Perils of Perception 2018

Global study highlights how our misperceptions about crime and violence, sex, climate change, the economy and other key issues.

Ipsos’ latest Perils of Perception study shows which key facts the online public across 37 countries get right about their society – and which they get wrong. Now in its fifth year, the survey aims to highlight how we’re wired to think in certain ways and how our environment influences our (mis)perceptions.



  • In several countries around the world, people are wrong about the scale of knife and gun crime in their country. Although in 13 countries the majority correctly guess which is the biggest killer out of firearms, sharp objects such as knives or other physical violence, in other countries people’s perceptions don’t match what the crime statistics say. 
  • For example, in Great Britain, 71% of people think knives cause the most deaths, when they actually account for just 25% of all deaths by interpersonal violence. The GB guess for knives is the highest across all 37 countries surveyed, but there is a similar pattern of overestimating their impact in other countries, especially other European countries such as Denmark, Germany, Italy, Poland and Hungary, and in Asian countries such as Japan, South Korea, Singapore and China.
  • Other countries overestimate the proportion of gun deaths, notably South Africa (where knives cause the most deaths), Netherlands and Sweden. And even some countries where the scale of certain types of violence dwarf others, large minorities do not realise this. For example, in the US – where firearms account for almost 70% of all deaths through interpersonal violence, only six in ten (59%) correctly identify guns as the biggest killer, and a similar pattern is seen in Columbia.
  • People in most countries think prisons are even more crowded than they actually are. On average people think prisons are 30% over full capacity (130%) when they are 9% over capacity (109%).  Having said that, the countries with the highest levels of overcrowding do tend to be the countries with the highest guesses.

Sexual harassment

  • In 13 countries for which there is data, all of them substantially underestimate women’s experience of sexual harassment in their nation.  On average, people think 39% of women have experienced harassment, but actually on average 60% have.  Denmark, the Netherlands, France and the United States are the most likely to underestimate the extent of sexual harassment in their country.
  • In each country, men give a lower estimate of sexual harassment than women.  On average, men guess 36% of women in their country have experienced sexual harassment, but women’s guess is 44% (although still an underestimate).

Climate change

  • 17 of the past 18 years have been the hottest since records began. However, every country in the study underestimates the global temperature rise over the past 18 years. The average estimate across the study was 9 years.
  • The majority of countries overestimate the amount of energy used that comes from renewable sources in their country. The average guess is 26% when it’s actually only 19%. Malaysia, Saudi Arabia, China and Singapore were the furthest out; some countries, though, actually underestimate how much progress they have made with renewables, such as Sweden and Montenegro.


  • Nineteen countries in the study include estimates for the amount of sex they think 18-29 year olds are having, and most of them get it very wrong. On average people think young women have sex 20 times every 4 weeks and the guess for men is slightly higher at 22 times. In fact, when asked in the survey, 18-29-year-old women and men report they have had sex 5 and 6 times every 4 weeks respectively.
  • People in Mexico, India, Brazil, Italy and Colombia are particularly far out in their estimates of how much sex young people are having. Sweden, Britain and Turkey have lower, slightly more realistic guesses. 
  • Across the study, there is little consistent difference in guesses by women and men.  And young people themselves are only slightly less likely to overestimate how much sex their peers are having than older people (on average, under 30s guess that other 18-29 year olds are having sex 20 times a month, among over 30s the average guess is 21).


  • All countries in the survey underestimate the near universal level of infant vaccinations in their country. The average guess is 73% when the actual figure is 94% according to World Health Organisation (WHO) figures. Furthest out are India, Mexico and China. 


  • Every country in the study heavily overestimates the proportion of people unemployed and seeking work in their country. The average guess across the study was 5 times greater than the actual (34% when in reality it is closer to 7%).
  • People tend to underestimate the size of their country’s economy relative to others. The majority of people placed their country’s GDP rank lower than the reality. This is particularly the case for emerging economies such as Argentina, South Africa and Romania.

Population facts

  • Every country massively overestimates the levels of growth of their elderly population. Across the countries on average, people think 54% of the population will be 65+ in 2050 when in reality the projection is less than half that (25%).
  • The majority of countries hugely overestimate levels of immigration; a pattern we have seen in previous studies. The average guess across 37 countries is that 28% are immigrants when the actual figure is less than half that (12%).
  • Nearly every country included in the study also over-estimates their Muslim population by a large margin. The average guess was more than double the actual figure (20% vs 8%).

How accurate is the Hong Kong public…?

