3. 5. Confidence intervals

I use the following procedure to construct 95 percent bootstrap confidence intervals. For each cohort of size N, the individuals are sampled with replacement N times to produce a new cohort. For each cohort, I then estimate the parameters following the same procedures as with the parameter estimates of the original data. This is repeated 10,000 times, and for each iteration, the parameter estimates are retained. For each parameter, the 10,000 estimates are sorted, and the estimates that fall at the 2.5th and the 97.5th percentiles are used to construct a 95 per cent confidence interval.