This paper contains two parts. First, we study the decoy effect (especially for the compromise decoy effect) by the experiments where the subjects face the trade-off of inflation and unemployment. As earlier studies show that the compromise decoy is not good as dominated decoy, we try to explore factors of determining compromise effect. Second, we investigate the factors affecting the subjects¡¦ preference over unemployment and inflation. In Part 1, we explore how to enhance the compromise decoy effect by changing the relative location among target, compromise, and decoy. It emerges that the distance between target and decoy, the distance between target and competitor, and the existence of the dominated decoy all affect the size of the compromise decoy effect. In Part 2, we explore the relation of subjects¡¦ preference over inflation and unemployment and their personal characteristics, such as location of hometown, the attitude toward risk, political participation, ideology, household income, knowledge of related terminology, whether to take related course or not, and friends¡¦ and relatives¡¦ unemployment status, and so on. Though the direction of effects is in line with our conjecture, the level of significance is not high enough.