| E1-H1 | Task-irrelevant perceptual matches cause more positive affect than task-irrelevant perceptual mismatches. |
| E1-H2 | Task-relevant perceptual matches cause more positive affect than task-relevant perceptual mismatches. |
Figure 1
Trial structure.
If participants needed longer than 2000 ms for either prime or probe task, “error too slow” appeared on screen until response. If participants made an error in either prime or probe task, the ITI was replaced with an error feedback message.
Figure 2
Probe RTs in Experiment 1.
Participants categorized probe words after completing color (a) or digit (b) tasks. We measured participants’ affective responses to the prime task via the RT difference between categorizing negative and positive probe words, separately for matching (Δm, for example digits “3” and “3”) and sequential (Δs, for example digits “3” and “4”) prime stimuli. Successfully identifying the color of matching prime stimuli did not elicit more positive affect than of sequential prime stimuli (a; see E1-H1). However, successfully identifying prime stimuli as matches did elicit more positive affect than identifying them as sequences (b; see E1-H2). This interaction effect was strongest in participants who reported that they had been looking for matches rather than sequences when engaged in the digit task (c; see Figure S1a for a more granular depiction). CI95PD indicates 95% confidence intervals of paired probe valence differences, computed separately for each prime stimulus in a and b (Pfister & Janczyk, 2013). CI95M indicates regular 95% confidence intervals.
| E2-H1 | For match seekers, matches cause more positive affect than sequential stimuli. |
| E2-H2 | For sequence seekers, sequential stimuli cause more positive affect than matches. |
Figure 3
Probe RTs in Experiment 2.
Participants categorized probe words after completing digit tasks with a match seeking (a) or sequence seeking (b) instruction. As in Experiment 1, we measured participants’ affective responses to the prime task via the RT difference between categorizing negative and positive probe words, separately for matching (Δm) and sequential (Δs) prime stimuli. For match seekers, successfully identifying prime stimuli as matches elicited more positive affect than identifying them as sequences (a; see E2-H1). For sequence seekers, this interaction pattern reversed but is more ambiguous (b; see E2-H2). For actual sequence seekers who not only were instructed but also reported to have obeyed the instruction to seek sequences, this interaction pattern is less ambiguous (c; see Figures S1bc for more granular depictions). Be aware that two confidence intervals for small subsamples with high variance are cut off at the upper end but that no information is lost because the depicted intervals are symmetrical. CI95PD indicates 95% confidence intervals of paired probe valence differences, computed separately for each prime stimulus in a and b (Pfister & Janczyk, 2013). CI95M indicates regular 95% confidence intervals.