Unraveling the personality trait behind
sustainable consumer behavior
a quantitative research
Erik van Gorsel (10655379)
Bachelor thesis
Universiteit van Amsterdam
Supervisor : Lita Astuti Napitupulu
24-06-2019
Abstract 3
Introduction 4
Theoretical framework 6
Environmental awareness 6
Big five personality traits 9
Methodology 13
Design, sample and procedure 14
Measurements 14
Analytical plan 16
Recoding 16
Results 18
Means, SD and Correlations 18
Testing existing literature 19
Testing the personality traits 20
Discussion 22 Findings 22 Limitations 23 Further research 24 Conclusion 24 Appendix 27
Appendix 1 - survey questions 27
Abstract
For this research a quantitative study was conducted to uncover whether the presence of one of the big five personality traits can cause a person to participate in more or less sustainable consumer behavior. Existing literature state that one’s actual commitment of buying green products is mostly determined by environmental knowledge, affect and verbal commitment. Affect hereby measuring the degree of emotionality relatedness to environmental issues Four hypotheses were established to test whether theories from existing research also apply for the data set used within this paper. H1 predicted a positive relationship between verbal commitment and actual commitment. H2 predicted a positive relationship between affect and actual commitment. H3 predicted a positive relationship between knowledge and actual commitment and H4 predicted a mediation effect between knowledge and effect with actual commitment. H1 and H2 showed a significant positive relationship. H3 and H4 showed no significant interaction and were therefore rejected.
A fifth and six hypothesis were established to uncover whether the presence of a certain or combination of different personality traits can predict someone’s actual commitment. H5 predicted that agreeableness and openness were positive related to actual commitment, with agreeableness being the strongest. Conscientiousness, neuroticism and extraversion were believed to have no significant interaction with actual commitment. H5 showed no significant relationship between individual or any possible combination of personality traits with actual commitment. H6 predicted a moderation effect of agreeableness and openness on the validated predictor variables for actual commitment and actual commitment itself. H6 however showed no significant relationship. The big five personality traits therefore seem to have no statistically significant influence on someone’s actual commitment.
However, further research is necessary to exclude limitations found within this research regarding the design of the outcome variable, the usage of the mini-IPIP scale and the composition of the participant group for the survey.
Introduction
Climate change has major consequences for animals, ecosystems, humans and the planet as a whole. One example of these consequences which has recently been in the news numerous times is the extinction of the honeybee. In a paper published by the Cornell university, Morse & Calderone (2000) express that the value of the increased yield and quality achieved through pollination by honey bees in the agriculture of the united states alone was equal to $14,6 billion in the year 2000. When the economic value of ecosystems and animal species becomes visible, it becomes apparent how dependent we are on them. Therefor one could argue that preserving these animal species and ecosystems is urgent.
In recent decades several initiatives have been made to accomplish a more sustainable society. One example is 'Het klimaatakkoord' from The Netherlands. ‘Het klimaatakkoord’, derived from the Paris agreement and aims to reduce all national greenhouse gases by 49% in 2030 compared to 1990 within The Netherlands (Rijksoverheid, 2018). Within the climate agreement from the Netherlands many different methods are discussed for reducing national greenhouse gas emissions. Several of these methods are aimed at the customer. For example, people are ought to buy and use more sustainable and eco-friendly products. The classification of eco-friendly and sustainable are visualised based on sustainability labels. These sustainable products often cost more than similar, non-sustainable products. However, it is believed that consumers are willing to pay this extra price for the well-being of the earth. However, this positive consumer attitude towards sustainability is non consistent with the behavioral patterns of customers (Griskevicius, 2010).
This non consistent behavior within the attitude toward sustainability is visualised within the united nation environmental program. UNEP (2015) suggests that 40% of the consumers are willing to pay for sustainable products but only 4% actually does so. Another example is from a research conducted by scholars. Within this research they found that 50-90% of the people are preferring green energy and are willing to pay the additional price for it (Pichert & Katsikopoulos, 2007). However in Germany only 16% of the people actually uses green energy. This difference between the percentage of people who are willing to pay the extra price and the percentage of people who are actually using the eco-friendly product is according to Griskevicius, Tybur and van den Berg (2010) due to underlying behavioral patterns.
Griskevicius, et al., suggest that many consumers might buy green products less for environmental or economic reasons and more for social reasons. Buying green products namely ensures to make you appear sensitive to the social and reputational aspects of conversation and can hereby improve your reputation. This act of self-sacrifice is performed as a motive to obtain status. The research of Griskevicius, et al., (2010) shows that often underlying behavioral patterns can be a deciding factor whether people choose to buy green products or not. Hereby making the activation of status motives a viable strategy for
promoting pro environmental behaviour. Goals from the climate agreement of The Netherlands could possibly be achieved more effectively, if status motivation were used to promote more sustainable and eco-friendly products.
Underlying behavioral patterns determine whether consumers buy green products or not. If these underlying behavioral patterns are recognized, policy proposals could be established in a way to fulfill these behavioral patterns and therefore accomplish a more sustainable consumption market. This paper will research the various causes why consumers participate in more or less sustainable consumption behavior. After finding the causes of sustainable consumption behavior, personality traits of consumers will be visualised in order to uncover whether certain personality traits can cause a person to participate in more or less sustainable consumption behavior. Finding positive or negative relationships between personality traits and sustainable consumption behavior may offer the possibility to enhance sustainable consumption behavior. By stimulating a personality trait that has a positive relationship with sustainable consumption behavior in, for instance, policy proposals or commercials. The purpose of this research is therefore to uncover whether the presence of one, or a combination of multiple, personality traits can have a explanatory effect on someone's actual commitment for sustainable consumption behavior.
