So the guy outright lied,

as shown in the chart above,

But the real strange thing?

— Whole comments section is filled with hundreds of people who just blindly accepted his BS false presupposition, that the chart shows nothing,

Nearly all instead just skipping over to more minor, secondary issues,

instead of confronting the core lie that the chart shows nothing

Humans are INCREDIBLY vulnerable to being manipulated by false presuppositions

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

Ok, I’ll spell it out

Dude brings charts on ACT scores,

uses those charts to say that the male curve doesn’t look fatter than the female curve enough to show sex differences

— But he refused to actually show that sex difference line

So plug in his own summary statistics and get AI to actually plot that sex differences line, and we see that his claim was BS

— Men hugely overrepresented at the the top and bottom of those curves

Just like the greater male variability hypothesis predicted

I.e. dude’s own numbers showing he was BSing

— And plotting the curves as he did is classic lying with statistics, because it hides away the parts of the data that show the most important parts of the data

There are several better ways to plot this data that are far less misleading, as we’ve shown

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

Why just plotting the probability density curves, as he has, misleading lying with statistics?

For one, when plotted this way,

it moves the biggest gap to the MIDDLE (red)

and then it spreads out the rest of the mass in the opposite direction along the sides (blue), to make it less noticeable

— when truthfully the sides are where the hugest differences are, as shown by the U-shaped line on the second chart

Classic lying with statistics

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

And would add that using normal distribution summary statistics, as that guy did, further distorts how extremely ~100% male the edges often become

— because, as you can see, the normal distribution approximations cause the Male/Female ratios to attenuate as you near the edges

— but that’s an error of the normal approximations

Really, the real empirical the female rates often go straight off a cliff, and the curves themselves are more pointy triangle things that hardly look like the normal approximation at all

(Except when there’s large
measurement error — which blurs things to look more like normal distributions. But again, that’s a problem with the measurement, not reflecting the underlying reality)

Basically numerous completely separate reasons why the reality is far more extreme even than thus guy’s charts show

— Though even with his numerous methods to hide the truth, he still couldn’t fully hide it

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