Horrible at N-Back
I read the Jaeggi paper, and if my inferences are correct, people in her study with higher initial Gf scores didn't have better initial (or final) working memory scores:
"There was an effect of training irrespective of initial Gf. Further, the gain in Gf was not dependent on initial working memory capacity."
Can you infer from that there isn't a strong correlation between high initial Gf and high initial working memory?
Also of interest is that the training has a bigger impact on Gf for people with lower initial Gf scores.
Anyway, cool study, cool game, and great website - thanks for maintaining it!
"There was an effect of training irrespective of initial Gf. Further, the gain in Gf was not dependent on initial working memory capacity."
Can you infer from that there isn't a strong correlation between high initial Gf and high initial working memory?
Also of interest is that the training has a bigger impact on Gf for people with lower initial Gf scores.
Anyway, cool study, cool game, and great website - thanks for maintaining it!
? | 3 years ago
Reply
""There was an effect of training irrespective of initial Gf. Further, the gain in Gf was not dependent on initial working memory capacity."
Can you infer from that there isn't a strong correlation between high initial Gf and high initial working memory?"
The quoted section doesn't address correlation between initial Gf and WM capacity. What it does say is that even if you know someone's Gf and WM capacity, you cannot tell how much of an improvement in measured Gf to expect from the training paradigm. The good news is that all trainees experienced improvements.
cognitivefun | 3 years ago
Reply
Horrible at N-Back
I'm awful at n-back tests. I'm struggling with the 2-back test. Does this mean I have a low IQ? I've always been bad at arithmetic.
Also, is there a version of the test that uses English letters as the audio stimulus? 3 or 4 of the sounds end in a "long e" and are difficult to distinguish, at least for my ear.
Also, is there a version of the test that uses English letters as the audio stimulus? 3 or 4 of the sounds end in a "long e" and are difficult to distinguish, at least for my ear.
fakingit | 3 years ago
Reply
Hi, I was horrible with the n-back test too, at first even the 1-back test was difficult enough and 2-back impossible. I think doing bad on the n-back means that your working memory is not so good. With lots of training I can now easily do the 2-back, and I am sure that with enough training 3-back and more is easily possible. You just have to push yourself hard enough to train until you get good. It takes some time, but if I can do it you can learn it too :)
I am not sure how this n-back test is related to intelligence, I work as a researcher in software development and according to my colleagues am quite good at it, so I am smart enough for the job. Now that I am training this n-back I am curious if I can do a better job, we'll see.
I am not sure how this n-back test is related to intelligence, I work as a researcher in software development and according to my colleagues am quite good at it, so I am smart enough for the job. Now that I am training this n-back I am curious if I can do a better job, we'll see.
martinus | 3 years ago
Reply
"Also, is there a version of the test that uses English letters as the audio stimulus? 3 or 4 of the sounds end in a "long e" and are difficult to distinguish, at least for my ear."
Have you seen the multimodal n-back test? You can choose between different kinds of auditory stimuli (like digits).
cognitivefun | 3 years ago
Reply
" Does this mean I have a low IQ?"
Don't get too hasty with your conclusions. Also, if your IQ is measured by the Raven's Progressive Matrix, then you can probably improve it (which is a reason why you shouldn't get caught up in the idea of "IQ").
cognitivefun | 3 years ago
Reply
Tricksy J-B tried to sneak that one by.
Is there anything research that you know of suggesting the more IQ tests you take, the better you perform for general tests? I'd be surprised if this wasn't true.
Is there anything research that you know of suggesting the more IQ tests you take, the better you perform for general tests? I'd be surprised if this wasn't true.
Whoopska | 3 years ago
Reply
Ya, but that's just "teaching to the test" which Jensen critisized many times in his famous book, the "G" Factor. Dual n back training is more about transfer of attentional processes from one GF loaded test to another.
? | 3 years ago
Reply
Yup: Lawson & Kirby, 1981 taught sixth graders strategies for the Raven and improved their scores. I'm sure there are others.
cognitivefun | 3 years ago
Reply
Ya, but that's just "teaching to the test" which Jensen critisized many times in his famous book, the "G" Factor. Dual n back training is more about transfer of attentional processes from one GF loaded test to another.
