Fluctuations in attention are related to fluid but not crystallized intelligence

Attentional Control and Fluid Intelligence

There are many defensible ways to slice the ability domain. In a previous post, I put fluid intelligence, working memory capacity, and processing speed together in a conceptual grouping called Controlled Attention. I did not do this capriciously but on my review of the available evidence. However, the precise nature of the ways in which these abilities depend on attentional control is still being explored.

In what I consider to be an important paper, Unsworth and McMillan (2014) provide direct evidence that fluid intelligence test performance is related to moment-to-moment fluctuations of one’s attentional state. The paper consists of three experiments designed to tease apart various explanations of the positive correlation between test item performance and self-rated attentional state measured before each item (ranging from 1 = not at all focused on the task to 10 = totally focused on the present task).

Overall findings

1. Test performance was not negatively affected by having to complete attentional state ratings.
2. Self-rated attentional state predicted performance on fluid intelligence test items but not on crystallized test items.
3. Participants with the most variability in self-rated attentional state from item to item performed more poorly on fluid intelligence test items than did people with more stable levels of self-rated attentional state. Thus, attentional control, in accordance with theory, appears to be an important component of fluid intelligence.

One of my suspicions was that is that participants might justify poor perceived performance on a previous item by claiming low levels of attention before the next item. It might be easier on one’s self esteem to claim, “That last item was hard because I am feeling scattered, not because I am not smart.” However, this explanation is undermined by the fact that self-rated attentional state predicted performance on fluid intelligence test items whether the items were in ascending level of difficulty or in random order. Even so, it would have been nice to have seen analyses showing that attentional state predicted performance on the next item more strongly than it “predicts” performance on the previous item.