Abstract
The purpose of this study is the standardization of I. P. A. T. Anxiety Scale by Cattell et al. in Japan. In order to accomplish this purpose the following procedures were taken.
1. Internal consistency: Items were translated into Japanese and were given to 255 high school and 340 college students. The response ratio and item-total correlations were computed (Table 1 & 2). Based on these results, two items which showed negative correlations were changed in their expressions, and one item was replaced by another from the same factor items of the Sixteen Personality Factor Scale. As a result, all the items came to have positive correlations with the total score: the mean r for the college students is 0.37.
2. Factor analysis: The test was given to 300 female college students. The tetrachoric correlations among items were computed: (Table 3). The correlation matrix was factor-analyzed by Thurstone's centroid method and six factors were extracted (Table 4). These factors were rotated orthogonally (Table 5). The first five factors were interpreted as corresponding to Cattell's by their content, although they were not very distinct and some items were different from the original ones. The reason for this structure may be that our analysis started from a rather homogeneous group of items from the Anxiety Scale with a few, unevenlysampled group of test items for each factor, while Cattell's first order factors were obtained from among the total pool of four to five thousand items.
The five primary factor scales were based on the factors found, consisting each of eight items. These were given to 300 male college students. The product-moment correlations among these primary factor scale scores were computed and factor-analyzed (Table6). The results showed that they consisted of almost one factor exclusively, which corresponds with the construction of the original scale. From these results we may say that our test has the similar structure as the original one in Japan.
3. The external validity: The test was given to those groups: 167 neurotics, 62 anxiety neurotics and 107 psychotics. The results were compared with that of the normal group. Significant differences were observed' between anxiety neurotics and normals, and between neurotics and normals in both anxiety scores and each factor scores at the 1% level. No significant difference, so far, has been observed between psychotics and normals.
Next, the test was given to some clinical neurotic cases twice or thrice with some intervals. Thechanges of anxiety and factor scores were found corresponding to the clinical observation.
4. Standardization: The test which was constructed based on the above-mentioned factor-analysis was given to 4229 high school students and 1864 college students, men and women, in Japan. A consistent difference beween men and women was found, but no difference among the districts or grades. Means, standard deviations, and distributions of the samples were computed and conversion tables were made for high school and college students, men and women separately (Table 7, 8a b & 9).
