Hiding Type 1 and Type 2 Error

The deception with tact
Just what are you trying to say
You've got a blank face, which irritates
Communicate, pull out your party piece
--INXS

When deciding to act under conditions of uncertainty, two types of errors can be made. A type 1 error is involves acting when you shouldn't (overreaction). A type 2 error involves not acting when you should (underreaction).

For example, in a health care context, a type 1 error would be a 'false positive' that leads you to believe that you have a medical condition that requires treatment when, in reality, you are fine. A type 2 error would be a 'false negative,' meaning that the test did not pick up the medical condition and leads you to believe that you are fine when you really needed to undergo treatment.

These pages recently considered the game theoretic position of public health officials and policymakers who must decide whether or not to act now to best address a potential but uncertain public health crisis in the future. Our analysis concluded that the dominant strategy is to act--because there is a high likelihood that public officials will realize favorable outcomes regardless of whether a future health crisis manifests or not.

Admittedly, however, this analysis depends to some extent on whether public health officials can control public perceptions of type 1 or type 2 policy errors. For example, if they feel that they can manipulate case and death counts numbers associated with a disease outbreak, then public health officials may be more emboldened to select a strategy reflective of 'active agency'--i.e., a response that fits their personal interest more than it fits public interest.

Over the past week we've seen examples of how public health officials are attempting to manipulate COVID-19 case and death counts to control public perception of type 1 and type 2 policy errors.

Accusations are surfacing that Chinese officials deliberately under-reported the severity of the coronavirus outbreak in China. They may have suppressed case and death counts to make the situation look far worse than it was. If the allegations are true, then Chinese public health officials have been engaging in behavior to suppress public perception of a type 2 error.

Reports are also surfacing that officials in some states, such as New York and Ohio, are beginning to count cases and deaths that have not been formally confirmed as COVID-19-based among their official statistics. Why make the numbers look worse than they truly are? To hide type 1 error. By inflating the official counts, public health officials seek to deflect claims that they overreacted--that they made a mountain out of a molehill.

Not sure fudging the numbers will work in this case. Regardless, we are witnessing textbook examples of officials displaying active agent behavior in attempt to hide type 1 and type 2 error.