Advertising fair-trade products

(1)

Master’s Thesis

Advertising fair-trade products.

Graduate  School of Communication

Master’s programme Communication Science

University of Amsterdam

Name: Marliek Weldring

Student ID-card number: 5909465

Supervisor: Bas van den Putte

(2)

ABSTRACT

The study focussed upon unravelling an effective advertising strategy for the promotion of fair-trade products. A distinction was made between internally driven consumers who act upon their moral values, and externally driven consumers who are striving to obtain a morally conscious image. Image-oriented and quality-oriented advertisement appeals were studied, as well as the role of purchase visibility on social media. An experiment was conducted and respondents were divided into groups of high and low self-monitors and groups with more dominant image purchase motives opposed to quality purchase motives. These characteristics were expected to fulfil a moderating role in the effectiveness of different advertising strategies and explain the differences in the degree of persuasion for the defined groups. Results indicated that independent of the advertising appeals used and independent from the possibility to share a purchase on social networks, advertisements let to an equally moderate attitude towards the advertisement, attitude towards the purchase and purchase intention for all defined groups. None of the posed hypotheses was supported.

Limitations of the research are discussed and recommendations for future research are made.

(3)

Index

Introduction --- 4

Theoretical Framework --- 6

2.1 Fair-trade consumption and identity --- 6

2.2 Motives for buying fair-trade --- 8

2.3 Social monitoring theory --- 10

Methods --- 14

3.1 Study design and sample --- 14

3.2 Manipulations --- 15 3.2.1 Advertisement appeals --- 15 3.2.2 Visibility --- 17 3.3 Measures --- 17 3.3.1 Independent variables --- 18 3.3.2 Dependent variables --- 19 3.3.3 Control variables --- 21 3.3.4 Manipulation check --- 21 3.4 Procedure --- 22 Results --- 23 4.1 Manipulation check --- 23 4.2 Control variables --- 24 4.3 Hypotheses --- 25

4.3.1 Self-monitoring and advertisement appeals --- 25

4.3.2 Purchase motives and advertisement appeals --- 26

(4)

4.3.4 Purchase motives and purchase visibility --- 30

Conclusion and discussion --- 31

References --- 35

Appendix A --- 39

Appendix B --- 42

Appendix C --- 44

(5)

Introduction

At the University of Amsterdam, in the movie theatre or at gas station counter, the colourful Tony’s Chocolonely chocolate bars are everywhere. Tony’s

Chocolonely is a Dutch fair-trade chocolate brand founded in 2005. The fair-minded but realistic view of the company is reflected in the shape of their chocolate bars: they aren’t cut in equal blocks but into a maze of different sized shapes. Their reasoning: “Every unequal bite is a reminder of the gross inequalities of the chocolate trade.” (“Our mission - the road...”, n.d.).

The company has been growing rapidly ever since their start. Net sales grew up to 7.3 million euros in 2013, an increase of 62.9 percent compared to the year before (Tony Chocolonely, 2014). Evidently, fair-trade is becoming more hip and happening (Andorfer & Liebe, 2013). A trend of critical consumers is appearing (Ingenbleek & Reinders, 2013). Purchase decision-making is revolving more around ethical responsibility (Andorfer & Liebe, 2013; Ozcaglar-Toulouse, Shiu & Shaw, 2006) and as a consequence, a growth of the fair-trade market can be detected (Ingenbleek & Reinders, 2013; Adams & Raisborough, 2011).

Although the success of Tony Chocoloney is undeniable, this isn’t the case for every fair-trade company (Van Dinther, 2015). There is still a lot of ground to be gained. While many consumers are critical in their decision-making process towards the origin of products (Alexander & Nicholls, 2004), lots of others are yet to be convinced about the importance of this aspect (Do Paco & Raposo, 2012) or are readily concerned about the matter but lack the actual behaviour of purchasing fair-trade products (De Pelsmacker, Driesen & Rayp, 2005).

While approaching this issue from a consumer-based perspective, it becomes inevitable that the role of fair-trade companies is considered. If consumers are to be

(6)

persuaded in buying more fair-trade products, the advertising of these products becomes a central aspect of interest. Like Tony Chocoloney, fair-trade companies are social enterprises that combine business goals with societal goals. Through the sell of their products they are trying to establish more equality in the world. It is therefore important that these companies develop effective advertising strategies for the promotion of their fairly traded goods. The current study is focused upon uncovering an effective approach that could serve as a starting point for companies to base their strategy on.

Fair-trade consumption is entangled with the fact that it is associated with an altruistic motive: by preferring a more expensive fair-trade product over a regular product, a consumer is helping third world country producers to earn a fair wage under reasonable conditions (Varul, 2008). This can be split up in two dimensions (Grauel, 2014; Varul, 2009): on the one hand, people can intrinsically be motivated to do the ‘right thing’ because they care for the world as a fair place. On the other hand, people can be externally motivated to buy fair-trade products: it serves as a means to express status as someone who is ‘doing good’ (Grauel, 2014; Varul, 2009). Building upon this distinction a connection is made with a concept from social psychology: self-monitoring (Snyder, 1974). This theory suggests that individuals differ in the extent to which they value expressive behaviour as a way of self-presentation. While categorizing consumers according to this concept, an advertising opportunity is created to approach them in a way that matches their motives and characteristics: for low self-monitors it is found that quality appeals are effective influencers opposed to high self-monitors who are persuaded more through the use of image appeals (Snyder & DeBono, 1985).

(7)

Most research regarding fair-trade consumption and the tension between inner morality and self-expression has been of qualitative nature (Alexander & Nicholls, 2004; Varul, 2008, 2009). These insights about fair-trade consumption have not yet been translated into an empirically tested advertising strategy. Neither has the concept of self-monitoring been studied in relation to fair-trade consumption. Nonetheless seems this theory pre-eminently applicable to the subject. Due to the altruistic component that it entails, fair-trade consumption offers a fruitful base for differences between high and low self-monitors. More research is needed to fill this gap

(McDonagh & Prothero, 2014). The current study aims to shed light on the dynamics underlying fair-trade purchases and form a new angle for segmentation, so that tailored advertising strategies can be used to effectively persuade consumers.

It is proposed that most people are potential fair-trade consumers (Grauel, 2014; Sha, Shiu & Clarke, 2000) and thus as a starting point, fair-trade advertising could be directed at anyone. Because individuals differ in the extent to which they find the expression and visibility of their moral consciousness important, it is explored how these differences can be taken into account when creating advertising strategies. The study revolves around the research question: While considering image-oriented advertising versus quality-image-oriented advertising and the role of purchase visibility, how can consumers be persuaded most effectively to purchase fair-trade products, when a distinction is made between internally driven, low self-monitoring consumers and externally driven, high self-monitoring consumers?

