Pick a Bayesian prior and roll with it. 20% of Gen Z is identifying as LGBT. Seems high to me. If we see another doubling for the next generation then something would seem to be obviously different. One doesn’t need to disentangle the effects per se