Probability of a shark attack - a statistical fallacy?

[afdgf]

What the OP was referring to is a basic problem in epidemiology or biostatistics. If you just use raw numbers, then that is misleading because you are not taking into accoount the number at risk of the event. This is the questions that epidemiologist are always asking which is "what is the denominator?" A prevalence would be the number of cases over the number at risk. Prevalence is referred to as a "rate" by epidemiologists, but as a "proportion" by biostats people. Prevalence is used for conditions that do not change. Once a person is infected with some diseases, then they stay positive on the lab test for life. In other diseases, the lab test reverts back to negative when the person is successfully treated. In the case of something like a shark attack, the appropriate rate to use would not be a prevalence, but rather an incidence. An incidence is a true rate and epidemiologists would agree with the biostats. However it is more complicated to calculate. The numerator is the number of people who had the event during a specific time period. The denominator is the number who were at risk during the same time period. Then there is usually a multiplier so that the numbers are easier to deal with instead of large number of digits after the decimal point. Even if you can get the numerator, it is very difficult to get the denominator. This is not a problem of statistics (if we use the term correctly). This is a problem of data. What data should be used for the denominator? In SCUBA this is extremely difficult. HTH
Dennis

 

[boulderjohn]

Hummm - I honestly can't tell if you're trolling here! ;-)

He is not trolling, since the verb is in the present tense. That post is 11.33 years old, decidedly in the realm of the past tense now.

It's easy to miss, but people responding to this thread should realize that most of the posts are more than a decade old, and the people who engaged in the original discussion are mostly not logging in any more.