When Math Lies

This just in:  I had a friend mention that she didn’t think gun research would work due to two reasons: (1) It would take too long and (2) people would still choose sides on the research results.

My less-than-clever retort was “2+2 will always equal 4”.

Then, I thought about this some more and decided there were times were math could be used to fool you.

Let’s look at a couple of examples:

A political example.

Newt Gingrich said, a month after Obama took over the presidency that “There are more people on food stamps than at any other time in history.”  Newt was right.

However…

The implication is that in a single month, Obama was responsible for every single person on food stamps.  What had actually happened was that the economy crashed at the end of the Bush administration.  It was continuing to sink for the first few months of Obama’s administration.  So, under Bush, a record was set for people on food stamps.  As soon as Obama took over, and one more person signed up for food stamps..a new record was set.

So, one person signing up for food stamps under Obama allowed Newt to rightfully say what he said.

A non-political example…

Suppose a drug company has a new product; an antidepressant called drug XYZ.  The FDA will only approve use of this drug if there are 3 successful trials showing this drug can outperform a placebo.  So, the drug company runs 100 trials.  In 80 of those trials, the drug does no better than a placebo.  In 17 of those trials, the placebo actually outperforms the drug.  In 3 trials, the drug outperforms the placebo.

The drug company then throws out 97 of the trials that didn’t give them the results they wanted.  Instead, they publish the 3 trials out of 100 that seemed to show their drug worked.

The FDA approves this because the 3 trial standard has been met.

A stock market example…

A broker sends out 100,000 texts.  In half of these he recommends buying stock in company XYZ as it will increase in value during the year.  In the other half of these, he recommends the exact opposite saying that the company will decrease in value next year.

Unless the company stock does not change at all, this broker will be right in half of his texts.

Next year, he does it again.  However, this time, he only sends the texts to the 50,000 that got the right recommendation last year.  He claims “I was right last year!”  In the new text, he has another company but dhe sends the say 50/50 recommendation; half say “buy” and half say “sell”.

Again, at the end of the year, half of these texts will be correct.

This means that 25,000 have gotten texts that accurately predicted the market.

He does this over an over; sending out 25,000 (50% buy and 50% sell) and now 12,500 will get the right recommendation. This is 12,500 that have only seen texts from this broker that have been 100% correct.

Every year he has a smaller pool to send to. Every year, half will get good recommendations and every year, half think he is a genius….even though he is wrong half the time.

Why is this?

Those that get bad recommendations are dropped from his mailing list.  Those that get good recommendations have no idea about the 50% that got bad recommendations.

Math counts and 2+2 will always equal 4.  However, we can all be fooled; especially when we are hearing what we want to hear.

Up, up and away…

Jim

Please follow and like us: