Stereograms

Research conducted by: Frisby, J. P. and Clatworthy, J.L.

Case study prepared by: Emily Zitek from DASL story contributed by Michael Friendly

Overview

The rectangles below appear to be composed of random dots. However, if the images are viewed with a stereo viewer, the separate images will fuse and reveal an embedded 3D figure. In this example, fusing the images of these random dot stereograms will reveal a diamond. (Another way for you to fuse the images is to fixate on a point in between them and defocus your eyes. This technique takes practice, but you can try it out with the links below.)

This experiment seeked to determine whether giving someone information about the embedded image can help speed up how long it takes to view this image. Seventy-eight participants were given no information, verbal information, and/or visual information (a drawing of the object) about what the embedded image should look like before attempting to fuse the images and actually view the 3D design.



Questions to Answer
Does giving someone information about an embedded image in a stereogram affect the amount of time it takes to see this image? More specifically, does the amount of time it takes to fuse the image in a stereogram differ when the person is given both verbal and visual information about what the image should look like as opposed to when the person is only given verbal information or no information at all?

Design Issues


Descriptions of Variables
Variable Description
Time

Time to produce a fused image of the random dot stereogram

Group Treatment group divided by type of information received:
1 = no information or only verbal information
2 = both verbal and visual information


References

Cleveland, W. S. (1993). Visualizing Data. Original source: Frisby, J. P. and Clatworthy, J.L., "Learning to see complex random-dot steregrams," Perception, 4, (1975), pp. 173-178.


Links
Exercises
  1. What is the independent variable? What is the dependent variable? (1.6)
  2. Create box plots comparing the time it takes to fuse the image for the NV group to the time it takes for the VV group. (2.7)
  3. Draw histograms of the fusion time for each group. (2.5)
  4. What do you notice about the distributions? Do they appear to be positively or negatively skewed? (1.11)
  5. Compute the mean, variance, and standard deviation of the time for each group. Do the variances appear to be equal? (3.6, 3.10)
  6. Perform an independent samples t-test (assuming equal variance) to compare the mean fusion time for each group. Is this p value significant at the .05 level? (10.2)
  7. Take the log transformation of the fusion time. Then draw box plots for each group. (2.7) How did the log transformation affect the distributions? (1.11)
  8. Conduct a t-test on the log data. What does this say now about the difference between groups? (10.2)