Psychology capstone *(tutor is needed throughout course)*

The Effects of Framing on Decision-Making

Decision-making is the process by which we choose between different options. Some decisions can be important (i.e., deciding whether to attend graduate school or not) others can be much simpler (i.e., deciding what to eat for dinner). Nonetheless, every decision, big or small, can have an impact on our lives. So how is it that we make these decisions and what factors influence our decision-making? Kahneman and Tversky (1973), explored decision-making and how the representativeness heuristic affects people’s decisions. The representativeness heuristic is a mental shortcut that relies on the similarities of a prototype, while ignoring true odds (Kahneman & Tversky, 1973). Since categorization of objects or members is based off resemblances, predictions are often wrong and decision-making is impacted.

Predictions are intuitive guesses that are made before decisions, and are affected by three factors: prior knowledge, specific information, and expected accuracy (Kahneman & Tversky, 1973). Previous knowledge is important because it influences what predictions people will make, based off their experiences. Specific information is also beneficial because it allows people to make predictions that support the provided material. However, excepted accuracy is the most vital because it relies on confidence levels (i.e., the probability that a prediction is correct). Notably, when confidence is high, individuals rely on intuition and when confidence is low, individuals rely more on given information (Johnson, 1987; Kahneman & Tversky, 1973). Although information is often provided, Johnson (1987) found that when information is missing, people use prior knowledge or experience such as heuristics and make inferences based off their personal experiences. These inferences cause individuals to make interpretations based on personal preference and to make decisions based off what they want or what they think they want (Johnson, 1987; McNeil, Pauker, & Tversky, 1988).

Tversky and Kahneman (1981) discovered that when personal preference is involved people partake in psychological accounting, which states that individual’s frame, and evaluate the outcomes of an act based off the consequences of their choice. Such accounting leads individuals to make decisions that benefit them the most because they want to get the most advantage (Kahneman & Tversky, 1973; Tversky & Kahneman, 1981). Tversky and Kahneman (1981) found that the expected utility theory is also important in the decision making process since individuals assess outcomes by evaluating the likelihood of each of them occurring. Additionally, the prospect theory, which is derived from expected utility theory, is also part of the process since people take into account the probability of a good outcome (Tversky & Kahneman, 1981). Primarily, Tversky and Kahneman (1981) discovered that losses cost more than gains, so decisions can be easily influenced by the framing effect. This is the idea that people are more risk-aversive (i.e., avoid risks) when an option is a certain gain and more risk-taking when an option involves a loss.

The framing effect has been tested in many scenarios and has been evaluated to be influential. Tversky and Kahneman (1981); Chien, Lin, and Worthley (1996); and Johnson (1987) presented participants with different money problems and framed them in terms of loss or gain. The researchers discovered that in the negatively framed scenarios (i.e., spending more money or losing money) participants were more risk-taking, while in the positively framed scenarios (i.e., winning money or gaining money) participants were more risk-aversive. Furthermore, Mikels and Reed (2009) also presented participants with a money problem and discovered that participants were affected by the framing effect, as they were more risk-taking (i.e., likely to gamble) when there was money at stake. Overall, prior research has evaluated that when presented with money scenarios, participants tend to be susceptible to the framing effect, since they are more likely to take risks when faced with a loss, and to avoid risks when encountered with a gain.

Notably, the framing effect has also been influential in mortality scenarios. Chien et al. (1996) and, McNeil et al. (1988) presented participants with a disease problem, and found that they were more likely to pick the certain option if the frame was positive (i.e., saving more people) and the risk option if the frame was negative (i.e., losing lives). Similarly, a different researcher evaluated a euthanasia mortality scenario and discovered that when participants had to choose between a positively framed scenario (i.e., not prolonging a patient’s life) and a negatively framed scenario (i.e., ending a patient’s life) the participants were more likely to choose the positively framed scenario (Gamliel, 2012). Overall, although mortality scenarios have been quite influential, morality scenarios have produced different results. McNeil et al. (1988) investigated genetic counseling and asked participants whether or not they would have a child even if there were a chance that the child would be born with a heart defect. The researchers found that whether the scenario was positively framed (i.e., normal heart) or negatively framed (i.e., abnormal heart), the framing effect was ineffective because morals are very powerful and not susceptible to framing. All in all, the framing effect has been influential in mortality problems, since participants have often taken risks when presented with a negatively framed scenario. However, the framing effect has not been effective in morality scenarios because morals are too strong.