Theoretical framework
This research will investigate whether the presence of a certain personality trait can cause a person to participate more or less in buying environmentally friendly products. Within this chapter a further understanding of the different concepts used in this study is provided. The most important concepts used within this research are the concept of environmental awareness based on the workings of Maloney & Ward (1973) and the big five personality traits, based on the workings of Goldberg (1980).
Environmental awareness
“The ecological crisis is a crisis of maladaptive behavior” - Maloney & Ward, 1973, page 1 Using technological solutions for environmental related problems has a certain intrinsic futility. Environmental related problems are often problems caused by humans trough overconsumption, overexploitation and pollution. Environmental related problems can therefor best be conceptualized as a problem of increasing population, increasing consumption and increasing production (Maloney & Ward, 1973). New technologies that are invented to reduce these environmental related problems will also enable us to increase our production and consumption and therefore ultimately increase environmental related problems. For this reason, the solution for environmental related problems will not lie in improving new technologies but more in changing human behavior (Maloney & Ward, 1973). Because the ecological crisis is conceptualized as a problem in terms of human behavior, the solution dictates in altering this same human behavior (Maloney & Ward, 1975). However, it is necessary to clarify someone’s motives before you are able to alter their behaviour. These alterations are not feasible until critical population behavior patterns are properly assessed. Maloney & Ward (1975) constructed the ecological attitude-knowledge scale in order to assess these critical behavior patterns.
The ecological attitude-knowledge scale indicates that environmental awareness is a multidimensional concept consisting of four different dimensions. Namely : verbal commitment, actual commitment, affect and knowledge. Verbal commitment indicates what a person states he is willing to do. Actual commitment indicates what a person actually does to improve the environment. Affect is the degree in which someone is emotional affected by ecological issues and knowledge is the specific, actual knowledge a person possesses related to ecological issues (Maloney & Ward, 1975).
Actual commitment
Actual commitment indicates to the measures someone is willing to undertake to minimize environmental issues (Maloney & Ward, 1975). The purpose of this research is to uncover whether one of the big five personality traits can be an explanatory factors for sustainable consumer behavior. Actual commitment within this study will therefore not refer to the entire broad concept, but only to someone's preference of green consumption over non green consumption. Green consumption is defined as the act of safeguarding the environment
through the use of sustainable consumer behavior (Haws, Winterich & Naylor, 2014). Someone’s actual commitment is strongly correlated with someone’s verbal commitment, affect and knowledge (Maloney & Ward, 1975 ; Frantz & Mayer, 2014 ; Bradley, et al, 1999). Meaning an increase in either someone's verbal commitment, environmental knowledge or affect will most likely cause them to participating in more actual commitment.
The definition and underlying correlations between verbal commitment, affect and knowledge are provided within the sub-chapters below. The value for actual commitment was derived through the use of the green scale, created by Haws, Winterich and Naylor (2014). Haws, et al. intention for creation the green scale was to express the importance of environmental protection through one’s purchases and consumption behavior (Haws, et al. 2014). The green scale was constructed by compiling 98 different measurement items. The composition of these items was based on established scientific research surrounding sustainable consumer behavior. Research they used for determining these 98 items consists primarily of the socially responsible consumption behavior scale by Antil (40 items)(Antil, 1984) and a collection of alternative environmental attitude scales (58 items). The correlation effect between the 98 collected items and five different indications for green consumption (frugality, price- and value consciousness, self-deceptive enhancement and impression enhancement) were measured next. Correlation results found 10 items with a high explanatory value. An additional four items were removed in order to achieve the minimal amount of items while retaining high validity. This resulted in the six highly reliable (α= .89) items used within the green scale. Additional research surrounding the green-scale demonstrates the possibility of predicting consumer preference for environmentally friendly behavior (Frantz & Mayer, 2014 ; Bradley, et al, 1999). The consumer preference for environmentally friendly behavior will be used for measuring this research’ actual commitment. The used methods for measuring actual commitment with the green scale are provided within the methodology section.
Verbal commitment
As explained earlier, verbal commitment indicates what a person states what he is willing to do to minimize environmental issues. The more verbal commitment someone has towards ecological issues, the more actual commitment there will be. So people who often indicate that they want or do participate in sustainable practices, often do indeed have a higher degree of actual commitment. This was evident both from the results of the ecological attitude-knowledge scale of 1973 and the renewed ecological attitude-knowledge scale from 1975 (Malony & Ward, 1973 ; Malony, et al., 1975).
Griskevicius, Tybur and van den Berg (2010) conducted a study to uncover underlying motives why people buy green products. Green products are in general often of lower quality than non-green products that are sold for the same price. Therefore it is quickly assumed that people who buy green products solely accept these extra costs because they place a high
value on nature. However, results from the research of Griskevicius, et al. indicate another motive, namely competitive altruism.
