Wednesday, 17 September 2014

Processing speed: the even quicker version, going fast

On 6 May I posted up my notes on a talk given in Edinburgh by Ian Deary on processing speed, and now here is the written up version of that talk by Ian Deary and Stuart Ritchie which gives much better descriptions of the concepts and the findings.

If you click on the third item “10 quick questions about processing speed” you can download the pdf. I should warn you it contains photos of the participants at the conference, but those can be avoided by readers of a nervous disposition.

As you read the paper, reflect on the fact that it may be the last time you read something from Edinburgh while it is still part of the United Kingdom. Thus, it will have historical as well as intellectual significance.


  1. thank you - excellent article - & amen:
    many diverse psychometric tasks get lumped under "processing speed" - e.g., the new WJ-IV combines the subtests "letter-pattern matching" (find the 2 letters/2 letter sets that are the same in each row & put a line through each as quickly as you can) & "# pattern-matching" (same but with #'s:) into "Perceptual Speed" (formerly known as Processing Speed in the WJ-III). the DAS-II's "speed of information processing" (put a line thru the biggest # in each row as fast as you can) is more a quantity speed thing. both involve paper pencil speed to some degree, but not as much as the wechsler. for my $, the wechsler's processing speed tasks are the silliest:)

    some of these tasks would better be termed "rapid 2D symbol scanning with added paper-pencil involvement!"

    a friend & i did a research article based on how a certain processing speed subtest was pretty high in "g" but had slightly less "gap" between certain groups (less group mean difference than expected for a fairly high "g" loading subtest) making it a potentially useful means of finding more "gifted" from the "traditionally lower scoring group" - it was PC pandering & best forgotten. it was my 1 try to fit in with the 90s zeitgeist:)

    some processing speed tasks use a lot of paper-pencil speed, whereas some look at language retrieval speed for specific/exact language (a more relevant thing to assess than paper-pencil speed - since it directly influences reading fluency).

    the wechsler is lucky that silly paper-pencil speed tasks actually correlate with meaningful language retrieval tasks:)

  2. Thanks for your comments. I think that you are, very correctly, in the territory that Pat Rabbit made his own at the conference: just because we label a task in a particular way, we cannot be entirely sure which parts of the brain (and in which order) are employed in the completion of the task. Mind you, with inspection time we ought to have less of a problem, ditto simple reaction time. I think that there is great relevance to a point I had not previously considered in enough detail: digit symbol task differ according to whether you have to write down a newly learned symbol (which has a higher demand on learning and writing unfamiliar forms) or a well known single digit (where you just have to learn to recognise the unfamiliar symbols, and then write down the familiar, over-learned digit, with far less demand on writing). Fun, isn't it, that pencil and paper tasks can reveal so much, and are now of almost exclusively historical interest.