Nevertheless, the framing effect has worked among different variables, such as age, education level, gender, and ethnicity. For instance, Chien et al. (1996) discovered that when presented with a mortality scenario or a money problem, adolescents were just as susceptible to the framing effect as adults. Similarly, Rönnlund, Karlsson, Laggnäs, Larsson, and Lindström (2005) evaluated age and presented younger adults and older adults with a disease problem and found that despite their age, participants were equally impacted by the framing effect. Conversely, a different study with a gambling scenario, led researchers to discover that in a negatively framed scenario older adults were much less susceptible to the framing effect as opposed to younger adults because they made much less risky decisions and often picked certainty (i.e., keeping money) instead (Mikels & Reed, 2009). Overall, although results tend to vary, in most cases researchers have discovered that younger individuals and older individuals are influenced equally by the framing effect.

When considering other variables, such as education, Chien et al. (1996) found that both math honor students and math non-honors students were equally influenced by the framing effect. Moreover, the researchers also found that gender was an additional unimportant factor because the framing effect influenced males and females equally. However, a previous study on gender differences with regard to the framing effect has led different researchers to find contrary results. The researchers discovered that across different scenarios (i.e., disease, cancer, school dropouts, job layoffs, and civil defense) females were more susceptible to the framing effect than males, since males answered contrary to the framing effect and were actually more risk-taking in the positively framed scenarios and risk-aversive in the negatively framed scenarios (Fagley & Miller, 1990). Aside from gender, another influential variable is ethnicity. Velez Ortiz, Martinez, and Espino (2015), evaluated end-of-life preferences and found different results among Caucasian and Latino participants. Latinos overall were more influenced by the framing effect because when they were provided with a positively framed scenario (i.e., chance of survival) they were more likely to accept resuscitation, while Caucasians were not. All in all, varying education levels tend to be equally susceptible to the framing effect. However, males and females and different ethnicities are not influenced as similarly, since prior research has demonstrated that females are more influenced by the framing effect than males, and Caucasians are less susceptible to the framing effect when compared to Latinos.

Varying levels of involvement (i.e., whose life is on the line) have also impacted how people make decisions. McNeil et al. (1988) discovered that participants were influenced by the framing effect, since they were more risk-taking when presented with scenarios that had to do with their own lives. On the other hand, the researchers also discovered that when participants were presented with a scenario that involved someone else’s life (i.e., a friend or a loved ones), participants were less influenced by the framing effect since they were more likely to choose a certain option because they wanted to be sure that they made the right decision. However, when making a decision on a stranger’s life, other researchers found that participants were susceptible to the framing effect because they were more risk-aversive in positively framed scenarios (i.e., saving lives), and risk-taking in negatively framed scenarios (i.e., losing lives) (Chien et al., 1996; & Tversky and Kahneman, 1981). Although individuals’ own lives, stranger’s lives, and loved ones’ lives, are susceptible to the framing effect, so are animal’s lives. Evidently, individuals display empathy towards animals, thus Bloomfield (2006) assessed the influence of the framing effect on human and animals’ lives and discovered that when participants were not presented with any additional information (i.e., pictures or names) they were more likely to take a risk in the human scenario, while in the animal scenario they were more risk-aversive. Overall, prior research has demonstrated that the framing effect varies across different levels of involvement and participants are typically more risk-taking with their own lives or strangers lives and more risk-aversive with loved ones’ lives and animal’s lives.

All in all, the framing effect has been evaluated to be influential under many circumstances, but there is still information lacking in the field. Specifically, more research should be conducted on stranger’s lives, in order to discover how individuals make decisions that don’t directly impact them. Therefore, the purpose of the current study is to examine stranger’s lives and how college students are influenced by the framing effect when presented with an adaptation of Tversky and Kahneman’s (1981) mortality scenario. The hypotheses propose that in the positively framed scenario, participants will be more likely to choose certainty over risk, and in the negatively framed scenario, participants will be more likely to choose the risk option over the certain option. Overall, it is predicted that when strangers’ lives are at risk, college students will be susceptible to the framing effect.

Method

Participants

Participants included a total of 234 psychology students from a California State University (m = 59; f = 175; age range = 19 - 59; average age = 24.59). Participants varied in education level (senior = 62.7%; junior = 34.7%; sophomore = 2.5%) and English proficiency (native English speaker = 69.1%; not native/very fluent = 26.3%; not native/fluent = 4.7%). No incentives were given and all participants were treated in accordance to the American Psychological Association Ethical Principles of Psychologists and Code of Conduct (American Psychological Association, 2002).

Materials

Handouts included an informed consent, a response sheet, and a debriefing statement. The top portion of the response sheet contained demographics (i.e., age, gender, level of education, and English proficiency) and the second portion included the response options for question 1 and question 2, indicating to pick either option X or option Y. A computer was used to present a PowerPoint on a projection screen. The PowerPoint presentation included five slides (i.e., consent, instructions, question 1, question 2, and debriefing statement). The scenarios were adapted and modified from Tversky and Kahneman (1981). Both question slides were presented with the following scenario:

Imagine that the United States is preparing for the outbreak of an unusual disease that is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact scientific estimate of the consequences of the program are as follows:

The positively framed scenario read:

If program X is adopted, 200 people will be saved.

