What is the Chance of Marriage Ending in Divorce in the US?

In Chapter 10 of my book, I note that "it has been estimated that about half of the marriages contracted since the 1970s will end in divorce," with a followup comment that the rate seems to have leveled off and so we are likely past the period of rising divorce rates. Today my attention was drawn to a website called "truthorfiction"claiming that the 50 percent figure is too high.

Let me say it straightforwardly: Fifty percent of American marriages are not ending in divorce. It's fiction. A myth. A tragically discouraging urban legend.

If there's no credible evidence that half of American marriages will end up in divorce court, where did that belief originate?

The author of this website goes on to note, correctly, that the US no longer routinely collects divorce data (a sign of its general irrelevance to society, I suppose), so we must rely on surveys and other data sources. The article provides a variety of other estimates, but actually does not come up with a solid number. One of the issues, of course, is that even if you were looking at vital statistics, you cannot compare divorces in this year with marriages, because most divorces occur after several years of marriage. So, we need a longitudinal type of analysis. We need something comparable to what the life table provides as a way of standardizing the average length of life. In this way we can capture both the timing and the tempo of divorce over the life course of marriages. The person who has done more of this type of analysis than anyone else is Robert Schoen, now Professor Emeritus of Sociology at Penn State. In 2006, he and co-author Vladimir Canudas-Romo of UC, Berkeley published a paper in the Journal of Marriage and Family (Volume 68, August, pp 749-758) in which they applied a life table approach to data for those years for which information was available, and then a set of simulations for years in which data are not available. There overall conclusion was that:

Adjusted values for recent years do not suggest a decline in the likelihood of divorce, with year 2000 values indicating a divorce probability of 0.43 – 0.46.

To be fair, this is a little less than 50 percent, but still pretty high and pretty close to "about half." Of course, in the absence of good data since 2000, you might well argue that the divorce rate has declined. But, in the absence of data, you could just as easily argue that it has gone up. In point of fact, most data suggest that there has been no change.