본 연구는 손실 영역에서 이론적 연구와 실무적 적용사이에서 불일치를 보였던 프레이밍 효과를 설명할 수 있는 이론적 틀을 제시하였다. Thaler(1985)는 소비자들이 손실 영역에서 통합 프레임을 선호한다고 주장하였으나 Gourville(1998)은 소 비자들이 가격을 지출할 때 분리 프레임을 더 가치 있게 생각한다고 주장함으로써 프레이밍 효과에 대해 상반된 견해를 보였다. 본 연구는 거래 상황에 대한 소비자들의 불확실성 지각이 이러한 불일치를 야기할 수 있다고 가정하고, 2(지각된 불확실성: 높음 vs. 낮음) X 2(선택 옵션: 통합 프레임 vs. 분리 프레임) 실험을 통해 지각된 불확실성이 선택 옵션에 대한 소비자들의 가치 인식에 어떻게 영향을 주는지 살펴보았다. 연구 결과, 거래 상황에 대한 지각된 불확실성이 높을 때는 소비자들이 통합 프레임을 선호했으며 거래 상황에 대한 불확 실성이 낮을 때는 분리 프레임을 더 선호하였다. 이러한 결과는 기존의 연구 결과들이 상호 모순되는 것이 아니라 소비자 들이 지각하는 거래의 불확실성에 따라 프레이밍의 효과가 달리 나타난 것으로 해석할 수 있다. 또한 이러한 결과는 Thaler(1985)가 제시한 쾌락적 편집(hedonic editing) 가설 이외에도 다른 심적 회계(mental accounting) 원칙이 존재할 수 있다는 것을 시사하고 있다.

Thaler(1985) introduces the rule of ‘hedonic editing,’ which posits that consumers maximize their utility based on the value function derived from the prospect theory; multiple loss events should be integrated to minimize consumers’ perceived disutility. He indicates that a consumer prefers to integrate a series of small costs into one large bundled cost(i.e., aggregation). In contrast, the widespread use of a PAD(pennies-a-day) strategy suggests that if the price is presented as multiple small units to consumers(i.e., segregation), their perception of loss is decreased. In Gourville’s(1998) study, the subjects with the $1 a day message showed more favorable responses toward a donation request compared to the subjects with the $350 a year message. According to Gourville, when consumers are asked to evaluate the transaction, they prefer to be anchored on small on-going expenses rather than on the total cost. This assertion for segregating costs contradicts Thaler’s rule of ‘hedonic editing’ indicating that many costs should be integrated into a single expense. With this apparent contradiction between existing theory and practice, we suggest a theoretical framework that can integrate the mixed evidence. We propose that the uncertainty in the choice task adopted in both Thaler and Gourville’s experiments might have been responsible for these differences. Based on this reasoning, it is hypothesized that an aggregated option is preferred in uncertainty, whereas a segregated option is preferred in certainty. Our study consists of a pre-test and two main studies. The pre-test replicated previous studies with the measurement of perceived uncertainty. The results of the pre-test replicated the outcomes of both Thaler and Gourville’s studies and confirmed our conception that the level of expected probability could be regarded as a proxy measure of perceived uncertainty. As the pre-test employed different choice frames and measurements under the assumption that there would be the difference in the consumers’ perceived certainty, the main studies (Study 1 and Study 2) were thus administered to get more reliable outcomes. The main studies explicitly manipulated the two variables: perceived uncertainty and choice frame. Besides, the experimental scenarios were selected from the real episodes regarded as pure loss events. According to the research outcomes, the main studies provided support for our hypotheses proposing that an aggregated option is preferred in the uncertain loss event and a segregated option is preferred in the certain loss event. Especially in Study 1, we checked whether each size of the loss event provided to the subjects in the segregation condition could be perceived as trivial or not. The results showed that the preference toward a segregated option might have little relationship with the relative size of the segregated loss event, which could be regarded as an alternative explanation about our hypothesis. The outcomes of the main studies imply that the difference in Thaler and Gourville’s experiments might not be a contradiction. Rather, one could say that the different decision making adopted different value functions depending upon the perceived uncertainty of the transaction. Since ‘hedonic editing’ under uncertainty assumes a convex curve in the loss domain, the perceived disutility in the same amount of loss could be minimized in the aggregated condition. However, what we call ‘utilitarian editing’ under certainty implicitly explains that the list price of a single transaction may correspond more closely to the evaluation of choice.