本文研究廣義精神相關疾病對個人就業的影響，論文分為二部分，資料分別取自2005年『國民健康訪問暨藥物濫用調查』中12-64 歲個人問卷(國民健康訪問調查)與18-64 歲成年人自填問卷(藥物濫用調查)。其一分析民國94年『國民健康訪問調查』 (National Health Interview Survey, NHIS)(2005)年的資料，利用Trivariate Probit Model with IVs 模型探討18歲至65歲的男女性罹患精神疾病(包括憂鬱症、躁鬱症、焦慮症等)對個人就業的影響，以飼養寵物與各區域精神疾病盛行率為精神疾病的工具變數，且有別於假設宗教信仰為外生的文獻，嘗試檢定與處理個人宗教信仰與精神疾病的內生性問題，並以各區域宗教信仰盛行率為宗教信仰的工具變數。實證結果顯示：不論男女，罹患相關精神疾病對個人就業有顯著負向影響，邊際效果分別為-48.04%與-55.62%。其二分析民國94 年『藥物濫用調查』的資料，利用Trivariate Probit Model with IVs 模型探討18歲至65歲的男女性重度藥物濫用與重度酗酒對男女性個人就業的影響，以身旁好友是否使用毒品此變數為重度藥物濫用的工具變數，處理重度藥物濫用的內生性問題，並且以各區域重度酗酒盛行率與18歲前酗酒為重度酗酒的工具變數，處理重度酗酒的內生性問題。實證結果顯示：男性重度藥物濫用對個人就業有負向影響，且達統計顯著水準，邊際效果為-66.59%，而重度酗酒對個人就業造成負向影響，但未達統計顯著水準；然而就女性而言，我們並未發現重度藥物濫用與重度酗酒對就業有任何顯著影響。由實證結果我們建議相關當局能加強教育程度並提升相關精神疾病療效，預期能提高勞動市場利益並降低相關醫療支出與社會成本。 This thesis analyzed the impacts of psychiatric-related illness on the individual employment outcome by gender for 18 to 65-year-olds by using the data from 2005 "National Health Interview Survey and Drug Abuse" in the 12-64 year-old personal questionnaire (National Health Interview Survey) and 18-64 years adults self-administered questionnaire (Survey on Drug Abuse). It is divided into two parts. The first part investigated the effects of psychiatric diseases, including major depression, manic depressive disorder and anxiety, on individual employment outcome by gender for 18 to 65-year-olds using data from "National Health Interview Survey" (NHIS). The Trivariate Probit Model with Instrumental variables used measures of the pet and the regional prevalence of psychiatric-related illness as instruments for mental illness in order to deal with the endogeneity problem between psychiatric-related illness and employment, and used the regional prevalence of religion as an instrument for regional variable for the sake of coping with the endogeneity problem between psychiatric-related illness and religion. Empirical results show that psychiatric -related illness is associated with lower rate of employment among males and females. The marginal effects are -48.04% and -55.62%. The second part analyzed what the impacts of psychiatric-related illness on individual employment outcome by gender for 18 to 65-year-olds using data from "Survey on Drug Abuse". The Trivariate Probit Model with Instrumental variables used measures of peer effect for substance abuse variable to deal with the endogeneity problem between substance abuse problem and employment, and used the regional prevalence of heavy drinking and alcohol abuse before the age 18 as instruments for heavy drinking variable for the sake of coping with the endogeneity problem between heavy drinking and employment. Empirical results show that psychiatric-related illness is associated with lower rate of employment among males but not females, the marginal effect of substance abuse on employment for male is -66.59%. Our findings suggest that improving education and expanding access to effective treatment of psychiatric-related illness may have significant labor market benefits.
