Maths - Defending what counts
What it's all about
I look around my statistics class. Brows are knitted as we tackle correlation. An exam question is on the board, writes Jonny Griffiths.
"So what do we need here?" I ask, hopefully. Looking innocent, Graham puts up his hand: "Is it Superman's rank correlation coefficient?" Suddenly, I am in Metropolis ...
Walters, Evans and Simcock, the three grey-suited men around the table - the nucleus of the National Security Statistics Board - are talking in tense whispers. On a projector screen is a scatter diagram, with 30 points arranged on it. The variables are so secret that no one has been permitted to label them on the diagram.
"It's no good," whispers Walters. "We can't use the PMCC for this test."
"No," agrees Evans. "We have to have a bivariate normal distribution for that to work and ..."
"The evidence for that would be an elliptical arrangement of dots," chips in Simcock. "And we hardly have that. There does appear to be some association ..."
"But it won't be linear, will it?" says Walters, ashen-faced. "And if we can't complete this test ..."
"It's the end of civilisation as we know it," says Evans, grimly.
"Wait!" says Simcock, suddenly spotting something out the window. "Is that a bird? Or is it a plane?"
"No, Gentlemen!" cries Walters. "It's a rank correlation coefficient!"
A few moments later, Superman dusts himself down. "Mr Spearman sends his apologies," he says. "He can't be here in person, but I've brought his non-parametric test along. Just what you need, I think."
Jonny Griffiths' resources will help you to conquer statistics. bit.lytesStatistics.