Competitive altruism is the act where people try to enhance their status within a group by being prosocial from a costly signaling perspective (Griskevicius, et al. 2010). Competitive altruism is associated with status because it demonstrates the ability and willingness to incur cost for public-welfare. Improving one’s status can help someone to achieve things that are difficult to attain with money, such as love and friendship (Griskevius, et al. 2010). The conclusion which can be drawn from this is that altruistic people improve their status for indirect self-interest (Griskevicius, et al. 2010). This reasoning of indirect self-interest will most likely cause the average outcome of verbal commitment to be higher than the average outcome of actual commitment. People who namely purchase green products solely for improving their status, will not purchase green products if no status improvement can be achieved. Therefor they will not participate in green consumption when the act is not visible for the outside world. In addition, the presence of verbal commitment is likely to be high for altruistic people because they want to give the implementation that they participate in as many sustainable initiatives as possible to further enhance their status. For these reasons the first hypothesis of this research will read :
H1: There is a positive relationship between verbal commitment and actual commitment. However, the degree of verbal commitment will be higher than actual commitment due to competitive altruism.
Affect
Affect is the degree in which someone is emotional affected by ecological issues. Humans tend to participate more in actual commitment when they are above average emotionally affected by, for example, the extinction of an endangered species. The connectedness to nature scale by Mayer & Frantz (2004) is a measurement scale of individuals’ trait levels of feeling emotionally connected to the natural world and is described as a measure of an individual’s emotional connection to nature. Results from the study by Mayer & Frantz conclude that connection to nature is an important predictor for ecological behavior and subjective wellbeing (Frantz & Mayer, 2004). Hereby supporting the findings of Malony & Ward. Reason for this positive relationship is explained by self-sacrificing behavior. Self-sacrificing behavior is namely motivated when someone feels connected to someone or something (Frantz & Mayer, 2014). Because affect has a positive relationship with actual commitment, the second hypothesis for this research will read :
H2: There is a positive relationship between affect and actual commitment.
Knowledge
An increase in someone’s environmental knowledge, will results in more environmental attitude. This is mainly apparent from the research by Bradley, Waliczek & Zajicek (1999). Within this research, Bradley, et al. distributed a questionnaire among high school students.
This questionnaire was distributed before the science course began and after the science course was over. Results from these questionnaires showed that the environmental knowledge of students had increased by an average of 22%. In addition, the environmental attitude and actual commitment of almost all students improved as well. This shows that a statistically significant correlation is present between pretest knowledge scores and pretest actual commitment and post-test knowledge scores and post-test actual commitment (Bradley, et al. 1999). This leads to the same conclusion Frantz & Mayer (2014) drew. A higher degree of Environmental knowledge results in an increase in affect, which in turn results in an increase of actual commitment.
Although studies show that environmental knowledge is found to be consistently and positively related to environmental attitudes, they also show that the level of environmental knowledge in surveys is often much lower than expected. This while men on average score higher for environmental knowledge than women (Arcury, 1990 ; NEETF, 2000). The often-occurring low value of environmental knowledge can cause major problems in environmental policy proposals. Voters will vote from a wrong perspective since they have very little knowledge about environmental topics (Bradley, et al. 1999). Scientific research indicates a positive relationship between environmental knowledge and actual commitment. The third hypothesis will therefore read:
H3: There is a positive relationship between knowledge and actual commitment.
Mayer & Frantz (2014) highlighted the essence of environmental education. They state that education provides the experience to change someone’s beliefs, attitude, and most importantly, behavior. Improving someone’s environmental knowledge, therefore often causes someone’s affect to increase as well (Mayer & Frantz, 2014 ; Bradley, et al. 1999). This rise in effect will results in more self-sacrificing behavior. Feeling connected to something or someone namely motives self-sacrificing behavior (Frantz & Mayer, 2014). This self-sacrificing behavior is than expressed as actual participation in the form of buying more green products. This interaction between knowledge, affect and actual commitment indicates a mediator effect. A fourth hypothesis in included to test this mediator effect. H4: An increase of environmental knowledge leads to an increase of affect, which in turn leads to more actual commitment.
Big five personality traits
The research goal of this paper is to uncover whether the presence of one or more personality traits contribute to actual commitment. The theory of the big five personality traits states that personality can be classified within five different personality dimensions. These five dimensions, also called traits, can be used to describe someone's character or personality. Each of these five traits are present in a person. However, people differ in the degree of presence of these personality traits. The five personality traits are divided into: (1)
extraversion, (2) agreeableness, (3) conscientiousness, (4) neuroticism and (5) openness to experience. (John, Naumann & Soto, 2008). The description and tendencies for a high and low value in each of the five different personality traits are shown in table 1 below.
Table 1 : All big five personality traits with corresponding tendencies and low/high value indicators. Source : (Digman, 1990)
The extraversion personality trait is primarily about engagement with the external world. People who score high in the extraversion property are often looking for interaction and social connection. People who score low on the extraversion personality trait are introvert. They prefer to avoid social interactions and enjoy being alone (Digman, 1990). Extraverse people are often spontaneous, cordial and cheerful and are often perceived more dominant in social settings, as opposed to introvert people (Friedman & Schustack 2016). Extraverts enjoy being with other people more than being alone. Introverted people, on the other hand, tend to avoid social situations and often find social pressure unpleasant. They appear less involved in the social world. However this does not mean that introvert people are unfriendly or unsocial. They just need less stimulation and prefer alone time more than extroverts (Rothmann & Coetzer, 2003).
The agreeableness personality trait is primarily about social harmony. People who score high on agreeableness are often interested in others and are considered more friendly (Digman, 1990). They are considered trustworthy, kind and generous. People who score high on agreeableness are often willing to compromise their interests with other (Rothmann & Coetzer, 2003 ). However they often have an optimistic view of human nature. They place great value on their relationships with others and are focused on harmony. People who score low in agreeableness often have a more businesslike approach. They are often more stubborn and more focused on their self-interest and are generally unconcerned with others’ well-being. Often people who score low in agreeableness do not take the trouble to devote
themselves to other people. People with a low score in agreeableness sometimes appear unfriendly and uncooperative because of their sketism (Toegel & Barsoux, 2012).
