@jacd Agreed! I still think the concept of randomization and mimicry has utility. Descriptive epi is a lot easier when we have a random sample
@mike_johansen it once took me 2+ months to figure out an error in a simulation. That error was a single line that converted a float to an integer
@Tim_P_Morris thinking a little more about it, the (non-)collapsibility difference probably has to do with the relationship between PH and AFT for the Weibull:
\beta_{PH} = -\gamma \beta_{AFT}
@gailbear my gripe is that the first word isn't capitalized!
@Tim_P_Morris is it 1, 2, or both?!
@mike_johansen I do wish the US would change our culture regarding going to work/school while symptomatic. I think that would be another great iterative improvement
@chelseaparlett MARS
@pwgtennant we also beat those pesky epistemiologists
@Tim_P_Morris @PWGTennant lol I can't work that fast
@PWGTennant it's a little dated, but https://github.com/pzivich/Python-for-Epidemiologists
@PWGTennant makes sense to me!
Paul Zivich. Computational epidemiologist, causal inference researcher, and open-source enthusiast #epidemiology #statistics #python