Any ETA on when this will resolve?

@AhronMaline oh man i'm sorry my life got hella insane and is still insane, i will try to resolve this next time i sit down with my data

@Aella Oh! I hope whatever's going on turns out well for you!

Betting on non-religious. For two reasons:

1. I think people with unusually high libidos are less likely to remain religious, even if that's where they started.

2. For religious people who attend church, that's one less morning free to stay in bed.

Are you factoring in consent/coercion? People (especially religious) aren't exactly going to be forward in admitting they might frequently have sex with their long term partners but only after exerting substantial pressure on them.

selection bias is gonna be nuts here

holy shit the options are so neck and neck

Please specify what you mean by "long-term", a decade or 3 decades

By long term I'm assuming between 5-10 years

Among people who report being in relationships 5 years or more

"long-term" is 5 years or more

How do you define “religion”? Is it just based on self-identification? The problem is that different societies/cultures have different standards of what counts as “religion”.

@Vesperstelo yes, also interesting is how religious the person is

@Vesperstelo, the description answers your question.

Sorry if this was posted elsewhere but how was this data gathered?

@StephenFowler28ac The question was asked, and the answer speculated, but no definitive answer was made yet.

@StephenFowler28ac Snowball sample; i made survey that gives people results, posted it on social media, it got shared around. It's also linked at the end of the Big Kink survey, which has been regularly capturing new ppl for years now (very good google SEO, going viral on tiktok, etc)

Usually when doing my own followers it caps out at 10k if i'm lucky, but more realistically like 5k. I would guess majority of people answering are not my followers.

The logical choice would be if it is not significant at the p = 0.05 level.

@LilyFleming everything is significant at the p=.05 level if you collect enough data. That doesn’t indicate a meaningful difference.

@MatthewKhoriaty Are you gesturing at the difference between statistical significance and practical significance, a la effect size? If a difference doesn't actually exist, gathering a lot more data shouldn't then allow us to reject the null hypothesis.

@Haiku In complex fields such as social science, the only variables that aren’t correlated are those that anti correlate. This is called the Crud Factor.

https://gwern.net/everything

@Haiku Enough data does allow us to reject any null. https://gwern.net/everything

Idk exactly the threshold I'd use to say it's 'equal', and I don't feel like figuring out an exact number to give right now

I think it would be a better market if you did. The probability of "no difference" depends entirely on your chosen threshold.

@Fion @Aella Seconding this. Also, you shouldn't bet on the market until after all decisions about threshholds, controls, etc have been firmly made.

@AhronMaline meanwhile, I have been happily volatility trading :-/ feel very dirty now that you two pointed that out. Though in my defense. I have never traded but the way I believed, so if I got trapped in a position, I would have been happy.

@JussiVilleHeiskanen We were addressing Aella, the market creator, asking her to clarify her threshholds. I added that she shouldn't trade before doing so, since that would create conflicts of interest - both for the market, and more importantly for her research integrity!

Are you involved in Aella's work? If not, I don't see any reason not to do volatility trading.

@AhronMaline no no, not in the field at all, just armchairing, like most everything here on my part

@JussiVilleHeiskanen so I'm a bit confused why you felt dirty about volatility trading, in response to our comments?

@AhronMaline just getting more serious about my choices at participation, with a bias toward meaningful calibration as a genuine search for truth. Not totally strict but a slight bias

@Fion I'm not sure how to determine threshhold here. Maybe i look at stdev of total responses and if the difference between religious and nonreligious is like. idk some fraction of that stdev?