To hell with UPs arbitrary percentage, I have an actual way to test if something is significant, something that you can't fucking argue with: actual, real MATH. A statistical test. More specifically, a Z test. I'll be marking each transphobic post as a "1," and any non-transphobic post as a "0." The list contains 155 elements, consisting of 37 1s and 118 0s.
mu: 0.37374
sigma: 0.48626
mu0: 0
Ha: mu != mu0
From this, we get:
Z = 7.64743
P(Z) = 2.0685x10-14
This means that the probability of this event happening by chance alone is infinitesimally small; we do not have enough evidence to reject the alternate hypothesis (Ha). Statistically speaking, this must then, therefore, be due to some outside influence, the most likely culprit being targeted transphobia. Oh, and the great thing about this? Literally anyone can check my work. Go ahead, if I cocked up something, let me know.
Fucking math. It works, bitches.