IPSOS VIEWS #2

March 2016

IPSOS
VIEWS
Perils of Perception
By Bobby Duffy
Global Director, Ipsos Social Research Institute
IPSOS
Perils of Perception

VIEWS

Ipsos has conducted a number of major studies in

What do
recent years exploring how accurate people are in
estimating a range of key social realities. And our
latest international survey 1 across 33 countries

we get
shows just how wrong we often are.

In Britain the public thinks that the top 1% wealthiest

wrong?
households own 59% of the country’s wealth, when
they actually “only” own 23%. Americans think that
33% of their population are immigrants, when in fact
it’s less than half that, at 14%.

Brazilians think the average age in their country is 56,

when it’s only 31. Russians think that 31% of their
politicians are women, when in fact it’s only 14%.

The British think an extraordinary 43% of young adults

aged 25-34 still live at home with their parents, when
it’s actually only 14%. In India, the online population
think that 60% of the whole country also has internet
access, when in fact only 19% do.

Israelis think that only 39% of working age women in

their country are in employment, when actually 68%
are. Saudis think that only 28% of their population is
overweight or obese, when in fact it’s a very worrying
71%.

An example of how wrong we can be is shown in the

chart (on the next page) on our (mis)perceptions of
obesity.

1
IPSOS
Perils of Perception

VIEWS

2
IPSOS
Perils of Perception

VIEWS

Looking across all the questions in the study, we created an “Index of Ignorance”, to identify which countries had the
best and worst understanding of these facts. As the chart shows, Mexico received the dubious privilege of being the
most inaccurate, while South Koreans are the most accurate.

3
IPSOS
Perils of Perception

VIEWS

Why are we so
often so wrong?
But the purpose of the study was not just to raise a wry smile at other peoples’ - or whole nations’ - expense. Even
the term “ignorance” was chosen carefully, not to imply lack of intelligence, just the absence of knowledge or information –
and, as we will see, some argue this ignorance may be rational. Instead the main aim was to raise questions on why
these errors arise, and what, if anything, we can and should be doing about them.

So why are people across the world so often so clueless about these realities?

A cros s the ye ars, countr ie s a nd topic s we have re searched, five key groups of influences stand out. These
will be in play to varying degrees for different individuals, nations and issues. They will sometimes be in tension with each
other, and sometimes reinforce or interact. But each is important in understanding why we’re so wrong.

The Ignorance Equation - Why are we so wrong?

4
IPSOS
Perils of Perception

VIEWS

1. Mathematical and 2. Biases and heuristics,

statistical ability including availability,
It’s partly that many of us just struggle with numbers – satisficing and inductive
we’re often very shaky on key mathematical concepts
that will be important for estimating realities. For generalisation
example, in a study for the Royal Statistical Society
in the UK, the large majority of the population can People will also be subject to all sorts of biases and
correctly identify that 50 is 25% of 200 – but even heuristics (or mental shortcuts) when answering the
with this very simple calculation, 8% get it wrong. questions. In particular, we tend to grab for easily
The large majority know that the average of 5, 10 available information even if it doesn’t quite fit the
and 15 is 10 – but 30% get this simple sum wrong.
2 question. For example, our huge overestimates of
the rural populations in most countries will be affected
And we really struggle on probabilities, which can by how much of the physical landmass rural areas
be key in estimating the likely incidence of an event make up, rather than a careful calculation of how
or characteristic: only a quarter of the UK population unoccupied it generally is. In Daniel Kahneman’s
correctly identify that the chances of getting two terms, answers to these sorts of questions are classic
3
heads when tossing a coin twice is 25%. examples of “fast” thinking, rather than “slow”.

Other studies also show we are particularly poor in In the survey research methods literature there is also
dealing with very big or very small numbers (which the helpful concept of “satisficing”, which reminds us
will impact our ability to think about whole populations, that people will put varying degrees of effort into
or relatively rare events like teenage pregnancy), and thinking about the questions, depending on their
4
we find it hard to distinguish between rates and levels own characteristics and how much effort it takes.
(on issues like immigration). For example, one of our questions asked people
to estimate how many teenage girls get pregnant each
year in their country – which resulted in wild

“We’re often very shaky on key overestimations (for example, the average guess in
the US was 24%, when in fact only 3% do). This will
be partly because people genuinely think it is much
more common than it is or are worried about it – but
mathematical concepts that will be it will also be partly because it’s easier to think about
how many teenagers give birth in total rather than an
important for estimating realities. “ annual rate. The total figure for all teenagers is much
higher than the annual rate, and closer to what the public
estimate.