Theoretical Framework

2.1 Fair-trade consumption and identity

There hasn’t been consensus on a clear definition of fair-trade consumption (Andorfer & Liebe, 2013). According to Harrison, Newholm and Shaw (2005) ethical

(8)

consumption is referring to consumers’ purchase decisions that are, besides the price and features, concerned with the political, social and environmental consequences. Fair-trade consumption is a component of this morally oriented consumer behavior, due to its altruistic aspect (Becktert, 2006). The general goal of fair-trade production in developing countries can be specified as the improvement of the conditions of small-scale producers, in which they are guaranteed fair prices for their goods (Andorfer & Liebe, 2013). Hence, the purchase of fair-trade goods is related to a wider concern than solely the usefulness of the product, but stems additionally from the concerns about the origin of the product. Consequently, fair-trade consumers are contributing to a higher wellbeing of the producer.

In line with this, the purchase of fair-trade goods comes with certain

implications about the nature of its consumers. Namely, that they are selflessly buying these products for a higher price, to ensure the wellbeing of others. The purchase is rooted in an ethical consideration and through the rejection of a non-fair-trade alternative, an altruistic image is created (Varul, 2008; 2009). Because of the

underlying moral values, this consumption is inevitably connected to the identity of a consumer (Andorfer & Liebe, 2013). Fair-trade consumption reflects the identity of a moral person, who voluntarily chooses to pay more in favour of distant others

(Andorfer & Liebe, 2013). More generally speaking, it is widely recognized that individuals use their purchase patterns as a means to construct a personal identity (Andorfer & Liebe, 2012; Giddens, 1991). Identities represent different aspects of the self that consumers use to socially categorize themselves and express who they are (Bhattacharjee, Berger & Mennon, 2014). Through the purchase of fair-trade goods, consumers socially categorize themselves as morally conscious and affirm their ethical identity.

(9)

Former studies have shown that identity plays an important role within fair-trade consumption. Sparks and Shepherd (1992) and Shaw, Shiu and Clarke (2000) found that self-identity contributed as the main determinant to the fair-trade purchase intention. Andorfer and Liebe (2013) also concluded that identity expression is one of the key factors underlying fair-trade consumption. When recognizing identity as a main aspect, a distinction can be made between internally and externally motivated consumers: those who buy these products because they are driven by their personal values and therefore inevitably reflect an ethical responsible identity, or individuals who actively strive to obtain the image of a fair-trade consumer, mainly for the sake of the image. The next chapter will further elaborate on this matter.

2.2 Motives for buying fair-trade

The fair-trade price plays an important role when zooming in on the motives for fair-trade consumption. Fair-trade products are generally higher in price than their regular counterparts, because of the guaranteed fair wages for the producers

(Andorfer & Liebe, 2012). Research suggests that consumers attempt to justify this price gap by more than altruistic thoughts: they create additional motives for buying fair-trade goods (Varul, 2008). Either they assign a higher quality to these products (Varul, 2008) or they add symbolic value to the use of these products (Andorfer & Liebe, 2012). The enhanced quality of a product is further defined as the perception of increased materials (such as better taste or longer lasting) and benefits derived from the product (such as the authenticity of the product opposed to mass production; Varul, 2009). The symbolic use value is specified as the extent to which the consumption is used to express certain beliefs (Varul, 2009). Consumers can be striving to be associated with the values a product stands for and use the product purchase as a means to express their identity (Varul, 2008). In light of fair-trade

(10)

consumption, the symbolic use value entails that the consumer is acting on a sense of ethical obligation. When purchasing these goods they facilitate fair-trade production and obtain a morally conscious image (Varul, 2009).

The division between quality purchase motives or symbolic use value purchase motives can be further categorized. In earlier research on ethical

consumption there has been much debate on the tension between fair-trade choices as a result of the inner self, opposed to being a result of the self-presentation to others (Grauel, 2014; Varul, 2008). A distinction is made between consumers who are either more internally or externally motivated to purchase fair-trade goods. The first

mentioned refers to consumers who are driven by their true inner values that reflect the importance of establishing equality in the world (Grauel, 2014; Varul, 2009). They are driven by motives originating from their pursuit to do the right thing. Internally motivated consumers are concerned with ethics and therefore the question ‘what is good?’ stand central to their purchase consideration process (Varul, 2009). In the consumption of fair-trade products, the self forms the main target audience in reaffirming their values and justifies their chosen behaviour. In justifying the higher price of fair-trade goods, intrinsically motivated consumers are thought to be

concerned with quality motives: they assign increased material characteristics to the product.

This opposed to externally motivated who are driven by obtaining the status of someone who is doing good (Grauel, 2014; Varul, 2009). They are driven by identity motives originating from the urge to express their selves as morally conscious. Extrinsically driven consumers are mainly concerned with aesthetics and

consequently revolve their behaviour around the question ‘what looks good?’ (Varul, 2009). While consuming fair-trade goods, others form their main target audience that

(11)

reaffirm and strengthen their behavioural choices. In validating the price gap between fair-trade and regular products, extrinsic motivations match with the fact that

consumers add extra symbolic value to their purchase: they use the product to create and express a certain image. Both types of consumers can feel equally committed and motivated to purchase fair-trade goods (Grauel, 2014).

2.3 Social monitoring theory

To shine more light on this tension between internal and external motives for purchasing fair-trade products, the social-psychological concept of self-monitoring (Snyder, 1974) is of valuable use. This theory suggests that individuals differ in the extent to which they observe and control their expressive behavior and

self-presentation (Myers & Sar, 2013). More specifically, a distinction is made between high or low self-monitors (Snyder, 1974). High self-monitoring individuals are focused on managing their self-presentation through expressive behavior. They are fairly concerned with projecting an image that allows them to meet the requirements of a social situation (Shavitt, 1992). This stands in contrast to low self-monitors who do not attempt to frame their behavior in a way that it reflects a certain image to their social surroundings (Myers & Sar, 2013). They are less concerned with social

appropriateness and more concerned about being consistent with their internal feelings and preferences (Shavitt, 1992).

In the current study, a relationship between the different motives for fair-trade consumption and the degree of self-monitoring is proposed. It seems likely that internally motivated consumers are low self-monitoring in nature and externally motivated consumers are high self-monitoring in nature. When categorizing consumers according to this concept, an advertising opportunity is created to

(12)

Petty & Wegener, 1998). In the context of consumer products, advertisements with quality appeals are more persuasive for low self-monitors, focusing on the inherent qualities and benefits of the product (Snyder & DeBono, 1985). This in contrast to high self-monitors who are persuaded better by advertisements containing image appeals, emphasizing the impressions created through the use of the product (Shavitt, 1992).

