Abstract
How do humans evaluate temporally accumulated evidence and arrive at a decision despite the presence of conflicting evidence? In the present study, we showed human participants a sequential presentation of shape exemplars drawn from two novel shape categories (A and B). All shapes were shown at fixation and the presentation order of the exemplars from the two shape categories was randomly intermixed. Participants had to decide whether a given presentation contained more exemplars from A or B. The ratio of exemplars from the two shape categories was either 1:2 or 2:3. Within a given ratio, the total number of shapes shown also varied. For example, for ratio 1:2, a sequence could contain 2 As and 4 Bs, 4 As and 8 Bs, or 6 As and 12 Bs. This allowed us to examine how decision making would be influenced by the ratio and the total number of shapes present. We found that both factors significantly impacted performance. At the same ratio, decision performance was better (i.e., shorter reaction time and higher accuracy) when the difference in the number of exemplars shown was greater between the two shape categories. When the difference in number was equated, greater difference in ratio produced better performance. Together, these results suggest that humans use both the absolute amount of difference and the ratio between two conflicting sets when evaluating temporally accumulated evidence.