‘I feel depressed because it is raining again’ or ‘I feel so happy because the sun is shining’. Does this sound familiar? Both are expressions, which indicate that people believe the weather could influence their humor and/or emotions. Does indeed the weather influence emotions? Is the weather able to influence certain decisions individuals make in their life? According to Simenon (2016) decisions are made based on brain development, experiences and activities. The amount of decisions an individual makes each day varies. The average amount of conscious decisions an adult makes every day is about 35,000 (Simenon, 2016). All these decisions have different consequences, varying from positive to neutral or negative. How or why individuals make certain decisions is dependent on different aspects such as personality, attitudes, circumstances, experiences and social pressure (Soane & Chmiel, 2005; Klein, 1999). In this research the main focus will be on the two factors attitudes and circumstances (in the sense of weather circumstances). Personality and other background variables of the individuals will be taken into account as well but other aspects are not included in this study.
Defining attitudes is difficult therefore many different definitions exist. In general, attitudes are defined as the way people tend to evaluate things which is affected by what they have learned to believe about the world, their selves and others, what they have learned to like or dislike, and how they have learned to respond to people, things and situations (ACS Distance Education, n.d.). A well-known attitude is the risk attitude of people in different situations and so a lot of research is already done about this attitude. A close linked attitude is the ambiguity attitude of people. The difference between the two is as follow: a situation with objectively known probabilities is called risky and a situation with unknown or uncertain probabilities is called ambiguous (Ellsberg, 1961). Real-life decisions are often characterized by unknown probability information (Brachinger et al., 2000) so that makes ambiguity attitudes interesting to investigate. The more is known about ambiguity attitudes and what can influence this attitude, the more is known about why people make certain decisions in their daily life and how the weather can influence these decisions. In the previous paragraph mentioned the weather could influence peoples’ mood so interesting to investigate if it also influence an individuals’ ambiguity attitude. Therefore the research question in this study will be: Does the weather influence ambiguity attitudes?
To find answers on the questions, literature research is used. Searching for relevant studies that provide information that could be interesting for the present thesis. Important focus is to find in which way(s) the weather could influence the ambiguity attitudes of individuals. Weather is a really broad concept so more specific weather conditions need to be defined. Only the weather conditions that seem to have an effect on the ambiguity attitudes of individuals will be used in the thesis. If all the information about weather and ambiguity attitudes is found in previous research the expectations need to be tested on a sample to see if they are correct or not.
After the introduction, chapter 2 will start with a more detailed explanation and information of the important concepts that are involved. Furthermore, the effect of different weather conditions on ambiguity attitudes will be described. Chapter 3 describes which data will be tested and how the data will be collected. Chapter 4 will consists all the methods that will be used to test the hypotheses as well a description of the results. The conclusion of the research will be discussed in chapter 5 and in the last chapter (6) the discussion and limitations will be described.
Decisions individuals make in their daily life are mostly uncertainty decisions because only vague information about the probabilities and the potential outcomes is available (Trautmann & Kuilen, 2014). A situation with unknown or uncertain probabilities and outcomes are called ambiguous (Ellsberg, 1961). According to Frisch and Baron (1988) ambiguity is the subjective perception of missing information. Also conflicting information can lead to ambiguity when people find it difficult to combine different types of information or to adapt information that they access from different sources (Cabantous, 2007; Einhorn, Hillel and Hogarth, 1985; Viscusi and Magat, 1992). All the different definitions of ambiguity (attitudes) have one thing in common: situations with one or more unknown factor(s). The attitudes individuals have towards ambiguity depend of different things like the likelihood of the uncertain situation, the source that generates the uncertainty, and the domain of the outcome (Einhorn and Hogarth, 1985).
In ambiguous situations people can react in two different ways because ambiguity consist of two dimensions. First, the motivational dimension, liking or disliking of ambiguity. This dimension is also called ambiguity aversion. If people are ambiguity averse they prefer risks (known probabilities) relative to uncertainty (unknown probabilities) (Dimmock, Kouwenberg and Wakker, 2015). The opposite or negative of ambiguity aversion is ambiguity seeking. For high likelihood situations most people are ambiguity averse and for low likelihood situations people are ambiguity seeking (Dimmock, et al., 2013).
