Scientific American - 09.2019

(Darren Dugan) #1
100 Scientific American, September 2019

GRAPHIC SCIENCE
Text and Graphics by Alberto Cairo

Life Expectancy at Birth (years)

Obesity Rate (percent of the population)

01020304050

U.S.

55

55

65

75

85

65

75

85

¹àày ̈D ́`¹yˆ`Ÿy ́ïi

 o er  iddle ncome

 pper iddle ncome High ncome

02040

76

78

80

82 Correlation

20 25 30 35 40

02040

02040

Correlation


Correlation


Correlation


Correlation


Iowa

Mississippi

Hawaii

D.C.

Life Expectancy at Birth (years)

55

65

75

85

Life Expectancy at Birth (years)

Life Expectancy at Birth (years)

o ncome

Obesity Rate (percent of the population)

Obesity Rate (percent of the population)

Obesity Rate (percent of the population)

76

78

80

82

40,000 60,000 80,000
Median Income per State (2017 dollars)

02040

Correlation


Each blue dot is a country

Hawaii
Iowa

Mississippi

D.C.

Each orange
dot is a state

C NS LTANT: HEATHER  RA SE

Datassist

 SRCES: “ASSCIATIN BETWEEN CLASS III BESITY (B I F 09 G

2 ) AND   RTALITY: A P  LED ANALYSIS  F 20 PR SPECTIVE ST DIES,”

BY CARI .  ITAHARA ET AL., IN

PLOS MEDICINE;

J LY 8, 201  CIA W RLD FACTB   (

worldwide obesity rates, 2016

) W RLD BAN (

life expectancy at birth rates and income group classification, 2016

)

CDC (

U.S. obesity rates, 2017

) HOW CHARTS LIE: GETTING SMARTER ABOUT VISUAL INFORMATION,

BY ALBERT CAIR , W. W. N RT N (IN PRESS) DATA F R WISC NSIN AND  AINE N T AVAILABLE

Does Obesity Shorten Lives?


Misreading data visualizations can reinforce biased perceptions


“A picture is worth a thousand words.” That saying leads us to be-
lieve that we can readily interpret a chart correctly. But charts are
visual arguments, and they are easy to misunderstand if we do not
pay close attention. Alberto Cairo, chair of visual journalism at the
University of Miami, reveals pitfalls in an example diagrammed
here. Learning how to better read graphics can help us navigate
a world in which truth may be hidden or twisted.
Say that you are obese, and you’ve
grown tired of family, friends and
your doctor telling you that obesity
may in crease your risk for diabetes,
heart disease, even cancer—all
of which could shorten your life.
One day you see this chart ( right ).
Suddenly you feel better because
it shows that, in general, the more
obese people a country has ( right
side of chart ), the higher the life
expectancy ( top of chart ). There-
fore, obese people must live longer,
you think. After all, the correlation


( red line ) is quite strong.
The chart itself is not incorrect.
But it doesn’t really show that the
more obese people are, the longer
they live. A more thorough des crip-
tion would be: “At the national
level—country by country—there
is a positive association between
obesity rates and life expectancy at
birth, and vice versa.” Still, this does
not mean that a positive association
will hold at the local or individual
level or that there is a causal link.
Two f allacies are involve d.

First, a pattern in aggregated data can disappear or even reverse
¹ ́`yĂ¹ùyāÈ ̈¹àyï›y ́ù®UyàåDïmŸ‡yày ́ï ̈yÿy ̈幆myïDŸ ̈Άï›y
countries are split by income levels, the strong positive cor relation
Uy`¹®yå®ù`›ĀyD§yàDåŸ ́`¹®yàŸåyåÎ ́ï›y›Ÿ‘›yå ́`¹®y
nations ( chart on bottom right ), the association is negative (higher
obesity rates mean lower life expectancy).

The pattern remains negative when you look at the U.S., state by state: life expectancy
at birth drops as obesity rises ( left ). Yet this hides the second fallacy: the negative
Dåå¹`ŸD ́`D ́UyD‡y`ïymUĂ®D ́Ă¹ï›yà†D`ï¹àåÎāyà`ŸåyD ́mD``yååï¹›yD ̈ï›`Dàyj
for example, are associated with life expec tancy. So is income ( right ). The fallacy is trying
to determine something about your individual risk by looking at aggregated data that
m¹ ́¹ïàyŒy`ïŸ ́mŸÿŸmùD ̈`Ÿà`ù®åïD ́`yåΆŸ ́åïyDmĂ¹ùåDĀmDïD¹ ́Ÿ ́mŸÿŸmùD ̈åĀŸï›Ÿ ́D
large sample of randomly selec ted people, you might discover that obesity may, or may
not, relate to life expec tancy for someone in your situation.

●^1 Try to see not just what
a chart shows but what it
may not be showing.
●^2 Don’t jump to conclusions,
particularly if a chart
åyy®åï¹`¹ ́Šà®Ā›Dï
you already believe.
●^3 Question whether you
are correctly verbalizing
the chart’s content.
●^4 Consider whether the data

represent the level required to
make the inferences you want.
†Ă¹ùĀD ́ïï¹ ̈yDà ́DU¹ùï
countries, say, consult data at
the country level, but if you want
to learn about your own health
àŸå§åjŠ ́mmDïDDU¹ùïŸ ́mŸÿŸmùD ̈åÎ
And either way, always remem-
ber that, in a chart or among
any data, correlation is not the
same as causation.

What to Do
Free download pdf