As Americans, we are in love with statistics. Sports commentators spew forth endless player facts during every game. Advertisers tell us that “four out of five doctors recommend.” Media organizations publish graphs and charts to illustrate impact. Yet without context, the vast majority of statistical observations are meaningless or even misleading. Think about the last time you said “Wow!” to some statistical claim. But then you discovered it was manipulated to serve someone’s political or fund-raising agenda?
In truth, most statistics are used to illustrate probabilities. But here is where the difference between probability and frequency becomes critical. Both probability and frequency rely on the question, “What are the chances?” But this is where we can draw the wrong conclusion. Probability is an interpretation. Frequency provides the exact number of times something will occur during a given period or population. If the chance of you getting the flu this season goes from one in five thousand to two in five thousand, for instance, that may seem infinitesimal. But if you read that the chances of getting the flu have doubled over last year, you would rush to the store to buy more soap, disinfectant and cold medicines.
This year, we’ve seen crime statistics that would make you think at Armageddon is imminent. Yes, murders in Chicago have risen to their highest level in 25 years. But when you drill down a bit, you discover that the vast majority of those murders have been the result of gang violence in three or four specific neighborhoods. I feel bad for the people in those areas. But that doesn’t dissuade me from riding the L from O’Hare Airport to the Loop.
I’ve jumped out of perfectly good airplanes. But knowing that the number of skydiver deaths has gone up 35% year over year doesn’t frighten me. Why? Because the actual number of fatalities went from 11 in 100,000 jumps to 15 in 100,000. It is still an extremely safe activity.
So, as a smart decision maker, which should you rely on when considering significant issues? Probabilities or frequencies?