Saturday, July 18, 2020

Lies, damned lies and statistics

Say there are 100 people in a community and 5 of them are poor. The poverty rate is 5%. Say one of them dies of starvation. There are now 99 people and 4 of them are poor. The poverty rate is now 4.04%. The Leader puts out a celebratory announcement: The Poverty Rate has Declined by nearly 20%!
But is this really a good thing?
A similar trap awaits the unwary in the other direction. Say, again, you start with a population of 100, where only 5 are poor. The poverty rate is 5%. However, let’s say that these poor people are the strongest and healthiest of the poor. The weaker ones, especially infants, already died. The Leader pumps money into healthcare for the poor, focusing on reducing infant mortality. The next year the census reports that the population is now 105, of which 10 are poor. The poverty rate is now 9.5%. The opposition party puts out an accusatory press release: Under Leader’s Inept Rule, Poverty Rate Increase by nearly 100%!
But is this really a bad thing?
The error, in both cases, is confusing a 3-state variable for a 2-state variable. There are really three states: poor-living, non-poor-living, and dead. Looking at poverty rates fails to consider the dead state.

No comments:

Post a Comment