The impact of music on uncertain decision making : an online mood induction experiment

(1)

The impact of music on uncertain decision

making

An online mood induction experiment

___________________________________________________________________________

Master Thesis

Msc Economics; Behavioral economics and game theory

Name Koen Maas

Student number 10665951

ECTS 15

Supervisor Dr. J.B. Engelmann

(2)

Abstract

An increasingly number of researchers have been getting interested in researching the impact of emotions on decision making. When the researchers included decision making under uncertainty they focussed mostly on risk. The subject of the effect of emotions on ambiguous decision making is less well explored in the literature. Therefore the objective of this thesis is to find the possible impact of emotion, specifically happiness and sadness, on ambiguity preferences. This was done by means of a online experiment using music to induce specific emotion randomly. Results show that there is no significant effect of happiness or sadness on both risk and ambiguity preferences.

Statement of Originality

This document is written by Koen Maas who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

1. Introduction

Before making a decision people weigh up their options if they can. They look at what the best possible outcome could be of their various possible alternatives. People do this by doing some sort of cost-benefit analysis in their head. Most of the theories concerning decision making, involve this use of a cognitive perspective (Bechara et al., 2000). Emotion is closely linked with decision making, but this can be either as a consequence of the decision or as the cause for a decision (Bechara et al., 2000). An example of experiencing emotions as a result of a decision can be feeling angry or sad after losing money because of the decision to buy a lottery ticket. However there might have been an emotion present previous or during buying the lottery ticket which influenced your decision in the first place.

In addition it is important to realise that the outcomes of decisions a person makes in everyday life are not always certain. The outcomes of these decisions can be either risky (known probabilities) or ambiguous (unknown probabilities). How a person reacts to risk is in generally depended on their own risk preferences. This is expressed in whether they are generally risk averse, risk neutral or risk seeking in their behaviour and decision making in particular.

Judging the probability of decisions being relatively risky or safe can be influenced by evoking emotional states (Westermann et al., 1996). One of the ways to evoke emotions prior to making a judgement about the probability of a risky decision is using music. Important to note is that the emotions evoked by mood induction have to be unrelated to the probability or decisions that have to be made after. Music can induce different emotions, Ladinig &

Schellenberg (2012) for example provide evidence that listeners generally like music that evokes happy feelings and dislike music that evokes sadness.

Multiple researchers have done research into the relation between risk preferences or attitudes and the influence of emotion (Schulreich et al.,2014; Conte et al.,2018; Yuen & Lee, 2003). However, probabilities of certain outcomes are not always known. Ambiguous

outcomes are not a rare occurrence in real everyday interactions. Therefore devising an experiment where ambiguous decisions are given as one of the options can be very meaningful addition to the existing literature.

The result of a certain type of music leading a person to make more ambiguous or riskier choices can be used in environments where these types of uncertain decisions occur. This could be used by shop owners or casino’s to only play certain type of music that leads people to take more risk and spend more. It can also serve as useful information for an

(5)

individual who can perhaps control their behaviour in expected uncertain situations by purposefully listen to a certain type of music beforehand. This topic of ambiguity has not been highlighted as much in research using music induction, therefore the research question posed in this thesis will be:

Does happiness and/or sadness influence ambiguity preferences using musical mood induction?

Using an online experiment it can be shown that happiness and sadness both do not seem to influence ambiguity preferences in a significant way. Moreover, risk preferences were also found not to be influenced by these induced emotions, which is contradictory to previously found results in the relevant literature.

This thesis is structured as follows. Section 2 will provide a review of the relevant literature on this topic. In section 3 the experimental design and methodology will be

provided, followed by the hypotheses, is shown. Subsequently, the results will be analyzed in section 4 giving the descriptive statistics of the data and the results of the statistical tests and regressions. The last part, section 5, provides the conclusion discussion of its limitations and provides recommendations for future research.

(6)

2. Literature Review

2.1 Risk versus Ambiguity in decision making

As stated in the introduction uncertain decisions are made by everyone during their lifetime. However it is important to distinguish between two types of uncertainty: risk and ambiguity. While making risky decisions the probability of the different possible outcomes can be estimated or are well defined (Bernoulli, 1738). To incorporate risk into economic modelling, expected utility of every outcome was calculated by giving each outcome a probability weighting and then taking their sums (von Neumann and Morgenstern, 1947). With

ambiguous decisions these probabilities of the different possible outcomes are unknown or not well defined (Camerer and Weber, 1992, Ellsberg, 1961). To model ambiguity in economic decision making the subjective expected utility by Savage (1954) can be used. Camerer and Weber (1992) reviewed this theory and research into ambiguity and point out that “it is hard to think of an important natural decision for which probabilities are

objectively known.” Or in other words ambiguous decisions are made more frequently by

people in the real world than risky decisions. However modelling and researching ambiguity is harder than studying risk since there are more unknown variables. Camerer and Weber (1992) conclude that people usually prefer to make a bet with more knowledge about the possible outcomes (risk) than to not know (ambiguity) while their beliefs may not even change. This means most people are ambiguity averse and thus at least somewhat value getting missing information, even if that doesn't change their decision after getting it.

Moreover, the fields of neuroscience and neuroeconomics provide evidence that risk and ambiguity are processed by different regions of the brain (Krain et al., 2006, Huettel et al., 2006). Therefore ambiguity is not just a special form of a risky decision making but a distinct form of decision making requiring a different brain mechanism. Neuroscience also helps explain how emotions can influence people's decision, specifically by influencing their risk or ambiguity preferences.

2.2 Emotion and decision making

Emotion affects decision making in everyday life, but how does this work exactly? When looking at the way emotion affects the decision making of humans an possible explanation is the concept of the somatic marker hypothesis. This hypothesis gives an neurological

explanation on how stimuli from a person's environment can affect their decisions. These stimuli, such as (the thought of) winning money, can trigger a ‘somatic response’. A somatic

(7)

response means that there is a psychological change in the body to the emotional stimulus Bechara (2003). Brain areas such as the amygdala and the orbitofrontal cortex are described as being particularly important in triggering this somatic response. Bechara et al. (2000) provide evidence for the importance of the orbitofrontal cortex. They also mention that the orbitofrontal cortex is not solely responsible for the decision making process, such as the amygdala, the somatosensory/insular cortices and the peripheral nervous system. Seymour & Dolan (2008) explain that the amygdala is not only the brain region most often linked to emotion, but that this area is also involved in guiding choice. The authors also refer to a study of Hsu et al., (2005) who compared decisions made under risk and ambiguity. Hsu et al., (2005) find that people choose the ambiguous option less often than the risky option. Moreover, recorded activity in the amygdala was found to predict this decision.

2.2.1 Emotion and uncertain decision making

Schwarz (2000) provides evidence from a collection of previous studies that a person's memory of a particular situation and the emotion they felt during that situation are often recalled when feeling the same emotion again later on. The response or decision someone makes in a situation is therefore influenced by those previous experiences. For example a positive decision is more likely when the person is happy than when they’re sad. Schwarz (2000) also points out that a person in a joyous mood is more likely to overestimate the probability of a positive outcome of a decision and underestimation of negative outcomes. The reverse is true for a sad mood and these estimations of decision outcomes. Moreover there is also evidence of people using their previous knowledge of a situation or problem more than the information and details given to them at the time of decision making when they are experiencing happiness. Whereas people who are experiencing sadness will use mostly the information in front of them and rely less on previous knowledge of the same problem. Anderson & Galinsky (2006) found that an increased sense of power can increase how optimistic a person is and thus how much risk that person is willing to take. The authors come up with three reasons why this might be the case. Firstly, people might become more optimistic in their risk estimates if they feel more powerful. Secondly, they are just more susceptible to be lured in by the potential rewards, apart from any risk probabilities. Thirdly, powerful people might be more confident in their capacity to capture the positives of risk or handle the downside if they fail. They come to the conclusion that it is optimism about risky decisions, rather than someone being inherently bad, that causes powerful people to

(8)

Leith & Baumeister (1996) studied the effect of bad moods on decision making. They conducted a series of studies but did not reach a conclusive answer. Looking at it at first glance there is evidence that bad moods increase risk taking and that positive feelings reduce risk taking. That bad moods cause more risk taking is the result of a loss of control. A person that feels angry or embarrassed act more impulsive and not as rational. However, looking in more detail at the results of Leith & Baumeister (1996), certain bad moods, such as anger or embarrassment, increase risk taking behavior, but sadness, caused risk avoiding choices. The authors suggest that this is because of the arousal that is involved with the first two and which is absent in the latter. They also find that people who are confident and have high self-esteem do not seem to make risky decisions but rather rational and cautious decisions.

Lerner & Keltner (2001) provide different conclusion to their research of fear and anger on risk perceptions. They find that people with fear make more risk averse decisions and that angry people are more risk seeking. They also point out that happy people made optimistic and more risk seeking choices. In their paper the authors propose a so-called “appraisal-tendency framework”, which they say can be used to research specific emotions. The basis for this theory is that a person experiences emotion based on the appraisal of their surrounding environment. This appraisal ‘rating’ is done in six cognitive dimensions, anticipated effort, self-other responsibility/control, certainty, pleasantness, and situational control, attentional activity. People who experience an emotion that consists of a certain set of dimensions, as specified within the framework, will make decisions in the future that correspond to those same dimensions (Lerner et al., 2015).

Schulreich et al. (2014) have done research into music-evoked change in risk attitude and probability weighting. They had participants listen to happy, sad or no music or random tones and then after had them repeatedly choose between a pair of lotteries with different risk associated. In their experiment the participants knew the probability of each lottery. The authors found that the participants chose the riskier lotteries more often in the condition were they listened to happy music, compared to the sad and random tones conditions. Schulreich et al. (2014) showed that, based on cumulative prospect theory, this was due to the participants assigning a higher decision weights to the larger payoffs in the happy condition compared to sad and random tones condition. They argue this result provides evidence that there is a causal effect of incidental happiness on risk attitude due to the change in probability weighting.

(9)

are willing to take more risk. Whereas Drichoutis and Nayga Jr. (2013) find that risk aversion is increased when having a positive or negative mood and Williams et al. (2003) show that unhappy managers are less risk seeking but happy managers are not more risk seeking. These mixed results about the effect of emotion on risk show that there is still a need for more research on this topic besides the effect of emotion on ambiguity preferences. Therefore this thesis will also incorporate a measure of risk attitude to see the effect of music induced happiness and sadness on emotion.

As discussed in the introduction the connection of ambiguity preferences and emotion is less abundant than the topic of risk preferences and emotion. However, there are a few such as Baillon et al. (2016) who find that feeling sad causes people to be more ambiguity neutral and payoff maximizing than when they are experiencing joy or fear. Thus in these latter cases people are found to be more ambiguity seeking. The authors also state a more general

conclusion whereby the differences in emotional state can lead to differences in ambiguity attitude. On average most people are ambiguity averse, but there is some variation in

ambiguity attitudes. An example of this variation in ambiguity attitudes can be seen from the results of a field experiment with Chinese small-scale stock investors conducted by Potamites and Zhang (2012). In this experiment most investors were ambiguity averse (57%), 15% were ambiguity neutral and 26% were ambiguity loving. Similarly, Dimmock et al. (2016)

measured ambiguity preferences by conducting a US household survey. The authors found that 10% of the people were ambiguity neutral, 52% were ambiguity averse, and 38% were ambiguity loving.

Lastly, an important aspect of decision emotional making is the prediction of future feelings. These are feelings that are a consequence of current decisions but are not

immediately felt. It is difficult for people to judge which decision will turn out the most emotionally beneficial. An example would be a choice of study, where the person making the decision might think that it will make him or her the most happy, but find out later when actually studying they are not happy at all with their choice. This shows that in this case there is also a case of uncertainty in the decision made. In general this uncertainty causes mistakes and suboptimal decision making in circumstances where future feelings are involved

(Loewenstein & Schkade, 1999). To be able to research the effect of emotion on decision making in an experimental setting, researchers can use a method of inducing a specific emotion in a participant and then letting them make decisions.

(10)

2.3 Emotion induction

There are many different emotion or mood regulation procedures possible such as using film, music, imagination, Velten technique or social interaction. The Velten technique refers to the technique used by Velten (1968) to induce a mood. This widely used technique consists of presenting someone with a number of positive/negative statements or somatic states and subsequently tell the person to feel the mood described by the statements. The results were then compared to the results of a control group answering neutral statements. Westermann et al. (1996) looked at the effectiveness and validity of 11 different mood induction procedures (MIPs) in total. They found that the effectiveness of inducing negative emotions was greater than for positive emotions and that film was the most effective MIP. However the difference with other MIPs when inducing negative emotions was not very large. When inducing positive emotion this difference became much greater. The other important finding

Westermann et al. (1996) point out is that effects are smaller when subjects know what the purpose of the study or experiment is. Västfjäll (2001) reviewed a large body of research using music as a induction method. Although there are some problems with this method, the overall conclusion is that music is an effective method of inducing moods or emotions.

