Transcript Quantitative methods for politics: an introduction

Quantitative methods for
politics: an introduction
Massimo De Angelis
•
•
•
•
•
•
“how to lie with statistics”
Graphs and Charts
Descriptive Stats
Measurements and the “invisible”
Growth rates
An application: the Political Economy of
growth rates.
“how to lie with statistics”
• 1954 Darrell Huff's perennially best-selling[1]
introduction to statistics for the general reader.
– Correlation does not imply causation
• numerous epidemiological studies showed that women
who were taking combined hormone replacement
therapy (HRT) were found to have a lower than
average incidence of coronary heart disease (CHD),
leading doctors to propose that HRT was protective
against CHD.
• Controlled trials, however, showed that HRT actually
caused a small but significant increase in risk of CHD.
• Re-analysis of the data showed that women
undertaking HRT were more likely to be from socioeconomic groups ABC1, with better than average diet
and exercise regimes. The two were coincident effects
of a common cause, rather than cause and effect as
had been supposed.[
– Statistical graphs can be used to distort reality:
• By truncating the bottom of a line or bar chart, one
makes differences seem larger than they are
• By representing one-dimensional quantities on a
pictogram by two- or three-dimensional objects to
compare their sizes, one makes the reader forget that
the images don't scale the same way the quantities do.
Two rows of small images would give a better idea than
one small and one big one.
Statement: Carpenters in Country 1 make 1/2 as much as Carpenters in
Country 2.
The first one is inaccurate and the second one is correct.
First Picture doubled the height of the smaller money sack, but then the artist
had to make the width larger for the money sacks to be proportional.
According to the first picture 4 of the smaller money sacks can fit into the
larger one. Therefore Carpenters in Country 1 make 1/4 as much as
carpenters in Country 2?
Basic descriptive stats
• 100, 100, 99, 98, 96, 95, 92, 90, 89, 73,
70, 10, 10, 10, 10, 9, 9, 8, 8
• Find the
• mean => 58.55
MEASURES OF CENTRAL TENDENCIES
• median => 81
• mode => 10
• standard deviation => 41.998
}
– How far on average the variable’s
observations are from the variable’s mean.
• Standard deviation =
• normal distribution
• Skewness
• Unimodal and
multimodal distribution
Index numbers
* index numbers are time series
summarising movements
in a group of related variables.
The best-known is the
consumer price index which
measures changes in retail prices
paid by consumers.
•What is included in the CPI?
•What is the base year?
Measurement and the “invisible”
• “GDP growth will solve our problems” . . .such as
– Reduce poverty
– Improve environment . .etc.
• Problem
– GDP is a monetised measure of social production
• What activity and who are the excluded from this measure?
• GDP per capital (often provided as a measure of “progress”)
is hides income distribution,
• See for example http://www.twnside.org.sg/title/2086.htm
http://www.econbrowser.com/archives/2007/01/sachs.png
Rates of Growth
• Rate of growth =
(Change in a variable/Value of variable t0)*100
Bacteria example =>
the Political Economy of growth rates:
the productivity deals of post-WWII
P = Profit
W = Wages
L = Labour hours
Y = Total Output
Y=P+W
g(Y/L) = g(P/L) + g(W/L)
Or g(P/L) = g(Y/L) – g(W/L)
If g(Y/L) = g(W/L) = g(P/L)
same income shares in
society
G => indicates growth
From http://homepages.uel.ac.uk/M.DeAngelis/EC213ch6b.doc