Debunked Legends

How to Lie With Statistics

You may have heard someone say that there are three types of falsehoods: lies, darn lies, and statistics.

Despite the imputation, statistics can be very compelling. Adding a statistic to your argument can be quite persuasive. You should always take statistical arguments with a grain of salt, however. It is well documented that 43.286% of all statistics are fabricated.

Even when a statistic is not made up, it may not be telling the entire story. Consider, for example, the wonder drug Lipitor.

Lipitor (generic name Atorvastatin) is a statin medication that is used to treat those who suffer from high cholesterol. When Pfizer marketed Lipitor, it was touted as a wonder drug. As society dealt with increasing problems of obesity and high cholesterol, Lipitor offered the eye-popping promise of a 36% reduction of heart attack risk.

Thirty-six percent! Who wouldn’t want to try something like that? All of us know people who suffered heart attacks while in the prime of life. Perhaps, with the help of Lipitor, you wouldn’t have to follow their example.

Advertisement for Lipitor, boasting a 36% reduction in risk of heart attack, while minimizing the clarification that the “36%” was really “1%.”

The number “36%” practically leaped off every advertisement that touted the effectiveness of the medication. The public was convinced. The medical community appeared to be persuaded as well. Prescriptions for Lipitor flew out of doctors’ offices at an unprecedented rate.

Lipitor quickly became the world’s best-selling medication. From 1996 to 2012, it generated more than $125 billion in annual sales, providing up to 25% of Pfizer’s annual revenue during that time.

If you took the time to read the fine print on the advertisements, you might have seen something that caused you to scratch your head. Overshadowed by the massive “36%” was a small footnote: “That means in a large clinical study, 3% of patients taking a sugar pill or placebo had a heart attack compared to 2% of patients taking Lipitor.”

Wait a minute…. Does Lipitor provide a 36% margin of protection, or is it 2%? Both numbers can’t possibly be correct, can they?

This is exactly why statistics have a reputation for fraudulence. As it turns out, Pfizer was telling the truth when it said Lipitor reduced the risk by 36%. It was also telling the truth when it said the risk reduction was 2%. What it didn’t specifically answer in each case was what each number was being compared to.

When you look at all of the numbers, you will see that the clinical trials for Lipitor looked at two groups. One was treated with Lipitor. The other received a placebo.

Heart attacks struck 1.9% of the Lipitor group. Of those who took nothing more than placebos, 3% had heart attacks. In other words, if you treat your high cholesterol with Lipitor, your likelihood of avoiding a heart attack improved by only 1.1% more than if you were taking totally-ineffective sugar pills.

Why don’t the Lipitor advertisements boast a 1.1% reduction in risk? Obviously, it is because not a lot of people would be willing to shell out the money for a prescription for such minimal benefit.

The reason Pfizer could justifiably claim a 36% reduction was by comparing the Lipitor group to the placebo group in a different way. Yes, there is only a 1.1% reduction in heart attacks, but when looking at the risk, statistics help frame things in a different light. If you took nothing by placebos, your risk of having a heart attack is 3%. By taking Lipitor, that risk drops to 1.9% — a difference of 1.1%. The comparative risk, therefore, can be found by dividing 1.1 by 3 — in other words, 36%.

Thirty-six percent describes the relative risk reduction of the results between the two groups. In terms of actual heart attacks, Lipitor drops the numbers by 1.1%, but when comparing the range between those taking placebos and the Lipitor group, there is a 36% relative difference.

Is that clear? If not, don’t worry. It is for that very reason that this kind of statistical manipulation is so effective.

Another way we see lying by numbers is in the way government budgets are reported. If you try to live within your budget, you understand that a “budget cut” means there is less money available to spend. If you had budgeted $100 to buy birthday presents but are told you have to implement a 20% budget cut, you know that you can only spend $80 on presents.

That’s not the way the government handles budget cuts. Ronald Reagan said, “Government programs, once launched, never disappear. Actually, a government bureau is the nearest thing to eternal life we’ll ever see on this earth.” Once a government program is established, it is assumed it will continue — at least into the next fiscal year. That being the case, there has to be money allocated to it for its budget.

Congress has established budgeting procedures that automatically increase the funding to programs. For discretionary programs, the procedure is called baseline budgeting. This assumes the spending in that program will increase at a level to keep pace with inflation. For non-discretionary programs, such as Social Security or Medicaid, the spending amount is established through legislation by determining how many people qualify for benefits and how much those benefits should be.

Let’s say there is a discretionary program called the Federal Initiative to Preserve Red Tape. It receives $100 million this year. Because the inflation rate is 8.5%, it will receive $108,500,000 next year — an increase in real dollars of $8,500,000.

Congressman Thrifty is concerned about the massive federal deficit. He thinks the Federal Initiative to Preserve Red Tape can get by with an increase of 1% ($1 million). Those who favor the Red Tape program are outraged. They hold a press conference and tell everyone that Congressman Thrifty wants to cut their budget by 7.5%. They say that this will cost them $7.5 million.

By way of compromise, Congress approves a budget for the department with a 5% increase. The advocates for the department are still upset that they got a 3.5% cut in their budget, even though they received $5 million more than they did the prior year.

In 2017, Congress adopted a budget that reduced Medicaid spending by $834 billion over ten years. When you crunch the numbers, however, you will see that total Medicaid spending increases from $393 billion in 2017 to $474 billion in 2026. Supporters of the measure could still argue that they had achieved a cut because the spending is less than what was otherwise planned.

It’s a lot like going on an uncontrolled spending spree but claiming that you saved so much money because of all the things you bought on sale. It is the same strategy as going on a diet and eating whatever you want, but opting for a diet soda, instead of the regular kind. You conveniently forget the 2,500 calories you just packed away by pigging out on pizza, but delight in the knowledge that you cut 150 calories by forgoing that can of Pepsi.


New Yorkers Are More Dangerous Biters Than Sharks

No one wants to run into a shark while swimming. A shark’s bite can definitely ruin your day. There is, however, another species that seems far more likely to take a chomp at you. It is a peculiar breed of creature found in the dangerous concrete jungle of New York City. It is called a…

Meet the Man Who Was Struck By Lightning 7 Times

Whoever said lightning never strikes twice never met Roy Sullivan. Sullivan was born on February 7, 1912. From 1936 to his retirement in 1977, he served as a United States park ranger in Virginia’s Shenandoah National Park. By all accounts, he performed his duties competently. It is was what happened to him on the job…

1 reply »

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.