"The concordance between identical twins, however, is rarely 100 percent - their IQ scores, for example, tend to be around 70 percent similar, compared with around 50per cent for non-identical pairs. By definition, inheritance therefore cannot be the only factor involved: if it were, identical twins would always turn out the same."
Surely "percent" is the wrong term here, which is puzzling, because why wouldn't the author of the submitted article, who is also the author of the book the article is based on, use the correct term?
Take a regression correlation, multiply by 100, and you have a crude sort of "percent". People intuit percent better than more complex (yet probably more mathematically rigorous) comparisons. If the book is written for a lay audience, this seems like an appropriate term.
That probably is where his "percent" is coming from, but I'm so used to reading even popular literature that doesn't dumb things down like that (preferring to explain typical ranges of correlations) that it was jarring to read that in the newspaper article. Maybe the newspaper editors imposed that on an author who does better in his book.
But what exactly does it MEAN to say "their IQ scores, for example, tend to be around 70 percent similar" when IQ scores have their own scale?
Surely "percent" is the wrong term here, which is puzzling, because why wouldn't the author of the submitted article, who is also the author of the book the article is based on, use the correct term?