2.3.1 Emotion induction using music

Västfjäll (2001) points out that there are individual differences in how people react or respond to music, which is a problem when doing studies that use mood induction through music. When doing research using musical mood induction it is important to also consider listener features such as musical experience, traits, emotional state and context or

environment of the music being played. Moreover, when using musical mood induction it is important to distinguish between what the participant feels and what the music sounds like to them. I feel happy and the music sounds happy are different don’t have to necessarily occur at the same time. A study by Taruffi & Koelsch (2014) investigates the seemingly

contradicting link between the usually undesirable emotion sadness and peoples search and appreciation for sad music. Taruffi & Koelsch (2014) find that listening to sad music can be beneficial by regulating negative emotions and mood and also comfort people when feeling down. Surprisingly, feeling nostalgic was the most common emotion felt by listening to sad music rather than sadness itself and your memory was found to be the most important instrument to invoke sadness. Finally, people who are most influenced by listening to sad music were found to be people who are highly empathetic and emotionally unstable.

(11)

3. Methodology and hypotheses

The research question posed in the introduction will be answered by setting up a online experiment using the survey program Qualtrics. The participants were recruited for the experiment by distributing a link via various social media, word of mouth and survey sharing websites such as surveycircle. The experimental design that was used to construct the

experiment will now be explained along with the hypotheses.

3.1 Experimental design

At the beginning of the survey the participants were divided randomly into three groups: a happy, sad or neutral group. Depending on which treatment group a participant was placed in, the participant was induced by the corresponding emotion. The induction of the incidental emotion will be through playing a 3-4 min piece of happy, sad or neutral music. The music selected for the experiment is taken from Mitterschiffthaler et al. (2007). These authors performed a fMRI study where participants were asked to rate different musical pieces on a VAS scale from 0 (sad) trough neutral (50) to happy (100). The musical pieces with mean scores most closely matched with the emotion on the scale were chosen for this thesis study. Meaning that A little night music (Allegro) by Mozart was chosen for the happy condition since it had the highest mean score with 82.22. Suite for violin & orchestra A minor by Sinding was chosen for the sad music condition since it had the lowest mean rating with 29.52. Lastly, Claire de lune by Beethoven was chosen for the neutral group since its mean rating was 50.53. After listening to this piece of music the participants were asked to perform two tasks.

The first task is used to measure ambiguity preferences of the participants using a Multiple Price List (MPL) matching that of ‘Task T’ by Gneezy et al. (2015). In this task the participant is presented with two urns. Urn 1 has 50 white and 50 black balls. Urn 2 also has a total of 100 white and black balls, but the distribution of those balls is not known to the participant. The MPL is presented as twenty rows of decisions with different monetary amounts. The payoffs for choosing the first urn stay constant, while the monetary amount for the second urn increases with each row. The participants are then asked to indicate the point at which they want switch from Urn 1 to Urn 2. There is thus only one switch point possible for the participants to make a decision about. Gneezy et al. (2015) provide evidence that this way of presenting the MPL, with only one decision of where to switch, does not lead to significantly different outcomes from the possibility of presenting the MPL where the

(12)

participant would need to make a decision on every row on which Urn to draw from. The choice was therefore made to only provide one switch point to make it easier to analyze and compare the forthcoming data. It is also important to note that these amounts of money presented in the MPL are purely hypothetical and no actual money was paid out in the experiment. Paying out money in an experiment has been found to not necessarily improve the quality of collection data (Mason & Watts, 2009).

The second task presented a MPL of 2 lotteries, lottery A and lottery B, this time taken from Holt & Laury (2002). The purpose of this task is to determine the risk attitude of the participant. In this task the decision table contained ten different situations. Each situation showed a choice between the two lotteries. The situations are similar in situation, except that as you move down the table, the chances of the higher payoff option increase. The

participants were asked to indicate the point at which they would want to switch from one lottery A to lottery B, similar to the first task. After the tasks a series of questionnaires were added to assess the person's emotional state and reaction to certain situations or statements.

The first questionnaire to be included in this experiment was the The Oxford

Happiness Questionnaire (OHQ, Hills & Argyle, 2002). This questionnaire was derived from the The Oxford Happiness Inventory (OHI) as a more compact instrument, which in turn follows the design and format of the Beck Depression Inventory (BDI, Beck, Ward, Mendelson, Hock, & Erbaugh, 1961). The OHQ consists of 29 statements about happiness where the participant has to indicate how much he or she agrees or disagrees with the statement on a uniform six-point Likert scale. The revised version of the BDI (BDI-II) itself was also used in this experiment. The BDI-II consists of 21 statements relating to symptoms of depression such as sadness, feeling alone, loss of appetite and suicidal thoughts. Thirdly a questionnaire to measure respondents tendency to regulate their emotions was included by means of the Emotion Regulation Questionnaire (ERQ) (Gross & John, 2003). The ERQ distinguishes between two emotion regulation strategies: Cognitive Reappraisal and Expressive Suppression and these will receive a separate score.

The last block of the experiment consisted of some questions about the piece of music the participant listened to in the first block. They were asked, among other things, if they had heard the music before, how much they enjoyed the music and what kind of music they would categorize the music as, happy sad or neutral, similar to the method used by

Mitterschiffthaler et al. (2007). The participants were also asked how the music made them feel using the Self-Assessment Mannequin method (SAM, Bradley & Lang, 1994). The

(13)

experiment was concluded with a couple of demographic questions about age, gender and level of education.

3.2 Hypotheses

It is difficult to derive an overarching conclusion from the previously discussed literature about the effect of emotion on decision making. The literature does not provide a widely accepted theory or stance on how happiness or sadness affects risk and ambiguity

preferences. Therefore the hypotheses presented below are simply the representation of one of the presented views in literature. The view that is picked as reasoning for the hypotheses is that joy or happiness causes people to be more certain of themselves which can lead to

overestimating their chances when making risky decisions. This conclusion is mainly based on the literature from Schwarz (2000), Lerner & Keltner (2001) and Schulreich et al. (2014). Although the literature has shown that risk and ambiguity are processed by different regions of the brain (Krain et al., 2006, Huettel et al., 2006), there is little substantial evidence of the specific effect of happiness or sadness on ambiguous preferences. Therefore the ambiguity preferences are assumed to be subjected to the same reasoning as the risk preferences. The hypotheses here are presented as a comparison to the neutral song treatment.

