Asymmetric item isolation effects: support for a process difference between absolute and relative judgments.

This study used an odd (isolated) item inserted into a homogeneous serial list to investigate process differences between absolute- and relative-order judgments. The serial list consisted of eight names of people ordered in height. These were all male or female names except the fourth name which was of the opposite gender. A name/rank pair accuracy recognition test was used as the absolute judgment, and a comparative judgment (comparing height ranks between two names) as the relative judgment. Reaction time (RT) was the dependent measure for the close-to-ceiling accurate performance. Although the isolated item gained a memory advantage in both the absolute and relative judgments, the magnitudes of the effects differed greatly between the two judgments. For the absolute judgment, the isolated item's RT dropped below the levels of the two end terms, transforming the homogeneous condition's bow-shaped serial-position curve into one with two fully blown bowings. On the other hand, the isolated item caused only a moderate dent on the relative-judgment function with the curve keeping the original overall single-bowing shape. A hypothesis suggesting that absolute judgments are based more on individual-item specific information processing whereas relative judgments more on relational-information processing was proposed to explain the asymmetric isolation effects.