My question is why?

I posted this in another thread and never received an adequate reply. Trisha replied this way:

quote:

---

The comment about marriages with girls being more likely to end in divorce made me think of the boys being the one to carry on the family name, which is the man's name. Most papers that have approached the subject recognize that that is pretty much the number one reason most men want sons. The reason for the importance of procreation is to continue the family line. It's that in the animal world and in the human world.

---

Thanks Trisha for the reply, but you haven't "solved" the conundrum.

Men may prefer boys to girls but the issue here is why women and men divorce more often when their child is a girl. Do men change wives in the belief a new one will give them a son? (Or is the cause of divorce rather women? Do they prefer to have a man around if there's a son?)

I don't know. Can other babblers provide a credible explanation for this phenomenon?

[ 12 December 2004: Message edited by: Scott Piatkowski ]

[ 13 December 2004: Message edited by: August1991 ]

This paper has some interesting theories:

quote:

---

...we argue that sons increase the fathers' involvement in parenting and, thus, are stability producing. Recent gender change has weakened the connection between a child's sex and the father's involvement with children providing an explanation for the attenuated effect in more recent periods.

---

quote:

---

Originally posted by Scott Piatkowski:
Somebody needs to learn the difference between a postive correlation and a causal relationship

---

Somebody needs to learn what random means.

If a random event (sex at birth) is correlated with a non-random event (divorce), the cause and effect is defined.

IME, the literary Left (like Tolstoy and Einstein) has a fundamental misunderstanding of probability. Some things can be explained, other things can at best be described.

----

Should I change the thread title to "Why do families with boys divorce less?" I will if it matters.

This issue confuses me.

I know that many parents with girls don't separate and many parents with boys do. But on average, there's a statistically significant bias toward divorce if the child is a girl.

I'm genuinely curious to know why.

[ 12 December 2004: Message edited by: August1991 ]

As they said on page 2:

The Effect of Sex Composition on Divorce and Custody

We begin our empirical analysis by investigating the relationship between the sex composition of children and the probability of divorce. We uncover a surprisingly consistent pattern for the effect of children’s gender on divorce. In particular, we document that parents who have only girls are significantly more likely to be divorced than parents who have only boys.

quote:

---

Originally posted by Sara Mayo:
But there is no evidence that the direction of the cause and effect is that girls cause divorce. As I have cited above, it seems that it is father's bad parenting skills with girls that cause divorce, not "girls being girls" causing the divorce. Again, August I ask that you change the thread title. I find it deeply offensive, especially as there are no researchers using those words.

---

I think I answered your points above. (I'm relatively new to editing on this forum - see below.)

"Father's bad parenting skills with girls"? Huh?

Change the thread title? To what?

[ 12 December 2004: Message edited by: August1991 ]

quote:

---

"Father's bad parenting skills with girls"? Huh?

---

What's your question? (Sorry about the apostrophy typo in fathers, btw). I am not the best writer, but that's the basic argument of the paper I cited. I am confused by your confusion...

quote:

---

Originally posted by Scott Piatkowski:
Somebody needs to learn the difference between a postive correlation and a causal relationship

To which August1991 replied:
Somebody needs to learn what random means. If a random event (sex at birth) is correlated with a non-random event (divorce), the cause and effect is defined.

---

You mean a random event like the weather on the day people were married? Shall we do a study to determine whether rain on your wedding day (besides being not at all ironic) is "the cause" of divorce? Or, maybe it's good weather that "causes" divorce. Or bad tuxedos. Or playing Phil Collins as your first dance.

It's like Republican voters who think that the fact that America was not attacked again after September 11, 2001 was a result of George W. Bush being President.

[ 12 December 2004: Message edited by: Scott Piatkowski ]

quote:

---

Originally posted by Sara Mayo:

What's your question? (Sorry about the apostrophy typo in fathers, btw). I am not the best writer, but that's the basic argument of the paper I cited. I am confused by your confusion...

(and see my post above for a suggested title change)

---

First, what is better? "Fathers prefer boys"?

Second, how do I change the title? (Do you know?)

----

But I still haven't seen a good explanation for why, on average, guys stick around/girls want a guy if the kid's a boy.

quote:

---

Originally posted by August1991:
Second, how do I change the title? (Do you know?)

---

Since you made the first post in the thread, you can change the title by editing that post. Or you can ask a moderator to do it for you. I've taken care of it.

I'm not picky about what you change it to. I have given you a suggestion already, but it's your choice.

quote:

---

But I still haven't seen a good explanation for why, on average, guys stick around/girls want a guy if the kid's a boy.

Is this genetic or merely a passing phase?

---

Genetic? There would have to be a lot more evidence (from studies in lots of other cultures to begin with) to even think there's a genetic link. The two studies we've discussed in this thread so far both indicate that the link of child sex with divorce rate is diminishing with time, so to me that would eliminate genetics as a factor.

As to the random issue, imagine we ask mothers to flip a coin at the birth of their child. The brown-eyed mothers will be as likely either way. After all, it's only a coin flip.

Now, do the same according to the sex of a child. Still the same, it's a coin flip. (How many Heads will have girls, and how many Tails will have boys?)

