How does the valence of an OCR and the social visibility of a product influence the consumers’ attitude and purchase intention?

How does the valence of an OCR and the social visibility of a product

influence the consumers’ attitude and purchase intention?

The moderating role of the need for conformity

2

How does the valence of an OCR and the social visibility of a product

influence the consumers’ attitude and purchase intention?

The moderating role of the need for conformity

-Lysan Cherelle van der Galiën University of Groningen Faculty of Economics and Business

MSc Marketing Management Master Thesis January 12, 2015 Leeuwerikweg 52 7971 DT Havelte Tel: +31 (0)6 28150840 E-mail: lysancherelle@live.nl Student number: 1883739

3

Management summary

The aim of this research is to examine the relationship between the valence of an online consumer review and the social visibility of a product on the attitude and purchase intention of a consumer. The moderating effect of the need for conformity is tested as well.

The effects of the variables are tested by a 2 (social visibility of the product: public and private) x 3 (valence of an OCR: positive, neutral and negative) between-participants design. A questionnaire was distributed, which was completed by 205 respondents. The participants had to rate several items that respresented their attitude and purchase intention after reading an online consumer review. Also, they had to respond to statements that indicate their need for conformity.

The results of this research show that the valence of an OCR has an influence on the attitude and the purchase intention of a consumer. Also, the attitude of a consumer has an influence on the purchase intention of a consumer. However, there was no empirical evidence found for the interaction effects of the social visibility of a product and the need for conformity. There were also no main effects found for the same variables.

4

Preface

Writing this Master thesis was the last step I had to take in my journey on the University of Groningen. It started with the Bachelor International Business and ends with the Master Marketing Management. I have learned a lot from my studies and about myself in the last few years. The same goes for the months while I was writing this thesis. Overall, it has been a good and instructive process. However, there have been times when I became frustrated and wanted to give up on it. Fortunately, the topic I was assigned to matches my own interests, which made it a lot easier to carry on. Also, the people surrounding me encouraged me to continue in the difficult times. First, I want to thank my family and friends for their ongoing moral support, endless patience, and encouraging words. Especially my parents who also made it financially possible for me to study all these years. Even though it took a bit longer than expected. I am grateful. Next to that, I want to thank my first supervisor Liane Voerman for her constructive feedback, enthusiasm, quick responses to e-mails, faith, and guidance during the whole process. Lastly, I want to thank my fellow students of my thesis group for their feedback on my progress and support during our thesis meetings.

Meppel, January 2015

5

Table of Contents

1. Introduction

7

1.1 Introduction of the topic 7

1.2 Online consumer reviews 7

1.3 High involvement products 8

1.4 Problem statement 9

1.5 Theoretical and social relevance 9

1.6 Structure of the thesis 10

2. Theoretical framework

11

2.1 Independent variable: The valence of an OCR 11

2.2 Independent variable: The social visibility of a product 12

2.3 Moderator: The need for conformity 14

2.4 Dependent variables: the attitude and purchase intention of a consumer 16

2.5 Overview of hypotheses and conceptual framework 17

3. Methodology

19

3.1 Research design and procedure 19

3.2 Participants 22 3.3 Operationalization 23 3.4 Construct validation 24 3.4.1 Factor analysis 24 3.4.2 Reliability analysis 25 3.5 Manipulation checks 27 3.6 Plan of analysis 29

4. Results

32

4.1 Descriptives of all conditions 32

4.2 Regression analyses 32

4.2.1 The main effects of the independent variables 33

4.2.2 The interaction effect of the social visibility of a product 33

4.2.3 The main effects of the moderators 34

4.2.4 The interaction effect of the need for conformity 34

4.2.5 The effect of the attitude on the purchase intention of the consumer 36

6

5. Discussion, conclusion, implications and limitations

38

5.1 Discussion 38

5.1.1 The valence of an OCR 38

5.1.2 The social visibility of a product 38

5.1.3 The need for conformity 39

5.1.4 The attitude and purchase intention 39

5.2 Managerial implications 39

5.3 Limitations and future research directions 39

6. References

41

7. Appendices

46

7.1 Appendix 1: Conditions 46

7.2 Appendix 2: English questionnaire 47

7

1. Introduction

1.1 Introduction of the topic

Traditionally, the opinions that people had about products, i.e. word-of-mouth (WOM), were orally and personally shared with people they met in real life (Eisingerich et al. 2014). Products were purchased in actual stores and people were able to see, feel, and try them before buying. The development of the Internet created the possibility of online retailing, which resulted in the fact that nowadays, people have the ability to purchase almost all products from all over the world online. Yet, not having the possibility to experience the product before purchasing makes the decision process harder (Pauwels et al., 2011). As a result, the opinion of other people who purchased the product in the past becomes more important, since they are the ones who already had the experience (Park et al., 2007). Fortunately, the Internet also provides the opportunity to share your opinion online at any time with everyone who is on the world-wide-web (Kwon and Sung, 2012). This is what we call electronic word-of-mouth (eWOM). Goldsmith (2006) defines eWOM as