Hypothesis 1: Happiness makes people less ambiguity and risk averse.

(14)

4. Results

STATA 15 will be used for the statistical analysis of the data gathered from the Qualtrics survey. The results will consist of a summary of statistics of both the ambiguity and risk task. Also the relation between the emotion induced and the results of the ambiguity and risk tasks will be looked at by comparing them to the neutral group results. To compare the means of the treatment groups, a one-way ANOVA will be used if the variable was normally

distributed. If a variable was found not to be normally distributed the nonparametric Kruskal-Wallis test will be performed. Various linear regression models (OLS) will also be

constructed using the data from the 2 tasks and the responses to the questionnaire.

4.1 Participants

The total number of participants that completely finished the survey was 35. Fourteen

participants were male (40%) and 21 participants were female (60%). One of the participants decided not to disclose their age. The average age of the remaining 34 participants was 35 years old (SD: 17.5). The youngest participant was 22 years old while the oldest was 85 years old. Most of the 35 participants (85.7%) had a university, college or equivalent qualification. Two participants (5.7%) had a intermediate between secondary level and university (e.g. technical training). Two participants (5.7%) had Secondary school as their highest

qualification and one person (2.9%) had primary school as their highest qualification. Lastly, 17 participants were students (48.6%), 6 participants were full- time employees (17.1%), 8 were part-time employees (22.9%), one was self-employed (2.9%), one was unemployed (2.9%) and the 2 remaining participants had been retired (5.7%).

4.2 Summary statistics

The ambiguity preferences of the participants can be deducted by looking at the results from the first task. The ambiguity preferences can be determined by looking at the switch point. If for example the participant switched at row 10 from Urn 1 to Urn 2 then he or she chose the risky option ten times and the ambiguous option also ten times. By converting this result into the share of ambiguous options chosen, compared to the total number of options, the

ambiguity preferences can be determined. So using the same example as before, 10 risky options out of a possible 20 give a share of (10/20) 0.5. If a person has a share lower than 0.25 he or she is classified as being ambiguity loving. A share 0.25 exactly is classified as ambiguity neutral and above 0.25 is ambiguity averse. The data show a overall mean of 0.56

(15)

and a standard deviation of 0.29. One person was ambiguity neutral (2,9%), two participants were ambiguity loving (5.7%), and the remaining 32 were all risk averse (91.4%). However a relatively large subsection of the ambiguity averse participants was only slightly risk averse, shown by the fact that 10 of the participants had a share of 0.3 as shown in figure 1. These percentages are not in line with findings in the literature on heterogeneity in ambiguity attitudes as reported in the literature section (Dimmock et al., 2016; Potamites & Zhang, 2012).

Figure 1: frequency graph ambiguity Figure 2: frequency graph risk

The data on ambiguity seems to be normally distributed from the figure, and this is also proven by doing a Shapiro-Wilk test for normality. In this case the p-value = 0.36838, which means that the hypothesis that the share of ambiguity avoiding choices is normally distributed cannot be rejected.

The risk preferences of the participants can be deducted by looking at the results from the second task. In this test the switch point again indicates the risk preference of the

participants. For example if a participant chose row 5 as their switch point their share of risk avoiding choices would be (5/10) 0.5. The risk task used in the survey classified participants with a score of 0.4 and lower as risk loving, a share of 0.4 exactly as risk neutral, and having a score higher than 0.4 risk averse. The data gathered show a overall mean of 0.63 and a standard deviation of 0.21. Figure 2 shows that that it is again likely that the results are normally distributed. Conducting another Shapiro-Wilk test for normality proves this because the p-value is 0.62719, which means that the hypothesis that the share of risk avoiding

choices is normally distributed cannot be rejected.

It is also very important to note that, the ‘sad’ and ‘neutral’ music clips used in the experiment were judged by the participants, on a similar VAS scale to Mitterschiffthaler et al.

(16)

(2007). Both clips were found to be around neutral with a mean score of 46.4, for the ‘neutral’ clip and a mean of 41.9 for the ‘sad’ song. This means the sad song used in this experiment was not considered to be particularly sad for the participants in that treatment. For the happy song the result was much clearer with a mean of 81.8, which indicates that the participants really did find this piece of music happy.

4.3 Results of the questionnaires

The Oxford happiness questionnaire (OHQ) answers were compiled and turned into a “happiness score” (Hills & Argyle, 2002). The questionnaire was presented as a six-point Likert scale: 1 = strongly disagree 2 = moderately disagree 3 = slightly disagree 4 = slightly agree 5 = moderately agree 6 = strongly agree

It is important to note however that some questions of the OHQ have to be scored in reverse. That means that if a participant answered strongly disagree to this particular question the score should be a 6 instead of a 1. (This modification concerns questions: 1, 5, 6, 10, 13, 14, 19, 23, 24, 27, 28, 29). After this modification of the results, the happiness score can be calculated by taking the sum of all points and then dividing by 29. According to Stephen Wright from Georgetown University’s Brain and Language Lab the score can be interpreted as follows: If a participant has a score between 1 to 2 he or she is unhappy, between 2 or 3 is somewhat unhappy, between 3 or 4 is neutral (not happy or unhappy), 4 is somewhat happy, between 4 and 5 is pretty happy, between 5 and 6 is very happy, and 6 is too happy (The Guardian, 2014). The results from the questionnaire show a mean happiness score of 4.19 with a standard deviation 0.51. This result indicates that the participants were generally pretty happy with their current situation. The happiness scores are normally distributed (Shapiro-Wilk test: p-value = 0.63068)

The BDI-II scores were also computed. In this case the scoring system used a 0–3 multiple choice scoring format. The score is simply calculated by taking the sum of all numbers, making the maximum score 63. The BDI-II indicates, as discussed before, if the person suffers from depression and how severe that depression is. A score of 0-13 indicates minimal depression, a score of 14-19 indicates mild depression 20-28 indicates moderate depression and a score of 29-63 is an indication for severe depression. The mean score for the 35 participants was 7.43 (SD: 6.6), indicating minimal depression. This result is again an

(17)

indicator that participants generally feel very comfortable with their general situation. The BDI-II scores are not normally distributed (Shapiro-Wilk test: p-value = 0.00152)

Finally, the Emotion Regulation Questionnaire (ERQ) scores were calculated. The ERQ was presented as a seven-point Likert scale going from 1 = strongly disagree to 7 = strongly agree. The scores were split into two emotion regulation scores: A cognitive reappraisal score and a expressive suppression score. The mean reappraisal score was 5.23 (SD: 0.79) and the mean suppression score was 3.74 (SD: 1.40). Both the reappraisal score (p-value=0.13276) and the suppression score (p-value=0.45136) were normally distributed.

