Developing Intelligence

Cognitive scientists are increasingly aware of how individual differences can confound experimental results. That is, differences in group means cannot always be interpreted clearly if, for example, only some subset of individuals in each group demonstrates the effect. Consequently, even the oldest paradigms in cognitive psychology are undergoing a revival with new mixed experimental/correlational methods.

Consider the Stroop effect, studied in detail since the 1930′s, which is today understood as reflecting not merely the frequencies of color-word incongruity, but also a dynamic constellation of other interacting variables like age, working memory capacity, and reaction time variability. One can safely argue that a complete understanding of the mechanisms generating such effects necessarily involves specifying how these effects interact with individual differences.

Newer discoveries in cognitive psychology are also being considered in the context of individual differences. One of the most exciting examples is the Deese-Roediger-McDermott paradigm, in which subjects are given a list of items to study (for example, nurse, surgeon, medicine, hospital, medic, pharmacist, stethoscope, ambulance, emergency, operation) and later asked to recall those items. The list items are selected to maximize their collective similarity to a critical word which is not presented (e.g., doctor). Unsurprisingly, people will often mistakenly recall this item as though it had been presented.

Why should this be the case? One account, known as Activation Monitoring, posits that neural activity may automatically spread from the presented words to the critical lure, conferring some activation on it. False recall may be more likely when this spreading activation is more likely, and may even be exacerbated when subjects are not effectively engaging control over the retrieval process to differentiate presented from related words.

Accordingly, when subjects are warned that a certain word will be strongly associated with all of those presented, but will not itself be presented, they are able to reduce their incidence of falsely recalling the critical lure word, perhaps through activation monitoring.

Individual differences in false recall on the DRM paradigm have been studied mostly in the elderly and clinical populations, demonstrating that the elderly show lower veridical recall but equivalent or higher false recall and equivalent or elevated false recall levels among both Alzheimer’s patients and schizophrenics.

As pointed out by Watson, Bunting, Poole & Conway, however, relatively little work has focused on the individual difference variables influencing false recall among young adults. (One study found that false recall is elevated among those with self-reported vivid mental imagery.) To fill this gap in the literature, the authors administered a DRM task to young adults with high and low working memory capacity, as measured through an operation span task. Half of each group of subjects were warned that strongly related words would not be presented; all subjects were told not to guess when recalling the lists.

The results showed that those with higher working memory capacity (WMC) correctly recalled more words on average, and those who were warned about the critical lures correctly recalled fewer words on average. In contrast, those with low WMC showed higher rates of false recall on average, as did those who were not warned about the critical lures. This negative correlation between veridical and false recall is consistent with previous work.

Importantly, span differences in false recall were most apparent when subjects had been warned about critical lures – in other words, in the absence of a warning about the critical lures, high and low spans showed similar rates of false recall. When they were warned, however, high spans were able to decrease their probability of false recall, but this warning was not as effectively utilized by those with low span. This result suggests that high spans were more able to maintain this warning and control retrieval from memory appropriately, but that low spans may have lapsed in their maintenance of this goal, or in their ability to use it in controlling retrieval.

A second experiment demonstrated that high spans again correctly recalled more words from each list, particularly when warned about critical lures. In addition, high spans’ performance improved more as the experiment progressed relative to low spans, when the subjects had been warned about the critical lures. A qualitatively reversed trend was observed in terms of false recall, in which false recall decreased as the experiment progressed, and high but not low spans showed decreased false recall when warned about critical lures.

These results demonstrate that veridical recall and false recall seem to show inverse trends in terms of how they are affected by warnings of critical lures (increased & decreased respectively), by individual differences in span (increased & decreased by high & low span respectively), by trends as the experiment progressed (increased & decreased respectively), and by interactions among these variables in young adults.

The authors argue that these results are compatible with Kane & Engle’s findings from the Stroop task indicating that those with low working memory capacity are more likely to succumb to word-reading rather than maintaining the goal of color-naming when the probability of incongruent trials is low (therefore not providing much environmental support for the maintenance of the goal). The authors argue that low spans are unable to maintain and use the warning about critical lures to affect their performance in this task in the same way that they are unable to strongly maintain the color-naming goal in Stroop tasks, when there’s little environmental support for that goal (i.e., a high prevalence of congruent trials).