I written several blogs on the necessary minimal statistical knowledge needed for the modern world. I may have been writing for the visiting Martian aliens.
One of my two oldest granddaughters was doing her 3rd year high school math for an IB class. IB is similar to AP advanced placement, except it is an international standard. She wanted my help on a statistics project.
I was astounded. She was using statistics with nearly the same agility as I. And she loved it as much as I do.
So I am not writing this for my progeny, maybe for Martian aliens.
Knowing about alpha and beta errors is important. It probably saved my life several decades ago, at least kept me from ruining it.
Sometimes called type I and type II errors, these are erroneous test outcomes. An alpha error is finding a result that doesn’t exist, and beta is not finding a result that does exist.
When I was first married and had a blood test for something-or-other, the test showed I had syphilis. Further testing showed that I didn’t. If that idiot doctor who did the test and the technician who told me the results had told me that the test (the Wasserman) had a 23% alpha error I would never have taken it. That is a bad alpha error. My wife and I knew it was a mistake but it could have seriously disrupted a happy marriage for a few weeks while the more reliable test (TPI) was done.
Nearly 20 years ago a doctor recommended a PSA test for prostate cancer. I looked up the alpha and beta errors. Both were too high for me to trust the test outcome. Mostly the alpha was too high. I’ve never taken the test.
Several of my friends took the PSA test, merely because they were getting older. They had the surgery and the various follow-ups. Nearly all of them had incontinence problems and all of them had erection problems. That defines ‘near death’ for me.
Did they have prostate cancer? Who knows, the PSA test is crap. If they had cancer was it dangerous? Again, who knows.
One of the few doctors who could pass for a competent scientist wrote a book on How We Die. He paid for 14 autopsies on men who died of ‘old age’. One third of them had prostate cancer. They didn’t die from it.
People should understand alpha and beta errors. It matters to our lives.