Tapsearch Com EditorDiversity and academic performance
January 21st 2007 01:45
A few days back I went through national IQ scores and compared them to school performance, finding that the two were related, but that there were many exceptions -- some countries performed better, and some worse, than they should have. I created a measure, called the IQ Utilization Score (IUS), to measure how a country's academic test scores measured up with its potential in IQ. The data is up here.
I e-mailed The Audacious Epigone about the post, because he has a much better grasp on statistics than I do. He remembered an old blog post I did comparing diversity (measured by the ELF, which asks, "what are the odds two random people in a country will be of different ethnicities?") to incarceration rates. That data here.
He hypothesized: Greater diversity might make education more difficult, driving down test scores relative to IQ. Therefore, there should be a negative correlation between the ELF and my IUS.
So, I ran the numbers. At first it seemed like yet another waste of spreadsheet-building time, with a correlation of .15 (that number squared is .02, meaning that diversity explains only 2 percent of the IUS data). What's more, it was positive, meaning that diversity actually increased test scores.
What I did was divide the countries into the top and bottom halves of IQ. The results are fascinating -- the top half has a correlation of .54, meaning that diversity explains 29 percent of the IUS data. The number is positive, meaning that greater diversity means better test scores relative to IQ in smart countries.
The lower-IQ countries get the short end of the stick, though, with a correlation of -.32. Diversity explains 10 percent of the IUS data, and greater diversity means worse test scores relative to IQ.
I e-mailed The Audacious Epigone about the post, because he has a much better grasp on statistics than I do. He remembered an old blog post I did comparing diversity (measured by the ELF, which asks, "what are the odds two random people in a country will be of different ethnicities?") to incarceration rates. That data here.
He hypothesized: Greater diversity might make education more difficult, driving down test scores relative to IQ. Therefore, there should be a negative correlation between the ELF and my IUS.
So, I ran the numbers. At first it seemed like yet another waste of spreadsheet-building time, with a correlation of .15 (that number squared is .02, meaning that diversity explains only 2 percent of the IUS data). What's more, it was positive, meaning that diversity actually increased test scores.
What I did was divide the countries into the top and bottom halves of IQ. The results are fascinating -- the top half has a correlation of .54, meaning that diversity explains 29 percent of the IUS data. The number is positive, meaning that greater diversity means better test scores relative to IQ in smart countries.
The lower-IQ countries get the short end of the stick, though, with a correlation of -.32. Diversity explains 10 percent of the IUS data, and greater diversity means worse test scores relative to IQ.
| 60 |
| Vote |
Add To: 
subscribe to this blog
