Thread by @FondOfBeetles

Over the past month or so, I have been testing the hypothesis ‘Doing X causes Y to happen’.

So I have been ‘Doing X’ a repeated number of times and scoring how often it ‘causes Y to happen’.

https://twitter.com/ElectricAgora/status/1365429346576588806

If I don’t do X, Y rarely happens, but there is a background rate of Y happening in the absence of X.

If I do X, Y almost always happens, but there have been a few times where it didn’t happen.

Not doing X = Y happens in 4/60 tests.

Doing X = Y happens in 57/60 tests.

It’s clear to me (and statistically) that ‘Doing X’ does indeed correlate with ‘Y happening’, and I have a well-known mechanism to assert not just correlation but cause.

Furthermore, I have perfectly plausible explanations for why there is a background incidence of Y happening when I didn’t do X and for why Y doesn’t always happen when I do X.

I am now testing those sub-hypotheses.

But the existence of those outlier observations doesn’t negate the conclusion that ‘Doing X causes Y to happen’ and undermine the model I’m working on.

The same principle applies to sex classifications in humans. The ‘outlier’ observations are even rarer than those described above, and we understand (for the most part) how those people with atypical biology have diverted from typical developmental trajectories.

But people with DSDs don’t ‘disprove’ anything about the development of males or females in humans. In fact, the study of such conditions has fed information in to our developmental models.

Clarity: someone gonna get snippy about my use of ‘causes’ in the second sentence. ‘correlates with Y happening’