I read the Diamond Princess study and the 19% figure is not based on following up with departed passengers. They're using some probability estimation based on t

I read the Diamond Princess study and the 19% figure is not based on following up with departed passengers. They're using some probability estimation based on the 50% discovered cases, I assume they're trying to account for people who had mild symptoms but not at the time they were tested. They identify flaws in the study though: the passengers were mostly old, so they are more likely to be symptomatic. They also didn't test everyone on the boat, only vulnerable, or symptomatic people, so their base numbers are missing some number of asymptomatic cases. It's not a great study really but I think you could take 19% as a low bound on the figure and in wider society 50% is plausible.