  1. Hong Kong people slightly overestimate the scale of overcrowding in prisons in Hong Kong. The average guess is that prisons are at 87% of capacity when the actual figure is 77%.
  2. Hong Kong people are too optimistic about the levels of energy consumed that comes from renewable sources. The average guess in Hong Kong is 16% when it is actually just 1%.
  3. Similarly, Hong Kong people underestimate the level of temperature change over the past two decades. In Hong Kong, people think 12 of the past 18 years were the hottest on record globally when in reality it is 17.
  4. They also guess that 84% have been vaccinated against Diphtheria, Tetanus, Whooping Cough, Polio and Measles when levels are much higher at 99%.
  5. Levels of unemployment are hugely overestimated in Hong Kong. People think 14% of working age Hong Kongers are unemployed and looking for work when the actual figure is less than a quarter of that (3%).
  6. When it comes to gauging our economic place in the world, Hong Kong people very accurate: the average guess is that Hong Kong has the 30th largest economy by GDP when in reality it is 31st.
  7. Hong Kong people estimate that half their population (54%) will be over 65 years old in 2050 which is much more than the World Bank projections (34%).
  8. People rather accurately estimate the proportion of immigrants in Hong Kong with an average guess of 37% when the actual figure is 39%.
  9. Hong Kong people overestimate the proportion of Muslims in Hong Kong. The average guess is more than double the actual figure (10% vs the 4% reality).

Looking across seven key questions where we get people to estimate factual realities, there are clear patterns in which markets have a more accurate view of their markets. To capture this, we’ve calculated the Ipsos “Misperceptions Index”, as shown in the table below.

Hong Kong people are the most accurate of all markets surveyed – followed by New Zealand, with the ever-accurate Swedes in third, in our “Misperceptions Index”.

This year Thailand receive the unwelcome prize of ‘least accurate’ in their perceptions, closely followed by Mexico and Turkey.


Misperceptions Index


Thailand   1 Least accurate
Mexico  2  
Turkey 3  
Malaysia 4  
Brazil 5  
Colombia    6  
Russia 7  
Saudi Arabia 8  
China 9  
Peru 10  
Belgium 11  
India 12  
Romania 13  
Germany 14  
South Africa 15  
Japan 16  
Serbia 17  
Italy 18  
Argentina 19  
France 20  
Netherlands 21  
Switzerland 22  
United States 23  
South Korea 24  
Chile 25  
Montenegro 26  
Canada 27  
Spain 28  
Australia 29  
Poland 30  
Singapore 31  
Denmark 32  
Great Britain 33  
Hungary 34  
Sweden 35  
New Zealand 36  
Hong Kong 37 Most accurate



Mick Gordon, Managing Director at Ipsos Hong Kong, said:

Ipsos’ global Perils of Perceptions study shows that, around the world, many people overestimate the real extent of some social issues, while underestimating the importance of others.  Our latest research reveals that in many markets certain signal crimes really grab the public’s attention more than official statistics suggest they should.  But on the other hand, on a major issue such as sexual harassment towards women, many people do not realise quite how prevalent it really is.  Reflecting concerns since the financial crash, people overestimate the level of unemployment in their market, and are overly pessimistic about the relative size of their market’s economy, while perhaps are too complacent when it comes to climate change and the use of renewable energy sources.

While the Hong Kong people seem have a better understanding of their city than people in other markets, they are still wrong about several aspects of daily life in Hong Kong, such as importance of the local Muslim population, the city’s reliance on non-renewable energy sources, the size of our older populations or the (very low) level of unemployment in the SAR.

There are many different reasons why we may be wrong about various facts of society. These can include external influences on us, such as what we hear in the media or online, but our own internal biases are just as important.  These biases include our tendencies to focus more on negative stories over positive ones, to believe that things were always better in the past, to put too much emphasis on our own individual experience, and simply not being very good with numbers.  But what is crucial to understand is that we overestimate what we worry about as much as we worry about what we overestimate - in other words, misperceptions can be a very useful pointer to people's real concerns.  It also means that trying to correct misperceptions by only repeating the facts is unlikely to work - instead we need to engage with the more emotional reasons that might be driving why people are worried about a topic.

For further information, please contact Keres Lee at +852 2830 2549 or [email protected].


Technical note:

These are the findings of the Ipsos Perils of Perception Survey 2018. 28,115 interviews were conducted between 28th September – 16th October 2018.

The survey is conducted in 37 countries 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 (SAR, China), Hungary, India, Italy, Japan, Malaysia, Mexico, Netherlands, New Zealand, Peru, Poland, Romania, Russia, Saudi Arabia, Singapore, South Africa, South Korea, Spain, Sweden, Switzerland, Thailand, Turkey and the USA. The following countries used either online or face-to-face methodologies: Montenegro, Serbia.

Approximately 1000 individuals aged 16-64 or 18-64 were surveyed in Australia, Brazil, Canada, China, France, Germany, Great Britain, Hong Kong, India, Italy, Mexico, Montenegro, Serbia, Singapore, Spain and the USA. Approximately 2000 individuals aged 16-64 were surveyed in Japan. Approximately 500 individuals aged 16-64 were surveyed in Argentina, Belgium, Chile, Colombia, Denmark, Hungary, Malaysia, Netherlands, New Zealand, Peru, Poland, Romania, Russia, Saudi Arabia, South Africa, South Korea, Sweden, Switzerland, Thailand and Turkey.

The “actual” data for each question is taken from a variety of verified sources. A full list of sources/links to the actual data can be found here

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.