Across Experiments 1 and 2, participants categorized picture stimuli while ignoring overlaid distractor words. Some images were more often paired with incongruent distracters (low PC items: high control-demand), while others were more often paired with congruent distracters (high PC items: low control-demand). Some were also paired equally often with incongruent and congruent distracters (PC-50). The receipt of performance feedback was manipulated between groups (A, B). In Experiment 1, we used different stimuli for the practice run (which acclimated participants to the picture-word Stroop task demands), the main task runs, and the transfer run (C). In Experiment 2, four of the stimuli were biased at the item and list level, associated with either a lower or higher likelihood of control-demand, while two PC-50 stimuli were biased only at the list-level (PC-72/28). Participants were exposed to all these images during the practice run (D). Note that all images are from http://cabezalab.org/cabezalabobjects/, but do not match the stimuli used in the Experiments (1: bird, deer, bear, turtle; 2: bird, deer, bear, turtle, cat, and dog).