Better to take one long or two quick looks? Independent processing of evidence from discrete visual presentations

Should we prefer one long look to two quick looks of equal overall duration? The authors systematically compared conditions in which a circular letter array was available either for a single look of 2d ms duration (onset asynchrony [SOA] from target to mask) or for two separate looks of d ms each. On the basis of the geometry of the underlying psychometric function relating the SOA to the probability of target identification, separate activation models with peaked hazard functions predict that for short SOAs single long looks would lead to better identification performance and that this advantage decreases (and may turn into a disadvantage) with increasing SOA. These predictions were tested, confirmed, and extended in 5 letter-identification experiments that focused on the mechanism by which information extracted from single arrays is integrated across different looks.