Question

Explain how statistics in the following statements may have been misrepresented through any
of the following: sources of data, data collection methods and/or data analysis and
representation. There may be more than one way in which the statistics could be
misrepresented. Give reasons for your choice(s).

a. A survey asks every other shopper at a local grocery store if they would prefer to buy
groceries online. Less than half agree to complete the questionnaire.

b. Why does chocolate A taste better than chocolate B? Rewrite this question to exclude bias.

c. Secure-It, a security system company, reports that homes that do not have an alarm system are
more likely than other homes to be broken into, as shown in the results below, of the
company’s recent study on reported break-ins in the city.*

*Margin of error plus or minus 6.5%.

Distribution of Break-ins (%) 20 22 alarm system no alarm system pets only pet and alarm

d. A survey of 17 students in a Gr. 12 Leadership class revealed that 90% of them are the first
child in their families. Therefore, the first child in the family has strong leadership skills.

Answer 1

(a)

The survey has non-response bias. Since there is no compulsion or incentive to complete the questionnaire, many respondents may just ignore the survey.

(b)

To question to exclude bias: Which chocolate do you prefer as better taste: chocolate A or chocolate B.

(c)

Since the y axis does not start from 0, the bar diagram misleads the viewer as though the differences are large. In case the bars are drawn correctly with y axis startingfrom 0, we vcan note that the differences are not large.

(d)

Just because a strong correlation is noted between the first child in the family and strong leadership skills, we cannot conclude that the first child in the family has strong leadership skills. Correlation does not imply causation.Correlation just informs us that there is a lnear relation between the first child in the family and strong leadership skills..Correlation does not say that one causes the other.