Conscientiousness relates to self-discipline. It indicates the extent to which you are focused and work according to a plan. It indicates the degree to which you are focused or flexible (Digman, 1990). People with a high degree of conscience are often dutiful and have a lot of self-participation. They are often efficient and work in a structured manner and are perceived more focused and stubborn (John, et al. 2008). People with a lower level of conscientiousness are often more flexible and unconstrained. However, they often are associated with sloppiness and lack of reliability (Toegel & Barsoux, 2012).
The neuroticism personality trait, sometimes called emotional instability, is primarily about the tendency to experience negative emotions such as anger, depression and anxiety. People who score high on neuroticism are often less emotional stable (Digman, 1990). Neuroticism relates to the degree to which you are sensitive to environmental stimuli and in the need for certainty. Neurotic people are often worried and nervous. They often experience emotional up and downs in comparison to emotional stable people. The other extreme of neuroticism is emotional stability. Emotionally stable people are more calm, relaxed and are less easily upset. However this freedom from negative feelings does not mean that people who score lower on neuroticism experience more positive feelings (Dolan, 2006).
People with a high degree of openness to experience are often curious and open to new views, people and experiences. They are perceived as unpredictable, risk-taking and something even lack of focus. They also tend to have a higher appreciation for imagination, art and emotions. People who do not score high on openness tend to prefer traditional views and do not like change. They are are perceived to be more data driven and pragmatic (Digman, 1990 ; Rothmann & Coetzer, 2003).
The five different personality traits are likely to have different relationships to actual commitment. Agreeableness and openness to new experience are likely to have a positive relationship with actual commitment. This while the other personality traits will probably have no statistically significant relationship with actual commitment. People who score high on agreeableness are more likely to participate in actual commitment because they desire to be be loved. A method to become more loved is to participate in competitive altruism (Griskevicius, et al. 2010). The hypothesis for the interaction effect between agreeableness and actual commitment therefor states that agreeableness is the highest explaining personality trait for actual commitment. The other positive related personality trait with actual commitment will most likely be openness to new experience. The decision of consuming and using more green products namely includes behavioural change. For example, choosing an organic supermarket over your local supermarket. The relationship between extraversion and actual commitment will most likely depend on your surrounding. Actual commitment and extraversion will have a positive relationship when your surrounding is concerned with
climate change and everyday topic revolve around climate agreements and minimizing emissions. However, actual commitment and extraversion will have a negative relationship when you live in a neighborhood where almost all residents deny climate change. Residences were however not included in this study. The expected correlation between extraversion and actual commitment will therefore be non significant. No substantiated research was found to substantiate a positive or negative relationship between conscientiousness, actual commitment and neuroticism, actual commitment. The correlation between conscientiousness and , actual commitment and neuroticism, actual commitment is therefore considered to be non significant.
H5 : Agreeableness and openness to new experience will have a positive relationship to
actual commitment. Conscientiousness, neuroticism and extraversion will have no statistically significant relationship to actual commitment. This while agreeableness will be the biggest deciding personality trait for actual commitment.
H6 : Agreeableness and openness will have a moderating effect on the validated predictor
variables for actual commitment and actual commitment itself. The moderation effect of conscientiousness, neuroticism and extraversion will not add more explained variance within actual commitment.
Methodology
This chapter contains the various used research methods for arriving at the results. First, it contains a short explanation about all the method and analysis used to answer all different hypotheses. A description about the different types of approached respondents during the survey is provided next. Followed by a summary of the survey question and an explanation on how these questions are able to compute the values for the nine different variables of the model. Hereafter, an explanation about the different statistical analyzes used for processing the hypotheses is provided. Lastly, the process of data re-coding and preparation for statistical analysis are explained.
Actual commitment will be used as the outcome variable in this research. The degree of someone’s actual commitment namely depends on someone’s affect, verbal commitment and knowledge. Actual commitment, within this research, refers to the act of participating in sustainable consumer behavior. The top three explanatory variables for actual commitment are (1) level of environmental sensitivity, also named affect. (2) Environmental knowledge and (3) verbal commitment (Sia, Hungerford & Tomera, 1986). The top three explanatory variables for actual commitment will function as predictor variables for actual commitment. The model used for this quantitative research is shown in figure 1 below.
Figure 1: model used for this research. Predictor variables are knowledge, verbal commitment and affect. Independent variable is actual commitment and the five different personality traits are moderators.
The values for all different variables are calculated based on a minimal of four till a maximum of nine 5 point likert-scale questions. The results from these questions were recorded in one mean variable after checking Cronbach’s alpha. The appliance of established
theories from existing literature were tested to see whether these theories also apply for this data set. The established theories were tested by calculating the correlation effect of all predictor variables (affect, verbal commitment and knowledge) on actual commitment. Afterwards a regression analysis with PROCESS by Andrew F Hayes was performed to determine whether a mediation effect between knowledge and effect on actual commitment exists. After this the effect of the five different personality traits on actual commitment were analyzed. The correlation between the five personality traits and actual commitment was first analyzed by looking at the correlation effect for each individual personality trait with actual commitment. Afterwards the joint correlation of openness and agreeableness was examined because this combination of personality traits was believed to have the highest prediction for actual commitment. The joint correlation effect of every possible combination of personality trait with actual commitment was analyst last to see whether an unexpected combination of personality traits exists with a high prediction value for actual commitment.
