Infant visual attention in the paired-comparison paradigm: test-retest and attention-performance relations.

The visual behavior of infants in the paired-comparison paradigm was assessed with multiple discrimination tasks week-to-week at 4 and 7 months and longitudinally from 4 to 7 months. Results indicated that although task-to-task reliability was extremely variable and typically low, most measures of infants' attention averaged across multiple tasks were reliable from 1 week to the next as well as relatively stable over the longer longitudinal period. Across all groups, infants who had shorter fixations (i.e., more fixations per fixed-exposure period) during the familiarization phase showed higher novelty preferences. While infants' shift rate during test phases was a reliable individual characteristic at 7 months, it was not at 4 months; rather, data suggested that the difficulty of the stimulus discrimination may be related to young infants' shift rate.