People who can design THAT level of testing get my complete admiration - and probably had experience on the front line as well as book learning. Marking those tests was as difficult as designing them, too - they had to be the result of a lot of iterations of adjustment, re-testing the test & checking what happened in reality.
But that's …
People who can design THAT level of testing get my complete admiration - and probably had experience on the front line as well as book learning. Marking those tests was as difficult as designing them, too - they had to be the result of a lot of iterations of adjustment, re-testing the test & checking what happened in reality.
But that's a test for a very specific situation. General intelligence tests have to be more, well, general. And it's so much harder to evaluate how accurate they are - there's no validation tests that can be guaranteed to be correct and objective, whereas your military test had one very specific and very objective deciding factor: were the guys who passed able to operate under fire?
There's a big difference between using psychological testing in such a very specific situation with such a specific success criterion, & using data from thousands of tests to make grand statements about whole populations.
When you do that, you can make reasonable claims about the statistics you generate, but you need to know the background of how the tests were carried out in some detail to evaluate their real value. I've seen many examples of "standardised" tests that could easily be misinterpreted, and seen papers showing that people who had practiced those tests & been given detailed feedback about their results were able to improve their scores significantly. In those circumstances, well-prepared candidates are going to appear more intelligent and the results are going to be completely unreliable but still have the veneer of respectable science.
People who can design THAT level of testing get my complete admiration - and probably had experience on the front line as well as book learning. Marking those tests was as difficult as designing them, too - they had to be the result of a lot of iterations of adjustment, re-testing the test & checking what happened in reality.
But that's a test for a very specific situation. General intelligence tests have to be more, well, general. And it's so much harder to evaluate how accurate they are - there's no validation tests that can be guaranteed to be correct and objective, whereas your military test had one very specific and very objective deciding factor: were the guys who passed able to operate under fire?
There's a big difference between using psychological testing in such a very specific situation with such a specific success criterion, & using data from thousands of tests to make grand statements about whole populations.
When you do that, you can make reasonable claims about the statistics you generate, but you need to know the background of how the tests were carried out in some detail to evaluate their real value. I've seen many examples of "standardised" tests that could easily be misinterpreted, and seen papers showing that people who had practiced those tests & been given detailed feedback about their results were able to improve their scores significantly. In those circumstances, well-prepared candidates are going to appear more intelligent and the results are going to be completely unreliable but still have the veneer of respectable science.