One small portion of the paper explained why composite/cluster scores from IQ tests often are higher (or lower) than the arithmetic mean of the tests that comprise the composite. This observation often baffles test users.
I would urge those who have ponder this question to read that section of the report. And THEN, be prepared to be blown away by an instructional video Joel posted at his blog where he leads you through a visual-graphic explanation of the phenomena. Don't be scared by the geometry or some of the terms. Just sit back and relax and now recognize, even if all the technical stuff is not your cup-of-tea, that there is an explanation for this score phenomena. And when colleagues ask, just refer them to Joel's blog.
It is brilliant and worth a view, even if you are not a quantitatively oriented thinker.
Below is a screen capture of the start [double click on icon to enlarge]
- iPost using BlogPress from my Kevin McGrew's iPad