Interesting Links for 22-05-2019

[andrewducker]

Rwanda has fought to stamp out cervical cancer - they might just be the first country to do it

(tags:cancer women rwanda GoodNews vaccination )

An interesting observation about horror film titles

(tags:horror movies names )

Protections for women online are woefully inadequate

(tags:police harassment crime uk Internet )

Terminally ill fans granted wishes to play Borderlands 3, Three Kingdoms

(tags:games death )

Schwarzenegger and Fassbender to star in Kung Fury 2

(tags:movies wtf )

Two dead after YouTube conspiracy video claims drinking bleach cures autism

(tags:autism poison youtube OhForFucksSake )

Two women programmers played a pivotal role in the birth of chaos theory

(tags:chaos programming women history )

The grocery trip which helped transform Russia

(tags:history russia usa shopping )

Foreign Minister Taro Kono to ask media to switch order of Japanese names

(tags:names japan )

People Have A Hard-To-Explain Bias Against Experimental Testing of Policies And Interventions, Preferring Just To See Them Implemented

(tags:research psychology testing )

LEGO reaches 100% renewable energy target three years ahead of schedule

(tags:lego renewables )

The physical kilogram is officially dead

(tags:science weight )

The art of time well spent

(tags:time advice life )

African samurai: The enduring legacy of a black warrior in feudal Japan

(tags:race Japan history )

AMC cancels sci-fi robot drama 'Humans'

Shame, I really enjoyed the first season.
(tags:ai robots tv )

Comments

[nancylebov]

I think the underlying thought behind the bias against A/B testing is the idea that people (perhaps especially people in charge) should just get things right.

Doing A/B testing means spending half your time doing things which are are wrong. Obviously a waste and possibly abusive.

[andrewducker]

I bump into that mindset intermittently and it's always deeply worrying.

[mtbc]

Humans had a decent first season, agreed. Developments toward the end of the latest season definitely made me fear we were heading into a distinctly daft arc so I'm not as sorry as I might otherwise have been.

[andrewducker]

Ah well, that's not so bad then.

[mtbc]

I like to think that they did get chance to say everything interesting they had to say (or at least raise/explore). (-:

[alithea]

Yeah, I confess I only watched the beginning of season 3 because I wasn't in the mood for where they seemed to be taking it, although I have meant to catch up at some point...

[calimac]

Yeltsin thinking the US grocery market was possibly staged for him: I've heard a story of some Russian visitor being convinced it must be staged, because if it weren't, the shop would be jammed with customers frantically buying everything. Isn't that how it would be if such a thing appeared in Russia? they were thinking.

Order of Japanese names: I can appreciate why they might want us to change the order, but it will create a headache as it's not always obvious to a Western eye which Japanese names are forenames and which surnames, so without a consistent policy it's hard to figure out which one to use for surname reference. This is already the case for some Chinese names, as some bearers turn the names around in Western writing and some don't. (If there's two hyphenated forenames, that's a clue, but that's not always the case.) Next up, Hungarian.

The researchers don't understand why people have an aversion to being subjected to A/B testing? They don't understand it? Really? The ignorance of human psychology here truly astounds me.

[naath]

Yeltsin went to that store on my 6th birthday; change happens so fast.

[andrewducker]

I was watching Pole to Pole with Jane the other day (we're watching a bunch of the Palin travel documentaries) and he goes through Russia, and it really was awful watching him try to buy some vodka in a shop. I can understand Yeltsin being blown away by western shops. It's such a massively different experience.

I liked a suggestion about capitalising family names. But then it would be odd to do that only for Japanese family names. I agree that it's a tricky one to solve.

[andrewducker]

Yeah. The world I was born into is radically different to this one. Although some are trying to turn the clocks back.

[doug]

Mmm, I don't think that's it. People have this objection even when that isn't the problem.

I've fairly recently had an argument where we had a shiny new thing that we were going to introduce. We were pretty confident it would be good (or we wouldn't have done it) but didn't have any hard evidence. We didn't have enough resource to give it to everyone, though, at least not at first. So the obvious solution was a phased roll-out. Everyone was happy with that. They were even happy with the idea that we might dish it out at random, to make it fair. But when I suggested we then compare their results to check that it really was worthwhile, many people got terribly upset and angry and thought that would be wrong. And at that point most people stopped being keen on randomising who got it, either, and were less worried about unfairness.

This is a real effect, and although I don't understand it, I do need to respect it, because it's simply not ethical to do things to people that they'd object to. Even if I think their objections are nonsensical.

[andrewducker]

I'd be interested to see if you could explain it to Doug.

https://andrewducker.dreamwidth.org/3738866.html?thread=25933554#cmt25933554

[doug]

The researchers don't understand why people have an aversion to being subjected to A/B testing?

I must confess I don't fully understand it either, but do want to know more.

What about a situation where there's a new drug, but it's difficult to manufacture so the supply is way less than the demand. Presumably you wouldn't think the ethical choice is to withhold the drug from anyone until everyone can have it? If there weren't a good medical reason to decide who gets it, would allocating it at random seem fair? And if so, would it then seem unfair or wrong if we did that and also compared the results for people who got it with people who didn't get it, in order to get good evidence about the drug's effectiveness?

I've met people who do think that last step would shift the scenario from being completely fine to being unethical and wrong, and I genuinely don't understand what the big difference is.

(For the avoidance of doubt, given that many people do think this way, it would be deeply unethical to run that experiment without explicit, informed consent, given in a no-detriment context.)

Anti-testing bias

[anna_wing]

In the public policy sphere there is a similar problem with trying to change policies that prove not to be working. Since most public policy in many countries is formulated or advocated more on the basis of ideology than on evidence, changing a policy to improve it (or dropping it altogether if it turns out to be ineffective) is then perceived as a "U-turn", "climb-down" , "giving in to pressure" or other politically-charged term.

I don't know if if this is a reflection of the bias, or if the bias arises from this attitude to policy-making/change.

Certainly the concept of "evidence-based policy formulation, outcome-based policy assessment" is still a new and bizarre one to many people in the public policy-making or advocacy sphere (a surprising number of them go on to ask "but what about your principles?" As if principles have any value divorced from the actual effects of their application).

[channelpenguin]

I don't understand it either.

Has anyone tried asking the "objectors" to articulate *why* they object? If so, what were the answers.

(p.s. I fully appreciate there is also a problem wherein many people also object to articulating the "why" of their opinions or behaviour - I would be interested if the subeset of "A/B testing objectors" and this subset show substantial overlap)

[calimac]

I have seen all caps in names in some academic contexts. It's useful. But I just reviewed a book by someone named DiAngelo, and the name was in all-caps on the book cover, and it wasn't until I looked it up that I was sure it wasn't Diangelo.

[calimac]

This is purely emotional, and has no basis in logic whatever, so please don't start explaining why there's no logical reason to feel this way:

But it's a deep-set emotional reaction that it's unfair to randomly differentiate people's rights this way. If nobody gets it, then the treatment being unavailable at all at least isn't unfairly distributed.