Let's now focus on measurement error due to errors in coding of the cause of death in the death record, based on hypothetical sensitivity and sensitivity values. Assuming no other sources of outcome misclassification and again, focus on IHD mortality. Presented below are hypothetical TRUE data from a simplified scenario examining census tract- level PM2.5 at baseline and cumulative IHD mortality over a 1-year period (this is not what the Hayes paper examined). In this hypothetical, simplified scenario, PM2.5 is categorized as low (=20 ug/m3); and outcome variable categories are IHD death and No IHD death. PM2.5 category IHD death No IHD death Total (ug/m3) >=20 5,599 66,884 72,483 =20 72,483 <:20>