第一篇：由於肝病在我國是非常流行的慢性疾病，因此我們利用NHIS的資料並使用計量方法來找出就業與肝病之間的關係。其中，我們考量了肝病與飲酒習慣分別對於就業的內生性問題，並且在trivariate probit模型中使用工具變數來得到一致性的估計結果，最後，我們利用WESML (weighted exogeneous sampling maximum likelihood)的估計方式來修正因為選取某特定樣本，所造成估計結果缺乏代表性的問題。實證的結果發現：無論對於男性或者女性，肝病對於就業皆存在一個顯著的負向影響效果。 第二篇：本文將同樣利用國家衛生研究院，民國90年國民健康訪問調查資料(National Health Interview Survey, NHIS)，建立一個兩階段具有樣本選擇(sample selection)的ordered probit模型來進行分析，其中第一階段為樣本的就業選擇問題，而第二階段為樣本在有就業的情況下，肝病將如何影響勞動者的收入，且由於問卷資料特性問題，本研究所取得的收入資料為區間資料，因此以ordered probit模型來進行分析。我們的實證結果發現：肝病對於運輸通訊業、政府服務業與專門服務業的男性薪資收入相對於農業部門的男性薪資有負向的影響，但對於其他行業的男性的薪資而言，則有正向的影響效果；在女性方面，我們則發現肝病對於政府服務業之女性薪資相對於農業部門之女性薪資有負向的影響，在其他行業中肝病則對薪資無顯著的影響。 Essay 1：Because the liver disease in our country has being serious epidemic disease, we analyze the NHIS data of 2001 to find out the relationship between the employment and liver disease. In consideration of endogeneity problems, we use the trivariate probit model which estimate three equations- employment, liver disease and drinking . Finally we used WESML (weighted exogeneous sampling maximum likelihood) method to correct the bias caused by choosing specific sample. Our empirical results indicate that liver disease has a significantly negative effect on both male’s and female’s employment probability.
Essay 2：This article used the NHIS data of 2001 to find out the relationship between the earnings and liver disease. Because the earnings data isn’t a continuous, we used ordered probit model with selection to examine the relationship between earnings and liver disease. Our empirical results indicate that liver disease has a negative effect on men’s earnings in transportation communication, government service and the professional service industry relative to agriculture. In other industries, liver disease has a positive effect on male’s earnings relative to agriculture. Finally, we found that liver disease only has a negative effect on women’s earnings in government relative to agriculture. We also dealt with the selection bias by using Heckman’s two stage method after transforming the earnings data into continuous data.
