Skip to main content
. 2013 Apr 4;177(9):904–912. doi: 10.1093/aje/kws340

Figure 1.

Figure 1.

Relationship between statistical power and sensitivity of the observed outcome measure in simulations with a 30% validation subgroup and a total sample size of 1,000 for the naive analysis, analysis limited to the validation subgroup, analysis using the direct maximum likelihood method, and analysis using the multiple-imputation method to account for outcome misclassification.