4.4 Regression analyses

The data from the ambiguity task as was normally distributed, therefore a one way ANOVA was conducted to see if the mean ambiguity avoiding choices differed between the three groups. The results show that there is no significant difference between the three treatment groups (p-value = 0.2495).

The data can be further explored by performing a OLS regression analysis. The results of these regressions are presented in table 1. Models 1 through 4 show that the song played did not have a significant effect on the ambiguity score. This means that inducing a sad emotion or happy emotion through the music clip did not lead participant to make more ambiguity avoiding choices. The regressions do indicate however that there is an effect of the Happiness score (p-value = 0.087) and the ERQ reappraisal score (p-value = 0.075) on the ambiguity score in model 2 and solely of the ERQ reappraisal score (p-value = 0.069) in model 4. This could mean that participant who were generally happier and better able to change the way they think about potentially emotion-eliciting events were more likely to make ambiguity avoiding choices.

The data from the risk task as was also normally distributed, therefore a one way ANOVA was conducted again to see if the mean risk avoiding choices differed between the three groups. The results show that there is no significant difference between the three treatment groups (p-value = 0.2510)

(18)

Table 1 : Summary regression results for ambiguity (1) (2) (3) (4) Ambiguity score Ambiguity score Ambiguity score Ambiguity score

Happy Song Treatment 0.184 0.133 -0.377 0.045

(0.118) (0.121) (2.012) (0.163)

Sad Song Treatment 0.148 0.141 0.833 0.106

(0.115) (0.118) (2.115) (0.146)

Happiness score 0.232* 0.311 0.245

(0.131) (0.288) (0.167)

BDI score 0.008 -0.001 0.009

(0.010) (0.015) (0.013)

ERQ reappraisal score 0.120* 0.049 0.145*

(0.065) (0.093) (0.075)

ERQ suppression score 0.019 0.105 0.004

(0.038) (0.091) (0.043) Treatment*Happiness score Happy Song Sad Song -0.040 (0.390) -0.012 (0.365) Treatment*BDI score Happy Song 0.045 (0.031) Sad Song -0.005 (0.030) Treatment*ERQ reappraisal score Happy Song 0.127 (0.165) Sad Song -0.024 (0.202) Treatment*ERQ suppression score Happy Song -0.103 (0.085) Sad Song -0.058 (0.095) (0.121) Female 0.008 (0.136) Age 0.001 (0.006) Vocational education 0.398 (0.549) Secondary school 0.230 (0.465) University 0.328 (0.534) Full-time employment 0.058 (0.378) Part-time employment -0.096 (0.351) Self-employed 0.411 (0.567) Student 0.126 (0.334) Constant 0.461*** -1.257* -1.517 -1.790* Obs. 35 35 35 34 R-squared 0.083 0.290 0.460 0.471

Standard errors are in parenthesis *** p<0.01, ** p<0.05, * p<0.1

(19)

The data was then again further explored by performing a OLS regression analysis. The results of these regressions are presented in table 2. Models 1 through 4 show that the song played did not have a significant effect on the risk score. This means that inducing a sad emotion or happy emotion through the music clip did not lead participant to make more risk avoiding choices. This results contradict the results of Schulreich et al. (2014), who did that participants made more risky choices after listening to happy music. The regressions also show no significant effects of the OHQ, BDI, and ERQ scores on the risk preferences. The only significant outcome (at a 10% confidence level) is the apparent interaction effect of the happy song treatment and the BDI-II scores. This indicates that participants who had a higher BDI score in the happy song treatment made more risk averse choices.

(20)

Table 2: Summary regression results for risk

(1) (2) (3) (4)

riskscore riskscore riskscore riskscore

Happy Song Treatment 0.106 0.113 -0.726 0.137

(0.085) (0.095) (1.569) (0.136)

Sad Song Treatment -0.042 -0.049 0.644 -0.016

(0.082) (0.092) (1.650) (0.122)

Happiness score -0.112 -0.154 -0.112

(0.103) (0.225) (0.140)

BDI score -0.001 -0.010 0.000

(0.008) (0.012) (0.011)

ERQ reappraisal score 0.001 0.010 0.011

(0.051) (0.073) (0.063)

ERQ suppression score -0.001 0.064 -0.014

(0.030) (0.071) (0.036) Treatment*Happiness score Happy Song Sad Song 0.342 (0.304) 0.047 (0.285) Treatment*BDI score Happy Song 0.044* (0.024) Sad Song -0.007 (0.024) Treatment*ERQ reappraisal score Happy Song -0.095 (0.129) Sad Song -0.114 (0.158) Treatment*ERQ suppression score Happy Song -0.103 (0.085) Sad Song -0.058 (0.095) Female 0.010 (0.114) Age -0.006 (0.005) Vocational education -0.150 (0.459) Secondary school 0.040 (0.389) University -0.274 (0.446) Full-time employment 0.165 (0.316) Part-time employment -0.007 (0.293) Self-employed -0.198 (0.474) Student -0.082 (0.279) Constant 0.614*** 1.095* 1.517* 1.021 (0.055) (0.561) (0.844) (1.197) Obs. 35 35 35 34 R-squared 0.083 0.144 0.356 0.225

Standard errors are in parenthesis *** p<0.01, ** p<0.05, * p<0.1

(21)

To check if the type of treatment affected the questionnaire response, the different questionnaires were all tested against both the happy song and sad song treatment. The OHQ and the ERQ were both normally distributed so for those a one way ANOVA was conducted to see if the questionnaire scores differed between the three groups. The results show that there is no significant difference between the three treatment groups for the OHQ (p-value = 0.8302). This was also true for both of the ERQ scores with a p-value of 0.2277 for the reappraisal score and a p-value of 0.2477 for the suppression score. The BDI-II scores were not normally distributed as was discussed earlier. Therefore the non-parametric Kruskal-Wallis test was used to test if there are significant differences among group means. The test indicated that there was no significant difference between the means (p-value: 0.8727). The results of these tests are visualized in figure 3 below.

