Professional Documents
Culture Documents
Can be very effective: False impression of accuracy Figures never lie, but liars figure. Make the son of a bitch deny it. There are three kinds of lies: Lies, Damned Lies, and Statistics. --attributed to Benjamin Disraeli Swift Boat Veterans for Truth Lyndon Johnson:
Need to know how to lie with statistics, to keep from being lied to with statistics.
Also talking about ways to say things that are true, but misleading
11/17/2008
"I guess I should warn you that if I turn out to be particularly clear, you've probably misunderstood me.
-Alan Greenspan at his 1988 confirmation hearings.
Siblings
Sibling Distribution
60 50
Median
# Respon ndants
Pull mean up
40
Mean
30 20 10 0 0 2 4 6 8 10
Number of Siblings
11/17/2008
Other Examples
Tax Cut 111 million taxpayers will save, on average, $1,586 off their taxes. Facts: 1) 25% receive NO cut (drops mean to $1,217)
Age Distribution
Median
30
# of Respondants
25 20 15 10 5 0
Mean
20
30
40
50
60
Age
11/17/2008
Guys Rule!
Gender Distribution
Honest presentation Full scale shown Bars same width, color Slightly more male students Not a big difference
100 % of Respondants
Gender Distribution
54 52 50 48 46
50
25
0 Male Female
44 Male Female
Average Income
150,000 100,000 50,000 0 19 20-29 30-39 40+ 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0 19
Average Income
20-29
30-39
40-49
50-59
60+
Average Income
Average Income
Average Income
300,000 250,000 200,000 150,000 100,000 50,000 0 19 20-29 30-39 40+ 300,000 250,000 200,000 150,000 100,000 50,000 0 19
Average Income A I
20-29
30-39
40-49
50-59
60+
Average Income
Average Income
11/17/2008
Bully Others Good of bad statement: About the same percentage of kids bully others and are bullied Bullied: The claim that 41.7% of students face a degree of bullying is not a lie However, it would have been equally true to claim that only 16.9% of students were victims of bullying sometimes or weekly
This map shows how people in different states of the US like pizza. (These data were made up). The code for the state color is: Red States = People Love Pizza Yellow States = People Like Pizza Purple States = People Hate Pizza
The map really doesn't say very much. We don't know
how it was determined that people like pizza...were people asked if they liked pizza? Were people asked how much pizza they ate in a week? a month? a year? Was the number of pizzas purchased at stores in different states counted? Was the number of pizza restaurants in different states counted?
What to Do?
Questions to ask about any statistic: 1) Who created it?
Do they have an agenda?
There are no scales or measurements to indicate any of this information. Although this type of graphic gives almost no information, it is used frequently in many popular magazines.
Another example, who cares if in our elections a state is blue or red since it is the electoral votes that count
11/17/2008
What to Do
Questions to ask about any statistic:
Conscious bias
Direct misstatements Ambiguous statement Selection of favorable data
Suppression of unfavorable Units of measurement may be shifted
5) Is it relevant?
Avoid misdirection
Are there enough cases to add up to any significance? Look for a measure of reliability (sources for error) What is missing:
Look out for average: variety unspecified, a matter where mean and median might be expected to differ substantially Percentages are given and raw figures are missing The factor that caused change to occur, implying that some other more desired factor is responsible (Ex. Looking at total deaths rather than death rate, dont forget there are more people now than there used to be)
Watch out for a switch somewhere between the raw figure and the conclusion Strange things crop out when figures are based on what people say -- even about thi things th t seem t b objective f t that to be bj ti facts
11/17/2008
A lot of presented data is truthful Most people are willing to present unbiased statistics We are learning how to read and interpret data (judgment)