Divorce rate

[B.J.]

I've seen it frequently said that in the US the chances of a marriage ending in divorce is 50%. Personally, I think this is a little bit high, and really doesn't get to the root of the question: What are the chances an individual will have to go through a divorce? I've looked at several different sites and various statistics, but none of them really address that since they mostly deal with the overall numbers.

For example, say you ask 4 married people if they've ever had a divorce, and 3 say no, but the fourth person has gone through 4 previous divorces. According to the way the data is usually presented, this would show a 50% divorce rate. But I think it should show that there's a 75% chance that an individual will have a successful marriage.

So the poll is just for our married or formerly married people, just to get an idea of how far off I am. BTW, I worded it as "Never divorced" rather than "Married once, not divorced" so that it also covers those who are widowed and may or may not have remarried later.

(Honestly, I'm not sure why I'm fascinated by this, since I'm very happily married, and have been for a long time.)

[farmkid]

What you're looking for is the divorce rate for first marriages vs. the divorce rate for later marriages. For first marriages, the divorce rate is something like 30-40%, whereas for 2nd marriages is is 50 something or 60 something % and for 3rd marriages, it is over 70%. Overall, almost half of marriages end in divorce, but the majority of people who get married never get divorced while those who do get divorced once are likely to get divorced again if they remarry.

As for me, 15 years into my happy first marriage with no expectation to ever divorce.

[SPOCKED]

The timing of this thread is interesting. I was divorced in '96 and as of this very day, I believe I am in a position to finally stop support payments (which is an amount for all 3 kids). All my kids are over 18 and my youngest has moved out. My 19 year old son, who is still living at his mom's, is enlisting in the Navy, and my oldest daughter is long out of her mom's house. I'm expecting a callback from my lawyer this week to discuss ending support. My plan is to pay for my youngest daughter's rent until she works full time and is out of 12th grade.

Anyway, divorce sucks - I hope to know death's sweet embrace before I ever have to go through that shit again!

[auntiehill]

Married over 7 years (8 years in March); hubby and I fit into a great demographic---it's the first marriage for both of us, and we married when we were both over 30. "They" say (the statistics) that we have the best odds.

[Kirby]

Married 17 years, never divorced.
Of my closest friends from High School, 50% of them are divorced (3 of 6). That number drops to 1 of 5 College friends that I still keep in contact with, and none of my post-college friends are divorced, although it looks like one of them may be headed that way.

[Holdfast]

B.J. wrote:

---

I've seen it frequently said that in the US the chances of a marriage ending in divorce is 50%. Personally, I think this is a little bit high, and really doesn't get to the root of the question: What are the chances an individual will have to go through a divorce?

This is actually a very interesting statistical question, not so much for the question itself, but because the thought processes required to generate an answer impact on many other (to me, more interesting) datasets, and how we interpret and apply statistics to life. In other words, how does a population-based statistic impact on an individual?

What you're essentially saying is that you would like to create a truly accurate Bayesian probability for a specific individual's chance of divorcing i.e. you want to know the probability of a given individual divorcing. However, in order to construct the algorithm required to derive that probability you need to know all the potential variables that contribute to the chances of a divorce occuring, and crucially, how much impact each variable has. Unfortunately, we don't have that data. So in an attempt to generate that data, we look for population correlations, and attempt to define the strength of the linkage. Even more unfortunately, in the absence of being able to know the causal relationship, it's entirely possible that such a method "double-counts" risks. e.g. is an alcoholic's pattern of secretive behaviour the contributing factor to a divorce, or is it the fact that they're drunk and so, unattentive to their partner's needs. Population analysis would suggest both; a sensitivity analysis would generate a different stimulus response curve to each, but at the end of the day, both may or may not impact. You can't conduct a prospective RCT to try to resolve these issues, so you have to rely cohort studies which causes uncertainty like this to arise.

So in constructing a Bayesian probability, we'd inevitably/implicitly draw on a more fiducial model to interpret the data. This is dangerous, and generates a difficult-to-quantify "fudge factor" in whatever percentage risk you generate from your analysis.

Bottom line? Multifactorial risks are extraordinarily difficult to model reliably statistically, and the narrower you try to refine the subset, the harder it is to generate a meaningful risk factor. Statistics falls apart into shades of grey; an individual assessment that relies more on assessor experience and heuristically-based judgement becomes as accurate (if not more accurate) than the professional statistician. Weirdly enough, because of this, the overall headline population risk we started from may end up being a more reliable predictor of an individual's risk within that population than anything else!

PS. I'm not a statistician, nor do I play one on TV. I just have some very rudimentary (and no doubt imperfect) familiarity with these issues from having to try to critically interpret studies on things like this in the medical literature. I'd love to hear what a professional statistician would say about it though; how would you guys try to resolve the uncertainty?

[DrCorby]

auntiehill wrote:

---

Married over 7 years (8 years in March); hubby and I fit into a great demographic---it's the first marriage for both of us, and we married when we were both over 30. "They" say (the statistics) that we have the best odds.

My wife and I fall into the same demographic (1st marriage for both, both over 30 when we married). And even more rare -- both sets of parents were also married to their first spouses until parted by death, late in life. We celebrated our 19th wedding anniversary a few months ago, and still happy about the decision all these years later.

[1001001]

Married 19 years, first marriage for both of us.

[Scout101]

farmkid wrote:

---

Overall, almost half of marriages end in divorce

Hey, look on the bright side: the other 50% of marriages end in DEATH!

[Kelso]

Divorced once.

Thank the gods.

[TorontoTrekker]

I voted for "Divorced at least once," though I might be skewing the statistics a little - even though I was married in the United Church of Canada, it wasn't legally recognized at the time. So when we split up, we had to notify the church and they had to issue some kind of proclamation (I'm not clear on exactly what they did to dissolve the marriage), but there was no legal involvement. I still consider myself divorced, though, just as I considered myself to be married from 1993-97.

[Danoz]

farmkid wrote:

---

What you're looking for is the divorce rate for first marriages vs. the divorce rate for later marriages. For first marriages, the divorce rate is something like 30-40%, whereas for 2nd marriages is is 50 something or 60 something % and for 3rd marriages, it is over 70%. Overall, almost half of marriages end in divorce, but the majority of people who get married never get divorced while those who do get divorced once are likely to get divorced again if they remarry.

As for me, 15 years into my happy first marriage with no expectation to ever divorce.

Interesting, I wasn't aware of these statistics. Do you have a source/study for this? I'm actually really interested by this because I had no idea the rates were that different along first/second/third marriages.

[ShamelessMcBundy]

Well I dunno if this counts or not, but my parents have been married for almost 40 years now. In my family, divorce is actually really rare.

[B.J.]

^Same here. Actually, I can't think of anyone that I'm related to that has been divorced.

[auntiehill]

My parents have been married for 60 years.

Hubby's parents divorced after about 17 years.

My oldest sister divorced after two years.
My second oldest sister divorced after 15 years.
Of hubby's 7 closest friends, 4 have gotten divorced and three never married.