[Yvain]

That post starts with "Please don’t read on until you’ve taken it, since this could bias your results."

But I checked the data I received before and after I posted that, and the birth order effects are about equally strong throughout.

[theptip]

I'd have preferred if Scott posted an encrypted timestamped message on Pastebin before the survey (to register his intended experiments), and released the keys after the survey; "this could skew your results please don't read" is not particularly well controlled.

[wging]

You don't really need to encrypt - just one-way hash. (Though if you count that as 'encryption' for this purpose, okay.)

[darepublic]

The prevalence of older siblings could be because of other reasons perhaps? Only people born with a certain period of time are likely to be interested in transhumanism and of those more are older siblings than otherwise.

[losvedir]

He explained that, too. The average age of all the older siblings that responded and the the average age of all the younger siblings that responded were within 1 year.

[AstralStorm]

The drawback of any survey driven study, right?

[searealist]

And what % was before and what % after ?

[Yvain]

5,523 before December 25, when I posted the pre-registrations. 2,254 after.

Of the before-posting people, 2199 had exactly one sibling, and 71.1% of those were the older.

Of the after-posting people, 776 had exactly one sibling, and 72.3% of those were older.

A t-test revealed no difference between the two groups, p = 0.527.

[itronitron]

it may also reflect that firstborns just have more of an interest in birth order effects so were more likely to respond