# How to Lie with Statistics

I’ll never forget a presentation I made several years ago during an Open House event at Villanova. The event was geared towards high school seniors and their parents who were visiting the campus for the day to learn more about the school.

I had volunteered to give about a 15-minute talk about the advantages of accounting as a major and possible career. One of the things we are proud of with our accounting graduates is their placement rate; usually at or close to 100% at graduation. I also noted that the average starting salary at the time was in the low \$50,000s.

After my presentation, the next faculty member gave her pitch for the Management Information Systems major. Such a major is usually not as popular in terms of the number of students pursuing such a degree when compared to majors such as finance or marketing or accounting.

The professor started talking about the placement rate for MIS majors and then stated that the average starting salary for MIS majors was over \$100,000.

When I heard that number I was in shock; it was incredibly impressive. She paused for a moment to let it sink in, and then noted that Brian Westbrook had just graduated as an MIS major. Westbrook did not pursue a career in technology, instead becoming a member of the Philadelphia Eagles with an initial contract of around \$600,000. When his salary was factored in with the starting salaries of the handful of MIS graduates (low to mid \$40,000), the mean salary for an MIS major was indeed over \$100,000.

This event came back to me today because of a story in the Wall Street Journal that starts with the line, “According to one recent study, 54% of children have below-average reading skills.” The reporter,  Eugenia Cheng (a math Ph.D.), pointed out the apparent paradox. “By definition, it seems that in any group 50% must be above average and 50% below average. But it all depends on what you mean by “average.” We often associate the word with the middle of a range of numbers, but in fact, there are different types of averages that represent different things. Cheng then goes on to explain the difference between the mean, the median, and the mode.”

It also reminded me of one of the top selling math books of all-time, “How to Lie with Statistics“, written by Dan Huff in 1954. The book offers an introduction to statistics for the general reader. Not a statistician, Huff was a journalist who wrote many “how to” articles as a freelancer. The book is a brief, breezy illustrated volume outlining errors when it comes to the interpretation of statistics, and how these errors may create incorrect conclusions.

The point of all of this is to remind you to be careful when interpreting statistics; the person presenting the statistics could be biased and using the numbers in a particular way to get his or her point of view across.

Heck, even I have been guilty of such behavior. I once noted that between Katy Perry, Justin Bieber, Taylor Swift, Barack Obama, and myself, we averaged nearly 57 million Twitter followers…

*image from CBS Sports

## 3 thoughts on “How to Lie with Statistics”

1. The classic example of why mode averages are needed alongside means and medians is when Bill Gates walks into a restaurant with 100 patrons the mean and median wealth of those patrons is over \$1 billion for a short while.

I provide over 400 illustrations of how to mislead with statistics at

Like

2. Oops, I meant to say why median and mode averages are needed along with mean averages. The median is not \$1 billion when Bill Gates walks into a bar. The mean is over \$1 billion.
It’s too early on a Sunday morning.

Sorry

Like

1. I agree, all three measures have their role in describing a population. Looking forward to scanning through all you’ve got on your site about misleading stats…

Like