本文使用來自1995年、2005年及最新公布的2014年的世界價值調查的資料，使用這三波的資料來探討變數對於生活滿意度的影響，從歷史文獻中我們可以得知所得對於生活滿意度具有很大的解釋能力，因此本文將資料分成OECD開所得國家以及非OECD高所得國家藉以觀測所得的差異是否會造成結果的不同，並使用了Ordered Probit Model建立含信任及不含信任兩種模型來討論其差異，以及使用決策樹建立含信任及不含信任兩種模型來討論變數在不含信任時，變數對生活滿意度的影響程度。 Ordered Probit Model實證結果顯示婚姻呈現負向關係，而家庭、健康、信任、所得則呈現正向關係，此外政治大部分為正向關係，最後，教育對生活滿意度的影響則有正有負，由此可看出東西方國家的差異。另外由決策樹的實證結果可以看出在不含信任的情形下，健康對生活滿意度的影響最大;而在去除信任之後，可發現模型與原先模型差異不大，表示信任的解釋能力可能已經含在模型中的其他變數裡。 而在結果可看出在OECD高所得國家中， 家庭、健康、信任、政治、所得大部分對生活滿意度呈現正向關係，婚姻呈現負向關係，而在非OECD高所得國家中與OECD高所得國家的結果一樣，家庭、健康、信任、政治、所得大部分對生活滿意度呈現正向關係，婚姻則呈現負向關係。 This study takes three waves of the World Value Survey data carried out in 1995， 2005， and 2014 to investigate the constituent of life satisfaction. Although literature studies shown that income claims large part of life satisfaction， we further examine the data from individual household and use ordered probit model and decision tree classification method to investigate the difference between OECD (high-income) and non OECD (low-income) countries. In general， the regression evidence appears a positive relationship between life satisfaction and income level， health state， education ， and perceived good governance ，but marriage appears a negative relationship. In addition to the variables above， the individual trust level is also taken into the model. Both of the ordered probit model and decision tree classification method with individual trust level does not improve its explanation power. This may be because individual families with higher levels of incomes had more involvement in communal groups and tend to have higher trust level. The interesting finding of this study is that in the OECD countries， life satisfaction is positively affected by income level， health state， education ， perceived good governance ， and negatively influenced by marriage; while in the non OECD countries， life satisfaction is positively affected by income level， health state， education ， perceived good governance， and negatively influenced by marriage.