The second dimension of ambiguity attitudes is the cognitive dimension, a-insensitivity. If individuals have a high a-insensitivity they tend to treat all ambiguous situations as a 50-50 gamble (Abdellaoui et al., 2011). This dimension measures how much a person can distinguish between different ambiguous situations. Low likelihood ambiguous outcomes are overweighed, and high likelihood ambiguous outcomes are underweighted (Dimmock, et al., 2013). Therefore these individuals are insensitive for normal signals but oversensitive for extreme signals (Dimmock, Krouwenberg, & Wakker, 2015). Due to loss aversion, the effect of overweighting will be stronger for the worst extreme outcome than for the best extreme outcome (Dimmock, Krouwenberg, & Wakker, 2015).
In this thesis both dimensions of ambiguity attitudes will be used to see if the weather has an influences on these two attitudes separately.
In this paragraph the term mood will be explained. Mood falls within the theoretical realm of ‘affect’, which can be defined as ‘the specific quality of goodness or badness someone experience as a feeling state (with or without consciousness) and demarcating a positive or negative quality of a stimulus’ (Slovic et al., 2004). According to Sizer (2000, 762), ‘Moods affect a wide range of our thoughts, feelings, and attitudes in ways that are not constrained by subject matter or inferential rules’. The specific mood of a person can influence in what way people remember things. It can also effect how people deal with social information. A positive mood decreases strenuous processing (Clark & Isen, 1982; Sinclair & Mark, 1992), while a negative mood increases strenuous and vigilant processing (Schwarz, 1990; Schwarz & Bless, 1991). Negative moods have a bottom-up processing; they focus on the details of the external world (Forgas, Goldenberg & Unkelback, 2009). In contrast, positive moods have a top-down processing and have a greater reliance on existing schematic knowledge and heuristics (Bless, 2000; Fiedler, 2001). Bottom-up processing is automatic, fast and non-volitional and generated through what is going on in the external world. This means it is an unconscious process. In contrast, top-down processing is controlled, slow and volitional and driven by inner processes so it is a consciousness process (Ramsøy, 2015).
2.3 Influences of daily weather on mood
Ambiguity aversion has been the subject of a large number of studies in psychology, economics, biology, neuroscience, and philosophy (Trautmann & Kuilen, 2014). From previous research it is know that the weather does influence the mood of a person (Grohol, 2016). Mood states are quite ephemeral and can easily be influenced by little things (Isen et al., 1982). For example, the majority of people think they feel happier on days with a lot of sunshine and miserable on days that are dark with a lot of rain (Denissen et al, 2008). So their mood is influenced by the weather of that moment or day.
When studying the association between daily weather and mood, seasonal influences of weather must be distinguished from day-to-day influences. This is important because mood reactions to day-to-day weather fluctuations may not generalize to reactions to seasonal weather fluctuations, and vice versa (Denissen et al., 2008). In this thesis the daily weather is used and the seasonal weather fluctuations are left out. Several studies already investigate the effect of daily weather on people’s mood. First, Keller et al. (2005) investigated the effect of temperature and barometric pressure on single-occasion explicit and implicit mood valence (positive and negative mood) and cognition (memory and cognitive style) (Denissen et al., 2008). Denissen et al. (2008) found no main effect of these two weather parameters on mood but they do found a moderator effect of season and the time participants spend in the open air. So on spring days when people spent a lot of time outside, their mood was positively associated with air temperature and also the barometric pressure was positively associated with mood. But on summer days when people spend more time outside on hot days it has a negative correlation and was associated with a decreased mood (Denissen et al., 2008). Second, Watson (2000) collected diary reports of subjects during the fall and spring between the years 1985 and 1993. He focused on the amount of sunshine and rain on a day, but found no consistent effect on the daily mood of people. To see the effect of sunshine on the mood of people he also investigated the difference of the effect of the amount of sunshine on mood. So he compared days with 0% sunshine with days with 100% sunshine and found that sunshine only influences the arousal (intensity) of the mood but not the valence (Denissen et al., 2008). Third, according to Bassi et al. (2013), feelings of joviality, self-assurance, and attentiveness show a statistically significant increase during good weather conditions, and are associated with greater risk tolerance. The research discussed above provided measures of different weather parameters and effected different types of mood but most showed that weather in general can have an influence on the mood of people.
The concept ‘daily weather’ is really broad; therefore this needs to be specified. Different researches are discussed to find out what the term ‘daily weather’ can exist of. A previous study on weather and its relation with psychological constructs took only one or two weather parameters into account (Bushman, Wang and Anderson, 2005; Keller et al., 2005). Though, it is important to examine a wide variety of weather parameters and, differentiate the effect of each parameter (Denissen et al., 2008). In the study of Denissen et al. (2008) the effect of six different weather parameters (temperature, wind power, precipitation, sunlight, photoperiod and air pressure) on three different mood conditions (positive affect, negative affect and tiredness) was investigated. No significant effect of all the weather parameters on positive mood situation was found. For the negative mood situation a significant positive effect of temperature and a negative effect of wind power and sunlight was found. For the last mood situation, tiredness, a significant negative effect of sunlight was found.
