The Crowhill Report - Content

crow
Views and opinions on the news, culture, politics, beer, art, science, education, religion and ethics

Sites endorsed by Crowhill:
Crowhill Publishing Homebrewbeer.biz
The Krehbiel Report on Publishing@gregkrehbiel


How to lie with statistics

by Greg Krehbiel on 8 November 2017

At approximately the same time, a Facebook friend complained about “spin doctors,” and Pentamom sent me an article about the misuse of statistics.

I defended the spin doctors.

There is value to using the right word at the right time in the right setting. It makes the message resonate with the audience, which is a good thing.

The problem with spin doctors is not when they spin, it’s when they lie.

Pentamom sent this: Statistics We Refuse to Collect, which is worth your time. I particularly liked this.

From that point on, I have given little credence to statistics that fly in the face of common sense and common observation, or that are vitiated by a flaw in the question.

Right. When someone says “this study (“Science, Science!!”) says such and so,” and such and so is clearly nonsense, I don’t pay attention.

Let me give an example. Common sense tells us that cohabitation is less stable than marriage, because each person knows that he or she can pack up and leave without legal consequence, and without qualms for having broken a sacred vow. It is also more volatile, since by the testimony of many who engage in it, it is a trial run. Youth, instability, volatility, and sexual passion make for quite a canister of nitro-glycerine. We know that a girl is far more likely to be beaten by her live-in boyfriend than is a wife by her husband. But if for ideological reasons you want to obscure this fact, and if you don’t care overmuch for the safety of the girls you are putting at risk, or rather if you do wish the girls well but you hate marriage even more, you will fold the two things together, and invent the category “domestic violence.”

I find it amazing that people fall for this stuff, but it happens all the time.

2017-11-08  »  Greg Krehbiel

Share your thoughts

Re: How to lie with statistics







Tags you can use (optional):
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>