search

fccoelho / paperLM1

5 stars 2 forks source link

Minor comment 9 #37

Closed maxbiostat closed 9 years ago

maxbiostat commented 9 years ago

page 7, paragraph 2: There is a need to propose remedies where numerators and denominators do no cancel when there is under-reporting.

fccoelho commented 9 years ago

Like in issue #22 , we need to use the simulation to argue this.

maxbiostat commented 9 years ago

Hey @fccoelho ,

On second thought, maybe simulation is not the [whole] answer. The theoretical take on this is that if we suspect underreporting is different for \sum_jY_j and S_0j, then we need to replace those with the appropriate integrals of different functionals of time (?). I fail to see a case where this scenario could be realistic, though.