“50% of marriages end in divorce…”

You can chalk this up to the list of things that I complain about once a year. So feel free to skip this if you’ve heard it before.

Here’s the deal: the divorce rate, as conventionally calculated, is a near-useless statistic. Perhaps totally useless. The divorce rate compares the total number of divorces in any given year to the total number of marriages. This has an almost entirely distorting effect on the number: the pool of potential divorces to be factored into the rate is all existing marriages. The pool of potential marriages is just new marriages. The number of divorces in a given year can remain stable, and the number of new marriages can drop (due to economic downturn, for instance) and the divorce rate can jump, despite the fact that the number of existing marriages ending in divorce hasn’t changed. It’s a ludicrous way to calculate the statistic.

An honest divorce rate would involve looking at individual marriages and tracking them to see how many unique marriages end in divorce. This kind of statistical work is labor-intensive, but it appears from the best available data that the honest divorce rate likely never exceeded 40% or so, and that it peaked in the early 90s and has been heading slightly downward since. And, as I have argued before, a divorced/stayed married binary says very little about the aspects of marriage we find important for society.

Part of the reason I bring this up in frustration so often is because this isn’t some deeply contrarian idea or a piece of secret, hard-to-tease-out knowledge. Rather, it’s a fact that’s been long acknowledged by statisticians and sociologists; it’s been talk about it the occasional magazine and newspaper article; and yet it can’t ever seem to penetrate the popular consciousness. Once something has penetrated the conventional wisdom, disputing it is wading upstream.

Incidentally, if you aren’t aware the Statistical Abstract is online, and a very powerful tool for research.