? | 3 years ago
Reply
Sure, but the very fact that you can "teach to the test" for IQ measurements means that people should be cautious in using it as a generalized indicator of intelligence or cognitive capacity. This is primarily in response to fakingit's concern.
cognitivefun | 3 years ago
Reply
Excellent counter. (I am not the ? above.)
IQ testing is simply that: testing. We need to keep this in mind at all times. A score is not an intrinsic quality to a person (contrary to self-serving, popular belief).
IQ testing is simply that: testing. We need to keep this in mind at all times. A score is not an intrinsic quality to a person (contrary to self-serving, popular belief).
? | 3 years ago
Reply
It's not really an excellent counter at all. Teaching to the test actually serves as a control measure that proves the validity of iq tests. Here's how:
You take a group of kids and teach them techniques to solve problems that are similiar to the ones they will take on a given iq test, eg. the Ravens. Then low and behold, they score well on the Ravens since the items are similiar to what they've been coached on. But then, when you present the same kids with a test that is just slightly different, their scores drop back to pre training levels, proving that the improvement isn't transfering as a generalized measure of "G" and is simply a reflection of having been coached on the particular item types encountered.
Another group of kids with higher initial ability will show the exact same rise and fall but the scores across the board will be higher and they will always score higher on a wide range of tests since the tests all have as their common element a signifigant "g" loading. The tests scores all correlate due to the "g" loading they contain.
The interesting thing about dual n back training is that you are supposedly, at least according to Jaeggi's initial research, training a generalized mental acuity that should transfer across the board for all "g" loaded tests and improve scores due to any number of improved cognitive functions-greater WM capacity, improved attentional focus and other frontal and cortico striatal increases whose improvement all play a role in boosting scores on intelligence measures.
You take a group of kids and teach them techniques to solve problems that are similiar to the ones they will take on a given iq test, eg. the Ravens. Then low and behold, they score well on the Ravens since the items are similiar to what they've been coached on. But then, when you present the same kids with a test that is just slightly different, their scores drop back to pre training levels, proving that the improvement isn't transfering as a generalized measure of "G" and is simply a reflection of having been coached on the particular item types encountered.
Another group of kids with higher initial ability will show the exact same rise and fall but the scores across the board will be higher and they will always score higher on a wide range of tests since the tests all have as their common element a signifigant "g" loading. The tests scores all correlate due to the "g" loading they contain.
The interesting thing about dual n back training is that you are supposedly, at least according to Jaeggi's initial research, training a generalized mental acuity that should transfer across the board for all "g" loaded tests and improve scores due to any number of improved cognitive functions-greater WM capacity, improved attentional focus and other frontal and cortico striatal increases whose improvement all play a role in boosting scores on intelligence measures.
? | 3 years ago
Reply
It wasn't really a counter. That improved cognitive ability (whatever that entails) leads to generalized improvements across different testing metrics is not under dispute. Your points are valid enough.
The point, particularly in reference to fakingit's original post, is that performance IQ tests should not be looked upon as an indication of general intelligence. Higher general intelligence implies higher IQ performance, but it is not correct to assume the reverse is true. Furthermore, and much more importantly, it is unproductive for one to be overly concerned about their IQ score.
Just imagine what would have happened if Richard Feynman saw his score, and was discouraged by it!
The point, particularly in reference to fakingit's original post, is that performance IQ tests should not be looked upon as an indication of general intelligence. Higher general intelligence implies higher IQ performance, but it is not correct to assume the reverse is true. Furthermore, and much more importantly, it is unproductive for one to be overly concerned about their IQ score.
Just imagine what would have happened if Richard Feynman saw his score, and was discouraged by it!
cognitivefun | 3 years ago
Reply
Super-high-IQ Paul Cooijmans says that the second attempt at one of his upper-range tests tends to be better than the first, but ever decreasing returns soon sets in. It was somewhere on this site: http://www.paulcooijmans.com/. I cannae find it now.
cevapcici | 3 years ago
Reply
