B.2.5.2. Results

In the auditory task, participants omitted 8 % of the target, in average. False alarms were exceptional. In the visual search task, Reaction Times (RTs) inferior to 200 ms or exceeding 2000 ms were discarded from the analysis, as reflecting anticipatory responses or inattention, respectively. Owing to RT outliers or errors, 3,4 % of the trials were discarded. An analysis of variance was carried out on mean correct RTs, with task (single task, dual task), condition (distractor, no-distractor) and display size (3 items, 6 items) as within-subject factors. The main effect of task was significant (F(1,19) = 22.7; SEM = 95,893; p < .001), the dual task resulting in longer RTs (543 ms) than the single task (494 ms). A unilateral Student t-test showed that this difference was significant when considering only the no-distractor condition (t(19) = 4.4; p < .001). Generally, RTs were longer when an onset distractor was present (532 ms) than not (505 ms; F(1,19) = 23.9 ; SEM = 29,349; p < .001), signalling an attentional capture. The display size effect did not reach significance (size-3: 521 ms; size-6: 516 ms; F(1,19) = 1,2; SEM = 1035; p > .29).

Figure 5: Correct RTs in Experiment 2.1 (left panel) and 2.2 (right panel), as a function of the cognitive load (low in single task vs. high in dual task), the perceptual load (low, 3 items vs. high load, 6 items) and the distractor condition (base vs. onset). Bars represent 1 S.-E.
Figure 5: Correct RTs in Experiment 2.1 (left panel) and 2.2 (right panel), as a function of the cognitive load (low in single task vs. high in dual task), the perceptual load (low, 3 items vs. high load, 6 items) and the distractor condition (base vs. onset). Bars represent 1 S.-E.

The task X condition interaction was significant (F(1,19) = 5.0; SEM = 4,655; p < .037). As a repeated measure analysis was used, this significant interaction reflected the fact that the RT increase due to the salient distractor (i.e. the attentional capture) was more important in the dual task (38 ms) than in the single task (16 ms). These results directly corroborated the hypothesis set out in this article, concerning WM and resistance to interference.

The two one-way interactions involving display size were both significant. First, the condition X display size (F(1,19) = 10.5 ; SEM = 7,142; p < .005) interaction reflected the fact that AC decreased when the display size increased from 3 items (40 ms) to 6 items (14 ms). This followed the perceptual load hypothesis proposed by Lavie (2005).

Second, the task X display size interaction also reached significance (F(1,19) = 5.7; SEM = 5,748; p < .03), suggesting that loading WM impaired more the visual search task when the number of items, and thus, the perceptual load, was low. The two-way interaction was not significant (F(1,19) = 0.0; SEM = 23; p > .83).