Hacker News Reader

> We need to add in to these calculations the news that the half-dose group included no patients older than 55, and to wonder what effect that had on the numbers, too.

I’m struggling to understand that. It can’t be random, with so many people in each treatment. So why are their treatment arms unbalanced? And how do they control for that in the analysis?

View on HN

When you run a large cohort experiment, i would imagine you balance the age profile to population after you do things. If you can get 1000 students in one day but spend three weeks hunting over 60s willing to be injected.. do you refuse to test 1000 students? No. You'd do something akin to Monte-Carlo random sampling into the age bins to get a distribution of experiments to match your needs. Maybe by the time they learned they had a fuckup on their hands, they'd only tested with the bodies available?

Half-dose group was unplanned and they switched everyone to full dose when they found out. It was in the beginning of the trial and they had only administered it to young healthy UK people by then.

Reply on news.ycombinator.com