Abstract: When making choices, people are sensitive to the way in which the problem is presented. This sensitivity was well exemplified by the framing effect, initially described by Tversky and Kahneman (1981). In the well-known “Asian disease problem”, they found that the majority were risk-averse when the options of the problem were framed positively, yet turned to be risk-seeking when the options were framed negatively. Though several studies of framing effects amply demonstrated that changes in the presentation of information would elicit potent effects on decision behavior, they were not as apt to demonstrate that these effects were due exclusively to changes from risk-averse preferences in the positive frame of options to risk-seeking preferences in the negative frame of options. In fact, researchers had found that people’s risk preferences vacillated not only in response to changes in frame but also to changes in other variables such as the perceived desirability of particular social groups, the surface structure of a problem, and the individual differences (Schneider, 1992).
Most of framing studies tended to use scenarios that were similar to the Asian disease problem. However, there were two limitations within this kind of materials. First, the focal problem always began with a threatening cover story, setting the whole problem in a bad situation, and then followed by a pair of options either positively framed or negatively framed. Second, the risky outcomes in these studies were all with a median probability. In order to examine whether the character of the cover story (e.g., good vs bad) and the probability of outcomes would affect people’s risk preferences, the stock market scenarios used in the present research were designed to begin with either a good cover story or a bad one, and to be followed by outcomes with three levels of probabilities (high, median, low) respectively. Therefore, it was a 2(the cover story: good vs. bad)*3(probabilities: high vs. median vs. low)*2(outcome framing: positive vs. negative) mixed design with probabilities as the within-participants variable and the other two as between-participants variables.327 shareholders were recruited from three stock exchanges, and 465 college students were recruited from a university.
Chi-square and logistic regression were performed to analyze the data. The results were as follows: (1) shareholder participants’ risk preference pattern differed from college participants in that shareholder participants remained risk aversion through all the experimental treatments; (2) college participants revealed different risk preferences under different cover stories; (3) under the traditional bad cover story, framing effects were found across the three probability levels, which indicated that risk preference was affected only by the framing of options, but not by the probability; (4) under the good cover story, framing effect was found only on high probability level, but reversed when probability was low, which indicated that risk preference was affected by the framing of options, the levels of probability, and their interaction.

Key words: framing effects, probability levels, characters of the cover story, prospect theory