Kliger and Levy (2003) also investigated the influence of mood in risk perception correlated with a weather parameter. They used the cloud coverage to control for the weather. Data reported by the National Climatic Data Center on a scale from 0 to 10, where 10 indicates total cloud coverage, is used. Almost all differences were found in comparing the two extreme cloud coverage groups. People were less risk tolerant under pleasurable weather conditions (no clouds or just a few clouds) and more risk tolerant during unpleasant weather conditions (overcast).
Bassi et al. (2013) found experimental evidence that sunlight and good weather have a positive impact on risk-taking behavior. So people become more risk tolerant on sunny days. They used three definitions for the quality of the weather: the amount of sunlight (objective weather condition), subjective weather conditions of people and the precipitation on a day. In order to objectively categorize weather conditions, they collected data on how many times the sky was clear, partly cloudy and overcast. They defined good weather as a day in which the sky was clear for the majority of the time (50% of the time) and so the sun was shining for more than 50% of the time (Bassi et al., 2013). The subjective weather condition was a questionnaire to analyze the perceived quality of weather. The last one measured was the precipitation in a day. A day, in which the amount of rainfall exceeds the daily average amount in that area, is defined as a rainy day or in other words a bad weather day (Bassi et al., 2013).
In the present thesis the same definition of Bassi et al. (2013) for good and bad weather will be used because only these authors provided definitions of both good and bad weather. Only subjective weather conditions can be measured because there is no possibility to do a questionnaire with the same participants. In the thesis various weather parameters will be used to find an answer on the research question. Since it can be concluded from the discussed research above that daily weather is not just one component but consist of several weather parameters.
2.4 Influence of mood on risk attitudes
Risk attitude is the attitude an individual has in situations with known probabilities and outcomes. In this paragraph the information that is already know about the effect of mood on risk attitudes will be discussed. According to various research a positive mood is expected to increase the risk-taking tendency (Isen, 1997; Isen et al., 1982; Nygren et al., 1996). The Affect Infusion Model (AIM) states that a positive mood is expected to increase the risk tolerance and a negative mood should lower the risk tolerance (Forgas, 1995). When people are in a good mood they focus on positive features of the situation while a bad mood shifts one’s attention to the negative cues in the environment (Grable and Roszkowski, 2008). In contrast, the Mood Maintenance Hypothesis (MMH) expects that a good mood will lead to caution and a bad mood will foster greater foolhardiness (Isen & Labroo, 2003; Isen & Patrick, 1983). People in a good mood want to stay in that mood, so they do not take risky decisions that could result in losses and would shift them into a bad mood. But when people are in a bad mood they will take risks in the hopes of taking a change and obtaining a reward, which would give them a good mood again (Grable and Roszkowski, 2008).
There are several researches that support the AIM. Wright and Bower (1992) found that people in a happy mood tend to be more optimistic and optimistic people are more likely to report higher probabilities for positive risk events and lower probabilities for negative risk events. They found that mood states have a greater influence on judging events that were less frequent (Grable and Roszkowski, 2008). Sizer (2000) added that people might be less cautious when in a happy mood because positive moods are associated with wide information focusing and reduces concentration on details. Grable and Roszkowski (2008) investigated the effect of mood on the risk tolerance and found that being in a happy mood was positively associated with having a higher level of financial risk tolerance as compared to a neutral mood. Furthermore, people who are in a bad mood have a lower risk tolerance compared to a neutral mood, but this effect was extremely small and not significant. So this research suggests that a positive mood has greater bearing on risk tolerance than people in a negative mood (Grable and Roszkowski, 2008). Another proof for the theorem that people in a good mood take more risky decisions is the research of Yuen and Lee (2002). They found that people in a depress mood (bad mood) would have a lower willingness to take risks than people in a neutral or in a positive mood.
Less evidence is found in the literature for the MMH. Kliger and Levy (2003) found that in real capital market decisions, investors were less risk tolerant under pleasant weather conditions (good mood) and more risk tolerant during unpleasant weather conditions (bad mood).
In the present thesis it is not clear yet which of the two theories hold for ambiguity attitudes. Expected also the AIM because more evidence is found for this theory with risk attitudes but after the research of the data and the hypotheses a conclusion can be drawn in a later chapter.
