obtained independently by two scorers. One was not associated with the research
project, and the other was the main researcher. Pearson product-moment correla-
tion coefficients between the two sets of scores are presented in table 3. The
correlations between scores vary from 0.793 to 0.913 with a median of 0.875.
The results suggest that the scoring procedure is adequately objective.
Validity
Two aspects of the validity were investigated. One is construct-related validity.
Guilford (1950) explicitly stated that the first step in validating creativity tests
should be factorial validity, a form of construct validity, determined by factor
analysis of test scores. When factor analysis with principal components was run
on the data from this test, only one factor was obtained. The solution cannot be
rotated. As shown in table 4, all items loaded sizeably (from 0.685 to 0.840) on to
one factor which explains 63% of the total variance. Kline (1993) suggested items
selected for a unifactorial test should load significantly on only one factor, and in
most cases the load should be larger than 0.3. According to Horn and Cattell
(1966) creativity is a second order ability factor titled a ‘retrieval capacity’ or
‘general fluency’, which loads on ideational fluency, association fluency and irre-
levant associations tests. It is the general retrieval power which accounts for a
variety of skills. In his three dimensional structure of intellect, Guilford (1956)
connected divergent production with creative thinking, where many alternative
ideas need to be brought to light with ease. According to these researchers, scien-
398
W. HU AND P. ADEY
Dostları ilə paylaş: