本研究為分析A 市99 與100 學年度國中階段一般智能資賦優異學生鑑定之團體初試與個別複試概況，及初試與複試工具相關情形，並分別探究初試測驗工具與複試各分測驗對鑑定結果之預測情形。採用交叉表描述統計、典型相關與決策樹分析，主要發現如下：(1)99 與100 學年之初試通過率無顯著差異，而99 學年度男生通過初試鑑定的人數超過女生，複試通過率則以99 學年度顯著高於100 學年度，性別未顯現差異；(2)99 學年度初試各分測驗與複試各分測驗未有顯著相關，100 學年度初試3 個分測驗與複試8 個分測驗有兩個典型相關達.00 之顯著水準，其中，初試的數量分測驗與複試的圖形推理分測驗具有正相關，初試的圖形分測驗與複試的圖形統合亦具有正相關；(3)100 學年度初試的數量分測驗對複試鑑定結果最具預測力；(4)99 學年度複試的語彙分測驗與理解分測驗對鑑定結果最具預測力，而100 學年度複試的數學推理與語詞刪異對鑑定結果最具預測力。本研究結果期能助於國中階段一般智能資賦優異學生初試與複試鑑定工具的選用與研發，提高鑑別的效度與鑑定安置後教學輔導的參考。

This study analyzed the A city's database of intellectually gifted junior high school students' identification from 99 to 100 academic year, adopted crosstable analysis, canonical correlation analysis, and decision tree analysis. The main findings were as follows: (a) The yearly differences showed the pass rate of 99 academic year was higher than 100 academic year in the secondary test. The gender differences showed the pass number of male students were more than females' in preliminary test on 99 academic year. (b) There were no relations between the preliminary and secondary subtests on 99 academic year. But there had two canonical correlations between 3 subtests in the preliminary test and 8 subtests in the secondary test on 100 academic year. There was typical relation between quantitative subtest in the preliminary test and graphical reasoning subtest in the secondary test. There was also typical relation between graphics subtest in the preliminary test and graphical integration subtest in the secondary test. (c) The quantitative subtest in the preliminary test had the best predictive power on 100 academic year. (d) The verbal and comprehend subtests in the secondary test had the best predictive powers on 99 academic year. The mathematics reasoning and words and phrases deleted subtests in the secondary test had the best predictive powers on 100 academic year. We expected the results contribute to the select of identification tools, the validity of identification, and nurturing intellectually gifted students.