I’m sure that’s just an estimate right? Did they actually prove it runs 28,000 years? 
They should do an experiment on married couples, see if any babies are born with extra thumbs.
All this time, nature has had it wrong? Or was that extra thumb evolutioned away because it got stubbed too often?
I’m just imagining…
Ow! …Well forget it I’m completely out of the mood now
Science supports getting stoned with your lobster before you cook it.
1 Like
If you’re going to fake your data, USE A REALISTIC DISTRIBUTION
I have referred to the original study, too, so I’m a bit peeved.
Here is the histogram of the faked data – the number of miles driven:
This is a more realistic distribution of number of miles driven:
and yeah, I’m going to contact the editor of the newsletter where I referenced this study.
This was my old article:
2 Likes
Holy cow. That is flat out amazing. Any word on whether he attested for data accuracy at the beginning or end of the paper?
OK, more seriously, here is the Ariely’s response to Data Colada:
Response
Interestingly:
•I did not suspect any problems with the data.
• I also did not test the data for irregularities, which after this painful lesson, I will start
doing regularly
An actuary that admits this almost certainly would fail ASOP 23 . Sounds like a professionalism case study in the making.
ETA: As actuaries we most often operate in auditable circumstances. I would never imagine I could deny auditors access to the underlying seriatim data. The Data Colada group notes:
There will never be a perfect solution, but there is an obvious step to take: Data should be posted. The fabrication in this paper was discovered because the data were posted. If more data were posted, fraud would be easier to catch. And if fraud is easier to catch, some potential fraudsters may be more reluctant to do it. Other disciplines are already doing this. For example, many top economics journals require authors to post their raw data [16]. There is really no excuse. All of our journals should require data posting.
2 Likes
Yup, I was going to write about professionalism (I am eagerly awaiting my article on Batman and professionalism, yes, really) – ASOP 23, data reliance, etc.
This is one reason I always link to my original data sources (as well as mention the date I extracted the data). Anyway, there are loads of things I can think of relating to this stuff:
- negative results vs. positive results
- data checks
- reliance on others
- motivations
- data science/AI/ML
- assumptions
- professional standards and discipline
etc.
An actuary used raw data and had to admit to that? That’s kinda bad
Sorry, to be clear, Ariely is not an actuary. I was musing about how this behavior would be analyzed if he were. And if he were an actuary hopefully he would have performed analysis under ASOP 23 that would have called the data into question (although I have a suspicion he faked the data himself).
I see, makes sense. If he’s faking the data then I imagine he wouldn’t be too bothered by ASOP 23. Glad to hear he’s not an actuary though.
Well, this is going to be part of the argument I make.
If an actuary were involved in this (actually, I should go check), there could be serious repercussions.
Sometimes in academia, there are repercussions, but they generally don’t rescind your PhD… they may yank grants, though.
Combo wtg science and wtg life insurance industry. I had a friend tell me that companies in Canada are now issuing life insurance to MtF trans people at standard female premiums (presuming underwriting is standard).
The cases that were described to me were on people that had fully (?) transitioned medically. I don’t know, but suspect if they haven’t transitioned medically they’d probably be declined due to outstanding medical treatment. And I don’t know how it’s handled if they’ve not medically transitioned at all. Probably stuff that’s still being worked through.
Nevertheless, wtg science for making this possible and wtg insurance for being progressive in this area. A great example of why I’m so passionate about the industry.
3 Likes
I’ve gotta believe that’s just a PR decision. I’d be amazed if life insurers had enough credibility on trans people to even have a statistical view.
1 Like
I believe it was just that, though they can’t really use it as PR. Nevertheless, it was a really good outreach decision.
Changing your linkedin background during pride month is one thing. Issuing female rates otoh is really putting your money where your mouth is.
Well, there aren’t THAT many cases, so, not that much money stuffed into its mouth, wherever that might be in an insurance company.
1 Like
Are they doing it the other way round as well? I’m guessing it mostly matters because women live longer and therefore have lower rates?
This.
How often do insurance companies write off a one off type claim where the cost to contest is greater than the cost to pay? Currently for trans people, that is what I would think the costs look like.
Okay, more on the Dan Ariely thing.
Here’s the original paper:
They actually tried to replicate (and couldn’t), which was published in 2020:
There were 5 authors on the first paper, and all 5 were on the second paper (plus 2 new authors), so if all those involved on the research side were noted, the fraudsters were not able to repeat the feat, I guess.
But… it could have been a grad student used on the research the first time who wasn’t around for the second time. Why they would fake data if they weren’t a listed author… hell if I know.
That Buzzfeed article you posted above indicates Ariely was the only person in contact with the insurance carrier and he received the data from the insurer. He can’t blame anyone else on the team. He blames the insurer.
The conflicting answers Ariely gave on the origins of the data in that article certainly have an odor to them.
1 Like