The framing effect refers to the phenomenon that, when the same problem is presented with different representations of information (frames), people make significant changes in their decision making, or even reverse their decision (Levin et al., 2002; Peng et al., 2013). One example is decision making in the context of the Asian disease problem (Tversky & Kahneman, 1981): Participants are presented with two options when 600 people face impending death; Option A is to save 200 people (400 will die), and Option B is a 1/3 probability that all 600 people will be saved (no one will die) but a 2/3 probability that no one will be saved (all will die). When this problem was described in a way that highlighted people being saved (positive frame), Tversky and Kahneman (1981) found that 71% of participants chose the certain option (Option A), whereas when it was described in a way that highlighted those who would die (negative frame), only 22% of participants chose Option A.

Many studies have found that the framing effect is stable across medical care, management, economics, and various other fields (Cao et al., 2017; Peng, Feng, et al., 2019; Stark et al., 2017). However, in existing research, frames have generally been passively offered rather than being actively created by decision makers. X. Wang (2004) innovatively characterized the options in the Asian disease problem in the form of ambiguous pie charts, asked participants to form a self-frame according to their own apprehension, and found that the effects of self-frames were consistent with those of passively offered frames. Participants tended to be conservative in the positive self-framing conditions, and tended to be risk seeking in the negative self-framing conditions. McElroy et al. (2007) suggested that participants’ apprehension of decision-making information is significantly affected by personality differences, which will lead to forming different self-frames. Further, Peng et al. (2014) found that trait differences in the level of anxiety resulted in differences in the editing of decision-making information, that is, high-anxiety participants tended to create more negative self-frames than did low-anxiety participants. Yu et al. (2015) found that the self-framing effect existed only when decisions were made immediately after self-frames were formed.

Optimism is a stable and positive personality trait related to emotion, cognition, attitudes, and behaviors, and optimistic individuals have positive expectations of the future (Jose et al., 2018; Y. Wang & Peng, 2017). The relationship between optimism and risky decision making has attracted the attention of researchers, and many have documented that optimistic individuals tend to be risk seeking, whereas pessimistic individuals tend to be risk averse (Luo & Isaacowitz, 2007; Xia et al., 2018). For instance, Luo and Isaacowitz (2007) found that people low in optimism paid more attention to threatening information during decision making as compared with those high in optimism. Gibson and Sanbonmatsu (2004) found that optimists held positive expectations for the future, and were, therefore, willing to make bets even in the case of failure, which is significantly different from pessimists. Further, Xia et al. (2018) found that optimists tended to make risky decisions in subsequent gambling behavior following near wins, whereas pessimists were more cautious.

On the basis of previous research results, it can be concluded that (a) self-frames are the active and subjective encoding of decision-making information and affect risky decision making in a similar way to passively offered frames, and (b) personality traits are a key factor influencing the formation of self-frames (Peng et al., 2014; Yu et al., 2015). Because optimists hold stable positive expectations of the future (Xia et al., 2018), we proposed the following hypothesis:
Hypothesis 1: Compared with low-optimism individuals, high-optimism individuals will tend to use more active words to describe decision-making information in a self-framing task, and will create more positive self-frames.

According to previous studies, people tend toward risk aversion in positive self-framing conditions, and tend toward risk seeking in negative self-framing conditions (X. Wang, 2004). Therefore, if optimists create more positive frames, they should tend to avoid risks, which is contradictory to most previous findings that optimists are more risk seeking than are pessimists. Thus, we proposed the following hypothesis:
Hypothesis 2: There will be two effect paths of optimism on risky decision making: a direct path and an indirect path (through self-framing), and the two effects will be opposite and competing.

The research model is shown in Figure 1.

Method

Participants and Procedure

The participants in this study were 416 undergraduate students at a general university in western China, aged between 17 and 22 years (M = 19.55, SD = 1.36), of whom 264 were men and 152 were women. Participants were all of Han ethnicity.

We sent out 416 copies of the survey, and all were returned. Of these, 32 were removed because the participants failed to understand the pie charts (e.g., they misunderstood the dotted lines standing for “die” as standing for “recover” or “rescue”). Ultimately, we obtained 384 valid copies for analysis. Participants were informed about the purpose of the research and provided their written informed consent before completing the measures in a classroom environment. Course credit was given as compensation for participation in this study.