Design, sample and procedure
This research was conducted on the basis of a quantitative research design. A quantitative research was used because this research aims to analyze whether the presence of a certain personality trait provides for more actual commitment. Gathering the information needed for this model was possible by using a survey with a limited amount of question. Qualitative research, used for further inquiries was therefore unnecessary. Therefore a quantitative study with a cross-sectional design was utilized.
The data used within this study was collected by the use of a survey. A total of 102 respondents completed the survey. The respondents were approached based on convenience sampling. Most participants were approached on social media. A gift voucher of €20 was distributed to approach additional participants. The average age of repondants is 24,86 year, ranging from 18 till 82 years old and is divided between 49,4% woman an 50,6% men. Most participants are students from either the university of Amsterdam, Hogeschool van Amsterdam or the Vrije Universiteit. Of the 102 respondents, 85 fully completed the survey. Resulting in a completion rate of 83.33%. The data was collected over the span of one month. Measurements
This study contains 9 different variables, each variable was calculated based on a minimum of 4 and a maximum of 9 questions. All survey questions used for this research are displayed in appendix 1. The variables of the 5 different personality traits were calculated based on the mini-IPIP scale of Donnellan, Oswald, Baird and Lucas (2006). The mini-IPIP scale is shortened version on the The BFI (Big five Inventory) scale from Goldberg (John, et al. 2008). It bases someone’s personality trait based on the outcome of 20 questions, instead of the original 42 item BFI scale. This reduction in in questionnaire items was used to benefit the survey response rate. Keeping questionnaires as small as possible often results in a higher response rate. The mini-IPIP scale consist of 20 questions. 4 different questions are indicated for each personality trait. For example, the personality trait agreeableness contains the following questions: (1) I sympathize with others’ feelings. (2) I am interested in other
people’s problems. (3) I feel others’ emotions. (4) I am not really interested in others (R). Each personality trait contains one reverse question. Each question is answered on a 5-point likert scale, which ranges from completely disagree to completely agree. The value of a certain personality trait is calculated by taking the mean of the four different questions that belong to the personality trait.
The questions used for the variable knowledge were based on the NEEFT/Roper Survey (2000). The NEEFT survey contains 12 multiple choice questions to measure environmental knowledge of random participants in America. The questions from the NEEFT/Roper survey are based on contemporary environmentally related topics from the news. A total of 9 questions from the NEEFT/Roper survey were used in this research. These questions were used for measuring environmental knowledge. The questions were made applicable to the Netherlands before they were added to to the survey. Examples of these questions are : (1) Where does most of the garbage in The Netherlands end up? (A) Oceans. (B) Incinerators. (C) Recycling centers. (D) Landfills. (2) Which of the following is a renewable resource?, multiple answers possible. (A) Oil. (B) Wind. (C) Trees. (D) Coal. (E) Iron ore. (F) Brown coal. The value of the knowledge variable was calculated by calculating the percentage of the number of correct answers.
The variable affect was calculated by asking to what degree participants would be emotionally affected when various elements of nature would disappear due to climate change or ecological problems. Examples of these questions are : From an emotional point of view, to what extent are you affected by (1) Animals threatened with extinction, such as the polar bear, actually die out. (2) The presence of the North Atlantic plastic garbage patch. Every question was answered on a 5-point likert scale, which ranges from not at all to very much. The variable of affect was calculated by taking the mean of the four different questions. The questions from the variable actual commitment were based on the green-scale of Haws, Winterich & Nayer (2014). The green scale is defined as the tendency to express the value of environmental protection through one’s purchases and consumption behavior (Haws, Winterich & Nayer, 2012). The green scale is constructed based of six different studies and contains a total of six items to capture green consumption in a reliable and valid manner. Examples of these items are : (1) It is important to me that the products I use do not harm the environment. (2) I consider the potential environmental impact of my actions when making many of my decisions. Every question was answered on a 5 point likert-scale, ranging from strongly disagree to strongly agree. The value for the actual commitment variable was calculated by taking the mean of the six different questions.
The variable verbal commitment was calculated on the basis of various propositions. The participants were asked to what extent these propositions applied to them. These propositions were constructed in such a way that they reflected new products participants had to purchase in the future and to what extent they would opt for the more sustainable variant. This
involved looking at the extent to which the participants indicated that their future purchases would be sustainable. Examples of these questions are : (1) When I buy a new car, I think it is important to minimize pollution. For this reason I will base my choice on electric or hybrid cars. (2) When I buy a new washing machine, I think it is important that it is sustainable. For this reason I will only base my choice on washing machines with energy label A or higher. Every question was answered on a 5 point likert-scale, ranging from totally disagree to totally agree. The value of the verbal agreement variable was calculated by taking the mean of the four different questions.
Analytical plan
A total of five hypotheses were tested within this study. Hypotheses 1, 2, 3 and 4 will test whether or not there is a positive relationship between the different predictor variables (verbal agreement, knowledge and affect) and the outcome variable (actual commitment). A multiple regression analysis will be used to test hypotheses 1, 2 and 3. This multiple regression analysis will calculate the explained variance the predictor variables have on the outcome variable. Hypothesis 4 states the presence of a mediation effect between knowledge, affect and actual commitment. PROCESS v3.2 by Andrew F. Hayes will be used to test this mediation effect.
