Erich vs. the Third Kind of Lie

Benjamin Disraeli (1804 - 1881) was a British statesman who once said, "There are three kinds of lies: lies, damned lies, and statistics."

Well, Ben, it isn't just statistics. Often the government, private interests, or the press misstate (accidentally or otherwise) statistics, mathematics, and science. It can be very misleading. We'll try our best to combat that problem here.

Every vs. Average

Yesterday, WITF reported the following:
Pennsylvania is home to 120,000 high school dropouts between the ages of 16 and 24. Now, the Pennsylvania Partnerships for Children is launching an effort to re-connect those young people by helping them get their GED and pursue post secondary education or an associates degree. President and CEO Joan Benso says high school dropouts are a drain on the economy because they are more likely to be unemployed, get arrested and live in poverty. "Every high school dropout under the age of 25 costs the taxpayers at least $683 a year more in state and federal and local supported services than they consume." she says. Benso says nearly 50 percent of the jobs in Pennsylvania require more than a high school diploma, but less than a bachelor's degree -- so, that's an area her group is hoping to move dropouts into. She hopes the next governor will make dropout re-engagement a priority.

Now, this is a pressing concern, and I don't mean to minimize it. But, it does bring up an important mathematical distinction that has been misused here. It is the difference between "every" and "average". Look at the quote by Ms. Benso above, and I will explain what I mean.

First, average can be used for several different measurements. In this case, the average being used is the mean, what most people think of as the average. It is where you add up all the values in a set and then divide by the number of values you had. I looked up the statistic Ms. Benso was quoting, and the number $683 is the mean of the amount that high school dropouts are consuming in services minus the amount they are contribuing in tax dollars.

So what's the difference, right? What Ms. Benso could accurately say is "The average high school dropout under the age of 25 costs the taxpayers..." rather than "Every high school dropout..." Saying every means that if you pick any single high school dropout in PA under the age of 25, that person costs the taxpayers $683 more for services than they give back in tax dollars. Every single one is exactly the same, and is costing taxpayers money.

But what's really true is that some high school dropouts in PA under the age of 25 cost the taxpayer a lot more than $683, and others cost less. Some might even be contributing more in taxes than they are consuming. Yes, on average, each dropout costs us money. But any individual dropout could be doing very well in terms of income and taxes paid in.

Is it important? In this instance, perhaps not. But take a different example. The average American costs $1066 in health care expenses annually. If this were "every American costs $1066 in healt care expenses annually", that means would have to charge at least $1066 in annual premiums, really much more to make a profit. Because they would be paying out this amount for everyone they covered. The people who are just going to the doctor for physicals are only spending a couple hundred dollars. But some people with massive medical needs are spending hundreds of thousands of dollars. If you are young and healthy, you don't want your insurance premiums to be based on everyone, you want it to be based only on people like you, young and healthy. So you want the insurance companies to recognize an average as an average, not as an every.

So, when you are stating numerical facts, always be careful about the word "every." Every may seem very similar to average, but on average, they are nothing alike.

2 million? Would they be unemployed?

In the President's State of the Union speech on 1/27/2010, he said the following, "Because of the steps we took, there are about two million Americans working right now who would otherwise be unemployed." Two million, wow. That's a big number. First, let's think about what that means.

Right now unemployment is 10% with 15.3 million people unemployed. So if President Obama's number is accurate, then without the Recovery Act or 2009, there would be 17.3 million people unemployed. That would be an unemployment rate of 11.3%. (Click here to see how I got this number.)

So if this number is correct, it is a significant number. That is a lot of change. To be confusing with percentages, the change from a 10% unemployment rate to an 11.3% unemployment rate is 13%. (Click here to see how I got this number.) Sorry, I know the percentages all look the same.

I decided to research the validity of the 2,000,000 jobs number. And where else better to start than the government's own Recovery Accountability and Transparency Board. According to their numbers, which were last updated in October of 2009, there have been 640,329 jobs saved/retained by the Recovery Act. Of course, it is possible that more have been created or saved since October, but from 640,329 to 2,000,000 is an increase of 1,359,671 jobs, or an increase of 212%! (Click here to see how I got these numbers.) That's a big increase, Mr. President.

But wait. The government has changed how they calculate the number of jobs created/saved. It's now based on the number of salary hours being paid for by firms who received recovery dollars. But it includes all jobs created since the company received the recovery money, all jobs they were going to eliminate but didn't due to recovery dollars, and all jobs that they would have kept anyway, but are still paying for using Recovery dollars. That's not just jobs created or saved. It is also jobs that are being paid for anyway. Not the same thing. I'm not saying this method is illegal or immoral, but I do believe that it's misleading and not a good example of full disclosure. But it wouldn't sound as good in a televised speech to say, "Because of the steps we took, there are about two million Americans working right now who are being paid by Recovery dollars we got from you, the taxpayers, some of whom would have kept their jobs anyway."

The lesson: when you hear big numbers sound big. That doesn't mean they actually are.


15.3 million + 2 million is 17.3 million.
Now if 15.3 million unemployed corresponds to 10.0% unemployment, what does 17.3% correspond to?

15.3/10.0 = 17.3/x
We can multiply both sides by 10.0x and we get 15.3x=173.
Divide both sides by 15.3 and get x=173/15.3=11.3.

Click here to return to the entry.

A percent increase is the amount of increase divided by starting number then multiplied by 100%.
(11.3%-10.0%) is 1.3%, that's the amount of increase.
Next divide by 10.0%, the starting amount. 1.3%/10.0% is 0.13.
Finally multiply by 100% and we get 13%.

Click here to return to the entry.

Again, a percent increase is the amount of increase, divided by the starting number, then multiplied by 100%.
(2,000,000 - 640,329) = 1,359,671. The amount of the increase.
Next divide by 640,329, the starting amount. 1,359,671/640,329 = 2.12
Finally multiply by 100% for a 212% increase! Big increase!

Click here to return to the entry.

divider image

Photo Creditswebsite design & development by: SiteStrux, Inc.