<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> The standard deviation in the maximum likelihood estimator for a binomial distribution is 1 / sqrt{n p (1 - p)}, so given the error that they implicitly claimed of 2%, and p being about 60%, solving for n gives about 10k.</p>