5
IPSOS
Perils of Perception

VIEWS
This has parallels with Tetlock’s distinction between hedgehogs and foxes5 in his study of the accuracy of predictions.
Tetlock’s research suggests that those who use lots of different information and are willing to be influenced by new
information are better predictors, compared with those who have one big idea and use it with gusto. There will be
similar parallels on estimating realities: those who stick to quick stereotypes are more likely to be wrong.

A further, related concept that will be important in explaining our error is inductive generalisation. Regardless of the
information we use, we are often tied to our own perspective, and many of us struggle to imagine the variety within
our own countries. Inductive generalisation – which in this context is broadly that we think the whole population reflect
our own narrow experience – will be particularly important in explaining why some countries score much worse than
others in our Index of Ignorance.

This is highlighted by our Indian sample massively overestimating their population’s access to the internet. Our study
was mostly carried out through an online survey – and in developing countries this will be representative of a more
affluent, connected group rather than the population as a whole. In some ways, we may have expected this more
educated sample to get closer to reality – as our first point above suggests, those with higher education levels tend
to be more accurate on these type of questions.

But what we find throughout the study is that people often grossly generalise from their own situation, forgetting how unrepresentative
they are. In fact, as the chart below shows, you can explain a lot of the variation in our ignorance scores across countries just by
knowing internet penetration, and by extension, how unlike the whole population these groups are.

6
IPSOS
Perils of Perception

VIEWS

3. Emotional innumeracy 4. The media and the

We also suffer from what social psychologists call
6
power of anecdote
“emotional innumeracy” when estimating realities. This
theory suggests we have two goals when answering Of course, the media are bound to have a role in
these questions: “accuracy” goals, where we want to forming these misperceptions. It’s notoriously difficult
get the right answer, but also “directional” goals, where to prove direct causal effects from media coverage,
we’re sending a message about what’s worrying us, but the media are undoubtedly important sources of
consciously or not. information and impressions that influence us.

This provides the very neat and vitally important reminder But we need to be careful here. For example,
that cause and effect run both ways - our concern may whenever we release results from these studies,
lead to our misperceptions as much as our misperceptions one of the first responses is always “that will be
creating our concern. a media effect”, often picking one or two more
populist or tabloid media sources (like the Daily
This is likely to be part of the explanation for the Mail in the UK).
widespread and huge overestimates of how much the
wealthiest own and what proportion of our populations But the fact that this happens everywhere suggests
are immigrants. We are worried about the concentration we can’t lay the blame entirely at one particular title
of wealth and immigration levels, and this is reflected in or even type of newspaper that not all of us will be
us overstating the scale of the issues. exposed to.

But the survey suggests there are also some issues The media undoubtedly contribute – but the real
where we’re not as worried as we should be. For example, driver is how we remember information. We are
most countries hugely underestimate how much of their far more likely to remember negative stories and
population is overweight or obese. vivid anecdotes stick, regardless of whether7 they
are describing something vanishingly rare. This
Our misperceptions can therefore be seen as an presents a far more difficult challenge – improving
important indicator of our levels of concern – and where our misperceptions is not just about correcting how
we underestimate, maybe we should be worrying more the media use “facts”, it’s about how we as humans
(although, as we’ll come back to later, scaring people remember stories.
about the scale of an issue could be counter-productive
in changing behaviours).

7
IPSOS
Perils of Perception

VIEWS

5. Rational ignorance
Finally, some suggest that our ignorance is not really
“Improving our misperceptions is not
about our failings but largely a response to the political
context around us. Professor Ilya Somin, from George
Mason University in the US, for example, argues that
just about correcting how the media
it is a consequence and vital flaw in our whole political
8
use “facts”, it’s about how we as “
system .
humans remember stories.
His explanation is at the opposite end of the spectrum
to the social psychologists, where our lack of knowledge
is not due to “thinking fast” but rather it is entirely rational.

