Perceptions are not reality: what the world gets wrong

The latest Ipsos Perils of Perception survey highlights how wrong the public across 40 countries and districts are about key global issues and features of the population in their country.

The key patterns are:

• Most countries think their population is much more Muslim than it actually is – and that the Muslim population is increasing at an incredible rate
• All countries think their population is less happy than they actually say they are
• Most countries are more tolerant on homosexuality, abortion and pre-marital sex than they think they are
• And nearly all countries think wealth is more evenly distributed than it actually is.

In Great Britain we get some things very wrong – but some things right…

1. Current Muslim population: we hugely overestimate the proportion of Muslims in the British population – we think that 1 in 6 Britons are Muslim, when actually fewer than 1 in 20 are. The findings show we think 15% of the population are Muslim, when actually only 4.8% are, less than a third of our guess.
2. Future Muslim population: we also think that the Muslim population is growing to a much greater extent than it actually is. We think that 22% of our population will be Muslim by 2020, when projections from the Pew Research Center suggest Muslims will only make up around 6% of the British population by then.
3. Happiness: we think our fellow Brits are much more miserable than surveys of happiness show. We guess that only 47% of Britons would say they are very or rather happy, when actually 92% say they are in a recent representative survey.
4. Homosexuality: we also think people are rather less accepting of homosexuality than they actually report in surveys. We think that 28% would find homosexuality morally unacceptable, when 17% actually say that.
5. Sex before marriage: we’re a bit closer to actual views when we guess at Briton’s opinion on sex before marriage – we think that 20% of the population find it morally unacceptable when in fact only 13% do.
6. Abortion: we also think more people are anti-abortion than they actually are – we think that 34% find abortion morally unacceptable, when only 25% say they do.
7. Wealth of the bottom 70%: we asked people to guess what proportion of total wealth the least wealthy 70% in Britain owned – and we’re one of the very few countries who have an accurate view of how unequal we are. This majority of the population actually own only 21% of wealth in Britain – and we guess that they own 19%.
8. Home-owners: but we are way out on the extent of home ownership. We think under half (47%) of household in Britain own their own home, when in fact 64% do.
9. Health spending: we think we spend twice as much as we actually do on our health. We think we spend 19% of our total GDP on health expenditure (including private healthcare as well as the NHS), when in fact we only spend 9%. If our guess was right, this would be the equivalent of an extra £225bn spending on health.
10. Current population: like many other countries, we’re pretty accurate on how large our population is in Britain – our average guess is 64 million and the actual population is around 63 million.
11. Future population: but we’re also unusually good at guessing our future population. We asked people what the population of their countries would be by 2050, to compare with UN projections – and Britain is one of a very few countries that guessed almost exactly right, at 75 million.
12. Trump: Fieldwork for the survey was conducted the month before the presidential election in the US – so we asked who people thought would win. And like the vast majority of countries – along with nearly all pundits – we were wrong: 61% thought Clinton would win and only 16% saw the Trump victory coming.

But Britain is far from the worst in identifying realities and predicting the future – in fact we are the third most accurate country in our “Index of Ignorance”. Looking across the 40 countries included, many are much more wrong…

