本研究之目的在驗證核保判斷為高風險的要保人是否會導致高理賠？並分析核保政策執行後對損失降低的效果。研究對象為台灣某家產險公司，期間為2009年～2010年，以順序邏輯斯迴歸為研究方法。研究結果發現，16個自變數中有11個變數對損失有顯著影響，7個變數為負向關係。模型的準確率約在79%~91%。研究結果可協助該個案公司建立一個健全的核保制度。

The purpose of this study is to examine whether high risk group can result in the high claims, and the effect of implementing underwriting policy on the reduction of loss. The data on group injury insurance from a Property-Liability Company in Taiwan is analyzed from 2009 to 2010 by using ordered logistic regression. There are 16 independent variables to analyze the relationship between underwriting making factors and claims. The result indicates 11 variables with significant impact and 6 variables with negative impact on the loss. The accuracy rate of the model is 79%~91%. The result of the study can be used to establish a sound underwriting system.