| 英文摘要 |
The purpose of this study was to develop a group intelligence test for children in grades 2-3, which could be used as an identification instrument for gifted students. The four major characteristics of the test were a theoretical basis on crystallized and fluid intelligences, a combination of verbal and nonverbal contents, parsimonious test time, and multiple test forms. It consisted of three subtests: Verbal Reasoning, Arithmetic Reasoning, and Figural Reasoning. It was a group test with a 40-minute time limit for each version. Results revealed that the Cronbach α coefficients of subscales were .712~ .835, while for the overall test it was .882. The test-retest reliability coefficients over a period of four weeks were .724~ .885, while for the overall test it was .893. The correlations between the different test versions were .725~ .794. It was found that the total scores of the test were significantly related to the student’s academic performance, the scores of a well-known verbal intelligence test and a nonverbal intelligence test. In addition, a confirmatory factor analysis yielded two factors （crystallized and fluid intelligences） for this test. Generally speaking, the group intelligence test for children（grades 2-3） exhibited acceptable indicator values for reliability and validity. To this end, the normalized t-score norms and percentile ranks for 2135 subjects in the test（Form A） were also developed. Finally, some suggestions were proposed for future research. |