These tailored advertising opportunities originate from the different focus high and low self-monitors tend to have when evaluating a product. They differ especially when it concerns a product that could, besides from its functional purposes, be used to fulfill a social identity function (Shavitt, 1992). The self-monitoring concept is

applicable to fair-trade products because it offers high and low self-monitors the opportunity to focus on different functional goals they aim to fulfill (Shavitt, 1992). In their purchase consideration process, low self-monitors focus on utilitarian goals (Snyder & DeBono, 1985): they mainly pay attention to the functionality of the product. In the case of fair-trade this could for instance be a high-quality material. High self-monitors focus on social identity goals (Snyder & DeBono, 1985): they are mainly concerned with the image related to the product. In the case of fair-trade products this could be a moral consciousness. Based upon this, fair-trade

advertisements containing an image aspect are expected to be more persuasive for high self-monitors and advertisements focusing on the quality are expected to be more persuasive for low self-monitors.

H1: The effect of quality versus image appeal is moderated by self-monitoring, such that the persuasiveness of an advertisement is to be higher for high self-monitors when it contains an image appeal, and for low self-monitors when it contains a quality appeal.

(13)

Because of the proposed synchrony between purchase motives and self-monitoring, it is expected that the same effects occur for those who are driven by symbolic use value motives opposed to those who are driven by quality motives.

H2: The effect of quality versus image appeal is moderated by purchase motives, such that the persuasiveness of an advertisement is to be higher for persons who find symbolic use value motives more important when it contains an image appeal, and for persons who find quality motives more important when it contains a quality appeal.

The concept of self-monitoring revolves around expressive behaviour, which highlights the importance of consumption visibility in case of high self-monitors. For low self-monitors this is likely to be of lesser influence because they do not frame their behaviour towards others. Expressive consumptions are more applicable to the social monitoring theory than those that take place in the private sphere. There used to be a clear line between expressive and non-expressive behaviour, with the first

mentioned referring to actions that are perceived by others than the person performing them (Snyder, 1974). Nowadays, consumers are able to share any product purchase on social networks (Kietzmann, Hermkens, McCarthy & Silvestre, 2011) so that even private consumptions can be made visible and therefore expressive in nature. It is expected that the possibility to share a purchase with others influences the purchase behaviour of low self-monitors. But, because high self-monitors do not value visibility, it is expected that such a possibility does not affect their behaviour. Relating this to fair-trade advertising, this possibility should positively influence the persuasiveness of high self-monitors, but make no difference for low self-monitors.

H3: The possibility to share a purchase on social networks leads to a higher persuasiveness for high self-monitors, but not for low self-monitors.

(14)

Again, it is expected that the same effects occur for those who are driven by symbolic use value motives opposed to those who are driven by quality motives.

H4: The possibility to share a purchase on social networks leads to a higher

persuasiveness for persons with more dominant symbolic use value motives, but not for persons with more dominant quality motives.

In sum, the current study is aiming to shed more light on how to approach consumers most effectively in purchasing fair-trade products. A distinction is made between low and high self-monitors. This is thought to run parallel to the type of motives consumers have for purchasing fair-trade products: quality motives or symbolic use value motives. Testing of the relationship between self-monitoring and purchase motives was not a purpose in this study because no one-on-one relationship exists between them. Both viewpoints are treated separately to distinguish the most fruitful approach for improving fair-trade advertising. Based upon both concepts it is expected that particular advertisement styles are more appropriate for certain groups. The concept of self-monitoring has not been applied to fair-trade consumption before but seems well suited to do so because it evokes different functional goals people can focus on. The effectiveness of advertisements containing either quality or image claims, is tested. The influence of purchase visibility on the persuasiveness of consumers is also studied. The focus lies upon visibility on social media and this has not been studied before in relationship to self-monitoring nor purchase motives. But, because of the growing presence of social media in people’s lives, this is an important aspect that should not be ignored when exploring fair-trade advertising opportunities. The persuasiveness is measured through the variables purchase intention, the attitude towards the purchase and the attitude towards the advertisement because these are the

(15)

most commonly used indicators when testing advertisement effects (Jae, 2011; Khandelwal & Bajpai, 2011). See Figure 1 for the conceptual model.

H1 H2 H4

H3

Figure 1. Conceptual model.

Methods

3.1 Study design and sample

To answer the research question, an experiment was conducted. The experiment had a 2 (advertisement focus: image appeal vs. quality appeal) x 2 (visibility: possibility to share purchase on social media vs. no possibility to share purchase on social media) factorial between subjects design.

A basic premise in this research was that anyone has the potential to be a fair-trade consumer (Grauel, 2014; Varul, 2009). Therefore, the target group of the study entailed all Dutch-speaking persons over the age of 18 years. Respondents were approached by the use of the researcher’s network, both through social media and

Advertisement Quality Appeal

vs. Image Appeal

Visibility

Sharing on social media vs. No sharing on social media Persuasiveness - Attitude Advertisement - Attitude Purchase - Purchase Intention Self-monitoring High vs. Low Purchase Motives Quality vs. Symbolic Use Value (Image)

(16)

email. In total 176 persons participated in the study. After clearing out the data set, in which respondents were removed who either answered one or more of the control questions incorrect or failed to complete the survey, a total of 153 participants remained. From these respondents 74.5% was female. The average age was 33.07 years (SD = 14.96).

Participants were randomly assigned to one of the four conditions: 36 persons in the image appeal/sharing condition, 37 in the image appeal/no sharing condition, 38 persons in the quality appeal/sharing condition and finally 42 persons in the quality appeal/no sharing condition. The level of self-monitoring was measured and

participants were divided into groups of high and low self-monitors within each condition. Respondents were also divided into groups who were either more concerned with quality or image purchase motives.

3.2 Manipulations

Chocolate was chosen as the product for the advertisement. The purchase of chocolate is in general not an expensive, high-risk expense. Therefore, it was expected that the confrontation with a single advertisement could influence the respondent’s purchase intention, attitude towards the purchase and attitude towards the advertisement. To avoid the influence of existing attitudes a fictitious brand was chosen, named ‘Fair Treats Chocolate’.

3.2.1 Advertisement appeals

When it comes to fair-trade advertising, two aspects are found to be important when creating advertisements (Andorfer & Liebe, 2012). First of all, a fair-trade label is essential because it is an indicator of the nature of the product (Andorfer & Liebe, 2012). Therefore, advertisements developed for this study contained the label of Fairtrade International. This one was chosen because it is the most well known label

(17)

and as it contains the word ‘Fairtrade’ it will leave no ambiguity to what it stands for. Secondly, facilitating a sense of a relationship between the consumer and producer is an important characteristic in fair-trade promotion (Alexander & Nicholls, 2006). This is achieved by offering background information about the producer. Therefore, all advertisements developed for this study comprised of a picture of a cacao farmer accompanied by his name and country of origin.

