Funny thing about statistics. Often, the people using them truly believe what they claim, precisely because they don’t really understand statistics. I used to give people this test:
Which statement is true: a) the highest concentration of cancer occurs in small communities; or b) the lowest concentration of cancer occurs in small communities. 80% of people will answer ‘a’ based on the idea of cancer clusters. Few, if any, will answer ‘both’, which is correct, because they don’t understand that, statistically, the extremes occur in smaller populations because they are small. Larger populations regress to the mean. And when explained. They don’t believe it.
I think most of our journalists live in that mentality.
Mr. Lorentzen, this is a timely piece. A coworker was t-boned just the other night. When he called the police to report the accident, they said they didn't have anyone to send. The woman that crashed into him didn't speak any English, but they were able to communicate in Spanish. It turns out, she didn't have a driver's license, registration or insurance. This inconvenient data will not affect the 'official' data
That actually brings up a whole new realm of variability compared to history. How many crimes go unreported because of the compromised legality of some increasingly large communities? Homeless, mentally challenged, or citizenship? Tough analytics…
“There are three kinds of lies: lies, damn lies and statistics.” Mark Twain, attributed to Benjamin Disraeli, who may never have said it.
Funny thing about statistics. Often, the people using them truly believe what they claim, precisely because they don’t really understand statistics. I used to give people this test:
Which statement is true: a) the highest concentration of cancer occurs in small communities; or b) the lowest concentration of cancer occurs in small communities. 80% of people will answer ‘a’ based on the idea of cancer clusters. Few, if any, will answer ‘both’, which is correct, because they don’t understand that, statistically, the extremes occur in smaller populations because they are small. Larger populations regress to the mean. And when explained. They don’t believe it.
I think most of our journalists live in that mentality.
Mr. Lorentzen, this is a timely piece. A coworker was t-boned just the other night. When he called the police to report the accident, they said they didn't have anyone to send. The woman that crashed into him didn't speak any English, but they were able to communicate in Spanish. It turns out, she didn't have a driver's license, registration or insurance. This inconvenient data will not affect the 'official' data
That actually brings up a whole new realm of variability compared to history. How many crimes go unreported because of the compromised legality of some increasingly large communities? Homeless, mentally challenged, or citizenship? Tough analytics…