(22)

In every graph the confidence intervals show an overlap with each other, which means that there is no significant difference at this level between the mean values of the respective questionnaire scores across the treatment groups.

5. Discussion and conclusion

5.1 Summary of findings

The goal of this thesis was to find out if music induced emotion, in the form of happiness and sadness, would affect people’s ambiguity preferences. The effect of these music induced emotions on risk preferences was also included to compare the results to the previous literature. The results show that there is no valid reason to assume that ambiguity or risk preferences are affected in a significant way by the two music induced emotions used in this type of online experimental setting. The results are not in line with hypotheses presented in paragraph 3.2, which is particularly interesting for the risk preferences since there is much more evidence available from the literature that does seem to be a effect of induced emotion on risk preferences than there is for ambiguity preferences. There are conflicting results in the literature on how the effects works, but there seems to be a consensus that there at least is an effect, which is thus contradictory to the results from this experiment. However, there are a substantial amount of limitations on the interpretation of these findings.

5.2 Limitations

The first and most obvious limitation is the low number of respondents, which was 35. With the number of total respondents so low, the number of respondents per treatment group was obviously even lower. Having less than 15 respondents per treatment group is not much to come to any substantial conclusions to possible effects if there would have been any in this experiment. Therefore if this experiment were to repeated, recruiting more participants would increase the usefulness of the study.

Secondly, the ‘sad’ and ‘neutral’ music used was shown to be to closely matched, and was classified as being just neutral music. This makes searching for the impact of sadness on risk and ambiguity preferences almost impossible. A recommendation for further research will therefore be to do this in two stages. First asking people to rate several music pieces and then pick three to use in the second stage, which is similar to the experiment used in this thesis, using the same participants. This two stage procedure would however be very difficult to conduct using online distribution and is more viable in a personal setting.

(23)

Thirdly, the subjects for the majority guessed correctly that the music clip was intended to change their mood before making the tasks. Therefore there may be a demand effect of the experimenter resulting in these particular findings. However, these types of experiments using emotion induction are commonly known to have this limitation and are hard to overcome without deception.

The fourth possible limitation is that the emotions induced by the music may have faded out by the time the subjects reach the risk task, which can be caused by their focus on answering the ambiguity task or simply the time passed. To combat that the participant was given only one decision to make in both tasks and having the music play for a extended period of time.

Finally, there is the possibility that people may not completely understand the task and make their decisions randomly. This is hard to combat other then describing the tasks as clearly as possible and providing an example. The experiment was designed to be as concise and fast paced as possible, without being scruffy and incomplete. Listening to the music for multiple minutes while not doing anything else may bring boredom to some people but there is again not much that can be done about that, since that is an integral part of this experiment.

5.3 Conclusion and future research

This thesis shows that there is no valid reason to assume that ambiguity or risk preferences are affected in a significant way by the two music induced emotions, happiness and sadness used in this type of online experimental setting.

Recommendations for future research are to use a two-step procedure as described in the limitations section and preferably use a lab experiment to be able to have more control over the behavior of the subjects. Furthermore, a larger subject pool should be used and lastly other emotions may be tested to see if they are, in contrast to happiness and sadness, can change the risk and ambiguity preferences of people.

(24)

References

Anderson, C., & Galinsky, A. D. (2006). Power, optimism, and risk‐taking. European journal

of social psychology, 36(4), 511-536.

Baillon, A., Koellinger, P. D., & Treffers, T. (2016). Sadder but wiser: The effects of emotional states on ambiguity attitudes. Journal of Economic Psychology, 53, 67-82.

Bechara, A., Damasio, H., & Damasio, A. R. (2000). Emotion, decision making and the orbitofrontal cortex. Cerebral cortex, 10(3), 295-307.

Bechara, A. (2003). Risky business: emotion, decision-making, and addiction. Journal of

gambling studies, 19(1), 23-51.

Beck, T., Ward, C. H., Mendelson, M., Hock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 7, 158–216.

Bernoulli, D. (1738). Specimen theoriae novae de mensura sortis. Commentarii Academiae

Scientarum Imperialis Petropolitanae, 5, pp. 175-192

Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: the self-assessment manikin and the semantic differential. Journal of behavior therapy and experimental psychiatry, 25(1), 49-59.

Camerer, C., & Weber, M. (1992). Recent developments in modeling preferences: Uncertainty and ambiguity. Journal of risk and uncertainty, 5(4), 325-370.

Conte, A., Levati, M. V., & Nardi, C. (2018). Risk preferences and the role of emotions.

Economica, 85(338), 305-328.

Dimmock, S. G., Kouwenberg, R., Mitchell, O. S., & Peijnenburg, K. (2016). Ambiguity aversion and household portfolio choice puzzles: Empirical evidence. Journal of Financial

Economics, 119(3), 559-577.

Drichoutis, A. C. and Nayga Jr., R. M. (2013). Eliciting risk and time preferences under induced mood states. The Journal of Socio-Economics, 45:18–27.

Ellsberg, D. (1961). Risk, ambiguity, and the Savage axioms. Quarterly Journal of

Economics, 75, pp. 643-669

Gneezy, U., Imas, A., & List, J. (2015). Estimating individual ambiguity aversion: A simple

(25)

Goudie, R. J., Mukherjee, S. N., De Neve, J. E., Oswald, A. J., & Wu, S. (2011). Happiness as a driver of risk-avoiding behavior. University of Warwick-Department of Statistics.

Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes: implications for affect, relationships, and well-being. Journal of personality and social

psychology, 85(2), 348.

The Guardian (2014). Take the Oxford Happiness Questionnaire. Retrieved 12-7-2018 at:

https://www.theguardian.com/lifeandstyle/2014/nov/03/take-the-oxford-happiness-questionnaire

Huettel, S. A., Stowe, C. J., Gordon, E. M., Warner, B. T., & Platt, M. L. (2006). Neural signatures of economic preferences for risk and ambiguity. Neuron, 49(5), 765-775.

Hey, J. D., & Orme, C. (1994). Investigating generalizations of expected utility theory using experimental data. Econometrica: Journal of the Econometric Society, 1291-1326.

Hills, P., & Argyle, M. (2002). The Oxford Happiness Questionnaire: a compact scale for the measurement of psychological well-being. Personality and individual differences, 33(7), 1073-1082.