Before constructing the stimulus material for the main experiment, several advertisements were developed. A pre-test was conducted to determine which layout was designed most professionally and which texts were best suited to reflect an image versus quality-oriented style. To offer respondents different options to choose from while judging the professionalism, three advertisement styles were developed. These designs (see Appendix A) varied on the following aspects: picture of a cacao farmer, country of origin, background picture and the size and place of the fair-trade label. To create either an image appeal or quality appeal, the advertisements differed in their textual encouragements to buy a fair-trade chocolate bar. For both appeals, three different texts were developed and incorporated in each of the advertisement designs, resulting in six different advertisements. In the pre-test (n = 17) participants judged all advertisement on the following items: ‘this advertisement is focussing on the image of possible buyers of the product’ and ‘this advertisement is focussing on the quality of the product’. Answers were measured on a 7-point scale ranging from 1 = completely disagree to 7 = completely agree. The texts of the two advertisements that were rated as most corresponding with the intended appeals were used for the final

advertisements. For the image appeal, the following text was selected (M = 5.35, SD = 1.69): ‘Do you want to show that you are someone who cares for other people? Then you buy Fair-trade.’ For the quality appeal, the following text was selected (M = 4.74,

(18)

SD = 1.77): ‘Do you like quality and a unique flavour? Then you buy Fair-trade.’ Participants were asked which of the three layouts was designed most professionally. The most often preferred design was used for the final advertisements. For the main experiment, new advertisements were created for each appeal type based on a combination between the best-preferred design and best reflecting text. Advertisements across conditions only differed in textual elements.

3.2.2 Visibility

The advertisements were presented in a simulated online context. They were shown in a browser in which the website ‘www.fairtreatschocolade.nl’ was displayed. The website consisted of a print screen with the advertisement as the main subject (see Appendix B). Aside from this, some browsing options were shown to make sure that participants experienced the print screen as a website. Both the advertisements appeals were presented in two contexts: with the possibility to share the purchase on social media platforms or without such a possibility. This sharing opportunity was created by the display of a button next to the advertisement saying ‘Buy Fair Treats Chocolate now and share on social media!’. Sharing buttons of Twitter, Facebook, LinkedIn and Google + were included because these platforms are known to be the most popular (Moreau, 2015). In the no sharing condition a button was displayed as well, but only contained the words ‘Buy Fair Treats Chocolate now!’.

3.3 Measures

Most variables were based on existing scales that were originally written in English. For the purpose of this research, these items were translated into Dutch (see Appendix C). The items measuring purchase motives form an exception because they were not based on pre-existing scales and developed especially for this study.

(19)

Respondents were approached by email or social media with the request to participate in a thesis study. A link led them to the online experiment. They were informed about the research and their anonymous participation, agreed to the terms and acknowledged their voluntary participation (see Appendix D). Questions were posed in the below described order. The demographics age, gender, educational level and income were measured. Participants were randomly assigned to one of the four advertisement conditions and were told that several questions concerning the content would be asked. Only those who answered these three knowledge questions about the content correctly were included in the analysis. Two questions regarding the nature of the advertisement followed, namely the extent to which they judged it to be image-oriented and quality-image-oriented. The purchase intention, the attitude towards the purchase and the attitude towards the advertisement were measured followed by the attitudinal and behavioural control variables and at last the items concerning self-monitoring. The questionnaire ended with thanking the respondents for their participation and the possibility to leave their email address in case they wanted to receive information regarding the research results.

3.3.1 Independent variables

Self-monitoring. This concept was measured by use of the revised self-monitoring scale (Lennox & Wolfe, 1984). The scale consisted of thirteen items in total, but comprised of two subscales measuring different constructs: the ability to modify self-presentation (seven items) and the sensitivity to expressive behaviour of others (six items). An item on the first subscale was for instance: ‘When I feel that the image I am portraying isn't working, I can readily change it to something that does’. An example item of the second subscale was: ‘I am often able to read people’s true emotions correctly through their eyes’. Items of both constructs were measured on a

(20)

bipolar 5-point Likert scale, ranging from 1 = completely disagree to 5 = completely agree. The dimensions were treated as separate indicators of self-monitoring. Low scores indicated a low degree of self-monitoring and high scores indicated a high degree of self-monitoring. Reliability analyses are discussed in chapter 3.4.

Purchase motives. To determine whether respondents were driven more dominantly by quality or symbolic use value motives, both motives were measured. First of all, respondents were asked to what extent they consider the quality of the product when purchasing fair-trade chocolate. Their placed importance on the following items was measured: product characteristics, user benefits, satisfaction of needs and the consideration of quality. These concepts were formulated in the

following format: ‘If I were to buy chocolate with a fair-trade label…’ followed up by the item descriptions, for example: ‘…I mainly pay attention to the factual

characteristics of the chocolate’. Secondly, respondents were asked to what extent they place importance on the symbolic use value when consuming fair-trade

chocolate. Four items were included that reflected the following concepts: personal expression, personality reflection, establishment of identity and self-judgement. The same format was used: ‘If I were to buy chocolate with a fair-trade label…’ followed by the four items descriptions, for instance: ‘…through this, I could show others the type of person I am’. Both motives were measured on 5-point Likert scales ranging from 1 = completely disagree to 5 = completely agree. Reliability analyses are discussed in chapter 3.4.

3.3.2 Dependent variables

To measure the persuasiveness of the different advertisements, three variables were measured.

(21)

Attitude towards advertisement. This variable was measured by use of a scale of Mitchell and Olson (1981). The scale consisted of four items, all measured on a bipolar semantic differential 5-point scale. The central statement was ‘I find the advertisement…’ with answers ranging from one to five on the following concepts: bad-good, negative-positive, unpleasant-pleasant and poor quality-high quality. The items unattractive-attractive and not informative-informative were added to obtain in total two general items, two items of affective nature and two items of cognitive nature. Altogether, the six items formed a reliable scale (Cronbach’s Alpha = .88).

Attitude towards purchase. The attitude towards the act of purchasing the product was measured through several items based upon a scale of Mitchell and Olson (1981). The scale consisted of three items revolving around the statement ‘I find the thought of purchasing the product...’. Items were measured on a bipolar 5-point semantic differential scale ranging from one to five on the following concepts: bad-good, foolish-wise, beneficial-harmful. Three more items were added to the scale, which covered the affective component of the attitude towards the purchase:

unattractive-attractive, unappealing-appealing and unlikeable-likable. Altogether, the items formed a reliable scale (Cronbach’s Alpha = .89).

Purchase intention. The purchase intention, to what extent someone is

considering to buy a certain product, was measured by the use of a scale of Spears and Singh (2004). The question ‘to what extent are you considering to purchase the

product in the advertisement?’ stood central in the measurement. The scale consisted of five items measured on a bipolar 5-point semantic differential scale, ranging from one to five on the following concepts: never-definitely, definitely no intention to buy-definite intention to buy, very low-high purchase interest, buy-definitely not buy

(22)

it-definitely buy it and probably not buy it-probably buy it. The items formed a reliable scale (Cronbach’s Alpha = .94).