2.5 Correlation between risk and ambiguity aversion
Having discussed the effect of daily weather on mood and the effect mood has on the risk attitudes of individuals; this research needs to explain the effect on ambiguity attitudes. But almost all studies about the relationship between risk and ambiguity report evidence of a positive correlation between risk aversion and ambiguity aversion. Like in the study of Charness and Gneezy (2010) they report that ambiguity seekers hold more risk portfolios and Kocher and Trautmann (2013) found that participants in ambiguous markets are more risk seeking than those in a risky market. So ambiguity aversion displays similar characteristics to risk aversion, but the effect is in a stronger extent (Trautmann & Kuilen, 2014). However, there are also some studies that find less or no positive correlation between risk and ambiguity aversion and Lauriola and Levin (2001) find only evidence of the positive correlation in the domain of losses. Akay et al. (2012), Cubitt et al. (2012), and Sutter et al. (2013) found a negative correlation between risk and ambiguity aversion. So not fully one conclusion can be drawn but most of the studies have overall evidence there is a positive correlation between risk and ambiguity aversion. In the present thesis it is therefore stated that there is a positive correlation between risk and ambiguity aversion. This information is useful to predict later in this research the effect of weather and ambiguity attitudes.
2.6 Weather influences ambiguity attitudes
As explained previously, there are two different dimensions of ambiguity: ambiguity aversion and a-insensitivity. These different ambiguity attitudes can also react different on situations and individuals. So there is the possibility they also react different on various weather conditions. Stated there is a positive correlation between ambiguity aversion and risk aversion. This is the last link that is needed to establish the relationship with the weather and ambiguity aversion. For clarity, the link is as follows: the weather influences the mood of individuals → the mood influences their decisions and risk aversion → a positive relationship between risk aversion and ambiguity aversion. In paragraph 2.4 already the relationship between mood and risk aversion is explained. Because there is a positive correlation between risk and ambiguity aversion, the same relationship between mood and ambiguity aversion holds.
Below hypotheses 1 and 2 are created with the information that is explained in the previous paragraphs and have to do with the first dimension, ambiguity aversion. Creating the hypotheses based on previous research shows bad weather/bad mood makes people more risk averse. But Baillon et al. (2016) had found that bad weather makes people more ambiguity neutral. They found that people who are in a sad mood make wiser decisions because of enhanced information processing. This means they are more ambiguity neutral compared with reduced information processing when they feel happy. This is the only research with an immediate effect on ambiguity attitudes so needs to be taken into account when creating the hypotheses. The only thing that is known from all the different research is that bad weather has an influence on the ambiguity aversion attitude of individuals but not in which strength. Therefore this will be tested in this thesis.
H1a: People are more ambiguity seeking when the weather is good.
H1b: People are more ambiguity seeking when the wind power is high.
H2a: Bad weather has an influence on the ambiguity aversion of a person.
H2b: People are more ambiguity averse when the temperature is high.
For the second dimension, no direct evidence is found that risk is correlated with the a-insensitivity. A-insensitivity is the cognitive dimension of the ambiguity attitudes so the decisions people make when doing cognitive tasks or using their cognitive skills. Decisions making under ambiguity are influenced by personal judgments and confidence, which may be influenced by emotional states (Bower, 1981; Forgas, 1995; Schwarz & Clore, 1983). A research of Cao and Wei (2005) has shown that temperature significantly affects mood, and mood changes in turn cause behavioral changes. They found evidence that fluctuations in the weather can have a reaction on the stock returns in the financial market (Cao & Wei, 2005; Hirshleifer & Shumway, 2003; Saunders, 1993). Cao and Wei (2005) also found that a lower temperature is related with a higher stock return because people are more aggressive and a high temperature is related with a higher or lower stock return because people can be aggressive but also apathy. In another research of Hirsleifer and Shumway (2003) in the financial market, they found that good weather is strongly significant correlated with stock returns. So less cloud coverage is associated with higher stock returns. The research of Baillon et al. (2016) can also be used to predict the effect of weather on the a-insensitivity because they mention that when people are sad they are more focused so probably these people make a more cognitive decision. Therefore this can also have a small influence of the understanding of ambiguous situations. These three research imply that weather can influence decisions of cognitive tasks or skills and therefore result in the following hypotheses:
H3a: Good weather influences a-insensitivity.
H3b: Temperature influences a-insensitivity.
H3c: Bad weather creates a neutral a-insensitivity.
