一般而言，再犯者的犯罪情狀及惡質性，均較初犯者為之嚴重，對社會治安也造成更大的危害。因此本研究的主要目的在探討影響受非機構式保護處分少年再犯的因素，研究者根據文獻探討的結果，分別檢驗背景變項、親朋犯罪經驗、少年個人犯罪經驗、少年性格、社會支持以及中立化技術等對再犯的影響，並以台中地方法院觀護人室接受假日生活輔導以及保護管束這二種非機構式保護處分少年為研究對象，共取得1,115份的有效問卷，經描述性分析、相關分析及多元區辨分析等統計方式處理後，茲將研究發現說明如下：
1. 背景變項方面：沒有唸書也沒有就業者、教育程度為國小以下者、居住於繼生家庭者較會再犯，而尚在唸書者、高中（職）及以上教育程度者以及和親生父母共同居住者則明顯較不會再犯，另已婚者也比較不會再犯。
2. 親友犯罪記錄：親戚（叔伯姨姑）有犯罪紀錄者較會再犯，至於父親犯罪紀錄、母親犯罪紀錄以及兄弟姊妹犯罪紀錄之有無則和再犯與否沒有顯著關係；同學或朋友非行情形在再犯者與未再犯者之間亦無顯著差異。
3. 個人犯罪經驗：非初犯者以及初犯年齡愈低、犯罪次數愈多者愈有可能再犯，另觸犯毒品罪者的再犯機率亦比較高。
4. 少年性格：再犯者與未再犯者在法律認同以及求樂衝動性傾兩個分量表上有顯著差異存在，法律認同愈低愈可能再犯，求樂衝動性傾愈高愈可能再犯，至於刺激尋求則再犯者與未再犯者並無顯著差異存在。
5. 社會支持：研究結果顯示社會支持對再犯與否並無顯著影響，但朋友支持分量表與同學或朋友非行情形有顯著正相關存在，即同學或朋友非行情形愈高，朋友支持程度愈高。
6. 本研究之再犯者與未再犯者在中立化技術上並無顯著差異存在。
7. 多元區辨模式：在事前機率為.21的情況之下，本研究包括繼生家庭、婚姻狀況、觸犯毒品罪、初犯年齡以及犯罪次數等變項之精簡模式有顯著的區辨力，預測正確率為78.89％。
根據本研究之研究結果，提出如下之建議：
（一）對司法體系的建議
1. 落實保護處分的執行
2. 特別注意觸犯毒品罪者
3. 落實法律教育、避免執行中的少年再犯
4. 篩選高危險群，落實少年輔導工作
（二）對少年輔導工作實務的建議
1. 重視家庭功能，對結構不健全之家庭提供協助
2. 設立適合受非機構式保護處分少年就讀的中途學校
3. 輔導未升學的少年就業
4. 建立正確的兩性交往態度以及婚姻觀念
5. 落實防毒教育
6. 落實輔導工作，減少同儕不良影響
7. 建立正確休閒觀念
（三）對後續研究的建議
1. 研究對象的增加
2. 研究工具及變項的增加
3. 長期追蹤研究
4. 對特殊個案進行質性研究
Generally speaking, the recidivists are worse than the initial delinquents and have the more serious damages to the society. Therefore, this study is to explore the factors related to the factors related to recidivism of juvenile delinquents who are under community supervision in Taichung District Court.
Based on the literature review, the researcher construct a model which includes the variables of juvenile demographic background, the criminal experience of the relatives and the peers, the juvenile’s offense experience, the juvenile’s personality characteristics, the social support from the family and the peer, and the neutralization techniques. The researcher designed a questionnaire to survey the 1,115 youths who are under community supervision of Probation Office, Taichung District Court.
Statistical procedures used in this study included descriptive statistics, factor analysis, Pearson’s correlation coefficiency, Chi-square, t-test and multiple discriminate analysis. The major findings of this study are shown as follows:
1. The Demographic Background: The youths who had senior high school diploma and above, lived with parents or married would less easily relapse into recidivism than who had no senior high school diploma, not lived with parents or unmarried.
2. Criminal Records of Relatives: There was no significant differences between the recidivists and non-recidivists about the criminal records of their parents and their siblings except their relatives (uncles and aunts).
3. Offense Experience of Subjects: The younger of the initial offense and the more of the offense record, the more probability of the juveniles to be recidivists. Especially, those who do drugs.
4. Personality Characteristics of Juveniles: There are significant differences between recidivists and non-recidivists in law identification and sensation seeking. The lower law-identification, the higher tendency of recidivism; the higher sensation-seeking, the higher tendency of recidivism.
5. Multiple Discriminate Analysis: According to the prior probability of .21, the reduced model including step-family, unmarried, drug abuse, the age of initial offense, and number of delinquent offense can predict 78.89% precision of categorization of recidivists and non-recidivists.
Also, the researcher proposes some suggestions about the practices of justice system, the practices of guidance, and the further research.
資料來源:https://ndltd.ncl.edu.tw/cgi-bin/gs32/gsweb.cgi/ccd=21S_ID/search?s=id=%22087NCNU0210012%22.&searchmode=basic