Holt, C. A., & Laury, S. K. (2002). Risk aversion and incentive effects. American economic

review, 92(5), 1644-1655.

Hsu M., Bhatt M., Adolphs R., Tranel D., & Camerer C.F. (2005). Neural systems

responding to degrees of uncertainty in human decision-making, Science, 310, 1680-1683

Krain, A. L., Wilson, A. M., Arbuckle, R., Castellanos, F. X., & Milham, M. P. (2006). Distinct neural mechanisms of risk and ambiguity: a meta-analysis of decision-making.

Neuroimage, 32(1), 477-484.

Ladinig, O., & Schellenberg, E. G. (2012). Liking unfamiliar music: Effects of felt emotion and individual differences. Psychology of Aesthetics, Creativity, and the Arts, 6(2), 146.

Leith, K. P., & Baumeister, R. F. (1996). Why do bad moods increase self-defeating behavior? Emotion, risk tasking, and self-regulation. Journal of personality and social

psychology, 71(6), 1250.

Lerner, J. S., & Keltner, D. (2000). Beyond valence: Toward a model of emotion-specific influences on judgement and choice. Cognition & Emotion, 14(4), 473-493.

(26)

Lerner, J. S., & Keltner, D. (2001). Fear, anger, and risk. Journal of personality and social

psychology, 81(1), 146.

Lerner, J. S., Li, Y., Valdesolo, P., & Kassam, K. S. (2015). Emotion and decision making.

Annual Review of Psychology, 66.

Loewenstein, G., & Schkade, D. (1999). Wouldn’t it be nice? Predicting future feelings. In D. Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: The foundations of hedonic

psychology (pp. 85±105). New York: Russell-Sage

Mason, W., & Watts, D. J. (2009, June). Financial incentives and the performance of crowds. In Proceedings of the ACM SIGKDD workshop on human computation (pp. 77-85). ACM.

Mitterschiffthaler, M. T., Fu, C. H., Dalton, J. A., Andrew, C. M., & Williams, S. C. (2007). A functional MRI study of happy and sad affective states induced by classical music. Human

brain mapping, 28(11), 1150-1162.

von Neumann, John and Oskar Morgenstern. (1947). Theory of Games and Economic

Behavior, 2nd ed. Princeton: Princeton University Press.

Potamites, E., & Zhang, B. (2012). Heterogeneous ambiguity attitudes: a field experiment among small-scale stock investors in China. Review of Economic Design, 1-21.

Savage, Leonard J. (1954). The Foundations of Statistics. New York: Wiley.

Seymour, B., & Dolan, R. (2008). Emotion, decision making, and the amygdala. Neuron,

58(5), 662-671.

Schulreich, S., Heussen, Y. G., Gerhardt, H., Mohr, P. N., Binkofski, F. C., Koelsch, S., & Heekeren, H. R. (2014). Music-evoked incidental happiness modulates probability weighting during risky lottery choices. Frontiers in psychology, 4, 981.

Schwarz, N. (2000). Emotion, cognition, and decision making. Cognition & Emotion, 14(4), 433-440.

Taruffi, L., & Koelsch, S. (2014). The paradox of music-evoked sadness: An online survey.

PLoS One, 9(10), e110490.

Västfjäll, D. (2001). Emotion induction through music: A review of the musical mood induction procedure. Musicae Scientiae, 5(1_suppl), 173-211.

(27)

Westermann, R., Stahl, G. U. N. T. E. R., & Hesse, F. (1996). Relative effectiveness and validity of mood induction procedures: analysis. European Journal of social psychology, 26(4), 557-580.

Williams, S., Zainuba, M., and Jackson, R. (2003). Affective influences on risk perceptions and risk intention. Journal of Managerial Psychology, 18(2):126–137.

Yuen, K. S., & Lee, T. M. (2003). Could mood state affect risk-taking decisions?. Journal of

affective disorders, 75(1), 11-18.

Appendix

Screen 1

Dear participant,

Thank you for participating in this survey on decision making, which is part of a Master Thesis in Economics at the University of Amsterdam.

This survey can be completed on your laptop, tablet, computer or smartphone. However, it is necessary that you have a stable and strong internet connection.

The survey will consist of three parts:

Part 1: A excerpt of a piece of music will be played, so please make sure you have access to headphones.

Part 2: You will be asked to make decisions in 2 tasks. The decisions you make in this part

of the survey result in hypothetical payments.

Part 3: the final part of the survey will consist of some questions about music and your own

preferences

We would like to ask you to finish the whole study in one sitting. Once you have started the survey by clicking next, please fill out all questions until you are informed that you are finished.

Please read all questions carefully and concentrate on the study. We kindly ask you to stay

fully focused on the task and not do anything besides the questionnaire. In order to minimize

distractions, please turn off your cell phone notifications, close all chat windows and any other media.

Your data will be treated anonymously and will not be passed on to third parties.

If you consent to this and have time to complete the survey attentively in one sitting (15 min), you can start by clicking "continue".

(28)

Screen 2a (shown to subjects in happy song group).

In this part of the survey you will be presented with a musical excerpt called: Wolfgang Amadeus Mozart - A Little Night Music (Allegro)

Please make sure your speakers or headphones are enabled in order to hear the audio. Please allow yourself to not be distracted while listening by closing your eyes. Only after the music is finished you are allowed to click continue. This button will therefore only appear on this page near the end of the song.

Please click here to listen to the music:

Screen 2b (shown to subjects in sad song group).

In this part of the survey you will be presented with a musical excerpt called: Sinding - Suite in A minor

Screen 2c (shown to subjects in neutral song group).

In this part of the survey you will be presented with a musical excerpt called: Beethoven - Sonate au Clair de Lune

Screen 3

This is the end of part 1.

(29)

Screen 4

Task 1

In this task, a decision table is presented below with 20 different situations. Each situation shows you a choice between two Urns: Urn 1 or Urn 2.

Urn 1 contains 50 white balls and 50 black balls

(30)

Screen 4 continued

How to read this table:

Look at the first situation at the top of the table.

If you choose to draw from Urn 1 your payoff for drawing a black ball is $2 if the ball drawn is white you earn $0. The chance that you draw a black ball is 50% and the chance the ball is white is also 50%.

(31)

Screen 4 continued

If you choose Urn 2 your payoff for drawing a black ball is $1.64 if the ball drawn is white you earn $0. The chance you draw a black ball or a white ball is unknown.

The other situations are similar, except that as you move down the table, the payoff possibilty of Urn 2 increases.