----

If there's a statistical difference, what's the cause? Imagine there are several thousand marriages. Is it random?

If you take a placebo but the other takes the new drug, is the difference chance alone? How many Heads will have girls, and how many Tails will have boys?

Thanks Sara for clear instructions on title change.

[ 13 December 2004: Message edited by: August1991 ]

quote:

---

Originally posted by Sara Mayo:
The answer to me is pretty obvious: sexism. In past decades, there was a much more codified role for father-daughter relationships, one that excluded true emotional sharing. Thankfully now fathers are much more involved in their children's lives, and in particular have closer relationships with their daughters and the girl-divorce link is diminishing. Chalk up another victory for feminism!

---

Sexist preferences => Preference for having sons
Preference for having sons met => Increased male parental involvement
Increased male parental involvemeny => Increased marital stability
Increased marital stability => Decreased Divorce

This hypothesis would explain the observed correlation, the middle term being male parental involvement.

Well, beavers tend to be monogamous but other mammals not. Penguins tend to be monogamous like other birds. Why?

quote:

---

The answer to me is pretty obvious: sexism. In past decades, there was a much more codified role for father-daughter relationships, one that excluded true emotional sharing. Thankfully now fathers are much more involved in their children's lives, and in particular have closer relationships with their daughters and the girl-divorce link is diminishing. Chalk up another victory for feminism!

---

It's all simply sexism? Huh?

Why?

Well, that is the question of this thread.

[ 13 December 2004: Message edited by: August1991 ]

quote:

---

Originally posted by Mandos:
But they gave you an explanation that is an obvious one generated by the fact that the correlation no longer apparently exists. Why then do you persist in seeking alternate explanations?

---

quote:

---

Originally posted by Chris Borst:
August, you seem to be ignoring the perfectly good hypothesis to which Sara Mayo has referred you:

Sexist preferences => Preference for having sons
Preference for having sons met => Increased male parental involvement
Increased male parental involvemeny => Increased marital stability
Increased marital stability => Decreased Divorce

This hypothesis would explain the observed correlation, the middle term being male parental involvement.

As sexist preferences can be hypothesized to be declining, and male parental involvement increasing, for exogenous reasons, the observed correlation should tend to zero, as your initial document (as remind notes) indicates it has. Note, this means that there is now no longer any correlation between child sex and rate of divorce. The question is of historical interest only.

---

quote:

---

Thankfully now fathers are much more involved in their children's lives, and in particular have closer relationships with their daughters and the girl-divorce link is diminishing. Chalk up another victory for feminism!

---

quote:

---

Originally posted by Jacob Two-Two:
Maybe you should give us your working definition of sexism, August, 'cause i'm having a hard time understanding your confusion.

---

That's all.

quote:

---

Sara, do you have evidence of that diminishing girl-divorce link?

---

I don't think you even have to read the paper, really, but you can if you want. The results are stated above.

[ 13 December 2004: Message edited by: Mandos ]

Divorce is a random affair. Boys have nothing to do with it.

The stats will change once everyone is no longer sexist.

I think perhaps that you are asking the deeper question of why men historically preferred boys? I think there are many reasons for that that get labelled under a single word sexism but are quite intriguing and complicated. But that wasn't what you asked at the beginning of this thread, so the premise was flawed.

[ 13 December 2004: Message edited by: Mandos ]

However, on to a randomly selected excerpt:

quote:

---

Since the predictions hinge on the convexity or concavity of θ , it is important to
understand the two conditions described in equation (12). Both conditions indicate the expected future benefit of a change in family composition c, depends on the amount of curvature in the utility function.

In the current setting, the terms M c λ and D c λ are simply two different inverse Mill’s ratios or hazard functions (i.e., ratios of density functions to survivor functions).

This quantity multiplied by the marginal utility of an effective child in the married and divorced states must be less than what is often referred to as the coefficient of absolute risk aversion, an expression which describes the curvature of the utility function.

---

"This quantity multiplied by the marginal utility of an effective child in the married and divorced states must be less than what is often referred to as the coefficient of absolute risk aversion..."

August, you've asked "why" on this thread half a dozen times and we have given you our theory which fits nicely with the data. If you don't think sexism is involved, what is your alternative hypothesis?

[ 13 December 2004: Message edited by: Sara Mayo ]

I only asked because central to most definitions of sexism is the belief that men are superior to women, which would clearly cause a sexist man to consider a male child to be worth sticking around for moreso than a female child. Since this straight line of logic seemed so hard for you to swallow, I figured maybe your definition of sexism was very different from my own.

quote:

---

Originally posted by audra trower williams:
Hi. I actually don't edit thread titles unless they have glaring typos or give NO idea what the thread is about, ie: "I CANNOT BELIEVE IT" or whatever.

---

#1. Effects are generally weak and decline over time. They are talking about what was rather than what is.

#2. The paper is empirically driven rather than theoretically driven and thus is more susceptible to the identification of spurious effects.

#3. Much of the data is internationally derived and there are differences in the ways that countries collect such data.

#4. I am always careful when looking at logistic regression coefficients because they are not really intuitively interpretable. They always seem inflated.