Sunday, December 23, 2007

Lies, Damn Lies and Statistics


Quote of the day:
“There are three kinds of lies: lies, damn lies and statistics.”
--Attributed to Mark Twain and sometimes to Benjamin Disraeli

“Lies, damn lies and statistics” is a comical way to dismiss survey results we don’t like. It’s also the title of a “West Wing” episode (which I haven’t seen), a book (which I haven’t read) and a blog (which hasn’t been updated in 3 years).

Most of us carry around a vague suspicion that this statement is true, though we’re not sure why. In spite of this, we still pay lots of attention to statistics, especially if there is some sort of contest involved. We like to see life as a horse race.

Also, in our drive to be rational, we put a positively irrational faith in numbers. We’d much rather believe the TV weatherperson telling us the temperature than step outside and feel the air.

Some statistics are aggregations of simple data. Yet bias can easily creep in as decisions are made about what data to aggregate. Should we stress changes in population makeup for the city, or the city plus suburbs? The data could be quite different.

Which brings me to one of the built-in problems with statistics. It is age-old. It is making use of selective data to support an argument that is already fully formed.

The “See I told you so!” phenomenon. We have trouble growing up and seeing the world the way it is. We'd rather see the world they way we've already decided it is.

In church work I regularly encountered this in a different way, not involving data but the bible. Often folks would “read back” into the bible. It’s known in academic circles as “proof-texting.”

You’ve seen it and heard it, probably today. It is using the bible (or any respected text, for that matter) to demonstrate the truth of a fully-formed point of view. When you hear a sermon that makes a point backed up by references to a variety of bible verses, this is what is happening. It’s backwards.

If we are trying to discern truth in statistics, the bible or anywhere, it doesn’t serve our purpose to inject truth into the data. The truth is already there.

2 comments:

Anonymous said...

Interesting that you should comment on this today, Craig. Just yesterday I was mentioning to Howard that some people use the bible just like they use statistics. It's a long-running adage that one can make statistics say whatever they want; pretty much same thing with the bible.

Craig Dorval said...

Great minds think alike! This is a vexing problem for me as a minister. On one hand, people outside the church are put off by very biased interpretations. On the other hand, some people inside the church like these interpretations and expect me to agree with them.