Red wine on a white bed

So reddit op was a bit misleading about what actually happened. This wasn't like a single page questionnaire that was just "1. Was your wife sick? 2. Did you divorce her?" and the problem didn't just affect the men.

This was a big huge public dataset which followed older married couples with a big huge survey on a billion things every 2 years for 18 years, then made the data available for people to analyze how they want.

These authors looked at how new health problems relate to later marriage outcomes. The outcomes being "still married," "divorced," "widowed," and "attrition" (=dropped out of the study before year 18)

PROBLEM IS they f*cked up the line of code that assigned couples to outcomes, and the "dropped out of the study" group got merged with the "divorced" group.

Friends. Do you have any idea how many people don't make it back for every single timepoint of an 18 YEAR LONG STUDY?

A LOT MORE THAN ACTUALLY GET DIVORCED

So most of the "divorces" in the original didn't actually happen, and the fake-divorce RNG happened to hit the sick-wife group hardest.

An honest mistake I'm sure, and the authors themselves were the ones who asked to retract the study (and republish their now-negative results) when they realized. So probably not malicious.

But this ALSO means that at least two authors, an editor, and 2-3 reviewers looked at a table that starts like THIS

And said "yeah less than 1% attrition seems right"

Like. like. We haven't even gotten to your gender biases yet.

You are looking at a result that says that, out of 2700 COUPLES, the study lost track of LESS THAN 27 OF THEM over the course of 18 YEARS.

And you're not going "hm that contradicts some basic things I know about humans"

Christ guys, I can't even keep a DnD group together for a year, and you're saying 99%+ of everyone turned in a survey every two years for EIGHTEEN YEARS? I don't believe you. YOU shouldn't believe you. You could have avoided this by just asking where all the dropout couples went.

Anyway, when you fix this the overall divorce rate is 6%.

You will notice that this is way the hell smaller than the widowhood rate of 24%. Four of every five marriages that ended during this study ended with a death, not a divorce. Way less bleak than the original!

And now they're left with 161 divorces, total, and are still trying to divide that into smaller groups to see changes in the rate for each group.

Unsurprisingly, they see nothing of note. The ONLY conclusions I actually trust in the corrected version of the paper are

The majority of 50+ year olds stay married for life

If illness in general affects divorce, the effect is too small to see here (or to matter in the real world. Divorce rate of 6% if both are healthy and 6-7% if one is sick? That's noise, baby!)

Husbands get sick more than wives (...in a population where the husband is more often the older spouse)

Life-threatening illness increases your risk of death

Being old increases your risk of death

Being "somewhat dissatisfied" or "very dissatisfied" with your marriage increases your risk of divorce

Groundbreaking.