本研究旨在探討國小教師對於智障者的態度與對於智能障礙的歸因情形，以及態度與歸因兩者之間的關係。研究者以修訂的「對智能障礙者態度量表」及自編的「對智能障礙歸因量表」二項研究工具，採問卷調查的資料收集方式，抽取台北地區國民小學普通班級任教師571人為研究對象。調查所得資料以描述性統計、單因子變異數分析、Pearson積差相關、多元逐步迴歸分析及典型相關分析等統計方法進行處理與分析。茲就本研究主要發現分別說明如下：
一、國小教師對於智障者的態度方面
（一）國小教師對於智障者的態度整體而言尚屬積極；但對於智障兒童的「學校融合」方面則持較保守的態度。
（二）國小教師對於智障者的態度整體而言並不因性別、年齡、教育程度而有所差異，但女性教師對於與智障者社會互動的開放程度較男性教 師積極。而接受特教專業訓練較多、和智障者接觸較多的國小教師， 對於智障者的態度也比較積極。
二、國小教師對於智能障礙的歸因情形
（一）國小教師對於智能障礙主要是傾向於「生物醫學」的歸因。
（二）不同性別、年齡、教育程度、特教專業訓練的國小教師對於智能障礙的歸因型態有顯著差異，對於智能障礙男性教師比女性教師更傾向於「宿命」歸因，而較年輕老、教育程度較高者、接受特教專業訓練較多者則比較傾向於「生物醫學」及「社會環境」歸因；此外，教師們對於智能障礙的歸因型態並不因和智障者不同的接觸程度而有所差異。
三、各背景變項的國小教師對於智障者態度的預測情形
（一）「特教專業訓練」及「接觸程度」能有效預測國小教師對於智障者的態度，然僅可解釋對智障者態度總變異量的5.48%，其中以「特教專業訓練」較具預測力。
（二）在其中的「社會開放」態度方面，以「接觸程度」具有預測力，可解釋1.28%的總變異量；在「學校融合」態度方面，以「特教專業訓練」具有預測力，可解釋6.34%的總變異量；在「社區融合」態度方面，以「接觸程度」及「年齡」具有預測力，此二變項可解釋3.68%的總變異量，且以「接觸程度」較具預測力。
四、各背景變項的國小教師對於智能障礙歸因的預測情形
在「宿命」歸因方面，以「性別」具有預測力，然僅可解釋1.14%的總變異量；在「生物醫學」及「社會環境」歸因方面，「年齡」、「教育程度」以及「特教專業訓練」均具有預測力，然此三雙項對該二分量表僅分別各解釋9.07%、10.07%的總變異量，其中均以「年齡」較具預測力。
五、態度與歸因的關係
國小教師對於智障者態度與對於智能障礙歸因之間有典型相關存在。由態度與歸因間重疊指標的大小判斷，歸因對於態度的解釋範圍較大。進一步分析態度與歸因間的關係組型發現：
（一）國小教師對於智能障礙若較傾向於「宿命」歸因時，其對於智障者的「社會開放」及「社區融合」態度則較為消極。
（二）較傾向將智能障礙歸因於「生物醫學」因素的教師，其對於智障兒童「學校融合」的態度則比較消極。
本研究除根據上述結果進行討論，並綜合結論及提出建議以為教育及未來研究的參考。

The purpose of this study was to investigate teachers' attitudes and attributions toward mental retardation and to understand the relationships between attitudes and attributions. The sample consisted of 571 elementary regular teachers from both Taipei County and Taipei City. Instruments used in this study were the revised "Inventory of the Attitudes Toward the Mentally Retarded" and "Inventory of the Attributions Toward the Mental Retardation". The obtained data were analyzed by descriptive statistics, one-way ANOVA, one-way MANOVA, Pearson correlation, multiple stepwise regression, and canonical correlation analysis. The main findings were as follows:
1. The Attitudes of Teachers Toward the Mentally Retarded
Data analysis in this study indicated the attitudes of teachers toward the mentally retarded were relatively positive, though the "School Integration" attitudes were somewhat negative. There were significant differences for "training in special education" and "contact with the mentally retarded". Teachers with more professional training in special education and more contact experiences with the mentally retarded, possessed more positive attitudes toward the mentally retarded. Besides, the female teachers showed more positive attitudes toward the mentally retarded in "Social Openness" than did the male.
2. The Attributions of Teachers Toward Mental Retardation
The main attributions of teachers toward mental retardation were "Biological-Medical". There were significant differences for "gender", "age", "education degree" and "training in special education". The male teachers possessed more Fatalistic attributions than did the female. The younger, higher educated and teachers with more professional training in special education, the more inclined toward Biological-Medical and Social-Environmental attributions.
3. Using Personal Background to Predict Attitudes
According to the results of regression stepwise analysis, although could just explain 5.48% of the variance, "training in special education" and "contact with the mentally retarded" could significantly predict the attitudes of teachers toward the mentally retarded. The best predictor of teachers' attitudes was "training in special education". In more specific words, "contact with the mentally retarded" was more powerful in predicting "Social Openness", although could just explain 1.28% of the variance. "Training in special education" was more powerful in predicting "School Integration", could just explain 6.34% of the variance. "Contact with the mentally retarded" and "age" were more powerful in predicting "Community Integration", could just explain 3.68% of the variance. The best predictor was "contact with the mentally retarded".
4. Using Personal Background to Predict Attributions
The results indicated that Fatalistic attributions was more significantly predicteq by "gender", although could just explain 1.14% of the variance. Biologieal-Medical and Social-Environmental attributions were both more significantly predicted by "age", "education degree" and "training in special education", could just explain 9.07%, 10.07% of the variance, respectively. The best predictor of both attributions was "age".
5. The Relationships between Attitude and Attributions
The result of canonical correlation analysis revealed that attitudes were significantly associated with attributions by two canonical variates. The first variate suggested that those teachers who exhibited stronger Fatalistic attributions toward mental retardation, were more likely to possess negative attitudes toward the mentally retarded in "Social Openness" and "Compunity Integration". The second variate suggested that those teachers' who exhibited stronger Biological-Medical attributions toward inental retardation, were more likely to possess negative attitudes toward the mentally retarded children in "School Integration". Besides, from the judgment of redundancy between attitudes and attributions, attitudes were explained more by attributions.
Limits of this study and its implications for education and future research were proposed.