Measures

The Life Orientation Test
Levels of optimism were measured using the Life Orientation Test, which was developed by Scheier et al. (1994). It consists of six items, such as “I usually expect the best outcome during uncertain times” and “Something will go wrong if it is meant to go wrong.” The items are evaluated on a 5-point Likert scale and a total score is obtained by summing scores on the items, three of which are reverse-coded. Li et al. (2013) translated the Life Orientation Test into Chinese, and found it has excellent reliability and validity in the Chinese cultural context. In our study we selected the median value for grouping, using the method employed by McElroy et al. (2007). Thus, individuals with scores above the median value were grouped into the high-optimism group, and those with scores below the median value were grouped into the low-optimism group.

Self-Framing Test
We adopted the self-framing test developed by X. Wang (2004), in which the Asian disease problem is represented by pie charts (see Figure 2), and the participants are asked to fill in the blanks of the following statement and make self-frames: “600 people have been infected by a deadly disease and face impending death. Now two options are available. If Option A is adopted, ___ people ___; If Option B is adopted, there is a 1/3 chance that ___ people ___, and a 2/3 chance that ___ people ___.”

Self-Framing Rating
Participants could use different words to describe the decision-making problem, for example, “If Option A is adopted, 200 people will be saved; If Option B is adopted, there is a 1/3 chance that all people will survive and a 2/3 chance that all people will die.” Our procedure was carried out according to the method used by X. Wang (2004), so that we collected all verbs and verb phrases used by the participants to describe the self-frames. These verbs were then randomly arranged, and 20 undergraduates who did not participate in the self-framing test were asked to evaluate the hedonic tone corresponding to each verb according to their own apprehension (−3 = very unhappy, 3 = very happy). The mean score of the 20 participants for each verb was determined and used as the final score for this verb. For instance, the mean score of “die” was −2.05. The scores for all verbs used in a frame were then added and regarded as the final hedonic tone of the frame. For instance, a participant who described the problem as “If Option A is adopted, 200 people will be cured; If Option B is adopted, there is a 1/3 chance that all people will be saved and a 2/3 chance that all people will die.” The scores of “cure,” “save,” and “die” were 2.17, 1.96, and −2.05, respectively, so that the final hedonic tone of the frame was 2.17 + 1.96 − 2.05 = 2.08. The participants’ scores for the hedonic tone of the self-frames were sorted from highest to lowest, and scores above the median value were classified as positive self-frames, whereas scores below the median value were grouped as negative self-frames.

Data Analysis

The independent variable was optimism (high vs. low), and the dependent variables included the proportion of positive self-frames in each group and the risky decision making in a certain self-frame. Data were processed with SPSS 16.0 software. The statistical methods included a chi-square (χ²) test conducted to examine the effect of optimism on the component proportions of self-framing and of risky decision making, and logistic regression analysis to test the suppressing effect of self-framing in the relationship between optimism and risky decision making.

Results

In Table 1 the proportions of positive and negative self-frames in high- and low-optimism groups are presented. As shown, fewer participants formed positive self-frames in the low-optimism group compared with the high-optimism group. Thus, Hypothesis 1 was supported.

The effects of self-framing on decision making were then tested. Results are set out in Table 2.

In Table 3 data are presented for the risky decision making of participants in the high- and low-optimism groups.

Logistic regression analysis was also conducted, as presented in Table 4. In Model 1, self-frames were considered the dependent variable, and optimism was regarded as a predictor. Optimism was found to significantly predict self-framing, and participants in the high-optimism group tended to make more positive self-frames (see Model 1).

In Models 2 and 3, decision making was considered as the dependent variable. Optimism was found to significantly predict decision making (see Model 2). When self-frames were added into the model as the independent variable, the regression coefficient of the effect of optimism on decision making increased (see Model 3), which suggests that self-framing suppressed this effect. In other words, there were two competing effects of optimism on decision making, and the direct effect and indirect effect through self-framing were opposite to each other. Thus, Hypothesis 2 was supported.

Discussion

In this study we found significant effects of optimism on self-framing and risky decision making: Compared with low-optimism individuals, highly optimistic participants tended to use more positive words in their descriptions, generated more positive frames, and were more likely to choose risky options. Consistent with previous studies, we also observed a significant self-framing effect: Participants tended to be risk averse when they formed a positive self-frame, and risk seeking when they formed negative self-frames (McElroy et al., 2007; Peng et al., 2014; X. Wang, 2004). Additionally, we found that self-framing suppressed the effect of optimism on risky decision making.

