Fig. 2

Illustration of the design of Experiment 1. a The baseline condition is an unrotated bar graph labeled either with words or letters. Relative to this baseline, we investigated three different angular rotations (90°, 180°, and 270°) and two different categories of labels (co-rotating vs. reformatted). b We manipulated the labels of the bar graph (words vs. letters). There were six different arrangements of the bars (each repeated three times), resulting in a total of 252 trials. c In one half of the trials (selected randomly), the participants had to indicate the smaller bar in comparison. In the other half, the participants indicated the larger bar. The two bars involved in the comparison were selected randomly. Each participant received a unique set of stimuli