3.3.3 Control variables

Some control variables were taken into account to possibly stabilize for the effects of these characteristics. The following demographic characteristics were measured: gender, age, education level and income. Secondly, some attitudinal variables were taken into account: the liking of chocolate, the attitude towards chocolate and the attitude towards fair-trade. For the first mentioned, respondents were simply asked to rate how much they agreed with the statement ‘I like chocolate’ on a 5-point Likert scale ranging from 1 = completely disagree to 5 = completely agree. The second variable was measured based upon an attitude scale (Shimp & Stuart, 2004). The scale consisted of four 5-point semantic-differentials (Cronbach’s Alpha = .91) measuring the degree of repulsiveness towards a product, ranging from one to five on the following concepts: unappealing-appealing, disgusting-not disgusting, distasteful-tasteful and revolting-not revolting. The third variable was measured based upon an attitude scale from Spears and Singh (2004) and revolved around the statement ‘I find fair-trade products…’. It consisted of the following five items (Cronbach’s Alpha = .88) measured on a 5-point semantic differential scale ranging from one to five: unappealing-appealing, bad-good, unpleasant-pleasant, unfavourable-favourable, unlikeable-likable. Lastly, several behavioural variables were taken into account: previous purchase behaviour of chocolate, previous purchase behaviour of fair-trade chocolate, general purchase behaviour of fair-trade products.

(23)

These variables entailed the following answer options: never, rarely, sometimes, most of the time and very often.

3.3.4. Manipulation check

To test whether the advertisements were experienced in the intended way, either image-oriented or quality-oriented, a manipulation check was conducted. The same questions that were used in the pre-test were included. Both the items ‘this advertisement is focussing on the image of possible buyers of the product’ and ‘this advertisement is focussing on the quality of the product’ followed after exposure to the advertisement. Respondents evaluated these statements on a 5-point Likert scale ranging from 1 = completely disagree to 5 = completely agree.

3.4 Procedure

Before the final analyses could be conducted, several scales needed further adjustment.

Self-monitoring. This scale consisted of thirteen items and was based upon two subscales: the ability to modify self-presentation (seven items) and the sensitivity to expressive behaviour of others (six items). These dimensions were treated

separately when analysing the data. The scales correlated significantly (r = .20, p = 0.01) but not high enough to be concerned with multicollinearity. For the ability to modify self-presentation scale the forth and sixth item were reversed because they were measured in the opposite direction. The scale had a moderate reliability (Cronbach’s Alpha = .74). Items of the sensitivity to expressive behaviour of others scale also formed a reliable scale (Cronbach’s Alpha = .81).

For both variables, a median split was conducted to create groups of low and high self-monitors. For the ability to modify self-presentation scale, 43.8% of the respondents were low self-monitors with scores of 3.57 or lower (M = 3.11, SD =

(24)

0.38) opposed to 56.2% of the respondents who were high self-monitors with scores higher than 3.57 (M = 3.87, SD = 0.25). For the sensitivity to expressive behaviour of others scale, 49% of the respondents were low self-monitoring in nature with scores of 3.67 or lower (M = 3.30, SD = 0.39) opposed to 51% of the respondents who were high self-monitors with scores higher than 3.67 (M = 4.12, SD = 0.33).

Purchase motives. Both quality-oriented motives and image-oriented motives were measured through four items. Because these were not based on existing scales, a principal component analysis was conducted with all eight items. Results showed that they clearly loaded on two different components: image motives (EV = 2.66; R2 = .33) and quality motives (EV = 1.75; R2 = .22). However, the quality motives scale had a low reliability (Cronbach’s Alpha = .56) that could not be improved. The items measuring image motives did form a reliable scale (Cronbach’s Alpha = .83). For further analysis a new variable was created in which both motives were combined. The image motives scores (M = 3.07, SD = .83) were extracted from the quality motives scores (M = 3.64, SD = .54), resulting in the new variable purchase motives. Respondents who scored below zero on this scale (34%) were more dominantly concerned with image-oriented motives (M = -.43, SD = .54) and those who scored above zero (66%) focussed more dominantly on quality motives (M = 1.06, SD = .62).

Results

4.1 Manipulation check

A one-way ANOVA was conducted to determine whether the manipulation of the image versus quality condition was successful. Respondents in all conditions were asked to what extent the shown advertisement was image-oriented and

quality-oriented. Respondents in the image condition judged the advertisement significantly (F (1, 151) = 10.37, p < .01, η2 = .253) more image-focussed (M = 4.10, SD = .85)

(25)

than those in the quality condition (M = 3.64, SD = .90). Respondents in the quality condition experienced the advertisement more quality-focussed (M = 3.20, SD = 1.02) than those in the image condition (M = 2.95, SD = 1.21) but this difference was not significant (F (1, 151) = 1.98, p = .16).

4.2 Control variables

To establish whether relationships existed between the control variables and the dependent variables, several bivariate correlation analyses were conducted. For all control variables measured at interval or ratio level, the Pearson correlation was measured and for the dichotomous variable gender, the Spearman correlation was estimated (see Table 1). With regards to the control variables measured at ordinal or nominal level, Chi-square tests were conducted (see Table 2).

Table 1.

Correlation analyses control variables

Age Gender# Liking Chocolate Attitude Chocolate Attitude Fair-trade Attitude Advertisement -.145 .206* .046 .059 .304** Attitude Purchase -.123 .254** .032 .111 .502** Purchase Intention .005 .208** .172* .200* .391**

Notes. ** Correlation is significant at .01 level, * Correlation is significant at .05 level # For gender, Spearman’s rho is estimated; for all other variables Pearson Correlation is estimated

Table 2.

Chi Square tests control variables

Income Education Previous Purchase Chocolate Previous Purchase FT-chocolate Previous Purchase FT-products Attitude Advertisement 138.16 65.70 88.05 110.69* 97.19 Attitude Purchase 141.30* 69.88 81.43 75.29 78.14

(26)

Purchase Intention

104.16 43.71 88.71 109.73** 113.45**

Notes. ** Difference is significant at a .01 level, * Difference is significant at a .05 level

The liking of chocolate, the attitude towards chocolate, the attitude towards fair-trade, gender, income, previous purchase behaviour of fair-trade chocolate and previous purchase behaviour of fair-trade products were related to one or more of the dependent variables and were therefore taken into account as covariates throughout the analysis of the data.

4.3 Hypotheses

Looking at the total average of all respondents, the advertisements caused moderate persuasion: attitude towards advertisement (M = 3.42, SD = .74), attitude towards purchase (M = 3.81, SD = .63) and purchase intention (M = 3.06, SD = .92).

4.3.1 Self-monitoring and advertisement appeals

The first hypothesis predicted that the persuasiveness of an advertisement would be higher for high self-monitors when containing an image appeal and for low self-monitors when containing a quality appeal. See Table 3 for an overview of the mean differences between conditions.

Table 3.

Means and Standard Deviations of the dependent variables on Self-monitoring and Advertisement Appeal conditions.

