Living Together

[Globalnomad]

Well, it doesn't prove causation, but that doesn't mean it's meaningless. It does rule out the possibility that cohabitation decreases the divorce rate in any meaningful degree. Since there's still a lot of people who seem to believe that it does, that's good to know.

Actually Bill, to be pedantic, it does not rule it out. What we would need is a sample of people who have the same propensity to divorce and then compare the divorce rate of those among them who lived together before marriage and those who did not (while controlling for other likely factors). But how are you going to assess "propensity to divorce"?

As it is, this statistic could mask the possibility that this group of people would have even higher divorce rates if some of them had not lived together before marriage.

I am not saying it is so (honestly, I have no opinion either way) - I am just saying that statistically, nothing is proved.

 

[BillH]

Actually Bill, to be pedantic, it does not rule it out. What we would need is a sample of people who have the same propensity to divorce and then compare the divorce rate of those among them who lived together before marriage and those who did not (while controlling for other likely factors). But how are you going to assess "propensity to divorce"?

As it is, this statistic could mask the possibility that this group of people would have even higher divorce rates if some of them had not lived together before marriage.

I am not saying it is so (honestly, I have no opinion either way) - I am just saying that statistically, nothing is proved.

Of course, it would be great if we had all of the data so we could try to control for other factors likely to affect propensity to divorce. And then professional econometricians might say, "Oooh! Look, we can say that if you do a probit regression on cohabitation on 8 other factors cohabitation actually has a negative coefficient with a t-statistic of 3.15! Therefore, cohabitation causes divorce to go down! [then in hushed tones] The fact that the estimated marginal probability is only -0.0001% isn't nearly so important as the fact that [in a loud voice again] the coefficient is negative!"

I stand by my assertion that it's not meaningless. It shows that if cohabitation causes divorce rates to go down, it does so in a small enough degree that the effect is largely swamped by other relevant factors.