Compared with those in the high-optimism group, participants in the low-optimism group were more likely to use negative words to describe the decision information, thus forming a negative self-frame. Optimism is a stable personality trait, and optimistic individuals have active expectations about the future in a way that crosses time and scenarios (McLennan et al., 2017). The self-frame is a decision maker’s subjective understanding of the decision information (McElroy et al., 2007; Peng, Cao, et al., 2019; X. Wang, 2004). Hence, low-optimism individuals are more likely to understand ambiguous information negatively, and will thereby form a negative self-frame. We found a significant self-framing effect in this study: Participants with negative self-frames tended to be risk seeking, and those with positive self-frames tended to avoid risks. This is consistent with previous study findings and suggests that self-framing is a stable phenomenon (Peng et al., 2014; X. Wang, 2004).

Optimism can significantly affect risky decision making, and our results show that high-optimism individuals were more willing to select high-risk options than were low-optimism individuals, which is in line with the findings of some previous researchers (Gibson & Sanbonmatsu, 2004; Luo & Isaacowitz, 2007). This result can be explained from two perspectives: First, high-optimism individuals have more active expectations about the future than do those lower in optimism; hence, highly optimistic people are more likely to overestimate the probability of active events occurring, and are better able to cope with negative events (Loh et al., 2017). Second, optimism can weaken threat perception during decision making (Pandey, 2018; Peng et al., 2018). As has been previously reported, decision makers under threat are more likely to take risks (Pandey, 2018).

We have demonstrated that self-framing suppresses the effect of optimism on risky decision making. Participants in the high-optimism group tended to code decision-making information more positively and then made more positive self-frames. This is noteworthy because people tend to avoid risks in positive frames. However, as high-optimism people have more active expectations toward the future and experience less threat perception, they are more likely to be risk seeking. The direct and indirect effects of optimism on risky decision making are opposite, and because the direct effect is stronger, both low and high optimists tend to make more risky decisions overall.

This study has both theoretical and practical significance, as we have expanded the exploration of self-framing to discuss the effects of optimism on self-framing and risky decision making. Tversky and Kahneman (1981) divided decision making into two cognitive phases: coding and evaluation. In the coding phase, information is simplified into a frame, and the framing effect should be stable whether the frame is formed passively or actively. In real life, frames are generally self-coded rather than provided. Therefore, the formation of self-frames is usually the first step of decision making (Fischhoff, 1983). Hence, research on self-framing not only expands understanding of the extent of the framing effect, but also accords with daily applications. Our results indicate that high-optimism individuals may code information positively, and have positive expectations about the future. Their attitude and behavior then become predictable to some extent.

Nevertheless, this study has some limitations. First, convenience sampling was adopted instead of random sampling, and only undergraduates were enrolled; this limits the external validity of the study. Second, we used a survey method instead of an experimental method, and not all key latent variables (e.g., other personality traits) were controlled for; thus, the causal relationships of personality traits and self-framing with risky decision making were not clarified. Third, personality traits and self-frames were evaluated simultaneously, so that the participants may have given tendentious answers. Thus far, psychology researchers have advocated for the heterochronic measurement of multiple variables. Future researchers might measure personality traits and self-framing at different time points to make the conclusions more convincing. Finally, decision making was evaluated by a one-of-two method, which is simple to use but did limit somewhat our analytic approach. These limitations should be overcome in future studies.

Cao, F., Zhang, J., Song, L., Wang, S., Miao, D., & Peng, J. (2017). Framing effect in the trolley problem and footbridge dilemma: Number of saved lives matters. Psychological Reports, 120(1), 88–101.
https://doi.org/10.1177/0033294116685866

Fischhoff, B. (1983). Predicting frames. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9(1), 103–106.
https://doi.org/10.1037/0278-7393.9.1.103

Gibson, B., & Sanbonmatsu, D. M. (2004). Optimism, pessimism, and gambling: The downside of optimism. Personality and Social Psychology Bulletin, 30(2), 149–160.
https://doi.org/10.1177/0146167203259929

Jose, P. E., Lim, B. T., Kim, S., & Bryant, F. B. (2018). Does savoring mediate the relationships between explanatory style and mood outcomes? Journal of Positive Psychology and Wellbeing, 2(2), 149–167. https://bit.ly/2RKMcL0

Levin, I., Gaeth, G., Schreiber, J., & Lauriola, M. (2002). A new look at framing effects: Distribution of effect sizes, individual differences, and independence of types of effects. Organizational Behavior and Human Decision Processes, 88(1), 411–429.
https://doi.org/10.1006/obhd.2001.2983

Li, Z., Yin, X., Cai, T., & Zhu, C. (2013). The structure of dispositional optimism: Comparing traditional factor models and a bifactorial model [In Chinese]. Chinese Journal of Clinical Psychology, 21(1), 45–47. https://bit.ly/31FBFW4

