Tuesday 29 April 2014

Processing speed and intelligence

Is processing speed simply a domain of intelligence testing, or is it a fundamental cause of the behaviours we call intelligent?

Do we get different results and insights about processing speed depending upon whether we use digit-symbol substitution tests, reaction time tests, and inspection time tasks?

Is processing speed a general and stable finding across sensory domains and central processes?

What are the neural foundations of processing speed differences?

These pesky questions have been set by Stuart Ritchie, and I am going to hear what 10 expert researchers think about these matters when they meet in Edinburgh tomorrow. However, you can help me by giving me some answers, or some questions you would like asked.

Or you could just count the number of f’s in this posting.

6 comments:

  1. The obvious parallel is to computing. If you look at a computer circa 1970 and a computer circa 2014, they appear to be qualitatively different things. But, it's not that big an exaggeration to say that it's all just speed---really it's all just minimum feature size on microchips, but that more-or-less translates into just speed.

    So, I think the important question is "How would we really know" whether it's just speed or not. If we did not know, because of detailed engineering knowledge of how computers work, that the differences there are pretty much down to just speed, we would undoubtedly have long arguments about multiple intelligences for them. In fact, if you read performance-oriented reviews of computers, they use a bunch of different tests for computer performance, and there are differences among computers of the same generation on various dimensions of performance. But nobody bothers much with testing between generations of computers because, between generations, it's just speed and the later generation always wins.

    ReplyDelete
    Replies
    1. Good points. Nick Mackintosh argued some 20 years ago that speed was the crucial cause of power differences in intelligence, and yesterday's meeting is going to champion that possibility again, with some important provisos.

      Delete
  2. and would the results have been different 100 years ago?

    probably.

    psychologists = really really really stupid people.

    ReplyDelete
  3. If the conclusion is that processing speed is a fundamental cause of the behaviours we/you call intelligent then how do you explain people with ADHD/the newly discovered disorder called SCT (sluggish cognitive tempo, see eg. on youtube: /watch?v=lR_rohjoCG4)
    Back in the days when I went to elementary school, they did a national test on all the 7th class students, my results: I scored in the very high end, with way above average understanding of math (my teacher even held a quick exclusive meeting with me, telling me something about how impressive it was), but on the other hand, I was way bellow average in the time it took to finish the test.
    I have tried to take the Mensa free iq test from iqtest.dk, I scored around 130 sd15, (most often "fail" at test because lack of time), but recently I took a high end untimed spatial IQ test called MACH where I scored 166 sd15, which is kind of confusing to me, it just doesn't really make sense ...... Also on reaction test found at humanbenchmark after doing over 100 clicks/takes I scored only slightly above average..

    ReplyDelete
    Replies
    1. We are dealing with group effects here, from which many individuals will differ. However, some of that difference will be caused by variance in the power and breadth of the various tests you have taken. To get valid results you need to have very representative samples of about 2000 subjects, and good reliability and validity data, which many specialist higher IQ tests do not have. The AH5 (Alice Heim) would be a good test to take, under proper supervised conditions.

      Delete
  4. Persian, perhaps anxiety causes varying results? I know anxiety hinders me quite often, particularly when I'm expected to prove myself in any capacity.

    ReplyDelete