudes may just be received wisdoms rather than carefully considered points of view.
This further examination of the data does not take away from the overall positive findings. However, it suggests that, on their own, those findings cannot be used to say that the public is already sufficiently engaged with the science and statistics that they encounter every day.
Lessons for science communication
In such a complex environment, what is the way forward for those whose job it is to communicate science or social science research to the general public? While attempts to “educate” the public on the scientific facts around an issue may flounder, that does not mean that scientists and researchers should stop talking to the public about their work. In fact, people want to engage directly with scientists and researchers more often. The PAS 2014 survey finds seven in ten agreeing that “scientists should listen more to what ordinary people think” (69%) and that they should “spend more time discussing the social and ethical implications of their research with the public” (68%). Half (53%) think scientists “should be rewarded for communicating their research to the public”.
Attempts to “educate” the public on scientific facts may founder, but that does not mean that scientists should stop talking about their work
The PAS studies suggest that the key to successful science communication is to understand that there are multiple “publics”. While people want to engage, not everyone will do so in the same way, or through the same channels. It is, therefore, worth policy-makers and researchers reflecting on whom they are aiming to engage and how. While some people will relish the technical details, others are more interested in the social and ethical implications of the work.
While UK public attitudes to science are overwhelmingly positive, this does not reduce the need for dialogue with the public. Regardless of whether the issues at hand are GM crops or big data, people want to hear more from the scientists and researchers working in these areas. While the public may already support various scientific and technological changes in principle, there are many concerns beneath this, and these are partly driven by confusion over how scientists and researchers go about their work, and what their intentions are. Scientists and researchers need to communicate these points, and need to tailor their communication for different audiences – very few people will engage with the hard scientific facts alone.
Statistics making sense
Scott Keir argues the case for statistical literacy in public engagement with science
In Chicago, on the day after the coldest night of 1950, Samuel Wilks delivered his Presidential Address to the American Statistical Association. He outlined future challenges for statisticians and society: the introduction of operations research in business, and the use of statistical analyses in research.
In government, Wilks noted that “statistical process is becoming more and more widely used as a scientific method to be utilized as a basis for developing policy”. Statistics in government is not just enumeration for historical study, said Wilks; it is a tool for decisions, judgements and controls1.
Fifty years later at the turn of the century, evidence-based policy-making was well established. Governments routinely consulted expert individuals and expert bodies on policy matters. But all was not well, at least in the UK. BSE in cattle, linked to variant CJD in humans, had severely damaged public trust in policy-making. Other science-based policies were considered to be “at risk” from a lack of public support.
A “crisis of trust” was declared in 2000 by a House of Lords Committee, led by Lord Jenkin. Their solution was clear: direct dialogue with the public should become a normal and integral part of policy-making. This should not just be advocacy or education – support does not necessarily follow from knowledge.
Today, the UK has Sciencewise, a government-funded expert centre to assist policy-makers in public dialogue involving science and technology issues; the UK's survey of the public and science is about attitudes, not facts; and open policy-making has been introduced. All of these are signs of progress, though a House of Commons Public Administration Select Committee report in 2013 noted that much work was still to be done.
This progress moves us from scientific literacy (a need for knowledge, improved by education), to public understanding of science (a need for appreciation, improved by promotion and education), to science and society (a need for trust, improved by public dialogue)2.
Given this, what is the place of statistical literacy in public attitudes to science today?
It is important to note that improved statistical literacy is not necessarily linked to more favourable attitudes toward scientific issues, as illustrated by a study of US opinions toward climate change and the public's capacity to comprehend and make use of quantitative information3.
However, a key aspect of statistical literacy is the ability to think critically about evidence and to understand concepts of risk and probability. This is relevant both to formal public dialogue, where support may be provided, and to wider public debates, where statistics are used to provide authority, to promote a point of view and challenge opponents2.
Today, the volume and accessibility of data relevant to science and policy-making is ever increasing – be it Chicago's historical weather records or the inner workings of government. Information is now the property of the citizen, not the state, as John Pullinger put it while President of the Royal Statistical Society4.
Back in 1950, Samuel Wilks paraphrased H. G. Wells – speaking almost 50 years before him – when he said: “Statistical thinking will one day be as necessary for efficient citizenship as the ability to read and write.” That day is today. We must ensure that all citizens can take advantage of this new wealth of data and evidence, to help us all make better decisions about our society.
Scott Keir is the Royal Statistical Society's head of education and statistical literacy.