Loh, J., Harms, C., & Harman, B. (2017). Effects of parental stress, optimism, and health-promoting behaviors on the quality of life of primiparous and multiparous mothers. Nursing Research, 66(3), 231–239.
https://doi.org/10.1097/NNR.0000000000000219

Luo, J., & Isaacowitz, D. (2007). How optimists face skin cancer information: Risk assessment, attention, memory, and behavior. Psychology and Health, 22(8), 963–984.
https://doi.org/10.1080/14768320601070951

McElroy, T., Seta, J., & Waring, D. (2007). Reflections of the self: How self-esteem determines decision framing and increases risk taking. Journal of Behavioral Decision Making, 20(3), 223–240.
https://doi.org/10.1002/bdm.551

McLennan, B., McIlveen, P., & Perera, H. (2017). Pre-service teachers’ self-efficacy mediates the relationship between career adaptability and career optimism. Teaching and Teacher Education, 63, 176–185.
https://doi.org/10.1016/j.tate.2016.12.022

Pandey, M. (2018). The opportunity-threat theory of decision-making under risk. Judgment and Decision Making, 13(1), 33–41. https://bit.ly/33I5qYZ

Peng, J., Cao, F., Zhang, Y., Cao, Y., Zhang, Y., Zhu, X., & Miao, D. (2019). Reflections on motivation: How regulatory focus influences self-framing and risky decision making. Current Psychology. Advance online publication.
https://doi.org/10.1007/s12144-019-00217-w

Peng, J., Feng, T., Zhang, J., Zhao, L., Zhang, Y., Chang, Y., … Xiao, W. (2019). Measuring decision-making competence in Chinese adults. Journal of Behavioral Decision Making, 32(3), 266–279.
https://doi.org/10.1002/bdm.2114

Peng, J., Li, H., Miao, D., Feng, X., & Xiao, W. (2013). Five different types of framing effects in medical situation [sic]: A preliminary exploration. Iranian Red Crescent Medical Journal, 15(2), 161–165.
https://doi.org/10.5812/ircmj.8469

Peng, J., Xiao, W., Yang, Y., Wu, S., & Miao, D. (2014). The impact of trait anxiety on self-frame and decision making. Journal of Behavioral Decision Making, 27(1), 11–19.
https://doi.org/10.1002/bdm.1783

Peng, J., Zhang, J., Zhang, Y., Gong, P., Han, B., Sun, H., … Miao, D. (2018). A new look at the impact of maximizing on unhappiness: Two competing mediating effects. Frontiers in Psychology, 9, Article 66.
https://doi.org/10.3389/fpsyg.2018.00066

Scheier, M., Carver, C., & Bridges, M. (1994). Distinguishing optimism from neuroticism (and trait anxiety, self-mastery, and self-esteem): A re-evaluation of the Life Orientation Test. Journal of Personality and Social Psychology, 67(6), 1063–1078.
https://doi.org/10.1037/0022-3514.67.6.1063

Stark, E., Baldwin, A., Hertel, A., & Rothman, A. (2017). The role of rational and experiential processing in influencing the framing effect. The Journal of Social Psychology, 157(3), 308–321.
https://doi.org/10.1080/00224545.2016.1198301

Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453–458.
https://doi.org/10.1126/science.7455683

Wang, X. (2004). Self-framing of risky choice. Journal of Behavioral Decision Making, 17(1), 1–16.
https://doi.org/10.1002/bdm.454

Wang, Y., & Peng, J. (2017). Work–family conflict and depression in Chinese professional women: The mediating roles of job satisfaction and life satisfaction. International Journal of Mental Health and Addiction, 15(2), 394–406.
https://doi.org/10.1007/s11469-017-9736-0

Xia, J., Wu, D., & Xu, X. (2018). Susceptibility to the gambling near-win effect in optimists versus pessimists: An event-related potential study. Personality and Individual Differences, 129, 159–165.
https://doi.org/10.1016/j.paid.2018.03.032

Yu, X., Zhang, H., Yu, X., & Zeng, J. (2015). Impact of self-framing on decision making: Timing matters. Social Behavior and Personality: An international journal, 43(4), 629–639.
https://doi.org/10.2224/sbp.2015.43.4.629

This study was funded by the major project of the 13th Five-Year Logistical Research Plan of PLA (BWS16J012).

Ran Zhang and Luming Zhao contributed equally to this research as co-first authors.

The authors thank Dr. Wei Guo for her kind help with data collection.

Xufeng Liu or Peng Fang, Department of Military Medical Psychology, Air Force Medical University, No. 169 Changle West Road, Xi’an 710032, People’s Republic of China. Email: [email protected] or [email protected]