Information Technology Reference
In-Depth Information
From the point of view of an AJS validation, they constitute a good reference,
since they are relatively easy to apply and provide reproducible and quantifiable re-
sults. They also allow the comparison of the “performance” of the computer system
with human evaluation, although this comparison is extremely delicate.
We will make a short analysis of two tests that are potentially useful for AJS
validation, namely the
Visual Aesthetic Sensitivity Test
of Götz et al. and Maitland
Graves'
Design Judgment Test
. Nadal (
2007
) provides further analysis of these and
other psychological tests.
The Visual Aesthetic Sensitivity Test (VAST)—created by Götz (an artist) and
Eysenck (Eysenck et al.
1984
,Götz
1985
, Eysenck
1983
)—consists of a series of
50 pairs of non-representative drawings. In each pair the subject has to express an
opinion as to which is the most harmonious design. Götz drew the “harmonious” de-
signs first and then altered them by incorporating changes that he considered faults
and errors according to his aesthetic views. The validity of the judgement was tested
by eight expert judges (artists and critics), making preference judgements and only
accepting pairs of designs on which agreement among judges was unanimous. When
groups of subjects are tested, the majority judgement agrees with the keying of the
items, which supports the validity of the original judgement.
There are easy, middle and difficult item levels. The difficulty level of items is
established in terms of the percentage of correct responses; the more subjects give
the right answer, the easier the item. Different groups of subjects, differing in age,
sex, artistic training, cultural background, and ethnicity have produced very sim-
ilar difficulty levels for the items. “The instructions of the test did not emphasise
so much the individual's preference for one item or the other, but rather the qual-
ity of one design” (Eysenck
1983
). The task is to discover which of the designs is
the most harmonious and not which designs are the most pleasant. The images re-
semble abstract art, minimising the influence of content on preference. There was
some cross-cultural comparison employing the VAST test. Iwawaki et al. (
1979
)
compared Japanese and English children and students. Frois and Eysenck (
1995
)
applied the test to Portuguese children and Fine Arts Students.
Graves (
1946
) presented “The Design Judgment Test” (DJT).
1
It was designed to
determine how humans respond to several principles of aesthetic order, presented in
his previous work (Graves
1951
). It contains 90 slides with pairs or triads of images.
In each of the slides, one particular image “is considered 'right' (and scored accord-
ingly) on the basis of agreement with the author's theories and the agreement of
art teachers on the superiority of that particular design” (Eysenck and Castle
1971
).
Thus, on each slide, one of the images follows the aesthetic principles described by
Graves, while the others violate, at least, one of these principles. Each slide is shown
for approximately 45-60 seconds to the subject, who chooses one image per slide.
The score of the test corresponds to the number of correct choices. All slides are in
black, white and green. All images are abstract. The images of each slide are simi-
lar in style and in terms of the elements present. The average percentage of correct
1
Photos
of
DJT
can
be
found
at:
