A Guide to Business Statistics

Chapter 2Populations and Samples

The summer of 2015 was a particularly menacing one for shark attacks on the coast of North Carolina. There were a reported 33 shark attacks in a 6-week span in a state that had seen a total of 25 attacks in the past 10 years. Living in North Carolina at the time, these stories were big news. The incidents received a lot of national attention too. In response to the panic during the height of the attacks, the Washington Post published a story online that attempted to put the recent high frequency of reported shark attacks into perspective.¹ To do so, the story reported on the average number of deaths each year in the United States from animal encounters. They found, on average, that sharks kill one person per year. Other creatures that were considered included snakes, spiders, bees, cows, dogs, bears, and alligators. Dogs, for instance, kill 28 people in an average year. Part of the take-away message was that shark attacks and fatalities are very rare indeed, even compared to other animal-related deaths. Continuing with the dog comparison, the article concluded that people are 28 times more likely to die from being attacked by a dog than being mauled by a shark.

If I can convince you of one thing in this chapter it is that the comparisons made in the Washington Post article are not very useful ones. The reason is that when comparing dog and shark fatalities simply by comparing their frequencies, the study implicitly defines the same population ...