When carrying out hypothesis 5, the relationship between each individual personality trait and actual commitment was calculated first. This was done with a multiple regression analysis. After this, the combined relationship between agreeableness + openness and actual commitment was calculated. According to scientific research (Griskevicius, et al. 2010) these two personality traits should have the highest prediction value for actual commitment. This calculation was performed with a multiple regression analysis. While hypothesis 5 tests the direct effect of personality trait on actual commitment. Hypothesis 6 will test the moderating effect of the different five personality traits on the relation between the predictor variables of actual commitment and actual commitment itself. A regression analysis with 6 different blocks will be used to measure this moderation effect. Block one will include all the validated predictor variables of actual commitment and block 2 - 6 will include all the validated predictor variables of actual commitment including one of the five personality traits. Adjusted R2 will be used to see whether the addition of a personality trait will explain more variance within actual commitment.
Recoding
The data set obtained through the survey of 102 respondents had to be recoded before statistical analysis could be done. First, all incomplete surveys were deleted. From the 102 respondents, 85 fully completed the survey. Resulting in a completion rate of 83.33%. Within the survey the variable “gender” had three options, namely : (1) male, (2) female and (3) prefer not to disclose. Because gender is a nominal variable (1) male was re-coded in the value 0 and (2) female was re-coded in the value 1. (3) prefer not to disclose was turned into a missing value.
The four different questions associated with each personality trait were merged into one mean variable. This resulted in one mean variable for agreeableness, extraversion, conscientiousness, neuroticism and openness. These scales show sufficient reliability as the Cronbach’s alpha for these variables are agreeableness : 0,750. Extraversion : 0,745. Conscientiousness : 0,693. Neuroticism : 0,670 and openness : 0,751. The six different questions associated with actual commitment were merged into one mean variable with a cronbach’s alpha of 0,721. Verbal commitment and affect both contained four questions. Both were merged in one mean variable contained a cronbach’s alpha of verbal commitment : 0,753 and affect : 0,758. The nine different questions associated with knowledge were merged into one variable by giving each participant a percentual value of the amount of questions they answered correctly. Results of this study will take into account that two two variables have a cronbach's alpha of lower than the required 0.7. The mini-IPIP scale from the research of Donnellan et al. scored a cronbach’s alpha of 0,70 for Neuroticism and 0.75 for conscientiousness while conducting the same survey used in this research (Donnellan, et al. 2006). Reason for the lower scored cronbach's alpha within this research will be discussed in the discussion.
Results
Means, SD and Correlations
The mean, standard deviation, reliability and correlation of the key variables are shown below in table 1. All data used to construct this table is displayed in appendix 2. The correlation effect is calculated by Pearson’s. Pearson’s R measures the strength and direction of a linear relationship between two variables on a scatter plot. The value of R always lies between -1 and 1. 0 means no relationship. Above 0,50 or below -0,50 a moderate relationship and > 0,70 or < -0,70 means a strong relationship (Field, 2009).
No strong correlations between the variables were found. However there are moderate correlations between verbal commitment, actual commitment (p <0.05, R = 0.65) and affect, actual commitment (p <0.05, R = 0.66). This is indeed in line with hypotheses 1 and 2. Affect, knowledge and verbal commitment are the top three predictor variables for responsible environmental behavior according to the research of Sia et al (Sia, et al. 1986). However, it is noteworthy that knowledge, which is one of the predictor variables for responsible environmental behavior hardly has any correlation (R = 0,09) with actual commitment.
Table 2 : Correlation matrix - presenting mean, standard deviation, correlation matrix and Cronbach’s alpha for each variable N=85.
* Correlation is significant at the 0.01 level (2-tailed) ** Correlation is significant at the 0.05 level (2-tailed)
Males were given the value 0 and Females were given the value 1 Cronbach’s alpha is displayed on the diagonal.
The legitimacy of all five hypotheses was analyzed, based on multiple regression analysis and the PROCESS procedure within SPSS created by Preacher and Hayes. However, all assumption for multiple regression analysis were first tested and all outliers within the dataset were removed. Multiple regression assumptions consist of normality, linearity, homoscedasticity and absence of multicollinearity (Pruppers, 2018). Normality implies whether the data of the variables has a normal distribution. A P-P plot was used to test normality. Linearity tests whether the values of the predictor variables in the regression have a straight-line relationship with the outcome variable. Homoscedasticity tests whether the data is equally distributed, a scatter plot was used to test this distribution. Multicollinearity refers to the possibility that the predictor variables are correlated with each other. This was tested by using VIF values (Pallant, 2010). The outliers were removed by removed all values with a Z-score of > 1,96 or < -1,96. The z-score indicates the amount of standard deviations a certain value lies from the mean (Pruppers, 2018).
Testing existing literature
H1: There is a positive relationship between verbal commitment and actual commitment. H2: There is a positive relationship between affect and actual commitment.
H3: There is a positive relationship between knowledge and actual commitment.
First all outliers were removed by looking for all values with a Z-score of >1,96 or <-1,96. A total of 7 outliers were found within the variables ‘verbal commitment’, ‘actual commitment’, ‘knowledge’ and ‘affect’. This resulted in a total sample size of 78. The assumptions for multiple regression were tested and the data set passed all assumptions for a multiple regression analysis. Elaboration of this is displayed in Appendix 2A.
