The use of numerical and graphical statistical methods in the analysis of data on learning to see complex random-dot stereograms.

Several numerical and graphical statistical methods are illustrated in an analysis of data from an experiment that investigated a hypothesis of Julesz that giving a person a priori information about the structure of a complex random-dot stereogram reduces the time needed to perceive it when it is viewed. The data are divided into two groups, one consisting of those observers who received no cue or verbal cues (NV) and the other consisting of those who received verbal-visual cues (VV). A quantile-quantile plot shows that the NV times (mean = 7.6) are longer than the VV times (mean =5.6). By using probability plots, it is shown that the perception times have an exponential probability distribution. A hypothesis test based upon this distribution is used to show that the difference between the NV and VV times has significance slightly below 0.05.