Visual attention and memory in professional traders

Abstract

Professional traders need to process a large amount of visual information in their daily activity to judge how risky it is to trade specific investment products. Despite some studies investigating the effects of display clutter on traders, visual attention and memory were never investigated in controlled experimental tasks in this population. Following a preliminary study with 30 participants, visual selective attention and visual working memory were measured and compared between two groups of 15 traders and 15 non-traders (salespeople, acting as a control group) from a large-scale banking group in three experimental tasks measuring selective attention in complex visual contexts, simulating display clutter situations (Visual search), cognitive interference (Stroop task), and a delayed recall visual working memory task. In the Visual search task, traders displayed faster response times (RTs) than non-traders for small display sets, while their performance overlapped for large sets. In the Stroop task, traders showed faster RTs than non-traders but were nevertheless affected by cognitive interference. The memory task highlighted no significant differences between the groups. Therefore, this study found an advantage in traders' attention when processing visual information in small sets with no retention. This result could influence trading activity-determining an immediate use of relevant visual information in decision making-and traders' display layout organization.

Figure 1

Visual search results. This figure represents the main results from the Visual search task. In both panels, the x-axis represents the set size (i.e., the number of figures in the display set, ranging from 1 to 15), while the y-axis represents participants’ log-transformed response times (RTs), predicted regression lines are represented and the shaded area represents 95% confidence interval. Panel A represents the group * set size effect: traders showed lower RTs than non-traders in small display sets, their regression line increased with a significantly steeper slope, and RTs from the two groups overlapped in large display sets. Panel B represents the condition *Set size effect: RTs increased with a smaller (yet significant) slope in the feature-colour condition than in the feature-shape and conjunction conditions.

Figure 2

Stroop task results. This figure shows the boxplot of participants’ log-transformed response times (RTs, on the y-axis), in two separate groups (traders vs. non-traders). The x-axis represents the Stroop congruity condition, i.e., C congruent, IC incongruent, N neutral. The boxplot shows overall faster RTs for traders compared to non-traders. The cognitive interference effect (i.e., slower RTs for incongruent compared to congruent and neutral conditions) is present in both groups.

Figure 3

Experimental procedure. This figure depicts the experimental procedure of this study. In the first phase (learning phase), participants were asked to memorize the items in the picture for one minute. Then, in a counterbalanced order across participants, they performed the Visual search task and the Stoop task. Finally, they were asked to write down all the items they remembered from the learning phase, in any order (delayed recall visual working memory task). The typical duration of the whole procedure was 20 to 30 min.

Figure 4

Visual working memory task. This figure reproduces the items display shown to participants for one minute during the learning phase. This figure portrays 35 items of most frequent use in daily life or common animals.