Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes
Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes
In randomized studies with missing outcomes, nonāidentifiable assumptions are required to hold for valid data analysis. As a result, statisticians have been advocating the use of sensitivity analysis to evaluate the effect of varying asssumptions on study conclusions. While this approach may be useful in assessing the sensitivity of treatment ā¦