Transcript 10waystomisleadwithstatistics

TEN WAYS TO MISLEAD
WITH STATISTICS
•
•
Data Collection
Presenting Data
1-4: Data Collection
1. Shape the answers with a
biased or leading question.
Don’t you think that we should
stop physical contact in the NHL
so that players like Sydney
Crosby don’t get concussions?
2. Shape the answers by
limiting the answer options.
What’s your favourite kind of music?:
a) country
b) pop
c) opera
3. Confuse respondents with
a vague question.
Do you think that the hat policy is
really not a very good idea?
4. Skew your results with a
small sample size.
Would you have been able to
predict the male and female fish
with a sample size of 2 (out of a
population of 50)?
5. Skew your results by
selecting a biased sample.
To find out about Facebook use
among GVC students, you survey
your 5 friends – would the results
be accurate?
6-10: Presenting Data
6. Leave out important information

Leave the y or x axis undefined.

Do not define key terms

Keep the variables vague
“Man. named child poverty capital”
Tuesday, November 24, 2009, CBC News
The Social Planning Council of Winnipeg
says Manitoba is the child poverty capital of
Canada once again.
In a report released Tuesday, the council
stated that nearly ONE IN FIVE Manitoba
children lives in poverty and more than 68% of
aboriginal children under the age of six in
Winnipeg live in poverty.
What info is missing?
How do they define poverty?
 Poverty is defined as an income of
$38,000 for a family of four.
 Cost of living in Manitoba is much less
than in most other provinces.
 People think of poverty like the starving,
naked children in Africa.

Vehicle Safety
Probability
of:
SUV
Minivan
Life
threatening
head injury
Life
threatening
chest injury
Life
threatening
leg injury
16%
2%
20%
4%
35%
1%
What info is missing?
What speed was used in the crash tests?
 What are other factors involved in
possible accidents?
 How old are the vehicles that were
tested?
 What kinds of vans and SUVs were
tested?

7. Skew the baseline of a bar graph.

Draw attention to small differences by
manipulating the y-axis.

Start the y-axis at a number that is close
to your results.
8. Use a misleading average.
Having one zero or one high mark skews
what the true average is.
Eg:The average mark for the poetry
assignment in Mr. G’s class is 66%
based on a select sample from the class.
Johnny: 81% Jimmy: 83% Jennifer: 85% Jake: 0% Jill 82%
Without the 0% the average is now 83% - a more
honest reflection of how this group performed.
9. Use percentages instead
of raw numbers.

Percentages tend to hide the sample size
For Example:
 47% of Canadians disapprove of Stephen
Harper’s job performance – CBC News (Dec 2009).
 Raw data: 141 out of 300 randomly selected
Canadians (total population approx 25 million)
10. Make something out of nothing.
Draw attention to big numbers.
Eg: “My opponent wants to spend two
million dollars on a new logo for
Manitoba.”
Two million dollars is only .01% of
Manitoba’s budget. Funding for Public
Schools in 2008 was 238 million and a
much higher % of the total budget.