Research into Visual Perception conducted by Casperson
Following the failure of the early investigators into visual perception
if simple two-dimensional shapes to reach conclusions consistent with
each other, and to explain their findings systematically,
experiments of the same form were still being undertaken twenty years
and Fehrer’s account of the rôle of form in perception.
Casperson’s experiment, conducted in 1950, developed the work
of Helson and Fehrer, Wilcox, and others, by determining the discrimination
thresholds of six different geometric forms (an ellipse, a rectangle,
triangle, diamond, cross and star) and by relating their relative discriminability
to three quantifiable aspects of their construction.
The three aspects were maximum dimension and perimeter
(five figures were used for each area investigated) and, in each case,
seven different areas. 30 different stimulus figures
were used. These included ellipses, rectangles, triangles, diamonds, crosses
and stars (as below):
Casperson’s account and data analysis are particularly thorough
and frequently cited in later research.
Casperson used 20 (male) observers, spanning an age range of 18 to 27
Each observer had seven experimental sessions, in each of which
he judged a complete set of 30 figures, for a particular area, 24 times.
The sessions lasted from one to two hours depending on the size of
the figures and the speed with which the subject responded. As is usual
experiments of this type, Casperson states that “No knowledge of
results was permitted and no learning was apparent other than familiarization
with the procedure”. The probability of recognition (‘percentage
of correct reports’) was calculated using a formula designed to
correct for chance errors according to the data collected, rather than
by assuming that if the observer could not correctly identify a shape,
he responded with any of the five possible incorrect answers with equal
probability. The formula used was:
Rcs = corrected number of right responses for form X;
Tx = total number of form X presented;
Ws = number of form X wrong;
Px = proportion of total errors made on all forms which were
called form X.
Although the functions of ‘probability of correct response’ against
each of the aspects of shape investigated were included in his paper,
Casperson was most interested in identifying ‘that variable which
produces discrimination functions that are alike for all of the figures’.
None of the variables investigated produced identical discrimination
functions for all figures but the question of which variable was
the best was
considered at length. However, Casperson’s results
were specific to a particular shape (or shapes).
Overall, the results suggested that the
Gestalt principle of ‘simplicity' was inadequate for predicting
the relative discriminability of the ellipses, rectangles, triangles,
and stars used in Casperson's Experiment. It
was found that different shapes are associated with different discrimination
thresholds, and that the differences among the observers are small
compared to the differences associated with the shapes.