本研究的主要目的,是分別探討抽菸或喝酒對勞動者薪資的影響。考量抽菸與喝酒者樣本有非隨機的問題,因此採用Heckman兩階段方法來估計抽菸或喝酒對勞動者薪資的影響。本研究以2004年的華人家庭動態資料庫(Panel Study of Family Dynamics)及1995、2000、2002、2005年的台灣社會變遷基本調查(Taiwan Social Change Survey)之資料為研究樣本。 對抽菸和喝酒決策的實證結果顯示,教育程度對抽菸決策的影響為負向顯著,對喝酒的影響則不顯著;潛在工作經驗對抽菸與喝酒的影響呈現正向且遞減的特性;宗教信仰中的道教與天主教信仰者有高度抽菸與喝酒的傾向;居住(出生)於東部的受訪者有高度抽菸與喝酒的傾向。而薪資的實證結果顯示,抽菸工作者的薪資高於非抽菸工作者的薪資,喝酒工作者的薪資亦高於非喝酒工作者,潛在工作經驗對薪資的影響為正向且遞減的特性。由Heckman兩階段估計結果得知樣本選擇問題應是不可忽視的。
The purpose of this thesis is to discuss the effect of smoking or drinking on labor income. The wage equations are estimated by Heckman two-stage method to take into account the sample selectivity bias problem. The resource of the analysis data used in this thesis is from Panel Study of Family Dynamics in 2004 and Taiwan Social Change Survey in 1995、2000、2002、2005. The empirical results show that education has significant negative effect on smoking, and the effect of education on drinking is not significant. Potential work experience has significant positive and decreasing effect on smoking or drinking. Taoist and Catholic have high propensity to smoke and drink. People in the eastern region have high propensity to smoke and drink. Smokers’ income is higher than Non-smokers’ income. Drinkers’ income is higher than Non-drinkers’ income. The findings suggest that sample selectivity problem can’t be ignored.