Working-Memory Not Processing Speed Determines Fluid Intelligence

As I was researching academic studies related to processing speed (with a view to perhaps including processing speed training in the Brain Fitness program) I discovered this fascinating paper:

A latent variable analysis of working memory capacity, short-term memory capacity, processing speed, and general fluid intelligence

(Andrew R. A. Conway, Nelson Cowan, Michael F. Bunting, David J. Therriault and Scott R. B. Minkoff)

Conway set out to see whether working-memory, short term memory, and processing speed could be correlated to fluid intelligence. He found, somewhat surprisingly, that while working-memory capacity has a very strong correlation to fluid intelligence, neither short term memory nor processing speed has a significant correlation.

Conway discusses the significance of this result at some length. He makes some compelling points:

1. His findings strengthen the argument that working-memory can be equated to fluid intelligence. The more items of information we can hold and manipulate, the better we can arrive at intelligent analyses.

2. When testing the effect of processing speed on intelligence it is important to keep the tasks very simple, to avoid any unintended overlap with working-memory. He cites this as the reason that previous studies found a link between processing speed and fluid intelligence.

3. In young children and aging subjects, processing speed may indeed have an impact on fluid intelligence. This would call for processing speed training for the elderly.

I found Conway’s methodology quite thoughtful and sound. Overall, I’m disuaded from designing exercises that train only processing speed, but instead to continue to focus on working-memory and perhaps include an element of processing speed as a way to provide novelty and reward (a faster working-memory task, perhaps).

Tags: , , , , , ,

2 Responses to “Working-Memory Not Processing Speed Determines Fluid Intelligence”

  1. Will W says:

    Interesting you posted this, as just yesterday I was perusing Arthur Jensen’s reply to their research on WM.

    Here is the link to Jensen’s reply:

    Here is the link to Conway’s review of Jensen’s “G factor” that promted the reply from Jensen:

    Since I’m not a psychologist or a specialist in the field at all, just one with amateur interest, I won’t say I agree or disagree with Jensen that fluid G is some combination of processing speed and working memory or whether Conway is right to point to WM as the substrate of G. However, I’m more convinced of the evidence that fluid G is a result of both processes — not one or the other. It’s worth noting chrystallized intelligence obviously is not to be taken lightly, either, and there are likely ‘higher order’ cognitive functions tying into learned knowledge that are extremely important for intellectual functioning beyond working memory and processing speed — (the most likely provenance of Gf).

    I do think Jensen makes a good point in paragraph # 6 on the issue that addresses the questions that I have, especially considering on the studies done on non-working memory loaded inspection time (IT) — which, as a stand alone task, study after study done by Dreary and others, has shown to have minimum correlation of .5 with fluid IQ.

    Quoted below is Jensen excerpted from the link given:

    “6. And what about tasks such as visual and auditory inspection time, which has a remarkably high correlation with IQ, yet seems to make no demands on WM capacity? The big question we are trying to answer is: What is the essential difference between persons with a high level of g and persons with a low level of g? We can show that groups of persons who differ in g differ on this, that, or the other measure of cognitive ability, and we can describe these measuring instruments in remarkable detail, and try to infer their common features as clues to the hypothetical constructs called “cognitive processes” that they presumably call upon. Part of the problem is the incredible ubiquity of g across the varied spectrum of cognitive tasks, whether conventional tests or specially contrived experimental tasks. And measures of MS taken on almost any kind of cognitive task above the level of complexity of simple reaction time shows some degree of g loading. It is the ubiquity of the MS-g correlations in so many different types of cognitive performance that maintains interest in MS as a key variable in this whole puzzle. Can it be dismissed from g theorizing simply because measures of WM are more highly correlated with g than are measures of MS? Even that is doubtful; because when various measures of MS are aggregated, they correlate almost as highly with g as conventional psychometric tests of g correlate with each other (Vernon, 1989). “

  2. martin says:

    Thanks, Will.

    Great stuff. Thanks for posting this.

    Certainly to some degree, processing speed must be involved in g. And I agree that Jensen makes an excellent point in arguing that g can’t be arbitrarily narrowed to working-memory just because this is more highly correlated than processing speed.

    Conway got me thinking about how we can remove working-memory from a processing speed task. It seems perhaps an impossible goal. Every mindful object or thought must pass through working-memory. In this way then, I agree with him that it must be quite difficult to measure processing speed without also involving working-memory.

    All in all, a discussion still in flux.

    Thanks again, Will,

Leave a Reply

You must be logged in to post a comment.