本研究主要目的，在比較電腦輔助教學上之競爭方式，對學習成就和
焦慮的影響，以及探討學生使用軟體的反應，並根據研究發現，提供製作競爭式電腦輔 助教學軟體時的參考。其中競爭方式是指以區域網路連線的人與人競爭，以及單 機上人與電腦模擬的人物競爭等兩種學習方式。
本研究以台北縣漳和國中96名三年級學生為對象，採等組前後測設計。以二階段的隨機分配的方式，首先在三年級的31個班級中，各隨機分派兩班到人與人競爭組、人與電腦競爭組及個別學習組，共6個班級。然後每班再隨機抽取16名學生進行實驗，實驗的課程軟體內容為「二次函數」。實驗處理後，以數學成就測驗、焦慮量表及軟體反應問卷作為研究工具，測量三組受試者的反應情 形。所得資料以共變數分析、t檢定及卡方，進行統計分析。
根據研究結果顯示，在學習成就方面，以區域網路連線的「人與人競爭組」， 其學習成就高於「人與電腦競爭組」及「個別學習組」，而「人與電腦競爭組」 和「個別學習組」的學習成就並無顯著差異；在焦慮方面，利用區域網路連線的「人與人競爭組」，其焦慮程度高於「人與電腦競爭組」；在軟體反應意見方面， 大部份學生最喜歡，也最能引起學習興趣的競爭方式，為以區域網路連線的「人 與人競爭」，而對「人與電腦競爭」和「個別學習」方式的喜好程度，並無太大 的差異。

The purpose of this paper is to compare the effects of learning performance and learning anxiety by the competition methods of computer aid instruction (CAI),furthermore, to understand students' response in using software systems and based on the paper result to provide references when people leverage the competition methods to design CAI. The methods of competition in learning include the competing among users in local area network (LAN) environment, and that between user in stand-along PC environment and simulative computer characters.
The study sample size of this paper is 96 third grade students from Taipei Chung-Ho Junior High School. The sampling model is using two-step's sampling method. First,the program selects two classes for each testing competition methods, including competing with real person, competing with computer, and learning independently,from thirty one classes by random sampling, totally 6 classes. Second, each class random sample 16 students to proceed the experiment, and the content of the experiment was quadratic function. After experiment, this paper compares with the test results from these three groups in the examination of Mathematics performance, anxious scale, and questionnaire of software response. All of the observations were analyzed by ANCOVA, t-test, and Chi-Square test.
The conclusion of this paper indicate three results. First, students in the group of competing with real person reflect greater accomplishment than the groups of competing with computer and learning independently. Also, there is no significant difference between the last two groups. Secondly, the anxious degree from the group of competing with real person is higher than that from the group of competing with computer.