Reports

R.A. Fisher's null hypothesis significance test transformed his field as few statisticians ever do. It goes like this. We run an experiment and assign half of subjects (like flipping a coin) to a wonder drug and another half to get a placebo. If we somehow show good results with our proposed drug, we have to prove that there's only one chance in twenty of getting results this good.

One out of twenty. Enter those numbers into calculator and we see that equals 0.05.

The probability threshold, or the p-value in statistician's lingo, of 0.05 is a magic number we are all obsessed with. Everybody I know, including myself, makes an anxious pilgrimage to this p-value whenever we run a statistics test on computer. To get a sense of what really happens when we execute the so-called signficance test, think about Cinderella's stepsister waiting for her shoe test. Hair on our arms stand straight up. Our hearts drop to somewhere in the bowels, waiting for the almighty p-value to pop up in the statistical package result page.

Those are the kind of moments when you wait and stare, and say, jeez, I've gotta go to the rooftops and shout hurray when the p is less than 0.05 (and you can still go there and jump if it's way more than 0.05).

Another classical example of such anxious moment comes from opening the envelope (or electronic version) of your kid's school report. I received mine this afternoon. That should have been an easy experience. But it wasn't.

I know there won't be p-value or ranking in my daughter's school report. Why, then did it take me pretty darn efforts to take a deep breath before opening the report? Admit it. It's because we are careworn parents.

One thing we know for sure: parents are anxious by definition.

As I read along the letter, the principal's disclaimer in the first few paragraphs made the report less stressful - and to good effect. To paraphrase the parenting expert Michael Grose, before you rip open the report do a little self-check to see if you are in the right frame of mind: Are your expectations for your son or daughter realistic and in line with their ability?

It doesn't take much of an imaginative leap to understand what the principal means by the tricky nature of expectation. His insight isn't new. Too much expectation can turn kids off learning. Too low and there is nothing to strive for. The curved, but not linear, relationship has played a central role in many things for centuries. It's called the Laffer curve. When we plot the child's effort on a vertical axis and the parental expectation on a horizontal axis, the curve turns out to look like a camel's hump. Definitely not a straight diagonal line.

The point is just this: the highest point on this curve is never to be on one end or the other. Usually somewhere in the middle. It's where we are aiming for.