Interesting Links for 10-03-2019

[andrewducker]

10 most awkward work moments

(tags:work funny )

5 Ways We Train Teenage Girls To Love Psychopaths

(tags:relationships abuse society )

The USA (and UK) is really messed up when it comes to platonic intimacy

(tags:relationships usa uk society )

Entirely automated recycling facility which processes 90 tons per hour

(tags:recycling automation video )

Japanese woman confirmed as world's oldest person aged 116

(tags:age Japan )

Humanist marriages least likely to end in divorce

(tags:divorce religion scotland )

Comments

[armiphlage]

That automated recycling facility is totally unlike how I would design it. Rather than take up manufacturing space with so many conveyors, I'd have the robots pick and separate the waste right on the tipping floor. Of course, my facility would probably take a year to process a single tonne of trash, so their system makes better sense.

[haggis]

I think the purpose of the system to turn a single stream of very very variable materials into multiple streams of similar materials (by weight, density, metal content) and then the robots sort them into 2-3 categories. Asking robots to search through the whole mound into many categories is probably too hard and time-consuming.

[doug]

I love Rev Norman Smith, the Church of Scotland person, saying, "Without understanding many of the variable factors affecting divorcing couples, any suggestion of causation between type of ceremony and divorce rate is entirely spurious."

That's not quite how I see it. To slightly reword the XKCD cartoon mouseover, correlation doesn't prove causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing "look over there". It's wrong to say its proven, but it's not spurious to suggest it might be the case. As I am fond of saying, when you see a correlation between two things, A and B, there are seven possibilities:

(1) A causes B
(2) B causes A
(3) C causes both A and B
(4) Some complex mixture of the above
(5) There's no causal entanglement and the apparent correlation is a mistake
(6) There's no causal entanglement and the apparent correlation is a fluke, the result of statistical chance
(7) Some complex mixture of all of the above

Most people, me included, and probably the Rev Smith, are happy to rule out retrocausality. Believing that causation only flows forward in time helps us here because we have a clear separation in time between A and B. We can be pretty sure that living in a long-term happy marriage does not ever cause the couple to have had a humanist wedding, and we can be pretty sure that divorces cannot make people be religious at an earlier stage in their relationship. But that only rules out one of the seven options.

If we accept the results, I think we can rule out the statistical fluke option, on size alone (without having done the sums to verify this, but by eye it looks so massively big there's no need for a retrospective p value calculation, even if you believe in them). I don't think we can rule out the "no causal entanglement" options entirely, though. For one thing, the results are so big that I suspect them simply on those grounds. (ost social science findings are nowhere near that size.

But if we treat it as a real data, we absolutely are in to discussing causation. This is not simple. I completely agree with Rev Smith that there are likely to be many factors at play here. But I disagree that any discussion of causality is spurious: I think it's absolutely reasonable to discuss it.

I would also agree with him that it seems prima facie absurd to suggest that the mere form of a marriage ceremony has such a large effect. I am slightly surprised to agree with a cleric here, since I might have guessed that a religious person with a sacramental view of what's going on in a wedding would take a very different view to a materialist atheist like me. "They twain shall be one flesh", "let not man put asunder," and all that. But he is Church of Scotland and they tend to be pretty practical, solid, non-mystical sorts. (I should know, I was raised Presbyterian, with "if you pray for potatoes, start digging".)

So the Rev Smith and I are, I think, largely in agreement that the form of marriage chosen is a proxy variable here for a number of factors that substantially affect the lifespan of a marriage. But what to do about those? We could perhaps unite around a call for better evidence, but to be honest the sorts of experiments you'd need to get really strong proof of causation here are wildly unethical and ludicrously impractical.

Given this imperfect information, we quickly get in to very murky waters of decision theory: problems like the 'smoking lesion' and David sending for Bathsheba. But for those problems to be tricky, you have to be sure about several aspects of the problem, in a way we are not here. So it does seem to me that if you are in doubt as to which form of marriage to choose, and you want to reduce your risk of divorce, you might be well advised to try to be like, and behave like, the sort of person who would want a humanist marriage. Quite how to do that (assuming you don't, or don't particularly) is challenging ... but it would at minimum include having a humanist marriage ceremony.

(Disclaimer: I am not married and do not want to be.)

[doug]

And it seems the Church of Scotland was also talking about this on Today this morning, but not the Humanists, because they're not allowed on Thought for the Day: https://twitter.com/Humanists_UK/status/1105026084758794241

[naath]

I expect, I may be wrong, that a large factor at play is that the type of couple likely to want a humanist wedding are celebrating an existing long term happy relationship and desiring some legal recognition (and tax breaks) (having perhaps had any number of previous relationships that didn't make it that far; and so never got 'divorced' because they never bother to marry) and the type of couple likely to want a religious ceremony (when humanist ceremonies are easily available) are celebrating the start of what they hope will be a happy long-term relationship in the belief that having the relationship first would be not the sort of thing the church would like. A rather fundamental disconnect in the idea of what marriage is *for*.

[haggis]

Your comment and Doug's final line bring up another issue - the number of people in marriages of any sort are a subset of the number of people in long-term relationships and the number of people divorcing in any given year are a subset of couple breaking up in any given year. They are the easiest group for bureaucracy to track but not necessarily representative.