On a failure to replicate: Methodologically close, but not close enough. A response to Hogbenet al.

Williams, Brannan and Lartigue (1987) (Clinical Vision Science, 1, 367-371) reported that poor readers took significantly longer to search letter arrays for a target than did good readers. In addition, they reported that blurring the letter arrays leads to faster search times for poor readers and a loss of the significant differences between the groups seen with unblurred displays. In a recent attempt to replicate these findings, Hogben et al. (1996) (Vision Research, 36, 1503-1507) found no differences in search rates between good and poor readers using unblurred arrays, and no differences in search rate between the groups when blurred arrays were used. In the present article, we have compared these two research efforts, and a third paper on the same topic, with regard to methodological factors in an attempt to understand how these two different results could occur. It is our belief that the letter spacing employed in the two studies may account for the difference and should be the focus of future studies of the original effect.