Self-monitoring Ability to modify

self-presentation

Sensitivity to expressive behaviour

of others

Low High Low High

Attitude Advertisement A d ve rtise m en t A p p eal Quality M = 3.49 _{SD = .82} M = 3.33 _{SD = .78} M = 3.42 _{SD = .74} M = 3.39 _{SD = .87} Image M = 3.38 SD = .68 M = 3.49 SD = .69 M = 3.29 SD = .72 M = 3.57 SD = .64 Attitude Purchase Quality M = 3.65

SD = .73 M = 3.85 SD = .60 M = 3.75, SD = .47 M = 3.73 SD = .86

(27)

Image M = 3.89 SD = .57 M = 3.95 SD = .55 M = 3.79 SD = .60 M = 4.03 SD = .49 Purchase Intention Quality M = 2.99 SD = .95 M = 2.82 SD = .92 M = 2.79 SD = .86 M = 3.02 SD = 1.01 Image M = 3.29 SD = .81 M = 3.19 SD = .94 M = 3.17 SD = .96 M = 3.28 SD = .82

A MANCOVA analysis was conducted to test whether differences between conditions were significant (see Table 4). All described covariates, the three dependent variables, the advertisement type and both subscales of self-monitoring were taken into account. Neither for the self-monitoring scales nor for the

advertisement appeal, significant main effects were found on the dependent variables. Additionally, results showed no significant interaction effects. Therefore, the first hypothesis was rejected. Although some of the effects were in the expected direction, differences between conditions were not significant. The persuasiveness of an

advertisement was not higher for high self-monitors when containing an image appeal or for low self-monitors when containing a quality appeal. The hypothesis was also tested in MANCOVA analyses where both self-monitoring scales were included separately, producing similar non-significant results.

Table 4.

Effects of Self-monitoring and Advertisement Appeal on dependent variables.

F-values MANCOVA Attitude Advertisement Attitude Purchase Purchase Intention Advertisement Appeal .59 1.05 2.51

Ability to modify

self-presentation .02 3.62 .01 Sensitivity to expressive behaviour of others .21 .00 .01 Ability to modify Self-presentation * Advertisement Appeal .61 .03 .01 Sensitivity expressive behaviour of others * .13 .42 .53

(28)

The following variables were included as covariates in the analysis: the liking of chocolate, the attitude towards chocolate, the attitude towards fair-trade, gender, income, previous purchase behaviour of fair-trade chocolate and previous purchase behaviour of fair-trade products.

4.3.2 Purchase motives and advertisement appeals

The second hypothesis predicted that for persons who judged symbolic use value motives as the most important purchase motive, the persuasiveness of an advertisement would be higher when containing an image appeal. But for those who found quality motives more important, the persuasiveness of an advertisement was expected to be higher when containing a quality appeal. See Table 5 for an overview of the mean differences between conditions.

Table 5.

Means and Standard Deviations of the dependent variables on Purchase Motives and Advertisement Appeal conditions.

Purchase Motives

Quality Symbolic use

Value/Image Attitude Advertisement Adver tise m en t A p p eal Quality M = 3.20, SD = .79 M = 3.73, SD = .71 Image M = 3.37, SD = .69 M = 3.63, SD = .67 Attitude Purchase Quality M = 3.58, SD = .73 M = 3.99, SD = .49 Image M = 3.85, SD = .58 M = 4.08, SD = .47 Purchase Intention Quality M = 2.66, SD = .81 M = 3.28, SD = .99 Image M = 3.22, SD = .82 M = 3.27, SD = 1.02

To test whether differences between these conditions were significant, another MANCOVA analysis was conducted. All described covariates, the three dependent variables, the purchase motives and the advertisement type were taken into account. Table 6 shows only one significant result. There was a main effect of purchase motives on the attitude towards the advertisement, indicating that respondents who scored high on this variable, meaning they more dominantly valued quality motives,

(29)

had a more positive attitude towards the advertisement. However the interaction effect showed that this was independent of the advertisement appeal (F (1,134) = 1.79, p = .18). Because none of the interaction variables had a significant effect, the second hypothesis was rejected. Respondents with more dominant symbolic use value purchase motives were not persuaded better by advertisements containing an image appeal. Neither were respondents with more dominant quality purchase motives persuaded better by advertisements containing a quality appeal.

Table 6.

Effects of Purchase Motives and Advertisement Appeal on dependent variables.

F-values MANCOVA Attitude Advertisement Attitude Purchase Purchase Intention Advertisement Appeal .07 1.05 1.30 Purchase Motives 5.05* 3.19 2.08 Purchase Motives * Advertisement Appeal 1.79 .35 3.69

4.3.3 Self-monitoring and purchase visibility The third hypothesis predicted that the possibility to share a fair-trade purchase on social networks would lead to a higher persuasiveness for high self-monitors, but not for low self-monitors. See Table 7 for an overview of all mean scores across conditions.

Table 7.

Means and Standard Deviations of the dependent variables on Self-monitoring and Possibility to share purchase.

Self-monitoring

Ability to modify self-presentation

Sensitivity to expressive behaviour

(30)

Attitude Advertisement Visi b il ity on S oc ial M ed ia Sharing Buttons M = 3.31 SD = .87 M = 3.38 SD = .81 M = 3.31 SD = .81 M = 3.40 SD = .86 None M = 3.54 SD = .66 M = 3.45 SD = .65 M = 3.42 SD = .64 M = 3.55 SD = .66 Attitude Purchase Sharing Buttons M = 3.70 SD = .81 M = 3.89 SD = .61 M = 3.80, SD = .56 M = 3.84 SD = .84 None M = 3.86 SD = .56 M = 3.82 SD = .55 M = 3.73 SD = .52 M = 3.93 SD = .57 Purchase Intention Sharing Buttons M = 3.07 SD = 1.07 M = 3.06 SD = .98 M = 3.02 SD = .95 M = 3.11 SD =1.07 None M = 3.17 SD = .75 M = 2.94 SD = .91 M = 2.89 SD = .89 M = 3.21 SD = .77 To test whether differences between conditions were significant, another MANCOVA was conducted. All described covariates, the three dependent variables, the presence or absence of social media buttons and both self-monitoring subscales were taken into account. Table 8 shows only one significant result. There was a main effect of the ability to modify self-presentation on the attitude towards the purchase, such that respondents who score high on this variable had a more positive attitude towards the purchase. But the interaction effect showed that this was independent of the presence of social media buttons (F (1,130) = .75, p = .39). Because none of the interaction variables had a significant effect, the third hypothesis was rejected. Contrary to expectations, the presence of social media sharing buttons did not lead to a higher persuasiveness for high-self monitors opposed to low self-monitors. The hypothesis was also tested in MANCOVA analyses where both self-monitoring scales were included separately, producing similar non-significant results.

Table 8.

Effects of Self-monitoring scales and Purchase Visibility on dependent variables.

