Comments on The Day in Correlation/Causation Fallacies, Part 2

UGH!

Your first point is irrelevant. It amounts to: Is it being single that causes these problems or is being single that causes these problems?

You simply gave a hypothesis for WHY being single causes the problems.

Your second point makes me want to scream. I can't believe how many people do this. Yes, Erika, there is a danger in confusing cause and effect, cause with correlation, and failing to attribute cofounding factors.

But guess what, Erika?

This "research" that was conducted by a "professional" at a "university" with a "degree" that was then "peer-reviewed" and then "published" in the "Journal of Health and Social Behavior" just maybe took your points into consideration.

I am sure there aren't a whole field of scientists who are now slapping their foreheads and going, "shit, how in the hell didn't we catch this? What were we thinking?!...we only went to college, and then grad school, and slowly built a career that established us as leaders in our field..you would have thought we would have taken such things into account. THANK GOD FOR ERIKA!!!"

This is my biggest pet peeve: People who don't liek the results of a well-designed, peer reviewed, and published study and so smuggly explain it away by making general challenges of the study that were most certainly taken into account.

Uh duh, don't hey know that it is all complicated and stuff? Gosh, dem dumb scintists didt even think that maybe mental illness a causes that there divorce not the other way 'round! Heck, I reckin that they had't evun thunk about da issha of cause-e-a-shun vs cory-a-shun!

Forgive me, but that is truly how ignorant you sound.

Damn, Johnny! Having PMS today?

By invoking the causation/correlation fallacy in this instance you would essentially be arguing that there is NO relationship between mental health and divorce/remarriage and that any statistical correlation is the result of random coincidence, or that the independent and dependent variables are reversed and the higher levels of psych problems resulted in the divorces rather than the other way around.

This is not what you are arguing. In fact, I don't see how your arguments conflict with the conclusions of the study as you have posted them.

It is not really appropriate to raise the correlation/causation fallacy here. If you take a biostatistics class you quickly learn that short of a double blind clinical trial (and even that has its problems)- which would be unethical and unfeasable in this case - causation cannot be proven by a single study.

The US Surgeon General came out with guidelines for determining causation:

Temporal relationship (does illness occur after divorce?)
Strength of association
Dose-response relationship (if you get two divorces are you more sick than if you get one)
Replication of findings (more than one study)
Biologic plausibility (based on what we know about the human body does it make sense)
Consideration of alternate explanations (They do this, as do nearly every peer-reviewed article - just look at the actual study)
Cessation of exposure (if they get married again does health improve?)
Consistancy with other knowledge
Specificity of association

Not that all of these is going to occur, but it shows that a preponderance of evidence is needed. That is how they proved smoking caused cancer. Just because you publish a study does not mean you are claiming causation. And if evidence comes out that you are wrong, well, time to change the theory (e.g. if it turned out that a gene caused people to smoke and caused them to get lung cancer independently).

But as CH points out, the only critique you give is actually supported in the study - one of their hypotheses is that stressful life events after divorce harm health - moving, switching jobs, kids moving schools, etc. They are not saying that the actual divorce itself makes you sick.

Finally, I think Johnny went overboard here - it is totally appropriate to criticize research even if you don't have a PhD. Scientists can be total idiots, biased, or close-minded. But you should look at the actual study, not the the artile about it, because the media don't care about correlation - they love reporting causation.

But, be careful what your motivations are - people pull out the correlation/causation argument when they don't like findings but less often when they support them. Example, it only takes me one artivle about a study that shows that coffee/red wine improve health, and I will be thrilled, as I like coffee and red wine.

One more thing, if I may rant on, is that a study showing that divorced women have worse health does not necessarily mean that we should decrease divorce. It could mean that we need more support services for divorced women. Or we could look upstream and ask, what do all divorced women have in common? Well, they got married. (OK, actually, lifelong single people tend to be less happy and less healthy than married people, but that has more to do with companionship, I think, than family values, although I can't prove it.)

The research was conducted at Iowa State University, not the University of Iowa. Big diff for those of us from the Hawkeye State.

I hate to comment on studies that there's no way in hell I'll be reading in its entirety...BUT being a slow work day and all...

I wonder how the age of the divorced in each of the couples corresponds to one another? Wide age group? I think it's safe to assume many or most first marriages end within a similar age range, say, at 20-35 years old, which is also a similar range for many of the most debilitating mental illnesses to first appear. Also, 10 years later, is about the time when many women start showing signs more common, serious physical illness, whereas men tend to develop them closer to their 50's.

Also, it’s no stretch to assume people that suffer from mental illness are more likely to be single.

Sorry, I’m giving my opinions based on someone’s serious paraphrasing, but like I said, slow work day.

I'm not even going to comment on the accuracy of studies of any sort. But let's assume for a moment that this study is right. Women who get divorced have mental and physical problems. Well, guess what? Divorce is stressful. In many cases it can be overwhelming, whether the woman intiated it or not. So if you take two women, one who is still married (let's assume happily) and one who is divorced, even if the divorced woman is happily divorced, chances are she had had a lot of divorce stress to get there. So overall, she might have a lot more stress in her life to deal with than the single woman.

ECB nailed this. Dead on.

All this study shows is the obvious fact that men are more likely to divorce women suffering from mental or physical health problems.

Or maybe that men drive women nuts and that's why women seek divorces.

There are two causation issues. The first is actual causation and the second is directionality. i.e. are divorce and mental health causally related, and if so what causes what. Ideally, and I really mean ideally, you isolate or control for every variable aside form the ones you are measuring, then take repeated measurements as the variables change. This is really fucking hard and is why social science unlike physical science does not often posit laws.

As soon as you think you are close to having done this, some punk gunning for a PhD comes along and says a but variable X correlates with both of them. hahahaha

Since we're digressing, I'm gonna go way out on a limb and in scratching the surface, you may not buy into this off the bat, but... many women see marriage as a life goal, the way one would, say, see graduation. It's an act of personal validation. So someone who objectifies their own life like that is bound to run into emotional and personal issues with a man who likely does not see things that way, or someone who is that self-absorbed. And HEY, divorces.

I admit this is a product of history. It's not like the past where a woman had nothing to look forward to but getting hitched and being a housewife for some sap. Women live in a modern world with choices, yet are trained in this mindset of, "You must get married in order to validate yourself." So there's this inherent conflict, and then we wonder why so many divorces and so many marital problems.