If program Y is adopted, there is a one-third probability that 600 people will be saved, and a two-thirds probability that no people will be saved.

Which of the two programs (X or Y) would you favor?

The negatively framed scenario read:

If program X is adopted, 400 people will die.

If program Y is adopted, there is a one-third probability that nobody will die, and a two-thirds probability that 600 people will die.

Which of the two programs (X or Y) would you favor?

Procedure

All participants were tested individually in a quiet group setting. Participants were given an informed consent, which they X’d and dated. Participants were then given a response sheet and filled out the demographics section, which was located on the top portion of the page. Next, a PowerPoint was presented which provided instructions informing participants that they were going to see the questions one at a time for 40 seconds each. The PowerPoint then proceeded to question 1 and question 2. To control for order effects, the presentation of the question slides was counterbalanced with half of the participants receiving the positively framed scenario first, followed by the negatively framed, and the other half receiving the negatively framed scenario first, followed by the positively framed. Participants indicated on the second section of their response sheet with a check mark for each question, whether they selected option X (i.e., certain) or option Y (i.e., risk). After the participants finished responding, they were debriefed and thanked for their time.

Design

The current study had a within subjects design. The independent variable was the framing effect with two levels, positive and negative. The dependent variable was the response options with two levels, a certain option (i.e., option X) and a risk option (i.e., option Y). A Pearson’s chi-square was used, with a significance desired of p < .05.

Results

A chi-square analysis was conducted to determine if the positively framed scenario would influence participant’s response options. Results from the positively framed scenario were not significant (2 (1, N = 234) = 3.35, p = .067) (see Table 1). In the positively framed scenario more participants selected the certain option (n = 131) over the risk option (n = 103), although the hypothesis was not supported.

Another chi-square analysis was used to determine if the negatively framed scenario would influence participant’s response options. Results from the negatively framed scenario were also not significant (2 (1, N = 235) = 2.66, p = .103). In the negatively framed scenario participants were more likely to select the risk option (n = 130) than the certain option (n = 105), although the hypothesis was not supported yet again.

A final chi-square analysis was conducted to determine the overall susceptibility to the framing effect. Significant results indicated that participants were not susceptible to the framing effect (2 (2, N = 233) = 75.00, p < .001). More participants were not susceptible (Option XX or YY; n = 138) than those susceptible (Option XY; n = 61) or those who responded opposite of prediction (Option YX; n = 34). As predicted, some participants were susceptible to the framing effect, however more participants were not.

Discussion

The purpose of the current study was to determine if participants were influenced by the framing effect when provided with positively and negatively framed mortality scenarios surrounding stranger’s lives. It was proposed that in the positively framed scenario, participants would be more likely to choose certainty over risk, and in the negatively framed scenario, participants would be more likely to choose risk over certainty. The results revealed that this was the case, since some participants did answer accordingly, but not enough to establish significance. Additionally, the overall susceptibility results demonstrated that the participants were not susceptible to the framing effect.

The hypothesis for the positively framed scenario was most likely not supported because the chosen problem involved strangers’ lives. Most commonly, participants are not as concerned with stranger’s lives as they are with their own lives or their loved ones’ lives. Veldwijk et al. (2016) presented participants with a mortality scenario in which they were told to imagine that they were prone to colon cancer, and could undertake preventative methods to combat the disease. The researchers discovered that participants were much more likely to be risk-aversive with their lives and thus pick the positively framed scenario (i.e., survival) over the negatively framed scenario (i.e., mortality). It appears that there is no time for risk-taking when it comes to an individual’s own life, but the results can be much different when considering strangers or loved ones’ lives. To validate this claim, Bloomfield, Sager, Bartels, and Medin (2006) evaluated how social relations influence the framing effect, and presented participants with different mortality scenarios (i.e., their own families, someone else’s family, friends, or strangers lives). The researchers discovered that a reverse framing effect occurred in the friends scenario since participants were more risk-taking when presented with a positively framed scenario, but with regard to the stranger scenario the participants were not influenced by the framing effect. Notably, the results from prior research demonstrate that individuals respond differently to the framing effect, according to whose life is on the line.

On a different note, the hypothesis for the negatively framed scenario was possibly not supported because the participants faced a time constraint (i.e., 40 seconds) and didn’t have enough time to read the questions thoroughly or think about their answers. In a study on time pressure, researchers provided participants with different mortality scenarios (i.e., cancer, heart operation, disease, and AIDS) and determined that the participant’s response options varied according to the amount of time they had to answer the questions (i.e., 40 seconds or no time) (Svenson & Benson, 1993). The researchers discovered that participants with no time pressure were susceptible to the framing effect; therefore, the results indicated that time restrictions reduced the impact of the framing effect. Certainly, individuals who are under a time constraint must make decisions quickly and might not have time to evaluate the scenario, thus guessing. Kocher, Pahlke, and Trautmann (2013) discovered that time pressure influenced participants response options in money scenarios. They determined that in the negatively framed scenarios and mixed frame scenarios (i.e., positively and negatively framed) participants were more risk-aversive (i.e., opposite of the framing effect). Overall, these findings demonstrate that when individuals are under time restrictions they are less influenced by the framing effect or tend to answer contrary to it.

