Jousting with the Lancet: Pajamas Media Interviews Professor Gilbert Burnham
Tom W said:
That seems to be the criterion for the contribution from main street households to dominate but even when they don’t dominate the bias can be as large as 100%.
That depends on the parameters. If the population living on main streets is only 1% of the total population, then if the rate of death there is as much as 20-fold higher than for the whole population, you’d still only underestimate the total deaths by 20% by ignoring the main street deaths.
Sam said:
What am I missing? Aren’t most of the reports of bombings in markets, on main commercial streets, near police stations, by lines for whatever?
Yes, and the victims would likely have their homes mostly on residential streets in the neighborhood, and would therefore be detected in the survey.