In this view, people have no reason to inform themselves,

with all the costs of time and effort that involves, if they
can’t influence anything through a political system
where their individual vote counts for virtually nothing.

What’s the point in finding out how the government

spends our money, whether crime is increasing or
decreasing or how many immigrants are coming to the
country if our vote doesn’t affect political outcomes and
decisions remain outside our control?

In this reading most modern systems of government

are inevitably flawed and people would be more likely
to get what they want if we cut central political control,
pushing decisions down to local areas, the private sector
and (ultimately) individuals, where choices are more
personal and therefore better informed.

This may seem extreme – but it does make one vital

point. Rather than concluding that our results demonstrate
that people are too dumb to be trusted with decisions, if we
want a better informed population, we need to give people
more, not less, power and control.

8
IPSOS
Perils of Perception

VIEWS

Does it
Given the scale of our errors and variety of explanations,
do our misperceptions really matter? And even if
they do, what should or can we do?

matter
Our view is that they do matter – and there are
important lessons for what we should do. We know
better than most that misperceptions have likely

that we’re
always been with us, in different contexts and countries,
and are incredibly difficult to shift.

so wrong
But that doesn’t mean we can’t learn vitally important
lessons for how government, media and businesses
should respond.

– and what “
do we do?
We know better than most that
misperceptions have likely always been
with us, in different contexts and
countries, and are incredibly difficult to shift. “

9
IPSOS
Perils of Perception

VIEWS

1. Our understanding of the 2. ...and we need to

norm is important to how we avoid normalising negative
behave: we need to normalise behaviours and attitudes
positive behaviours... But we face a different challenge on negative
behaviours that are less prevalent than we think.
We know that our mental image of normal behaviour Take the example of physical activity and exercise
influences how we ourselves behave – indeed, behavioural – we’re less active than we should be, but the public
science studies often find our understanding of social across countries massively underestimates how
norms is the most important influence on certain many people actually do the recommended amount
behaviours. Consistently underestimating voter turnout of physical exercise each week.
(as we’ve measured in previous studies) is a problem
then, as people have the wrong idea about the norm Lobbying and editorial campaigns have focused on
and are therefore less likely to vote themselves. 10
how we’re facing an “inactivity epidemic”. And this

“Behavioural science studies often

is a fair characterisation of the problem - if it was an
actual disease, physical inactivity would certainly be
considered an epidemic, as it costs more years of
11
life than alcohol consumption or excessive cholesterol.

find our understanding of social But the key question is does promoting that message
help or hinder behaviour change? American
norms is the most important influence “ 12
psychologist Robert Cialdini has long warned
about the dangers of normalising a negative behaviour,
on certain behaviours. showing the unintended consequences.

Similarly, underestimating the number of people who While we’re trying to send a message that many
do the recommended amount of exercise each day (as people are doing this undesirable thing, there
again we’ve found in a further study) is also a problem, clearly lurks the message that many people are
as we’re less likely to take part ourselves when we think doing that behaviour.13
others don’t.
Policy-makers and others trying to influence behaviour
This power of social norms has been harnessed as a therefore have a difficult line to tread between drawing
positive force in a number of ways by government and attention to widespread issues and normalising
others in recent years. For example, various experiments them. But that line will be much better trodden with
9
on tax payments are unequivocal: simply telling people a fuller understanding of our misperceptions.
that nearly everyone else in their area pays their taxes
on time increases payment levels by 15 percentage
points. This power of norms could be applied much
more widely across a range of issues.

10
IPSOS
Perils of Perception

VIEWS

3. Myth-busting is likely to
have limited impact –
because it misdiagnoses
why we’re wrong

“So, as we find in so much of our

A further important implication from the suggestion that
some of our apparent innumeracy is actually emotional
is that “myth-busting”, where we try to allay peoples’
concerns by telling them more about the facts, is likely
to have limited impact. It misdiagnoses a large part of
the issue, as our misperceptions are often an emotional communications work, just as
not a rational response.
important as providing a correct
This is not to say that we shouldn’t continue to
challenge the misuse of data by politicians and the picture of the facts is providing
media, through various fact-checking organisations that
have established themselves and done great work in an emotional narrative that appeals
the last few years. This may have a limited direct
impact on public perceptions, given it is working against to people, with its own role models
“
the weight and habits of the media and political rhetoric.
and vivid stories.
But the aims of these bodies are at least as much
preventative as corrective: the more those using statistics
badly are pulled up, the less likely they will think the risk
is worthwhile.

