Scenario 5 – Sensitivity to Visual Symmetry
The visual system is extremely sensitive to symmetry. This could be for a number of reasons and one popular hypothesis is that it is a cue for indicating mate quality. As part of your dissertation you are investigating our ability to detect facial symmetry. Face images were digitally manipulated with various degrees of symmetry across the vertical midline, ranging from none to perfect symmetry. In total, you have nine images: one with no symmetry, and eight with increasing levels of symmetry. You conducted a five-alternative forced-choice experiment utilising a single staircase approach. Five images were presented on the screen at once for one second, four with no symmetry and one with (ie one of the faces with some degree of symmetry), and the participants had to locate the more symmetrical face. The proportion of correct identifications are presented below for one participant.
Symmetry Index (1=low symmetry, 8=perfect symmetry) |
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
---|---|---|---|---|---|---|---|---|
% correct identification | 0.23 | 0.28 | 0.37 | 0.49 | 0.75 | 0.92 | 1.0 | 1.0 |
Plot a psychometric function of the results.
What is the guessing rate in this experiment?
Estimate the threshold of symmetry detection.
Some studies have shown that 5AFC studies are too cognitively demanding and that 3AFC approaches are more appropriate for better participant cooperation. What would the threshold have been if this was a 3AFC design?
Can you identify any problems with the experimental approach as described here?