Specifically, overall susceptibility was most likely not supported because the current study was conducted on undergraduate psychology students who have probably encountered the framing effect before. In this case, it is possible that the participants had previously taken a cognitive psychology course and were educated on the subject; thus influencing the results. In other cases, simple evaluation of decisions is enough to undermine the framing effect. For example, Almashat, Edelstien, Ayotte, and Margrett (2008), presented two experimental groups with a decision evaluation form, and the control group with a generic questionnaire. The researchers discovered that mere manipulation was significant enough for the experimental groups to not be susceptible to the framing effect. These findings demonstrate that knowledge is powerful and if rationalization can challenge the framing effect, then experience is far more likely to decrease its influence.

For instance, a researcher discovered that undergraduate psychology students were not affected by the framing effect, since only seven out of thirteen scenario trials were found to be influential, while the rest ran contrary or demonstrated no effect (Wang, 1996). The researcher concluded that the findings arose partially from social context, since this factor can influence an individual’s perception. Notably, since college students are commonly in an educational setting and are surrounded by other knowledgeable peers, their social environment can be very influential. Even if the students themselves do not have direct experience with the framing effect, they can still gain knowledge on the subject through word-of-mouth. Consequently, the participants in the current study were possibly not influenced by the framing effect since they have most likely taken a cognitive or social psychology course before and have been directly educated on the framing effect or they have heard about the framing effect from their peers.

In particular, the current study had many limitations such as, gender. Specifically, the current sample of males in the experiment was relatively small compared to the population of females. Therefore, future researchers should include an equal sample in their studies in order to determine if male and female college students are influenced differently by the framing effect, and if any generalizations can be made. Notably, in a meta-analysis researchers evaluated the influence of gender on the framing effect and discovered that overall males are much more risk-taking than females, in many different scenarios and despite how the problem is framed (Byrnes, Miller, & Schafer, 1999). Another critical limitation is that the current study utilized college students, specifically psychology majors. As previously mentioned it is possible that social environments, and knowledge on the subject can influence participant’s decision-making. Past research by Cao et al. (2017) tested the framing effect on different college majors (i.e., physics, history, and English) by presenting students with two mortality scenarios (i.e., bridge problem and trolley problem) and discovered that the participants were impacted by the framing effect. Thus, if a study is going to be conducted on psychology students, researchers should develop a new scenario in order to reduce the probability that students have previously encountered the classic framing model. Overall, future research should examine how prior knowledge on the framing effect can influence susceptibility, and should particularly beware of participant sophistication, the pre-exposure effect, time constraints, and chosen level of involvement.

By and large, research on the framing effect is required to understand what factors influence individual’s decision making. This type of information can be valuable for any morality or mortality based scenarios, in which doctors must frame questions accordingly in order to not influence the individual into making a decision based off wording. Research on the framing effect is also beneficial for decision makers in order for them to recognize that framing affects their response options. Since framing is used in many sources, from advertisements to news articles it’s important for individuals to gain knowledge on the framing effect, in order to increase their competence and make their own honest decisions without focusing on how a question or scenario is framed. The current research partially demonstrated how important prior knowledge is in influencing individuals to make decisions that they find best. Overall, although the framing effect itself is not harmful, in the hands of the wrong people it can be, and with more knowledge, individuals are able to take back control of their decision-making.

References

Almashat, S., Ayotte, B., Edelstein, B., & Margrett, J. (2008). Framing effect debiasing in medical decision making. Patient Education and Counseling, 71(1), 102–107. doi: 10.1016/j.pec.2007.11.004

American Psychological Association. (2002). Ethical principles of psychologists and code of conduct. American Psychologist, 57(12), 1060–1073. doi: 10.1037//0003-066X.57.12.1060

Bloomfield, A. N., Sager, J. A., Bartels, D. M., & Medin, D. L. (2006). Caring about framing effects. Mind & Society, 5(2), 123–138. doi: 10.1007/s11299-006-0013-3

Bloomfield, A. N. (2006). Group size and the framing effect: Threats to human beings and animals. Memory and Cognition, 34(4), 929–937. doi: 10.3758/bf03193438

Byrnes, J. P., Miller, D. C., & Schafer, W. D. (1999). Gender differences in risk taking: A meta-analysis. Psychological Bulletin, 125(3), 367–383. doi: 10.1037/0033-2909.125.3.367