2.7 Effect of gender and age on ambiguity attitudes
A lot of research can be found about the differences in risk attitudes of males and females. In general, females react more risk averse compared to males (Brachinger et al., 2000). This is generally known for risk attitudes. It is already stated that there is a positive correlation between risk and ambiguity aversion. It could therefore simply be said that females are also more ambiguity averse compared to males. However, first some research is analysed about the relationship of gender and ambiguity attitudes. Borghans et al. (2009) did a research about ambiguity attitude differences in gender and found that females are more ambiguity averse than men in the investment context, but the other way around in the insurance context. So according to that research there is not one conclusion that females or males are more ambiguity averse (Borghans et al., 2009). Because there is a positive correlation between risk and ambiguity aversion in that research and decisions in financial markets denote to cognitive tasks or skill, the same theorem is stated for both ambiguity attitudes. Females are more ambiguity averse and a-insensitive than males. This results in hypotheses four.
Furthermore, Fehr et al. (2007) found a gender difference in the impact of mood on decisions. Females in a good mood assigned higher subjective probability weights under gain and loss scenarios, consistent with AIM. In contrast, males were not influences by good mood. To research this, hypotheses five is formed.
H4: The effect of good weather is stronger on the ambiguity aversion of females than of males.
H5: Good weather influences a-insensitivity for females.
The age of people can also influence the different approach of taking decisions in different mood states. Young people have the tendency to focus more on the negative rather than the positive aspects in a certain situation (Chou et al., 2007). Young people are defined as teenagers (between the ages of 13 and 19 years old) and young adults (between the ages of 18 and 32 years old). In another research it is found that elderly are more sensitive for different weather conditions (Kööts, Realo, & Allik, 2011). They also found that there is no difference in risk taking among young people between positive versus neutral moods, but there is a difference between the negative and neutral states (Chou et al., 2007). The opposite occurred among the elderly. According to Roebuck (1979) the general definition of an old person is any person of 50 years and older. The difference in risk taking was greater between the positive and the neutral mood states than between the neutral and the negative mood states. But when the neutral point is neglected, then for both the young and old people, greater risk taking was evident among those in happy mood than those in a sad mood. In order to test these assumptions, the following hypotheses are formed:
H6a: Weather has a bigger impact on the ambiguity aversion of elderly compared to middle age and young people.
H6b: Bad weather increases more the ambiguity aversion of young people than of middle aged and elderly.
2.8 The moderator effect of personality
Personality will be examined as well to understand if this has an effect on the sensitivity to weather. Personality can be seen as a moderator between daily weather and mood. It has already been discussed that mood can have an effect on the ambiguity attitudes of persons. In many researches there is no significant effect found between personality traits and the sensitivity of weather (Denissen et al., (2008). But there are some studies that suggest a link between seasonality and personality. In a study of Murray et al. (1995) it is found that the personality trait neuroticism is relevant to Seasonal Affective Disorder (SAD). Seasonality and SAD are a type of depression that is related to changes in seasons. Most of the time the symptoms start in the fall and continue into the winter. The symptoms people experience are tiredness and moodiness. Also the study of Ennis & McConville (2004) found that an increased level of the neurotic personality trait is associated with more profound seasonal disturbances in mood and behavior. So it is interesting to examine whether there is a link between sensitivity to daily weather and personality.
To adopt an approach to measure personality between daily weather and ambiguity attitudes, personality will be estimate at the broad level of the Five Factor Model (FFM) or also called “Big 5 personality traits” (Cherry, 2016). The five broad personality traits described by the theory are extraversion, agreeableness, openness, conscientiousness and neuroticism.
• Extraversion: positive emotions, excitability, assertiveness and sociability. High extraversion people are often characterized, as attention-seeking and low extraversion people are more reserved.
• Agreeableness: friendly, affection and trust. People who are high agreeableness are more cooperative and can be seen as naive. Low agreeableness people are often more competitive and manipulative.
• Openness (to experience): curious and a broad range of interests. People with a high openness are more adventurous, creative and pursue self-actualization. Low openness people are more pragmatic and traditional (close minded).
• Conscientiousness: efficient, easy-going and thoughtful. People who are high conscientiousness prefer planned rather than spontaneous behavior and more obsessive. Low conscientiousness people are flexible and spontaneous.
• Neuroticism: nervous, sadness and emotional instable. People with a high neuroticism experience mood swings, irritability and sadness. People with a low neuroticism are more stable and have often more dynamic persons.
In the present thesis only the personality trait neuroticism will be taken into account because this is the only trait where evidence is found that it can have an influence on the ambiguity aversion attitudes. The following hypothesis will test this assumption:
H7: The effect of weather on ambiguity aversion index is increasing in the degree of neuroticism.