A better (but difficult method) of tracking divorce rates is by following a cohort over time and tracking who gets divorced. There is a statistical technique called hazard modeling (which I have never used, since i do qualitative research), where you could get actual likelihood ratios.

Rachel S. Writes:
October 20th, 2006 at 7:22 am

Neil K said, “The graph shows that the marriage rate is converging onto the divorce rate. That does look alarming. That means more marriages end in divorce each year. By my eyes your graph shows it was 5/10 (50%) somewhere in the 80s, and we’re approching 4/7 (57%) today.”

Those stats cannot be determined from that graph. All these rates (per 1000) track are the number of marriages and divorces in a given year, but we cannot assume that those are the same people getting married or divorced. Moreover since these are rates you should noted that the y-axis of the graph only goes to 25 per 1000.

It’s true we have no way of knowing from that graph which couples are getting divorced. But the graph does, nonetheless, suggest a conclusion that a higher percentage of all marriages are ending in divorce. You can see that fairly easily by the “divorce per thousand marriages” section. It also, incidentally, means that there’s a larger absolute number of divorces yearly, as 1) the divorce rate is going up, and 2) population is climbing.

(I don’t mind the axis issue so much here, because it’s a relative comparison between two things on the same graph and is not as misleading.)

The truth of the matter is that the “50% of all marriages will end in divorce” figure that is commonly cited is most likely inaccurate.

This statement is sort of meaningless without more detail. Inaccurate by how much? What other number would you suggest? Without an alternate from which to choose, I’m not sure what your “most likely inaccurate” comment is supposed to say. I agree that the chances of it being exactly 50% are exceedingly small.

We don’t really know how many marriages today will end in divorce tomorrow.

No, of course not. We cannot predict the future with perfection.

We can make aneducated guess, though, based on the existing statistics. We can make an even “better” educated guess if we are open about our assumptions: “if current trends continue, we would expect _____” Are you suggesting we should never make future predictions based on current trends? If not, what do you mean by this sentence?

In the random sample surveys I have seen, the number of divorced people is somewhere between 10-30%, but random samples include people who have never been married in the first place.

OK. It sounds like those studies are inapposite so I won’t respond to them.

My personal guess based on my own background as a family researcher is that the figure of marriages that end in divorce is probably considerably lower than 50%.

Because of your studies, I KNOW you’re not expecting reliance on your anecdotes in this respect. If you’re a family researcher, can you provide studies/statistics that support your statement? And can you suggest what “considerably lower” means in your statement? 47%? 45%? 20%?

Neil K said, “The fact that the divorce rate is falling in absolute numbers per 1000 people looks like a function of the falling marriage rate per 1000 people.”

Rachel said: No, not really because these are rates.

Actually, I am pretty sure Neil is right here. Over the long term, the divorce rate per unit of population is limited (for obvious reasons) by the marriage rate per unit of population. You have to get married before you can get divorced.

Yes, there can be smaller blips where divorce can exceed marriage because people can divorce who were married earlier–but in a population dynamics sense, they are linked. Nick used the word “absolute” which sort of confused the issue, but he used it in the context of a rate.

There is only one measure of divorce shown here which is independent of the rate of marriage per unit of population. That measure is the “divorce rate per thousand marriages which is reflected on the chart.

Reducing the marriage rate per unit of population will, all other factors being equal, reduce the divorce rate per unit of population. It will not necessarily reduce the divorce rate per 1000 marriages.

It certainly stand to reason that fewer people are getting divorced because the marriage rate is down, but there are probably more marriages today in absolute numbers than there were in 1860 because the population is about 3Xs more people.

I completely agree. The absolute numbers (not a rate) of both marriage and divorce have gone way up because of population increases.

Those stats cannot be determined from that graph. All these rates (per 1000) track are the number of marriages and divorces in a given year, but we cannot assume that those are the same people getting married or divorced. Moreover since these are rates you should noted that the y-axis of the graph only goes to 25 per 1000.

The truth of the matter is that 50% of all marriages will end in divorce figure that is commonly cited is most likely inaccurate. We don’t really know how many marriages today will end in divorce tomorrow. In the random sample surveys I have seen, the number of divorced people is somewhere between 10-30%, but random samples include people who have never been married in the first place. My personal guess based on my own background as a family researcher is that the figure of marriages that end in divorce is probably considerably lower than 50%.

Neil K said, “The fact that the divorce rate is falling in absolute numbers per 1000 people looks like a function of the falling marriage rate per 1000 people.”

No, not really because these are rates. They are not absolute numbers. It certainly stand to reason that fewer people are getting divorced because the marriage rate is down, but there are probably more marriages today in absolute numbers than there were in 1860 because the population is about 3Xs more people.

The graph shows that the marriage rate is converging onto the divorce rate. That does look alarming. That means more marriages end in divorce each year. By my eyes your graph shows it was 5/10 (50%) somewhere in the 80s, and we’re approching 4/7 (57%) today.

The fact that the divorce rate is falling in absolute numbers per 1000 people looks like a function of the falling marriage rate per 1000 people. For obvious reasons those two trends are linked. ;)

what caused the explosion in the divorce rate between 1960 and 1980?

Hairy-legged man-hating feminists, of course! ;)

I could forgive them that. I could even forgive them Janet Jackson’s nipple ring. But if it turns out that they’re responsible for hip-hop music…

Daran Writes:
October 18th, 2006 at 3:04 pm
what caused the explosion in the divorce rate between 1960 and 1980?

Hairy-legged man-hating feminists, of course! ;)

Amp: That first satirical paragraph was so perfectly written it almost (really!) made me spit coffee on my monitor I was laughing so hard. Genius.