本研究主要目的在探討聽障學生之國語語詞聲調問題、聽障學生與耳聰學生語詞聲調聲學特徵之差異情形、聲調在電腦分析與人耳評分的相關程度。研究樣本為十四名就讀一般學校之聽障學生與來自相同年級、性別、學校之十六名耳聰學生。 研究主要工具包括自編的國語語詞聲調評量表、自行研發之聲調語音訊號分析方法與軟體。主要的研究結果發現： 1.聽障學生之四聲調得分會因聲調不同而有差異，第二聲和第三聲錯誤率最高。 2.聽障學生之第二聲聲調在前時比在後時困難，第三聲詞語聲調，第三聲在後時比在前時困難。 3.聽障學生之聲調得分，當目標聲調固定在前或後的位置，先行或後隨之語詞聲調不同時，目標聲調的得分並無差異存在。 4.在聲學特徵方面，聽障學生與耳聰學生之第一聲與第四聲聲調軌跡大致相似；第二聲與第三聲，在聲調軌跡走勢則有顯著不同。第二聲不論是在第二字詞的前面或後面，均缺乏上揚的趨勢，第三聲降的趨勢很緩慢，與耳聰學生差異特別大的是二聲在前與三聲在後的聲調軌跡圖。 5.人耳聲調評分與電腦分析結果相當一致，人耳評量與電腦分析都發現第二聲與第三聲錯誤率最高。

In this pilot study, Mandarin disyllable words' tone pronunciation of hearing impaired students is investigated. Human listeners' evaluation and computerized tone-contour analysis are compared to find their correlation. The speech sample uttered by 14 mainstreamed hearing impairment students were compared to those of 14 students, with normal hearing, of the same grade, gender and schools. We developed the following tools for this study: 1) A Mandarin disyllable words' tone-shape evaluation scale, 2) Speech-signal tone analysis method and software. The main results of this study are as follows: 1. The scores of four Mandarin tones are significantly different. 2. For the second and third tones of hearing-impaired students, the positions within a disyllable word result in scores of significant difference. 3. When a target tone's position is fixed within disyllabic words of different tone combinations, this target tone's listener evaluation scores do not have significant difference for hearing-impaired students. 4. From acoustic analysis, tone-contours of the first and fourth tones for both hearing-impaired and normal students are same in transitions trend, but for the second and third tones, tone-contours have different trends. The contour of the second tone is slowly falling instead of normally rising contour in both front and back positions. Also, the contour of third tone is more slowly falling than normal falling contour independent of its position. 5. The results of both human listener evaluation and computerized analyzing of tone-contour are very consistent. Both human listeners' evaluations and computerized analyzing of tone-contours show that the second and third tones are the most difficult for haring-impaired students.