Read on Twitter
a little lesson in statistics! let's say there's a group AB, and it is made of group A and group B. 10% of group AB is group B, the other 90% is group A. we are looking at the prevalency of factor C. factor C is true of 82% of group AB. if you had a statistic which...
looked only at factor C among group AB, you'd see that 82%. you'd then use that predictively to estimate that if you were faced with any 100-person sample of AB, you'd likely see something close to 82 cases of factor C. this just isn't true. let's continue looking at this...
100 person sample, and look at what our perfect theoretical sample would be. we'd have 90 members of group A, 10 members of group B, and of those we'd have 82 cases of factor C, and 18 members without factor C. the problem comes when our sample does not equally represent...
A and B. let's say we took a sample of all A and all B. out of 100 members of group A, the larger group, we see 90 cases of factor C. a bit more than we expected, but not too far out there. 100 members of group B, and we find only 10 cases of factor C. this is FAR...
from what we would expect using the AB statistic. that's because the significantly smaller sub-group has less of an effect on the statistic prevalence of factor C in group AB, BECAUSE it is smaller. however, a 90% prevalency within 90% of the group and a 10% prevalency...
within 10% of the group actually does come out to 82%. these numbers were chosen ahead of time. and the thing is- 90% in A and 10% in B actually IS representative of the situation i had in mind when creating this. let's also look at group XY, which is 10% group X and...
90% group Y. let's say group X has 90% prevalency and group Y has 10% prevalency, this time around. this is 18% prevalency in group XY. let's assume XY is the same size as AB, so factor C affects exactly 50% of the total population, ABXY. now, i would think, if estimating...
things about a particular member of any group, it would make sense to have groups based on sub-groups of similar prevalence of factor C. group AX has 90% prevalence of factor C, group BY has 10% prevalence of factor C. group B and group X appreciate this! they have...
a better idea of what's actually facing them, and everybody else has a better idea of what to expect from them. but A and Y don't really like counting B and X as being in a group with them sometimes. that's fine. group A (45% of the population with 90% chance), group...
X (5% of the population with 90% chance), group B (5% of the population with 10% chance), and group Y (45% of the population with 10% chance) can all be listed out separately! this also ensures clarity, so no one wonders if now it's just Y masking B like it used to...
mask X, or A masking X like it used to mask B, and everyone knows that they're seeing correct statistics about X and B, not just learning and arguing over A and Y while pretending to know things about other groups, and not making incorrect generalizations based on...
misleading statistics that ignore demographics. what factor C actually is isn't important here. what is important is to understand that if group A was 49% of the population with 100% factor C and group B was 1% of the population with 0% factor C, group AB would look like...
it had 98%. if group X was 1% of the population with 100% factor C, and group Y was 49% of the population with 0% factor C, group XY would look like it had 2% factor C. with B and X even smaller than A and Y in this scenario, the C prevalency rates are skewed even more towards...
the larger sub-groups within a group. this also makes a group "BX" exactly 2% of the population, which, let's stop beating around the bush here, i have heard as a GENEROUS estimate of trans people's percentage of the population. when you make a statistic about people with...
penises, or group AB (cis men being A and trans women being B), or one about people with vaginas, or group XY (trans men being X and cis women being Y) it can be HIGHLY misleading about trans people. and sometimes, yes, people complain about us making statistics about...
men (group AX) and women (group BY), and if you genuinely want statistics for trans women (group B), for instance, group AB is extremely misleading and group BY has potential to be so. you don't actually know which statistic you can find, one that sorts us among...
people with penises and one that sorts us among fellow women, is more misleading. i don't know what my breast cancer risk is from either of those, except for guessing based on which my biology is closer to. if a trans woman has breasts grown from estrogen, her biology...
is closer to that of a cis woman than of a cis man by every measure except chromosomes, which, hey, when it comes to congenital diseases, may be relevant. but that's not an excuse for transphobes to say we might as well be in AB. that's a damned good reason for our...
statistic in particular, that of group B, to have its OWN statistic. one which communicates to us, clearly, what our risk in particular is. however, for some reason, every TERF thinks a statistic about group AB simply *must* be applicable to group B, and that is...
quite frankly, an unjustifiably irrational assumption. you can't even pretend you're trying to figure out the truth at that point. and then they forget all about group X because they're so focused on belonging to group Y. this should hardly surprise at all, however, given...
the fact that they somehow think having a group BY would allow group B to completely overshadow group Y, despite Y being more prevalent than B as it as X, and be as much more prevalent over B as A is, to the same or roughly the same degrees on each consideration. basically,...
STOP DOING STATISTICS ABOUT PEOPLE WITH PENISES AND PEOPLE WITH VAGINAS. THEY ARE EASILY MISCONSTRUED OR WEAPONIZABLE. IT IS IRRESPONSIBLE. DO STATISTICS ABOUT CIS MEN, CIS WOMEN, TRANS MEN, AND TRANS WOMEN SEPARATELY, BECAUSE THEY'RE SEPARATE DEMOGRAPHICS. AND THAT'S...
WITHOUT EVEN CONSIDERING THE FURTHER COMPLICATION OF THESE STATISTICS BY THE EXISTENCE OF NON-BINARY PEOPLE, OR EVEN SPECIFICALLY BETWEEN AGENDER, GENDERFLUID, AND "THIRD" GENDER PEOPLE WITHIN THAT. JUST BE AT LEAST A LITTLE MORE PRECISE, BECAUSE AS I'VE SHOWN HERE,...
STATISTICS FOR GROUP AB ARE MEANINGLESS TO GROUP B, EVEN THOUGH PEOPLE WILL WEAPONIZE THEM AGAINST GROUP B. IT IS IRRESPONSIBLE. GET YOUR SHIT THEFUCK TOGETHER.
