When given the positive frame, males were just as likely to choose frame A and B. This supports utility theory as they have equal utility. However females were three times more likely to select option A, the safer action. This could demonstrate loss aversion as the participants were less susceptible to the loss proposed in option B, so chose, the gaining option (option A). There is no evidence when provided with the negative frame to suggest the framing effect because more females took the riskier option B than males. My hypothesis proposes that males are more likely to take risks than females, however this is not evident in my results. The results contrast my hypothesis. This subsequently may dismiss Farthing’s conclusion that males are more likely to take risks, to impress females.
Figure 1 – Number of males and females who chose positive and negative frames The above graph shows that in the positive frame, more females took the safer option, illustrating loss aversion. Males are equally as likely to choose option A or B, showing the utility theory, as both options have the same utility. In the negative frame, the graph shows that more females took risks, contrary to my hypothesis that males are more vulnerable to risk taking behaviour.
Treatment of Results I used a chi square to analyse my results, because I used nominal data, which is also definite because I used gender for grouping. P < 0.05 Critical value = 7.82 Degree of freedom = 3 X2 = 2.4 To give explanation for my hypothesis, the value of X2 should be equal to or greater than the critical value. However as the value is only 2.4 and the critical value is 7.82, I am required to refute my experimental hypothesis and accept my null hypothesis. My null hypothesis states that gender does not have an affect on the susceptibility to take risks.
Discussion My findings show that my results do not support my experimental hypothesis; that males are more likely to take risks. My X2 value was only 2.4 when the critical value is P=0.05=7.82. Therefore I accept my null hypothesis, that gender does not have an effect on risk taking behaviour. Statistically the correlation is not significant. The results show that in the positive frame males choose at random, and females chose the more secure option (loss aversion) but overall males did not take more risks than females.
However, in the negative frame more females chose option B than males. This does not support my hypothesis on the foundation that it is the framing which affects how we make decisions. Therefore, it seems that females are more susceptible to loss, thus demonstrating loss aversion. On the contrary it appears that males are not as sensitive to the framing effect and do not tend to take more risks than females, which is illustrated in the negative frame.
Farthing (2005) conducted an experiment into gender and the framing effect. My results do not support this. When presented with the positive frame, more males tended to take risks than females, with the negative frame, the results were inverted, females being the major risk takers. Farthing also put across the idea that males are more likely to take risks to attract female attention. I am unclear if this theory can be applied to my research. The decision participants chose perhaps should not be influenced by this factor as it was performed in private.
To some degree, my results represent utility theory. Males show equal probability of choosing A and B when faced with the positive frame. This supports utility theory because participants have randomly selected their response, thus calculated the risk and concluded that there is no difference. This could also be applied to the negative frame. Females however, do not seem to support the utility theory.
Furthermore, there is some intellectual variation since Kahneman and Tversky carried out their study. Participants probably have a higher mathematical ability and are probably more able today, meaning that they can calculate with ease, the risk of the two options This could mean that they are more likely to conform to utility theory. Taking this into account, mathematical ability could be a confounding variable. It is possible that the male participants made their decision because they had superior mathematical skills.
In this experiment, participants had no time limit to provide their answers, which is not representative of every day situations. Consequently, there may be other factors which influenced the results, meaning that it is not typical of a decision which would be made on an everyday basis, to which the theory concerns. There were some matters which may have limited the validity. Primarily, the sample itself had awkward issues. My sample comprised 32 participants, which is significantly small to generalise from, thus causes the study to lack ecological and external validity. This could have been modified by using a larger sample.
Also, I used an independent groups design. This eliminates some problems, however does create others such as individual difference. Therefore a matched pairs design may have been more appropriate. Participants should be matched in accordance with age and mathematical capability. The participants in my experiment were all of the same age (year 10 students). This means that my results cannot be generalised to the entire population as it is only specific to students. They may have different experiences than an older person, or may have better maths skills. This could have been modified by using a wider age span for the sample.
I previously mentioned the mathematical ability of participants. It may be that younger people have a greater mathematical ability because of more modern teaching methods and that they are still actively using these skills. Thus, they may be more able to calculate the utility. This could be adjusted, to make the experiment more valid by using participants with the same mathematical ability; hence, a matched pairs design.
Another concern could have been influenced by social desirability, as a consequence of me knowing some of the participants. Therefore, participants may have provided their response based on my expectations of them. To rectify, participants should be chosen who are unacquainted with the experimenter and it should be repeated to participants that they should consider both situations before choosing their response on a serious level.
Some participants may not have taken the experiment seriously because of informality, resulting in them not considering both actions fully. Subsequently, the experiment should have been carried out in a more formal environment and style. Another problem which became recognisable whilst conducting the experiment is that participants may have already done the experiment, as another experimenting group could have used the same or some of the participants previously. If this was the case, then they would have already been debriefed, giving them knowledge of the study and its aims and would probably influence their answer, should they carry out the experiment again. To rectify this I could have discussed with the other experimenting groups which group of people we would use in our experiments, ensuring that the experiments did not use the same participants.
This experiment was not representative of decisions we may have to face in everyday life, as many are quick judgements. Because of the unlimited time participants had, the experiment was not realistic. A time limit would correct this, making the experiment more valid, as it would be true to such judgements that we face in everyday life. If I was to carry out this experiment again, there would be a number of aspects which I would adjust. Instead of investigating gender there may be other factors such as mathematical ability, age or social status. Mayhorn (2002) found that older people are more affected by the framing effect. This could be the foundation of a future study. I could explore whether older people are more likely to take risks. It may be that older people are less likely to take risks because they have their families to think about (setting good examples etc), whereas younger people do not have this worry.
A final factor which could be explored is social status. It is possible that the working class take more risks in order to survive unlike the upper and middle class who do not have this burden. One way of judging this would be by looking at the participants income, however this may cause difficulties in attaining from participants such personal data. It may be a good idea to compare the working and upper class and their attitudes of a risk. Furthermore, the answer sheet could include reasons why participants’ should chose programme A or B. This would provide qualitative data that could be used to determine what really influences decision making.