Please indicate at which decision you would like to switch from Urn 1 to Urn 2 by by selecting the corresponding number below.

Example: If you choose number 3 as your switch point you will choose Urn 1 for choices 1 + 2 and Urn 2 for choices 3-20. You can thus only switch once.

(32)

Screen 5

Task 2

In this task, a decision table is presented below with 10 different situations. Each situation shows

you a choice between two lotteries: Lottery A or Lottery B.

How to read this table:

Look at the first situation at the top of the table. Lottery A pays $2 with 10% and $1.60 with 90% chance. Lottery B yields $3.85 with 10% and $0.10 with 90% chance. The other situations are similar, except that as you move down the table, the chances of the higher payoff option increase.

Please indicate at which number you would like to switch from Option A to Option B by selecting the corresponding number below.

Example: If you choose number 3 as your switch point you will choose Lottery A for choices 1 + 2 and Lottery B for choices 3-10. You can thus only switch once.

(33)

Screen 6

This is the end of the decision making tasks.

The last part of this questionnaire will consist of some questions about yourself.

Screen 7

Below are a number of statements about happiness. Would you please indicate how much you agree or disagree with each by selecting the corresponding option.

You will need to read the statements carefully because some are phrased positively and others negatively. Don’t take too long over individual questions; there are no ‘right’ or ‘wrong’ answers and no trick questions. The first answer that comes into your head is probably the right one for you. If you find some of the questions difficult, please give the answer that is true for you in general or for most of the time.

(34)

Screen 7 continued

Screen 8

Below are 21 groups of statements. Please read each group of statements carefully, and then pick out the one statement in each group that best describes the way you have been feeling during the past two weeks, including today. If several statements in the group seem to apply equally well, tick the highest number for that group. You can only choose one statement for each group.

Sadness

0. I do not feel sad

1. I feel sad much of the time 2. I am sad all the time

(35)

Screen 8 continued

Pessimism

0. I am not discouraged about my future

1. I feel more discouraged about my future than I used to be 2. I do not expect things to work out for me

3. I feel my future is hopeless and will only get worse

Past Failure

0. I do not feel like a failure

1. I have failed more than I should have 2. As I look back, I see a lot of failures 3. I feel I am a total failure as a person

Loss of Pleasure

0. I get as much pleasure as I ever did from the things I enjoy 1. I don't enjoy things as much as I used to

2. I get very little pleasure from the things I used to enjoy 3. I can't get any pleasure from the things I used to enjoy

Guilty Feelings

0. I don't feel particularly guilty

1. I feel guilty over many things I have done or should have done 2. I feel guilty most of the time

3. I feel guilty all of the time

Punishment Feelings

0. I don't feel I am being punished 1. I feel I may be punished 2. I expect to be punished 3. I feel I am being punished

Self-Dislike

0. I feel the same about myself as ever 1. I have lost confidence in myself 2. I am disappointed in myself 3. I dislike myself

Self-Criticalness

0. I don't criticize or blame myself more than usual 1. I am more critical of myself than I used to be 2. I criticize myself for all of my faults

(36)

Screen 8 continued

Suicidal Thoughts or Wishes

0. I don't have any thoughts of killing myself

1. I have thoughts of killing myself, but I would not carry them out 2. I would like to kill myself

3. I would kill myself if I had the chance

Crying

0. I don't cry any more than I used to 1. I cry more than I used to

2. I cry over every little thing 3. I feel like crying, but I can't

Agitation

0. I am no more restless or wound up than usual 1. I feel more restless or wound up than usual 2. I am so restless or agitated that is hard to stay still

3. I am so restless or agitated that I have to keep moving or doing something

Loss of Interest

0. I have not lost interest in other people or activities 1. I am less interested in other people or things than before 2. I have lost most of my interest in other people or things 3. It's hard to get interested in anything

Indecisiveness

0. I make decisions about as well as ever

1. I find it more difficult to make decisions than usual

2. I have much greater difficulty in making decisions than I used to 3. I have trouble making any decisions

Worthlessness

0. I do not feel I am worthless

1. I don't consider myself as worthwhile and useful as I used to 2. I feel more worthless as compared to other people

3. I feel utterly worthless

Loss of Energy

0. I have as much energy as ever

1. I have less energy than I used to have 2. I don't have enough energy to do very much 3. I don't have enough energy to do anything

(37)

Screen 8 continued

Changes in Sleeping Pattern

0. I have not experienced any change in my sleeping pattern 1a. I sleep somewhat more than usual

1b. I sleep somewhat less than usual 2a. I sleep a lot more than usual 2b. I sleep a lot less than usual 3a. I sleep most of the day

3b. I wake up 1-2 hours early and I can't get back to sleep

Irritability

0. I am no more irritable than usual 1. I am more irritable than usual 2. I am much more irritable than usual 3. I am irritable all over the time

Changes in Appetite

0. I have not experienced any change in my appetite 1a. My appetite is somewhat less than usual

1b. My appetite is somewhat greater than usual 2a. My appetite is much less than before 2b. My appetite is much greater than usual 3a. I have no appetite at all

3b. I crave food all the time

Concentration Difficulty

0. I can concentrate as well as ever 1. I can't concentrate as well as usual

2. It's hard to keep my mind on anything for very long 3. I find I can't concentrate on anything

Tiredness or Fatigue

0. I am no more tired or fatigued than usual

1. I get more tired or fatigued more easily than usual

2. I am too tired or fatigued to do a lot of the things I used to do 3. I am too tired or fatigued to do most of the things I used to do

Loss of Interest in Sex

0. I have not noticed any recent change in my interest in sex 1. I am less interested in sex than I used to be

2. I am much less interested in sex now 3. I have lost interest in sex completely

(38)

Screen 9

Below you will see a number of statements. Read each statement and then choose the option

that you agree with the most. There is no right or wrong answer. Do not spend too much

time on any one statement; the first answer that comes to your mind is probably the best. Some of statements may seem similar to one another, but they differ in important ways.

(39)

(40)

(41)

(42)

Screen 10 continued

(43)

(44)

(45)

Screen 12

What is your age?

________________________________________________________________

What is your gender? Male

Female

What is your highest qualification? Primary school only (or less) Secondary school

Intermediate between secondary level and university (e.g. technical training) University or college or equivalent

What is your current employment status? Full-time employment Part-time employment Unemployed Self-employed Student Home-maker Retired

The impact of music on uncertain decision making : an online mood induction experiment