A multiple regression analysis was used to predict actual commitment from verbal commitment, knowledge and affect. The three predictor variables (knowledge, affect, verbal commitment) seem to have a significant effect in the prediction for actual commitment (F(3,74) = 28,147, p < .0005), with an R2 of 0,533. The R 2 value of ,533, indicates that 53,33% of the total variance within actual commitment can be explained by the combination of the predictor variables (knowledge, affect and verbal commitment). Verbal commitment significantly predicts actual commitment (p < 0.0005, unstandardized β = ,453). β indicates that an increase of verbal agreement by 1, results in an actual commitment increases of 0.453 (Pallant, 2010). Affect seems to be a significant predictor variable for actual commitment as well (p < 0.0005, unstandardized β = ,321). Knowledge however seems to have no significant effect (p = 0.325, unstandardized β = -0.003). The 0-hypothesis cannot be rejected, because the p-value is higher than the maximum allowance of 0.05. The 0-hypothesis states that no relationship between knowledge and actual commitment exists. Hypothesis 1 and 2 are confirmed and seem to have a significant effect for actual commitment. This non-significant effect for knowledge can occur due to the mediation effect between knowledge, affect and acute commitment as indicated within scientific research. Hypothesis 4 will test this mediation effect.
H4: An increase of environmental knowledge leads to an increase of affect, which in turn leads to more actual commitment.
The mediation effect on the relationship between knowledge and actual commitment was measured using the PROCESS procedure within SPSS created by Preacher and Hayes. The mediation effect showed no statistical significant effect. The effect of knowledge on actual commitment (c = -.0017, CI[-.01, .01] p = .68) was mediated by affect (a*b = 0.001, CI[-.004, .007]). No mediation can be concluded since both the direct (c) and indirect (a*b) have a p-value higher than the maximum allowed 0,05. The 0-hypothesis cannot be rejected because the mediation effect of affect on knowledge and actual commitment seemed to have no significant effect. The 0-hypothesis states that no mediation effect from affect on the relationship of knowledge and actual commitment occurs. The SPSS output used for this calculation is shown in Appendix 2B.
Testing the personality traits
H5 : Agreeableness and openness to new experience will have a positive relationship to
actual commitment. Conscientiousness, neuroticism and extraversion will have no statistically significant relationship to actual commitment. This while agreeableness will be the biggest deciding personality trait for actual commitment.
First all outliers were removed. A total of 22 outliers were found after searching for all the outliers within the variables ‘actual commitment’, ‘agreeableness’, ‘extraversion’, ‘conscientiousness’, ‘neuroticism’ and ‘openness’. This resulted in a total sample size of 63. The assumptions for linear regression were tested and the data set passed all assumptions for linearity. Elaboration on this can be found in Appendix 2D.
First a simple linear regression analysis was performed for each individual personality trait. None of the five personality traits showed any significant prediction for actual commitment. Agreeableness (F(1,77) = 1,798, p > 0,184), with an R 2 of 0,023. Extraversion (F(1,77) = 0,264, p = 0,609), with an R 2 of 0,003. Conscientiousness (F(1,77) = 0,668, p = 0,416), with an R 2 of 0,009. Neuroticism (F(1,77) = 0,203, p = 0,653), with an R 2 of 0,003. Openness (F(1,77) = 1,190, p = 0,279), with an R 2 of 0,015.
A multiple regression analysis was performed next to test the prediction of Griskevicius, et al. They predicted that the amount of actual commitment should be highest when people have a high degree of agreeableness and openness. However, the combination of agreeableness and openness showed no significant prediction for atual commitment (F(2,76) = 1,101, p = 0,338), with an R 2 of 0,028.
H6 : Agreeableness and openness will have a moderating effect on the validated predictor
conscientiousness, neuroticism and extraversion will not add more explained variance within actual commitment.
A regression analysis with 6 different block was performed to test the moderating effect of individual personality traits on the validated predictor variables for actual commitment. Block 1 contained affect and verbal agreement as independent variables and actual commitment as dependant variable. Block 2 used the same model, but included conscientiousness as one of the predictor variables. Whereas block 3 used agreeableness instead of conscientiousness, block 4 used openness, block 5 used neuroticism and block 6 used extraversion. Results from all the different blocks however showed no increase in adjusted R2 (model 1 = ,512 ; model 2 =,506 ; model 3 = ,499 ; model 4 = ,490 ; model 5 = ,488 and model 6 = ,480 ) in comparison to block 1. Therefore no increase in explanatory power by adding the different personality traits as moderator was concluded. Elaboration on this can be found in appendix 2D
Discussion
Findings
Within the first five hypotheses of this research, theories from existing scientific literature were tested for this dataset. The purpose of this was to see whether these theories also hold true for the dataset collected during the surveys. In addition, it was possible to examine whether special findings could possibly be made from the results that have not yet been found in previous research.
The first four hypotheses tested whether the three predictor variables for actual commitment show a significant correlation for the dataset used within this research. The three predictor variables for actual commitment were derived from scientific research, which is explained within the theoretical framework (Maloney & Ward, 1975; Frantz & Mayer, 2014; Bradley, et al, 1999). The first hypothesis says that a higher degree of verbal commitment will indeed lead to more actual commitment. This is based on the research by Maloney et al. from the ecological attitude-knowledge scale. Verbal commitment is namely seen as one of the biggest predictions whether people participate in actual commitment. The results from this study indicate a strong relationship between actual commitment and verbal commitment. Therefore it can indeed be concluded that people with a higher degree of verbal commitment also have a higher degree of actual commitment.
