Lies and Statistics
The most frequently quoted remark about statistics is probably: "There
are lies, damned lies and statistics." The danger of interpreting percentages
is beautifully illustrated by the invention of an imaginary advertisement for a
"Rabbitburger". Everyone knows what a Hamburger is. The Rabbitburger, it was
claimed, was 50 percent rabbit and 50 percent horse, which investigation
reveled as meaning that it was made from ration of meat of one rabbit to one
horse!
Percentages do not present us with much difficulty so long as we remember what percentage of what we are talking about. Averages may be more difficult in that we have to know what kind of average our statistician is feeding us with.
For example, if we have ten guinea-pigs, and we divide the total weight
of the animals, we should arrive at the mean weight. However, if in our sample
group there happened to be a giant-size guinea-pig, he would distort the
average (i.e. mean), so as to make it meaningless statistics.
Similarly, one underfed guinea-pig in the sample would cause a deviation
from the mean, although it would not distort the mean as much as the giant
weighing many times more than the average. In order that the average should not
be unduly distorted causing a distortion as against normal distribution curve,
we can ignore the exceptions but then we no longer have a mean average, but a
median.
In addition, should we for some reason want to fit all our guinea-pigs
with woolly jackets and not want to have to buy too many different sizes, we
must find out which size occurs most frequently and that size will be neither
the mean not the median; it will be the mode.
Yes
No
Flag
Comments
50 percent of all statistics are lies...
You must be logged in and verified to post a comment. Please log in or sign up to comment.