F-values MANCOVA Attitude Advertisement Attitude Purchase Purchase Intention

Visibility Social Media 2.96 .68 .10

Ability to modify

self-presentation .26 5.45* .07

Sensitivity to expressive

(31)

Ability to modify self-presentation * Visibility Social Media

.66 .75 .58

Sensitivity expressive behaviour of others * Visibility Social Media

.16 1.42 3.01

4.3.4 Purchase motives and purchase visibility

The fourth hypothesis predicted that the possibility to share a purchase on social networks would lead to a higher persuasiveness for persons with more

dominant symbolic use value motives, but not for persons with more dominant quality motives. See Table 9 for an overview of all mean scores across conditions.

Table 9.

Means and Standard Deviations of the dependent variables on Purchase Motives and Possibility to share purchase.

Purchase Motives

Quality Symbolic use

Value/Image Attitude Advertisement Visi b il ity on S oc ial M ed ia Sharing _Buttons M = 3.21, SD = .84 M = 3.57, SD = .78 None M = 3.35, SD = .65 M = 3.81, SD = .55 Attitude Purchase Sharing Buttons M = 3.71, SD = .77 M = 3.99, SD = .53 None M = 3.85, SD = .58 M = 4.07, SD = .41 Purchase Intention Sharing Buttons M = 3.00, SD = .94 M = 3.16, SD = 1.12 None M = 2.90, SD = .80 M = 3.42, SD = .82

To test whether differences between these conditions were significant, another MANCOVA analysis was conducted. All described covariates, the three dependent variables, the purchase motives and the advertisement type were taken into account.

(32)

Table 10 shows only one significant result. There was a main effect of purchase motives on the attitude towards the advertisement, indicating that respondents who scored high on this variable, meaning they more dominantly valued quality motives, had a more positive attitude towards the advertisement. But the interaction effect showed that this was independent of the advertisement appeal (F (1,134) = .39, p = .53). Because none of the interaction variables had a significant effect, the fourth hypothesis was rejected. The presence of social media sharing buttons did not lead to a higher persuasiveness for respondents with more dominant symbolic use value purchase motives opposed to those with more dominant quality purchase motives. Table 10.

Effects of Purchase Motives and Purchase Visibility on dependent variables.

F-values MANCOVA Attitude Advertisement Attitude Purchase Purchase Intention

Visibility Social Media 3.73 .07 .26

Purchase Motives 6.29* 2.92 1.89

Purchase Motives *

Visibility Social Media .39 .01 1.56

Conclusion and discussion

In the current study it was attempted to segment consumers according to the differences between internal versus external motivations to purchase fair-trade products. Consumers were categorized by use of the self-monitoring theory and different purchase motives. The role of image versus quality appeals in

advertisements and the purchase visibility were studied to answer the research question: while considering image-oriented advertising versus quality-oriented advertising and the role of purchase visibility, how can consumers be persuaded most

(33)

effectively to purchase fair-trade products, when a distinction is made between internally driven, low monitoring consumers and externally driven, high self-monitoring consumers? Relating the findings back to this question, it is concluded that both high and low self-monitors are persuaded to an equal extent by

advertisements regardless of the appeal type used. The same applies to respondents with either symbolic use value purchase motives or quality purchase motives. In contrast to hypothesis one and three, image-oriented advertisements were not better in persuading high self-monitors or respondents with symbolic use value motives, nor were quality-oriented advertisements better suited to persuade low self-monitors or respondents with quality motives. Additionally, the optional visibility of the purchase on social media did not increase the persuasiveness for high self-monitors or

respondents that were driven by image purchase motives, rejecting hypothesis two and four. The expected differences did not occur in this study. Based on the results of this research, fair-trade companies are advised to approach their target group as a whole because segmentation according to internally versus externally driven consumers did not appear to be fruitful.

There are several limitations in this research that might explain the lack of the expected differences between groups. First of all, respondents were not representative for the general population of people of 18 years and older. For instance, many more women (74.5%) than men participated in the study. Furthermore, the main majority was highly educated (69.3% WO; 23.3% HBO). Research results are therefore not a correct reflection of the general population. Secondly, the manipulation check was only half successful. Only the image-oriented advertisement was experienced in the intended way, but results were not significant for the quality-oriented advertisement. With the manipulation being a crucial element in the success of an experiment, this

(34)

might have caused the final results to deviate from expectations. Thirdly, results concerning purchase motives might not have been reliable. These scales were not based on pre-existing items and the scale measuring quality motives had a low reliability. For the final scale construction of the variable purchase motives, both scales were merged into one. Analyses were conducted using this scale despite of its reliability problems. The majority of the respondents scored quite similar on this scale, with all scores being close to zero. Only a few of them scored utterly different resulting in a low variance between groups, which might have caused the

non-significant results. A solution would have been to exclude all respondents with scores between -1 and +1, but this would have left a small sample and it was therefore chosen not to. Similar problems occurred concerning the variable self-monitoring. Here also appeared to be little variation in the scores on these scales. Again, if respondents with scores between -1 and +1 were to be excluded, not enough

respondents would have remained to conduct analyses. Nonetheless, groups of high and low self-monitors were created through a median split, which did not differ a great deal from each other. Mean scores of low self-monitors were above midpoints of the scales, indicating that most of them in fact were high self-monitors. The possible solution of using a midscale split was averted because too few respondents scored below three. A final limitation is that the perceived importance of social media was not taken into account. The possibility to share represented the placed value on purchase visibility. It seems however plausible that for some, social media plays a more important role than for others. This was not accounted for and may have caused insignificant results with regards to the purchase visibility hypotheses.

These limitations form the pillars for recommendations for future research. Future studies should aim to get a more diverse pool of respondents that would allow

(35)

results to be generalizable to the target population. Also, researchers should be cautious when developing their advertisements. In the current research it was chosen to have the advertisements as similar as possible to avoid influences of other

advertisement aspects on the persuasiveness for different groups. Therefore, the difference between quality and image-oriented advertisement was solely textual. In a follow-up study, visual elements could also be taken into account when creating the differences between image or quality appeals, which might lead to a more successful manipulation. Furthermore, a new scale measuring quality purchase motives should be developed with more extensive pre-testing so that it does not lack in reliability, as was the case in the current research. New research should also aim to make sure that the respondents really differentiate in the degree of self-monitoring and purchase motives. A bigger sample should be studied in order that only those that have extreme scores on these scales are selected for the analyses. Lastly, when studying purchase visibility on social media, the focus should be on a sample of respondents that are representable for active social media users.

It is premature to make theoretical assumptions based on this research, due to the many limitations it entails. Neither form these results a steady base to form practical recommendations for fair-trade companies. Earlier research has proven that self-monitoring is a worthy concept for segmentation and tailored advertising. For instance, Snyder and DeBono (1985) did find significant effects for the effectiveness of image appeals for high self-monitors and quality appeals for low self-monitors. Additionally Zuckerman, Gioioso and Tellini (1988) found that image appeals were more effective influencers for high self-monitors opposed to quality appeals. In conformity with this, Petty and Wegener (1998) concluded that matching such an appeal type to someone’s degree of self-monitoring enhanced attitude change

(36)

effectively. The moderating role of self-monitoring on the effectiveness of different adverting strategies was also demonstrated in more recent research (Myers, 2013).