The second hypothesis states that a higher degree of affect will lead to a higher degree of actual commitment. This assumption is based on the connectedness to nature scale from the research of Mayer & Frantz. People who are above average emotionally affected by ecological issues will as a result participate in more actual commitment. Results from this study indeed showed a strong relationship between affect and actual commitment.
The third hypothesis states that people with more environmental knowledge will also participate in more actual commitment. This assumption derived from the research from Bradley, et al. Results from this study however showed no statistically significant effect between knowledge and actual commitment. The amount of knowledge someone has about environmental topics therefore does not indicate whether a person participates more or less in actual commitment. Notable was the very small negative relationship between knowledge and actual commitment. Indicating that people with more knowledge actually participate in less actual commitment. Various sources indicate that affect would have a mediation effect on knowledge and actual commitment (Mayer & Frantz, 2014 ; Bradley, et al. 1999). this effect was tested within the fourth hypothesis. Results however showed no significant effect and therefore the degree of someone’s environmental knowledge does not seem to increase one’s affect which would in turn increase their actual commitment.
The purpose of this study was to uncover whether the presence of one, or a combination of multiple, personality traits can have a explanatory effect on someone's actual commitment.
This was tested within hypotheses 5 and 6. Hypothesis 5 first tested whether the individual presence of an personality trait has a direct relationship to someone’s actual commitment. However, none of the five individual personality traits proved to have any significant relationship with atual commitment. Therefore it can be concluded that a high individual presence within one of the big five personality traits by goldberg does not explain whether someone would participate in more or less actual commitment. The second part of hypothesis 5 tested whether the combination of personality traits, which according to literature (Griskevicius, 2010), should be the highest predicting for actual commitment did have a significant relation to actual commitment. Results however showed no significant effect and therefor having a high degree in both agreeableness and openness does not seem to have a direct influence on whether someone participates more or less in actual commitment.
Hypothesis 6 states that the moderation effect of agreeableness and openness on the relationship between the validated predictor variables for actual commitment (affect and verbal agreement) will cause more explanatory power for someone’s degree of actual commitment. This while the moderation effect of conscientiousness, neuroticism and extraversion will not cause for more explained variance within actual commitment. Results from the six model regression analysis however showed no increase in adjusted R2 for model 2 till 6. This means that none of the five individual personality traits showed an increase in the prediction whether someone participates in more or less actual commitment. For this reason, it can be concluded that the big five personality traits by Goldberg have no significant effect in predicting whether someone will participate in more or less actual commitment.
Limitations
Within this research it is concluded that the big five personality traits of Goldberg are not an explainance factor for someone's degree of actual commitment. However, there are some limitations within this study that may have caused wrong conclusions. These limitations mainly arose from the design of the outcome variable, the way in which the personality traits were calculated for the participants and the group of participants for the survey.
The outcome variable actual commitment was calculated based on the green scale by Haws, et al. The green scale used to a six item questionnaire to measure someone’s preference for environmental friendly products. Although someone's preference for greener products often has a strong relationship with someone's actual purchase of green products, they are not the same. For this reason, a follow-up study should research whether real actual commitment gives another outcome than the preference for actual commitment. Measuring consumers buying behavior of green products in relation to the non-green products could therefore be used to give a more extensive outcome variable for actual commitment. Because actual commitment was measured by the preference for green consumption, verbal commitment and actual commitment are very similar within this research. Wrong conclusions can arise because these two variables are very similar. Unfortunately, changing the variable actual
commitment into a variable that actually measures the purchase of green products from consumers was no longer possible due to time restrictions.
The degree of presence in the various personality traits of the participants was calculated using the mini-IPIP scale by Donnellan, et al. This mini-IPIP scale reduces Goldberg’s original questionnaire for measuring personality traits 20 questions. Although the mini-IPIP scale has been proven by scientific research to be a valid measurement for personality traits, it is less accurate than the original scale of Goldberg. Using the original questionnaire of Goldberg itself could achieve a more valid indication of the participant’s personality traits. Using the mini-IPIP scale within this paper, however was necessary to ensure a higher completion rate within the survey.
The participants used for this study were for a large part (approximately 90%) students from the University of Amsterdam or VU University with an age of 18-26 years old. Using a more diverse group of participants might show a clearer picture of the difference in environmental knowledge between participants. Results of this survey namely showed that most participants had a fairly high degree of environmental knowledge. This coherence of environmental knowledge could be one of the reasons why knowledge turned out not to be a correct predictor variable for actual commitment.
In addition, the survey was conducted in English. This while the native language of all participants was Dutch. Although most of the participants participating in a high level of education (HBO and WO). Language barriers may have caused misunderstanding within the questions of the survey.
Further research
Further research will first have to test whether the five personality traits of Goldberg do or do not have a relationship with actual commitment when the above limitations are solved. When relationships are found, it is important to uncover the underlying reason why a certain personality trait causes for more or less actual commitment. For example, stores that sell a lot of green products could hire fewer staff that inform consumers about environmental issues when it turn out that people with a high degree of neuroticism dislikes green stores because they dislikes social interaction. In this way, policy proposals, advertisements and social actions could be adapted to the reasons why consumers do or do not participate in actual commitment. When these adaptations are correctly implemented in practice, they can lead to a higher degree of positive, and lower degree of negative, relations with actual commitment. This will cause for a higher degree of overall sustainable consumption behavior and as a results policy proposal such as ‘het klimaatakkoord’ could be achieved more easily.
