Tuesday, March 17, 2009

Guinness, Statistics & Christianity: A St. Patrick's Day Meditation

By Richard Beck.

As a research psychologist I spend a great deal of my life teaching undergraduate and graduate statistics courses. Which I love. I rarely discuss statistics on this blog but there is so much I'd love to share with you about factor analysis, multiple regression, and One-Way Analysis of Variance.

But since it is Saint Patrick's Day I thought I'd share a story about statistics and Guinness beer and how my telling this story in class one day revealed an intriguing link between Guinness and Christianity.

If you don't know anything about Guinness beer know this: A lot of it is being consumed today. Guinness is, perhaps, the most recognizable beer in the world. And as the quintessentially Irish beer it is the drink of choice on St. Patrick's Day.

The connection between Guinness and statistics is this. In 1899 Young Lord Guinness wanted to bring scientific principles into the brewing business. Toward that end he hired a recent Oxford graduate, William Gosset. Gosset had a degree in chemistry and mathematics.

Now you might think that a mathematics degree would be wasted working at a brewery. But while working at Guinness Gosset developed one of the most frequently used statistical tools in use today: The t-test. I use t-tests all the time. My thesis students this year are using t-tests. It's a wonderful statistical tool.

The problem the t-test solved was this. To make beer you need yeast. Yeast, as a living and growing thing, was stored in jars in the Guinness brewery. Given that yeast is dynamically growing one never new how much yeast was in a given jar. Thus, the brewmasters would have to take a sample of yeast from a given jar and examine it under the microscope. Based upon this sample the brewmaster would try to estimate how much yeast was in the full jar. The problem Gosset solved was this: Just how accurate were the samples in estimating the contents of the jar? The t-test is the tool he developed that can help answer that question.

(The application in the social sciences is fairly straightforward. We want to study "the jar", otherwise known as the human species. But we only ever get to study a small sample of the human species, the people who visit our laboratories. Thus, once we study the behavior of this sample of people we are faced with Gosset's Guinness problem: How well does this sample estimate what is going on in the larger population/jar?)

Whenever I get to t-tests in a semester I tell Gosset's story. It helps break up the monotony of a statistics lecture. Well, one day I was telling this story in a class out on Dyess Airforce Base here in Abilene, TX. (I used to teach night classes out there.) After I told the Gosset story one of the airmen who had spent time in England and Ireland told the class this story:

Arthur Guinness was a Christian. And he was appalled and saddened about the toll Irish whiskey was having upon his countrymen. Thus, Mr. Guinness set about making a drink that was so heavy and filling (Guinness is a thick, stout beer) that his countrymen would drink less and more slowly and, thus, reduce drunkenness, intoxication and addiction.

I have no idea if this story is true (it is true that Guinness was a Christian). But I like it.

Guinness beer: One of the lesser known spiritual formation efforts Christianity has offered the world.

Happy St. Patrick's Day.