Even so, these steps will always struggle to get to a key

part of the problem. There are many instances where
the information provided by politicians or the media may
be perfectly factually accurate – but vanishingly rare.
The vivid anecdote is the only thing people remember,
and that drives their views of the importance or
prevalence of an issue. So, as we find in so much of our
communications work, just as important as providing
a correct picture of the facts is providing an emotional
narrative that appeals to people, with its own role models
and vivid stories.

11
IPSOS
Perils of Perception

VIEWS

4. But where people think This obviously won’t completely solve the problem – but
it is an important step. When people think worse of you
things are worse than they than is really the case, you should have little to fear from
telling the truth.
are, communicate the facts
more openly
The Perils of Perception work also helps to highlight where
“And these misperceptions matter: 23%
there is often little to lose from communicating the facts
more openly and transparently. In particular, where people claim to have boycotted a product or
think worse of you than the reality, you are better off at
least trying to tell them the truth. service because they think the company
This was highlighted in a further study14on perceptions of behind it has not paid the full amount of “
how much business contributes to the total tax-take in a
country – which is a key reputational issue, given the
tax they should.
increased focus on tax avoidance by major corporations.

People massively underestimate how much business as

a whole pays in – and massively overestimate how many
avoid tax. And, in the UK at least, the majority of people
think the “tax gap” between what the authorities expect
business to pay and what they actually pay has increased,
with hardly any (correctly) saying it has actually decreased
in the last few years.

And these misperceptions matter: four in ten say that not

paying the full amount of tax is one of their top three
reasons for thinking badly about a business. And more
directly, 23% claim to have boycotted a product or service
because they think the company behind it has not paid the
full amount of tax they should.

So in these circumstances, most businesses have little

to fear from transparency. The top two answers by some
distance are for businesses to publish the amount of tax
they pay (not buried in annual accounts, but prominently
on their website) and to provide an explanation for why
they pay that amount.

12
IPSOS
Perils of Perception

VIEWS

5. And while emotions rule,

improving understanding
and questioning of statistics
is still important
Finally, while myth-busting may have much more limited
“HG Wells said “Statistical thinking
impact than many might expect, the importance of the use
and understanding of statistics should not be downplayed.
Decades before the focus on big and omnipresent data,
will one day be as necessary for
data journalism and the rise of “sexy statistics”,15HG Wells
said “Statistical thinking will one day be as necessary for efficient citizenship as the ability “
efficient citizenship as the ability to read or write”.16

This shows incredible prescience on Wells’ part – but

to read or write”.16

whether we’ve lived up to the challenge is doubtful. We

still seem to undervalue skill with numbers, with, for
example, people in the UK over twice as likely to say they
would be ashamed about having poor reading or writing
skills than poor numeracy. And probably more tellingly,
people are four times as likely to say they would be proud
of their children for being good at reading and writing than
they would if their children were good at maths.

Of course (thankfully), we can’t all be “sexy statisticians”

– but there are practical things that can be done. On the
“supply side”, training those who are most likely to convey
statistics to us – particularly journalists and politicians – in
a more skilled and balanced way can only be a good thing.

But this will do little on its own, as the incentives need to

come from the “demand side” too, in our increased
understanding and questioning. This needs shifts in how
we view and teach mathematics, that’s more about the
active, sceptical citizenship that Wells envisioned, rather
than dry, abstract science. We have a long way to go.

