Bias in Presenting Data
OBJECTIVE To learn to avoid misleading ways of presenting data
In this chapter we studied how data can be used to discover hidden relationships in
the real world. But collecting and analyzing data are human activities, so they are not
immune from bias. When we are looking for trends in data, our goal should be to dis-
cover some true property of the thing or process we are studying and not merely to
support a preconceived opinion. However, it sometimes happens that data are pre-
sented in a misleading way to support a hypothesis that is not valid. This is some-
times called fudging the data, that is, reporting only part of the data (the part that
supports our hypothesis) or even simply making up false data. Mark Twain suc-
cinctly described these practices when he said, “People commonly use statistics like
a drunk man uses a lamppost; for support rather than illumination.”
Here’s a simple example of how we may choose to present data in a biased fash-
ion. Tom and Harry compete in a two-man race, which Tom wins and Harry loses.
Harry tells his friends, “I came in second, and Tom came in next-to-last.” Although
correct, Harry’s statement gives a misleading impression of what actually happened.
But misrepresenting or misinterpreting data is not always just silly—it can be a
very serious matter. For example, if experimental data on the effectiveness of a new
drug are misrepresented to bolster the claim of its effectiveness, the result could be
tragic to patients using the drug. So in using data, as in all scientific activities, the
goal is to discover truth and to present the results of our discoveries as accurately and
as fairly as possible. In this exploration we investigate different ways in which data
can be misrepresented, as a warning to avoid such practices.
I. Misleading Graphs
Although graphs are useful in visualizing data, they can also be misleading. One
common way to mislead is to start the vertical axis well above zero. This makes small
variations in data look large. The following graphs show Tom’s and Pat’s annual
salaries. Does Pat make a lot more money than Tom, or do they make about the same?
Look carefully at the scale on the y-axis before answering this question.
EXPLORATIONS
■
EXPLORATIONS
■
EXPLORATIONS
■
EXPLORATIONS
1
60,000
59,000
61,000
62,000
63,000
64,000
65,000
Tom Pat
10,000
00
20,000
30,000
40,000
50,000
60,000
70,000
Tom Pat
1. The business manager of a furniture company obtains the following data
from the accounts department. She needs to present a report on the financial
state of the company to the executive board at their annual meeting.
128 CHAPTER 1
Winner: Tom; Loser: Harry
www.grinningplanet.com