G. H. Fisher, Materials for experimental studies of ambiguous and embedded figures,
Res. Bull. No. 4, 1996, Dept. of Psychol., University of Newcastle upon Tyne.
NOTE: The documentation of this experiment is NOT complete!
Last edited: 14-09-1999
I conducted an experiment with three other people (Robert Morrison, Matthew Cocker and James Castleman) to
investigate if there are effects of prior exposure of one of the aspects before an ambiguous figure was shown on
the probability that both aspects are have equal chances of being seen..
The hypothesis was: The participants exposed to one of the aspects prior to the ambiguous figure
will be more likely to see this aspect in the ambiguous figure.
The participants were students in the 10th year level of a school offering secondary education
in Melbourne, Australia.
There were 83 participants in total. However, 7 protocols had to be discarded because the participants
failed to make a distintion between the two different aspects in the ambiguous figure.
NUMBER OF OCCASIONS UPON WHICH THE FIRST SHOWN ASPECT
OF THE AMBIGUOUS FIGURE WAS SEEN UPON EXPOSURE.
|
|---|
| Image | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 |
| Class A | 17 | 17 | 16 | 15 | 15 | 15 | 14 |
13 | 12 | 9 | 7 | 4 | 3 | 3 | 2 |
| Class B | 23 | 23 | 23 | 23 | 23 | 22 | 21 |
19 | 19 | 17 | 15 | 12 | 5 | 5 | 4 |
| Class C | 20 | 20 | 20 | 20 | 20 | 20 | 20 |
19 | 19 | 17 | 17 | 12 | 7 | 7 | 5 |
| Class D | 16 | 16 | 16 | 16 | 16 | 16 | 16 |
16 | 15 | 13 | 12 | 11 | 8 | 8 | 8 |
| Sum of Classes | 76 | 76 | 75 | 74 | 74 | 73 | 71 |
67 | 65 | 56 | 51 | 39 | 23 | 23 | 19 |
NUMBER OF OCCASIONS UPON WHICH THE SHOWN ASPECT OF
THE AMBIGUOUS FIGURE WAS (IN-)CORRECTLY IDENTIFIED.
|
|---|
| Image | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 |
| correctly identified | 76 | 76 | 75 | 74 | 74 | 73 |
71 | 67 | 11 | 20 | 25 | 37 | 53 | 53 | 57 |
| incorrectly identified | 0 | 0 | 1 | 2 | 2 | 3 | 5 |
9 | 65 | 56 | 51 | 39 | 23 | 23 | 19 |
Mean and Variance for Binomial Distribution
Mean = N * p
Variance = N * p * q
Standard Deviation = (N * p * q) ^ ½
|
where N is the total number of participants
p is the probability to see the first aspect
q is the probability to see the second aspect
|
In this experiment, the values for the variables are:
N = 83, p = 0.5, q = 0.5
Mean = 83 * 0.5 = 41.5
Variance = 83 * 0.5 * 0.5 = 20.75
Standard Deviation = (83 * 0.5 * 0.5) ^ ½ = (20.75) ^ ½ = 4.56
Therefore to encorporate 95.44% of the participants, if they behave according to a Normal Distribution
we substract and add two standard deviations to the calculated mean.
Thus the range of the number of responses lies between 32.38 and 50.62 if the picture were shown
by itself with no prior exposure to any of the aspects.
The results from the tables above show that the reponses indicating the number of occasions the
first-named aspect was seen was 67.
This lies clearly above the range calculated by the methods involved with a Normal Distribution.
Thus we can say that the prior exposure of the images from 1 to 7 has caused the participant
to be more likely to see the first-shown aspect in the ambiguous figure.
You may not reproduce this text or tables in any way unless you have
permission to do so.
Copyright 1999 by Hayko Riemenschneider