Abstract

Unmodeled heterogeneity in survival rates can bias point and interval estimates of survival rate, leading to erroneous inference. I examined the effect of heterogeneous recovery and survival probabilities on confidence intervals computed for S i and S using a model that assumed no heterogeneity in the population. The rates at which nominally 95% confidence intervals constructed using Ŝ i and Ŝ#58, and their respective standard errors, covered the true population expected values were lowest for positively correlated subgroup recovery and survival probabilities

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call