The theoretical foundation of this research is still relevant to maintain.

Differences between internally versus externally motivated consumers in relationship to fair-trade purchases remains an effective starting point in the exploration of new advertising opportunities. Future research should investigate the role of

self-monitoring and purchase visibility more extensively and in further detail. References

Alexander, A., & Nicholls, A. (2006). Rediscovering consumer-producer involvement: A network perspective on fair-trade marketing. European Journal of Marketing, 40, 1236-1253.

Andorfer, V. A., & Liebe, U. (2012). Research on fair-trade consumption - a review. Journal of Business Ethics, 106, 415-435.

Andorfer, V. A., & Liebe, U. (2013). Consumer behavior in moral markets. On the relevance of identity, justice beliefs, social norms, status, and trust in ethical consumption. European Sociological Review, 29, 1251-1265.

Bhattacharjee, A., Berger, J., & Menon, G. (2014). When identity marketing backfires: Consumer agency in identity expression. Journal of Consumer

Research, 41, 294-309.

De Pelsmacker, P., Driesen, L., & Rayp, G. (2005). Do consumers care about ethics? Willingness to pay for fair-trade coffee. Journal of Consumer Affairs, 39, 363-385.

Giddens, A. (1991). Modernity and self-identity. Stanford: Stanford University Press.

(37)

Grauel, J. (2014). Being authentic or being responsible? Food consumption, morality and the presentation of self. Journal of Consumer Culture, 14(2), 1-18.

Harrison, R., Newholm, T. & Shaw, D. (2005). The ethical consumer. Thousand Oaks: Sage publications.

Ingenbleek, P. T., & Reinders, M. J. (2013). The development of a market for sustainable coffee in the Netherlands: Rethinking the contribution of fair-trade. Journal of Business Ethics, 113, 461-474.

Jae, H. (2011). Cognitive load and syntactic complexity of printed

advertisements: Effects on consumers' attitudes. Marketing Management Journal, 21, 152-159.

Khandelwal, U., & Bajpai, N. (2011). A study on green advertisement and its impact on consumer purchase intention. Journal of Creative Communications, 6, 259-276.

Kietzmann, J. H., Hermkens, K., McCarthy, I. P., & Silvestre, B. S. (2011). Social media? Get serious! Understanding the functional building blocks of social media. Business Horizons, 54, 241-251.

Lennox, R. D., & Wolfe, R. N. (1984). Revision of the self-monitoring scale. Journal of Personality and Social Psychology, 46, 1349-1364.

McDonagh, P., & Prothero, A. (2014). Sustainability marketing research: Past, present and future. Journal of Marketing Management, 30, 1186-1219.

Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitudes? Journal of Marketing Research, 18, 318-331.

(38)

Moreau, E. (2015). Top 15 social networking sites you should be using. About.com. Retrieved from

http://webtrends.about.com/od/socialnetworkingreviews/tp/Social-Networking-Sites.htm

Myers, J. R., & Sar, S. (2013). Persuasive social approval cues in print advertising: Exploring visual and textual strategies and consumer

self-monitoring. Journal of Marketing Communications, 19(3), 168-181.

Ozcaglar-Toulouse, N., Shiu, E., & Shaw, D. (2006). In search of fair-trade: Ethical consumer decision making in France. International Journal of Consumer Studies, 30, 502-514.

Petty, R. E., & Wegener, D. T. (1998). Matching versus mismatching attitude functions: Implications for scrutiny of persuasive messages. Personality and Social Psychology Bulletin, 24, 227-240.

Shavitt, S., Lowrey, T. M., & Han, S. P. (1992). Attitude functions in advertising: The interactive role of products and self-monitoring. Journal of

Consumer Psychology, 1, 337-364.

Shaw, D., Shiu, E., & Clarke, I. (2000). The contribution of ethical obligation and self-identity to the theory of planned behaviour: An exploration of ethical

consumers. Journal of Marketing Management, 16, 879-894.

Shimp, T. A., & Stuart, E. W. (2004). The role of disgust as an emotional mediator of advertising effects. Journal of Advertising, 33(1), 43-53.

Snyder, M. (1974). Self-monitoring of expressive behavior. Journal of Personality and Social Psychology, 30, 526-537.

(39)

Snyder, M., & DeBono, K. G. (1985). Appeals to image and claims about quality: Understanding the psychology of advertising. Journal of Personality and Social Psychology, 49, 586-597.

Sparks, P., & Shepherd, R. (1992). Self-identity and the theory of planned behavior: Assessing the role of identification with "green consumerism". Social Psychology Quarterly, 55, 388-399.

Spears, N., & Singh, S. N. (2004). Measuring attitude toward the brand and purchase intentions. Journal of Current Issues and Research in Advertising, 26(2), 53-66.

Tony Chocolonely. (2014). Annual Report 2014. Retrieved from

http://www.tonyschocolonely.com/wp-content/uploads/2014/07/Tonys-Chocolonely-Annual-Fair-Report.pdf

Tony Chocolonely. (n.d.). Our mission - the road to 100% slave-free. Retrieved from http://www.tonyschocolonely.com/en/our-mission/the-road-to-100-slave-free/

Van Dinther, M. (2015). Waarom het echt beter is om meer te betalen voor eten. De Volkskrant. Retrieved from http://www.volkskrant.nl/de-tien-geboden-van-goed-eten/waarom-het-echt-beter-is-om-meer-te-betalen-voor-eten~a3922818/

Varul, M.Z. (2008). Consuming the campesino: Fair-trade marketing between recognition and romantic commodification. Cultural Studies, 22, 654-679.

Varul, M. Z. (2009). Ethical consumption: The case of fair-trade. Kölner

Zeitschrift für Soziologie und Sozialpsychologie, 49, 366-385.

Zuckerman, M., Gioioso, C., & Tellini, S. (1988). Control orientation, self-monitoring, and preference for image versus quality approach to advertising. Journal of Research in Personality, 22(1), 89-100.

(40)

Appendix A. Advertisements Pre-test

Figure 2. Advertisement one: image-appeal.

(41)

Figure 3. Advertisement two: quality-appeal. Figure 4. Advertiseme nt three: image-appeal.

(42)

Figure 5. Advertisement four: quality-appeal.

(43)

Figure 7. Advertisement six: quality-appeal.

(44)

Figure 8. Final advertisement: image/sharing on social media condition. Figure 9. Final advertisement: image/no sharing on social media condition.

(45)

Figure 10. Final advertisement: quality/no sharing on social media condition.

Figure 11. Final advertisement: quality-appeal/sharing on social media condition. Appendix C. Questionnaire

[Demografische kenmerken]

1. Hoe oud ben je? …

2. Wat is je geslacht?

o Man o Vrouw

3. Wat is je hoogst genoten (huidige) opleiding?

o Basisonderwijs o Voortgezet onderwijs o MBO

Advertising fair-trade products

Master’s Thesis