本研究主要的目的在根據Bowlby的依附理論及Horney的理論，探究國小六年級學童兩個社會世界（家庭、同儕）之間的關連性，探討的重點為：(一)學童依附關係的性別差異；(二)不同性別及不同父母依附品質的學童在同儕依附、敵意、社交地位上差異的情形；(三)父親依附、母親依附與同儕依附、敵意之間的關係；(四)三種依附關係與敵意同儕社交地位的預測情形；(五)三種依附關係與敵意對五種社交地位組別的預測情形。
本研究採問卷調查法進行資料的蒐集，並以台北縣、台北市九所國民小學共984名六年級學生為對象，進行研究。使用「父母與同儕依附量表」、「兒童敵意量表」和「社交關係計量表」等三種量表作為研究工具。將所得的研究資料以□考驗、二因子多變量變異數分析、典型相關分析、多元迴歸和區別分析等統計法，進行資料的分析，所得重要的發現如下：
一、國小六年級男女學童在依附關係上有顯著的差異。在父親依附、母親依附上並沒有顯著的性別差異，但在同儕依附上，女童顯著的高於男童。
二、不同性別及不同父母依附品質的學童在同儕依附、敵意及社交地位上有顯著的差異。在同儕依附及社交地位的表現上，高安全父母依附女學童最佳；高安全父母依附男學童及低安全父母依附女學童次之；而低安全父母依附男學童的表現是最差的。在敵意的表現上，高安全父母依附學童的敵意低於低安全父母依附的學童；女學童的敵意低於男學童。
三、國小六年級學童父親依附、母親依附與同儕依附、敵意分別有典型相關的存在。透過典型因素變項，父親依附與母親依附分別可以解釋男學童18.68%、女學童10.12%的同儕依附總變異，除此之外，也可解釋男學童7.7%、女學童9.2%的敵意總變異。
四、三種依附關係與敵意能有效的解釋國小六年級學童社會喜好總變異的21.53%。
五、三種依附關係與敵意能有效的區別五種社交地位組別。根據區別指標所形成的區別函數，預測956位國小六年級學童的社交地位組別，結果顯示：436位學童的落點被正確的預測，準確率達到45.61%。
根據以上的研究的結果加以討論，並提出親職、輔導上的應用與未來研究的建議。

The propose of this study were: (1)to investigate the gender differences in sixth grade students' attachment; (2)to detect the differences of attachment quality and gender on sixth grade students' peer attachment, hostility, and social status; (3)to explore the relationships among father attachment, mother attachment, and peer attachment, hostility; (4)to examined the relationships among social preference, attachment, and hostility; (5)to understand the relationship among sociometric status, attachment, and hostility.
This study was a survey study. The sample consisted of 984 sixth grade students from 9 elementary school in Taipei. Instruments used in this study were: The Inventory of Parent and Peer Attachment, The Chinese Youth Hostility Scale, The Sociometric Assessment. Data obtained in this study were analyzed by □ statistic, two way ANOVA, canonical correlation analysis, stepwise regression analysis, and discriminant analysis.
The main findings were as follows:
1.With regard to attachment relationship, there were significant difference between different gender 6th grade student. Female students were higher than male students on peer attachment relationship.
2.With regard to peer relationship, hostility, and sociometric status, there were differences between high safe parent attachment and low safe attachment, boys and girls.
With regard to the level of peer attachment and sociometric status, high safe parent attachment girls were highest; low safe parent attachment boys' were the lowest; the others' were the middle level.
Girls' hostility level was lower than boys, and high safe parent attachment student' hostility level was lower than low safe parent attachment students',
3.There was canonical correlation relationship between parent attachment and peer attachment. Parent attachment factors could account for 18.68%(boys), 10.12%(girls) from the variance of students' peer attachment.
4.There was canonical correlation relationship between parent attachment and hostility. Parent attachment factors could account for 7.7%(boys), 9.2%(girls) from the variance of students' hostility.
5.Three kinds of attachment and hostility and could predict students' social preference, account for 21.53% from the variance of students' social preference.
6.Three kinds of attachment and hostility could discriminant the five kinds of sociometric status. According to the indexes of discriminant and discriminant functions, 436 of 956 students' sociometric status could be predicted correctly.
Based on the results, implications of this study for parenting and guidance were discussed, and suggestions for future research were proposed.