Lies, Damn Lies, and Statistics

Mark Twain popularized the quote long ago, and I was reminded of it today when reading this article about drug testing in North Carolina.

At issue is this quote:

Here are the stats so far: Over the first five months of the program, 7,600 people applied for help. Of all those, social workers sent 89 to be screened.

Twenty-one from that group tested positive for drugs. That means of the total number of welfare applicants, less than 0.3% were denied benefits.

“I think it’s a total waste of money,” said Kight. “It’s expensive and unnecessary.”

Notice the percentage, 0.3%, used?

Why, exactly, would a reasonable person pull out the number of people testing positive for drugs, 21, versus the overall number of people applying, 7600? What useful information does that give us?

If each test costs $55, testing 7900 people costs a whopping $434,500, and you’re telling me that, for nearly half a million bucks, we only caught .3% of people using drugs?

No. No one at all is telling you that. But a reasonable person could conclude that’s exactly what the person who came up with the percentage wants you to think.

In reality, the state tested a total of 89 people, spending $4,895, and caught 21 of them, 23.6%. So if not giving benefits to those 21 people (or reducing benefits to some of those because they are responsible for dependents apparently) saves more than five grand, North Carolina just managed to save some money.

To be fair, the article gave me all the information that I needed to come to that conclusion, but the way that it presented the information was misleading at best. I have to rate it as … unreasonable.