13
IPSOS
Perils of Perception

VIEWS

End notes

Finally… 1. www.ipsos-mori.com/researchpublications/researcharchive/3664/Perceptions-are-not-
reality-what-the-world-gets-wrong.aspx?oUniqueId=13566

2. https://www.ipsos-mori.com/Assets/Docs/Polls/rss-kings-ipsos-mori-trust-in-statistics-
topline.pdf

3. http://www.youtube.com/watch?v=qzJxAmJmj8w

We believe that there’s a lot we can learn and apply to 4. https://pprg.stanford.edu/wp-content/uploads/Manuscript-A-Comparison-of-Mindless-

real issues from our understanding of misperceptions – ness-and-Survey-Satisficing-Vannette-D.-Krosnick-J..pdf

it helps government and business avoid communication 5. Tetlock, P.E. “Expert political judgment: How good is it? How can we know?”

mistakes, and to positively affect behaviour in smarter, 6. https://www.psychologytoday.com/blog/the-psychology-behind-political-debate/201008/

simpler ways. But there is one final benefit from the Perils immigration-innumeracy-how-well-informed-are

of Perception work – it reminds us of the incredible variety 7. http://dev.rickhanson.net/wp-content/files/papers/BadStrongerThanGood.pdf

of “reality” on different issues. 8. https://www.washingtonpost.com/news/volokh-conspiracy/wp/2014/11/03/political-igno-

rance-around-the-world/

In many ways it’s not our misperceptions but these 9. https://www.gov.uk/government/publications/fraud-error-and-debt-behavioural-insights-

team-paper
realities across different countries that are the most
interesting and important aspects of the study: 10. http://www.ukactive.com/turningthetide/

11. http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(12)61689-4/abstract

12. http://www.influenceatwork.com/robert-cialdini-phd/biography/
• The fact that the top 1% in Russia own 70% of the 13. http://www.mdpi.com/2071-1050/5/5/1829
nation’s wealth when the top 1% in New Zealand only
14. https://www.ipsos-mori.com/researchpublications/researcharchive/3633/People-signifi-
own 18%. cantly-underestimate-how-much-tax-revenue-comes-from-business.aspx

15. http://www.mckinsey.com/industries/high-tech/our-insights/hal-varian-on-how-the-web-
• That two-thirds of Americans are overweight or challenges-managers

obese, but only 23% in Japan. 16. Wells, H. G., “Mankind in the Making”

• That half of Italians aged 25-34 still live with their

parents, when it’s only 4% in Norway. Further information
• That the average age in India is 27, while it’s 47 in Japan. For the full results of the latest Perils of Perception study
and all previous studies, please see our dedicated
• That only 10% of politicians are women in Brazil when microsite at www.ipsos-mori.com/perilsofperception
44% are in Sweden.
Or contact:

When the reality is so extraordinary and varied, it’s Bobby Duffy

no wonder we’re so wrong. Global Director, Ipsos Social Research Institute
Bobby.Duffy@ipsos.com

14
IPSOS
Perils of Perception

VIEWS

Bobby Duffy is the Managing Director of the Ipsos MORI Social Research Institute and Global
Director of the Ipsos Social Research Institute. Bobby joined MORI in 1994 after graduating in
Economics/Social and Political Science from St John’s College, Cambridge, and now leads
a team of around 200 researchers in London, Manchester, Edinburgh and Brussels.

While at Ipsos MORI, Bobby has spent time on secondment at the Prime Minister’s Strategy
Unit and as a User Fellow at the Centre for Analysis of Social Exclusion at the LSE.
He set up and ran our Research Methods Centre, and his methodological development work
has included two papers shortlisted for paper of the year in the Market Research Society’s
journal.

He is a frequent public speaker and commentator, having appeared on most of the key UK
outlets, including the Today programme, Newsnight, ITV News at 10, BBC News at 6 and
many others. He has written widely on social policy issues, including in the Guardian, Sunday
Times, Wall Street Journal, Prospect, Daily Telegraph and many others.

He is a Visiting Senior Research Fellow at King’s College London and is a Trustee for the
LankellyChase Foundation.

The Ipsos Views white www.ipsos.com

papers are produced by the @_Ipsos
Ipsos Knowledge Centre.

GAME CHANGERS

<< Game Changers >> is the Ipsos signature.

At Ipsos we are passionately curious about people,
markets, brands and society. We make our changing
world easier and faster to navigate and inspire
clients to make smarter decisions. We deliver with
security, simplicity, speed and substance. We are
Game Changers.