Psychometric Testing
It is clear that activity informs the brain. The impact of different activities on the behaviour of the brain can be measured using functional MRI and other methods, in what is known as functional neuroimaging.
This means that the person who spends his time on real-time strategy gaming and the person who spends her time rock-climbing would be expected to have very different neural networks. In fact, the neural networks developed by these people would probably not be human-normal. Then again, given the uniqueness of human experience, the only way to assess human-normal would be in terms of frequency. If there are 2 billion farmers in the world, they'd probably set some sort of standard for brain function – an alien super-census might decide that we were on average a planet of slightly screwed-up agriculturalists.
But what this implies to me is that modern psychometric testing (as applied nowadays to students, for various dubious purposes) may well be unreliable, invalid, or both. It might be unreliable in the sense that what it measures may be of such low significance to a modern student's neural net that the student will give random answers and hence never be measured the same way twice. It might be invalid in that the modern student's neural net has talents that the test is not designed to measure, and which subvert the talents which the test measures by overriding them.
For example, a psychometric test might require a student to execute mental visual-spatial manipulations in order to select an image to complete a sequence. This test has measured visual-spatial manipulations reliably, with some adjustments, up to 2003. However, in the last four years, increased use of graphic manipulation software has produced a group of students who manipulate images in terms of other mental paradigms (e.g. Photoshop filters). What the test then ends up measuring, rather unevenly, for these students, is their ability to create mental sequences of Photoshop commands and visualise their effect.
Such a thing might be close enough to the targeted ability to pass for the ability itself. However, the mystery of the Flynn Effect prompts the question of how accurate a test last calibrated four years ago (for example, WISC-IV, last calibrated in 2003) might be.
My intuition is that humans love reification and ranking too much to get rid of psychometric testing and its more dubious secondary constructs. But eventually, in more and more areas of human endeavour, psychometric testing will become irrelevant. This will depend on how diverse the range of human brain function will become in this rapidly complexifying world.
This means that the person who spends his time on real-time strategy gaming and the person who spends her time rock-climbing would be expected to have very different neural networks. In fact, the neural networks developed by these people would probably not be human-normal. Then again, given the uniqueness of human experience, the only way to assess human-normal would be in terms of frequency. If there are 2 billion farmers in the world, they'd probably set some sort of standard for brain function – an alien super-census might decide that we were on average a planet of slightly screwed-up agriculturalists.
But what this implies to me is that modern psychometric testing (as applied nowadays to students, for various dubious purposes) may well be unreliable, invalid, or both. It might be unreliable in the sense that what it measures may be of such low significance to a modern student's neural net that the student will give random answers and hence never be measured the same way twice. It might be invalid in that the modern student's neural net has talents that the test is not designed to measure, and which subvert the talents which the test measures by overriding them.
For example, a psychometric test might require a student to execute mental visual-spatial manipulations in order to select an image to complete a sequence. This test has measured visual-spatial manipulations reliably, with some adjustments, up to 2003. However, in the last four years, increased use of graphic manipulation software has produced a group of students who manipulate images in terms of other mental paradigms (e.g. Photoshop filters). What the test then ends up measuring, rather unevenly, for these students, is their ability to create mental sequences of Photoshop commands and visualise their effect.
Such a thing might be close enough to the targeted ability to pass for the ability itself. However, the mystery of the Flynn Effect prompts the question of how accurate a test last calibrated four years ago (for example, WISC-IV, last calibrated in 2003) might be.
My intuition is that humans love reification and ranking too much to get rid of psychometric testing and its more dubious secondary constructs. But eventually, in more and more areas of human endeavour, psychometric testing will become irrelevant. This will depend on how diverse the range of human brain function will become in this rapidly complexifying world.
Labels: Complexity, Psychology, Students, Testing
etched with acid at 5:07:00 am
2 Comments:
i've done the spatial test at least 3 times. once in sec 1, once in sec 2 and then last week
what surprised me was how i thought about the third one in a completely different way
Well, what surprised me was that I tested as 140 standardised in 1979, 135+ in 1982, 165+ in 1990, and 145+ in 2001. I'm pretty sure it would be different again now.
Obviously intelligence is far more variable than we suspect; I am certain that the Multiple Intelligences theory is wrong – it is like assigning seven colours to a spectrum when there in reality a far larger range, plus stuff you can't see without enhancements.
Post a Comment
<< Home