1. Current Muslim population: many countries over-estimate their Muslim population by a staggering amount. For example, in France the average guess is that 31% of the population is Muslim when it is actually 7.5%. Other Western European countries such as Germany, Italy and Belgium are also massively out. The US and Canada are similarly inaccurate with both having average guesses of 17% against actual figures of 1% and 3.2% respectively. At the other end of the spectrum, very high Muslim population countries underestimate their Muslim population: for example, in Turkey the actual Muslim population is 98%, but the average estimate is 81%.
2. Future Muslim population: similar countries hugely overestimate the levels of growth in Muslim populations over the next four years. The average guess in France is that 40% of the population will be Muslim in 2020 when the actual projection is 8.3% (an increase of just 0.8% from the current level of 7.5%). Italy, Belgium and Germany also greatly overestimate the growth in their Muslim population. The US are also miles out, with an average guess of 23% compared with the actual projection of 1.1%.
3. Happiness: Globally, we’re very pessimistic about levels of happiness. Across the 40 countries in the study the average guess was that 44% of people say they are happy, when the actual figure is nearly double that at 86%. The average guesses in Hong Kong (28%) and South Korea (24%) are incredibly low, when actual happiness levels in recent surveys are 89% and 90% respectively.
4. Homosexuality: people tend to underestimate the acceptance of homosexuality, particularly in European countries. In the Netherlands the average guess is that 36% of the population find homosexuality morally unacceptable, when actually only 5% do. There is a similar finding in the Czech Republic, Germany and Spain. In Indonesia, a country which reports almost universal opposition to homosexuality (93%), the average guess of 79% falls 14% short of the actual figure. The study highlights the wide range of views on homosexuality between countries both in terms of actual views and what people think their country thinks. For example, in Russia, people guess that 79% find homosexuality unacceptable compared to the actual figure of 72%. Conversely in Norway, on average people think that only 22% of their populations would find homosexuality unacceptable (the actual figure is much lower at 5%).
5. Sex before marriage: people in the Netherlands are also most likely to underestimate how tolerant their country is on sex before marriage. On average people in the Netherlands said that they think 34% of the population would find unmarried sex morally unacceptable compared with the actual figure of just 5%. In Turkey, a country where opposition to unmarried sex is very high (91%), the average guess was considerably lower at 72%.
6. Abortion: when asked what percentage of their country’s population believe having an abortion is morally unacceptable people tend to overestimate the actual figure. For example, in the Netherlands people think 37% of the population view abortion as morally unacceptable, when it’s actually only 8%. The US is unusual in having a very accurate view of public opinion: people guess that 48% are against abortion, when surveys say 49% are. This accuracy is perhaps due to the high profile discussion on abortion in the US.
7. Wealth of the bottom 70%: people tend to overestimate what percentage of a country’s wealth is owned by the poorest 70%. On average, just 15% of total wealth is owned by the bottom 70% across these countries - but the average guess is almost twice that at 29%. Some countries are incredibly inaccurate: Indians think this group owns 39% of the country’s wealth when actually only 10% do. The US is also significantly out: Americans think the bottom 70% own 28% of the country’s wealth, when it’s actually a quarter of that at 7%.
8. Home-owners: Globally, we tend to massively underestimate how many households own or are buying their own home. We think just 49% of people own or are buying their home, when in fact 68% are owner-occupiers across these countries. India is the most wrong, thinking that only 44% own/are buying their home, when in fact 87% are. This is likely to reflect the online nature of the sample in the study: there are significant and growing pressures on home ownership in the middle class urban centres in India, but ownership is very high outside of these.
9. Health spending: Generally, people think our national spending on health is much higher than it actually is. Across all the countries in the study, we think that 21% of our GDP goes on health spending, when actually only 8% does. The least accurate countries are Indonesia and Malaysia who each think that 30%+ is spent on health, when the actuals are low single figures, for example only 3% in Indonesia.
10. Current population: one of the few areas of the study where people were very accurate was guessing at their population size. In most cases people were within 5% of the actual figure and German respondents on average, exactly identified their actual population of 82 million. Singapore and Hong Kong are both out by 10% with Hong Kong guessing 10% over and Singapore 10% under than the actual figures.
11. Future population: While respondents tended to be very accurate guessing current population figures, it was a different story for population projections for 2050. For example, Taiwan expects to have a significantly higher population in 2050 than they do now – but the UN expects the country’s population to shrink: they therefore expect to be 32% bigger than the UN expects. The US on the other hand is one of the countries expected to grow more than Americans expect: the UN estimate is that there will be 389m in the US by 2050, but the average guess is they’ll only have 351m residents, 38m fewer people, a 10% lower population. Britain and the Philippines are the only two countries to get projections spot on: for example, Britain expects its population to grow to 75m from its current base of around 63m, and the UN projection says the same.
12. Trump: Only three countries had more people saying that Donald Trump would win rather than Hillary Clinton: Russia, Serbia and China. And Russia stands out as particularly sure of a Trump victory: 50% said Trump and only 29% picked Clinton. The shock that greeted the result around the world is no surprise when you see how sure many countries were that Clinton would win: 80%+ thought Clinton would win in Mexico, Colombia, Chile, South Korea and Norway. The US itself was among the least sure – but even here people were twice as likely to pick Clinton (50%) than Trump (26%).

