Numerous researchers have investigated the compromise effect, according to which a middle option of a consideration set is assumed to be perceived more attractive by consumers, thus becoming more likely to be chosen than the extreme options. However, a closer examination of the experimental designs that were used in previous studies on compromise effects clearly reveals a lack of realism in terms of forced choices between fictitious options in hypothetical choice settings of student samples. This article reports two consecutively implemented studies demonstrating that the compromise effect is robust even in an enhanced design that incorporates basic conditions of real purchase decisions in laboratory-based experiments. Specifically, the relative share of the middle option increases significantly in an overall analysis when experienced consumers make unforced decisions between real brands in a binding choice context. However, segmented analysis indicates substantial differences, meaning that (1) the compromise effect is strong and significant among quality-seeking consumers, whereas (2) the compromise effect is weak and insignificant among price-conscious subjects.