一諾牽禁：法定毒品施用者自我污名、復元資本與維持清醒承諾關係之研究

本研究旨在探討可預測毒品施用者維持清醒承諾之因素，以作發展社工處遇之用。研究採量化研究，運用立意抽樣，併行實體與網路問卷資料蒐集。以380份有效樣本進行T檢定、皮爾森積差相關、階層迴歸分析後，可知：（1）自我污名與維持清醒承諾為正相關；（2）正向復元資本與維持清醒承諾為正相關；（3）正向復元資本對自我污名與維持清醒承諾可發揮調節效果。研究結論可知，自我污名與正向復元資本皆對維持清醒承諾的預測力達統計顯著，正向復元對維持清醒承諾之預測力不僅優於自我污名，且對自我污名與維持清醒承諾的關係具調節效果。控制變項中的年齡及監禁經驗對維持清醒承諾亦有顯著預測力。研究建議未來應更謹慎探討毒品施用者年齡、監禁經驗、自我污名與維持清醒承諾等變項間的關係，瞭解優化監禁制度與更生人服務，將有助於臺灣成年毒品施用者累積正向復元資本，以穩定維持清醒承諾之復元狀態。

關鍵字

維持清醒承諾；自我污名；正向復元資本；毒品；施用者

並列摘要

Enhancing commitment to sobriety among drug users can lead to longer and healthier recovery. This study aims to explore factors that predict the commitment to sobriety among drug users. According to Recovery Capital Theory, we examined the associations among self-stigma, positive recovery capital, and commitment to sobriety. A questionnaire survey and a web-based survey were used simultaneously. SPSS was employed to analyze 380 questionnaires through independent t-tests, Pearson correlations, and regression analyses. The results indicated that: (1) commitment to sobriety had a positive correlation to self-stigma; (2) commitment to sobriety had a positive correlation to recovery capital; and (3) positive recovery capital significantly moderated stigma and sobriety. These findings suggest that self-stigma and positive recovery capital are significant predictors of commitment to sobriety. Additionally, positive recovery capital appears to have a stronger impact than self-stigma and can significantly moderate the relationship between stigma and sobriety. Demographic factors such as age and incarceration experience also significantly predict commitment to sobriety. Future local research should aim to carefully interpret the relationship between self-stigma and commitment to sobriety among drug users. Optimizing community services within the criminal justice system and broader social reintegration programs to support ex-offenders are important components of positive recovery capital and would prove beneficial for drug users seeking to increase their commitment to sobriety in Taiwan.