What are the odds for a married couple to reach their Golden Wedding anniversay?

---

Request for Question Clarification by pinkfreud-ga on 07 Apr 2004 16:23 PDT

"About 52 percent of currently married couples had reached at least
their 15th anniversary in 1996, and 5 percent of them had reached at
least their golden anniversary (50 years)."

Census Bureau News
www.census.gov/Press-Release/www/2002/cb02-19.html

5 percent would be 1 in 20.

If this is the kind of material you're needing, I'll be glad to repost
this information as the answer to your question.

---

Request for Question Clarification by hailstorm-ga on 14 Apr 2004 22:49 PDT

Since lifestyles and lifetimes vary greatly throughout the world, I
would think that this question would only be valid for a particular
country or reason.  For the United States, you could probably get the
rough odds for achieving this feat by doing some calculations with the
average lifespan for males and females, the average age of marriage
for males and females, and the average divorce rate of married
couples.

Pinkfreud,

There is a reason why the data to which you refer does not provide the
information requested: the census data refers to "currently married
couples". It does not take into account, for example, all the couples
in which one or both spouses died before the census, nor does it take
into account the couples who dropped out of the sample because of
divorce, then did not remarry.

The probability of a given married couple reaching their golden
anniversary is the multiplicative probability of each spouse living
another fifty years from the time of the marriage times the inverse of
the probability of divorce for the couple over fifty years.

For couples where both partners are young, the chances of each
surviving fifty years from the time of marriage is fairly high, so
these probabilities would not diminish the overall probability of
reaching their golden anniversary very much.

For couples where at least one partner is older, however, age becomes
a very important factor. If both partners are in their sixties at the
time of marriage, the probability of their reaching their golden
anniversary is approximately nil, barring unexpected improvements in
medical technology. The age of the partners at marriage must be known
before answering this question.

Furthermore, comprehensive studies have revealed that divorce rates
depend upon many factors, such as ethnicity, cohabitation, age at
marriage, economic and social conditions, and a host of other factors.
These influences are significant rather than being mere academic
curiosities. Therefore, you would need to know a great deal more about
the characteristics of the couple before predicting a divorce rate
over such a long period as fifty years.

Another big problem with the census data is that it represents
historical, rather than current cohort data, thereby neglecting to
account for changing trends in divorce rates. Anyone married fifty
years at the time of the census would have endured much of their
marriage during a time when divorce rates were significantly lower
than they are now. This, coupled with the fact that divorce rates
decrease significantly with increasing length of marriage, would
inflate the number of couples married fifty years in the sample, as
compared with the predicted results for a later cohort.

If anything, the census data establishes only an absolute upper limit
on the probability of a given couple reaching their golden
anniversary. The information I saw was 4.8 percent of currently
married couples, as of the most recent census, had celebrated a
fiftieth anniversary. If we take into account mildly increasing life
expectancies versus drastically increasing divorce rates (more than
double versus 50 years ago) a totally empirical calculation would
suggest that a reasonable upper limit on the probability of reaching a
golden anniversary for a couple starting out today would be no better
than one in forty. For a particular couple, the probability might be a
great deal lower than that, simply based upon demographic
characteristics rather than individual temperament.

Here are some links to studies of marriage and divorce in the U.S.A.
done by the CDC.

http://www.cdc.gov/nchs/fastats/marriage.htm
http://www.cdc.gov/nchs/data/series/sr_23/sr23_022.pdf
http://www.cdc.gov/nchs/data/series/sr_23/sr23_015.pdf

Unfortunately, they are almost maddeningly ageist, in that they focus
almost exclusively on individuals under the age of 44, and marriages
of less than twenty-five years duration. There may be better material
out there for marriage in later life, but I cannot find it. In any
case, they do highlight the enormous variability in divorce rates with
demographic variables.

Best regards,

Alan Kali

Pinkfreud,
You have misunderstood the statistics involved in this situation.  I
assume you will not try to post your incorrect information as the
official answer.  Let's presume, just as an example, that we lived in
a wonderful world where 100% of people get married and reach their
50th anniversary someday.  However, out of the people who are
currently married, let's say 20% had reached their 50th already.  This
does not mean that the odds of reaching the 50th is 20%!

Alan,

All your points are valid, and there are many other factors that could
be considered also.

However, this is a two dollar question, suggesting that the customer
is only looking for a general answer, rather than an exact one based
upon every possible consideration.

As nobody knows what the customer wants except the customer, my
colleague pinkfreud-ga has requested clarification. This example is
precisely why clarification is built into the Google Answers system
and pinkfreud-ga is using it properly.

poe-ga
Google Answers Researcher

To Researcher Poe:  Well-stated!    

To alkali-ga: you may want to reconsider posting your true name in
light of the privacy issues that might occur.

Best regards,
journalist-ga

Ouch! How can I pour oil on these troubled waters?

I am terribly sorry, I did not mean to denigrate your answer,
pinkfreud, only to point out a fallacial error in the method of
answering the question. I would have refrained from commenting at all,
and in fact I read your comment and moved on, but a nagging in the
back of my mind caused me to return.

True, joeaud does not specify whether he means a couple who started
out fifty years ago versus one starting now, but either way the
methodology is incorrect, as research_help points out. Joeaud clearly
asks for the odds, a statistical measure of probability, and you have
given him prevalence restated as if it were a probability. This is
misleading, since the number has been arrived at by fallacious means.
Unless joeaud has a background in statistics, he is unlikely to notice
the error, agree that the answer "one in twenty" is in the form he was
looking for, and agree to the answer. My initial reaction was to wait
for joeaud's clarification, but I realized that the request for
clarification itself was misleading.

Regardless of the price of the question, it is probably unfair to
present a quick but incorrect answer as if it were correct. Perhaps it
would be best to have another researcher with a background in
statistics (surely there must be one?) referee this.

As regards poe's comment, the answer given is not, I'm afraid, a
"general answer" with a request for clarification, nor is it inexact.
It is the result of a distortion of the question as asked. It is not
my intent to nit-pick about minor factors that might affect the
correct answer to the question, but to point out the major issue of
the approach to the question. I only mention the other factors as a
matter of completeness.

I am not at all convinced that pinkfreud is using the request for
clarification appropriately. Rather than requesting clarification of
the original question, pinkfreud presents a response and asks if it is
an acceptable answer. A better approach would be, "Is this the kind of
methodology you're needing?", Even then, I'm not sure the questioner
can be expected to critique the methodology. Ensuring that the correct
methodology is used in answering the question is the researcher's,
rather than the questioner's prerogative.

I am, futhermore, not certain that the price of the question reflects
accurately the questioner's expectation of the quality of the answer.
Many questions here on Google set low prices for questions that
require enormous amounts of work. If the price is too low for the
amount of work required for a correct answer, perhaps it is best to
leave the question unanswered, rather than to purvey a response as
"the most correct, given the amount of money involved".

As regards the egos of unpaid researchers, I assure you that I do not
post unpaid answers for the sake of my ego (whatever that is). The
main reason I post them is because I feel that I can contribute. An
important but secondary reason that I post is because I wish to know
if I am correct. With all of the eyes around here scrutinizing
responses, if I am out to lunch I shall learn it in an instant. I
should have thought that the paid researchers would feel the same.

And "Alan Kali" is not my real name anyway. I'm actually a dog.

Best Regards,

Alan "Woof" Kali