Looking across the five questions on factual realities, there are clear patterns in which countries have a more accurate view of their countries. To capture this, we’ve calculated the Ipsos “Index of Ignorance”, as shown in the table below.

India receive the dubious honour of being the most inaccurate in their perceptions on these issues with China and the US also high up the list. The Netherlands are the most accurate, followed by Great Britain, with South Korea in third.

Bobby Duffy, Managing Director of Ipsos Social Research Institute, London, said:

“Across all 40 countries in the study, each population gets a lot wrong. We are often most incorrect on factors that are widely discussed in the media, such as the proportion of our population that are Muslims and wealth inequality. We know from previous studies that this is partly because we over-estimate what we worry about. But in this new study we also show that we’re often unduly pessimistic about how happy people are and our tolerance on controversial issues such as homosexuality, sex before marriage and abortion. In many countries, particularly in the West, we have a picture of our population that is unduly miserable and intolerant. This is important: we know what people think of as the norm is important in affecting their own views and behaviours. We also get facts wrong that will make us focus on some issues more than they perhaps deserve: for example, we tend to think our populations are much less likely to own their own home than they actually are. In many countries we have received the message loud and clear that pressure on housing and affordability are serious issues, but we’ve underestimated how many still own their home. There are multiple reasons for these errors – from our struggle with simple maths and proportions, to media coverage of issues, to social psychology explanations of our mental shortcuts or biases. It is also clear from our “Index of Ignorance” that the countries who tend to do worst have relatively low internet penetrations: given this is an online survey, this will reflect the fact that this more middle-class and connected population think the rest of their countries are more like them than they really are.”

Technical note

• These are the findings of the Ipsos Perils of Perception Survey 2016. 27,250 interviews were conducted between 22nd September – 6th November 2016.
• The survey is conducted in 40 countries and districts around the world, via the Ipsos Online Panel system in Argentina, Australia, Belgium, Brazil, Canada, Chile, China, Colombia, Denmark, France, Germany, Great Britain, Hong Kong, Hungary, India, Indonesia, Israel, Italy, Japan, Malaysia, Mexico, Peru, Philippines, Poland, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Taiwan, Thailand, Turkey, USA, Vietnam. The following countries used either online or face-to-face methodologies: Czech Republic, Montenegro, Netherlands, Norway, Serbia.
• Approximately 1000 individuals aged 16-64 or 18-64 were surveyed in Australia, Brazil, Canada, China, France, Germany, Great Britain, Italy, Japan, Spain, USA, and approximately 800 individuals aged 18-64 were surveyed Czech Republic, Montenegro, Netherlands, Norway, Serbia. Approximately 500 individuals aged 16-64 or 18-64 were surveyed in the remaining countries and districts.
• The “actual” data for each question is taken from a variety of verified sources including The World Values Survey and Pew Research Center. Download a full list of sources/links to the actual data (PDF).
• Attitudinal data (where existing sources were not available) was collected in separate surveys conducted between September and November 2016. Data was collected using a combination of online, telephone or face-to-face methodologies.
• Where results do not sum to 100, this may be due to computer rounding, multiple responses or the exclusion of don't knows or not stated responses.
• Data are weighted to match the profile of the population.
• Download the charts (PDF)
• Download the calculation note (PDF)
• Download the questionnaire (PDF)