Lying with Statistics

Lying with Statistics

Also called statistical fallacies or misunderstanding the nature of statistics.

This can be seen as an entire class of fallacies that result in presenting statistical data in a very biased way, and of course, interpreting statistics without questioning the methods behind collecting and presenting the data. Darrell Huff’s How to Lie with Statistics, published in 1954, discusses ways data can mislead.

The excerpt does not include a logical form or the original example arguments. It provides the following commentary.

Example #1

The first question that should immediately come to mind is, how on earth can one measure morality? With such a loose definition, it is not hard to imagine one collecting and measuring the data that only supports her desired outcome for the “numbers”. Furthermore, what is a “huge spike”? Visually, you can play with graphs to make numbers seem much more dramatic by not starting at zero, or by doing that little “chopped section” thing. For example, let’s accept that last year 20% of all people were immoral. This year it is 22%. This is a change of two percentage points. If shown on a graph with a vertical axis of 0% to 100%, the line connecting the 20% to the 22% would be barely inclined. However, if shown on a graph with a vertical axis of 20% to 25%, the line connecting the 20% to the 22% would appear to be a huge spike. The same exact data, a very different presentation.

Example #2

First, atheism is not a belief – it is a lack of one. Second, many non-believers are not even familiar with the term, “atheist”, and often consider themselves Christian, Jewish, or some other religion, based on their culture and family tradition, not necessarily their beliefs. A survey’s questions and category definitions determine whether it captures these distinctions.

Exception: At times, careful and honest explanations of the data and the presentation can void statistical fallacies, but like virtually all exceptions, this can be debatable.

Variation: Misunderstanding the nature of statistics is related to this fallacy, but the fallacy rests on the person interpreting the statistics. For example, you might be very troubled to find out that your doctor graduated in the bottom half of her class. But that is half of all the doctors in the world, and to be expected.