Task switching among two or four tasks: effects of a short-term variation of the number of candidate tasks

The two-task control condition differed from the condition in which the set of four tasks was reduced to two candidate tasks with respect to the fact that in the former condition, only combinations of one numerical and one letter task were allowed, whereas there was no such restriction in the latter condition. We checked whether this divergence resulted in different proportions of response-(in)congruent trials, which was not the case (50.7 vs. 50.5 % incongruent trials, χ ²(1) = 0.04, p > 0.8). We also checked whether participants’ performance was affected by whether the pre-cue indicated two tasks from the same domain (two-letter or two-digit tasks), compared to conditions with one task from each domain. There were no significant differences either in RTs (904 vs. 908 ms, p > 0.6) or ERs (4.9 vs. 5.6 %, p > 0.3), nor was there an interaction with Task Transition [F(1, 15) = 1.59, p > 0.22, MSe = 3,080 for RTs and F(1, 15) = 0.14, p > 0.71, MSe = 0.00122 for ERs).

It should be noted that the between-block variation of the number of tasks implemented by the control condition is confounded by differences with respect to the frequency and recency of the performance of individual tasks. This point is taken up in “General discussion”.