‘’word-of-mouth communication on the Internet, which can be diffused by many Internet applications such as online forums, electronic bulletin board systems, blogs, review sites, and social networking sites’’. The

difference of eWOM versus traditional WOM is that it can come from anonymous individuals, it is always written in text (Wu and Wang, 2011), and people can share their opinions from wherever and whenever they want (Eisingerich et al. 2014). Also, consumers perceive eWOM easier to decipher, more balanced and unbiased, and more powerful and effective compared to the traditional WOM (Floyd et al., 2014).

1.2 Online consumer reviews

eWOM contains all possible ways of word-of-mouth communication on the Internet. This is a very broad subject to examine. Therefore, this research will focus on a specific part, namely online consumer reviews (OCRs). Park et al. (2007) define OCRs as “new information presented from the

perspective of consumers who have purchased and used the product. It includes their experiences, evaluations and opinions”. This could be both reviews on the website of the company itself and

8 OCRs have a large influence on consumer behavior. First, Kwon and Sung (2012) mention that many consumers find OCRs important in their buying decisions. Senecal and Nantel (2004) state that online recommendations influence the choice of a product of a consumer. Next to that, Park and Park (2008) and Purnawirawan et al. (2013) state that the influence of OCRs on the sales of online retailers is significant, and that a consumer’s attitude and behavioral intention is influenced by the direction of a review. Literature researches different aspects of an OCR that are important such as the direction (Pan and Zhang, 2011), the source credibility (Floyd et al., 2014), the source expertise (Bansal and Voyer, 2000), and the volume (Liu, 2006). This research focuses on the direction of an OCR, also called the valence, which defines whether the content of an OCR is positive, negative, or neutral. The effect of the valence of an OCR is discussed many times. Some state that negative OCRs have more influence, while others think the opposite (Pan and Zhang, 2011). Also, it is stated that consumers tend to follow the majority of the opinions of other people about the product reviewed (Purnawirawan et al., 2013), due to the conformity tendency principle. This principle means that the judgments of others are used to form own opinions. Therefore, the need for conformity is another reason for change in consumers’ attitudes (Wood, 2000). Hoyer et al. (2013) defines conformity as

“the tendency to behave in an expected way”. The need for conformity is higher when someone

identifies himself with a group (Schlosser, 2009). Furthermore, the social visibility while using a product (public or private consumption) is a factor that influences a consumer’s attitude towards that product. Wakefield and Inman (2003) state that consumers are less price sensitive when it comes to the purchase of a product in a social context, so they might adapt their purchase behavior to that. An argument explaining this could be that there is a difference between the private reality of a person and their public appearance (MacDonald and Nail, 2005). Wood and Hayes (2012) argument that this is due to the social motives that consumers might accredit to products. This results in the fact that consumers use different criteria for evaluating their choices in a public situation than in a private situation (Lamberton et al., 2013).

1.3 High involvement products

Next to that, the type of product category could also have an influence on the attitude of a consumer towards a product. In this research, the focus lies on product involvement. This concept can be defined as “a consumer’s enduring perceptions of a product category’s importance based on that

consumer’s inherent needs, values, and interests” according to Zaichkowsky (1985). The degree of

9 on positive or negative cues during the peripheral route. The central route is taken when product involvement is high, while the peripheral route is more effective for low product involvement (Levy and Nebenzahl, 2008). The involvement theory states that consumers engage in extensive online search using e.g. OCRs for high involvement products, so the use of OCRs tends to be higher when it concerns high involvement products (Wu and Wang, 2011). High involvement products are associated with a higher perceived risk. On the other hand, consumers engage in limited online search for low involvement products, because low involvement products are associated with a lower perceived risk (Floyd et al., 2014). Therefore, high involvement products are used in this research.

1.4 Problem statement

This research investigates the influence of the valence of an OCR on the attitude and the purchase intention of a consumer. However, this relationship is already researched many times by other authors, which makes the relationship in itself not very interesting. Therefore, a distinction is make between a social visible product and not social visible product, focusing on a high involvement product, because consumers tend to use OCRs more by purchasing high involvement products (Wu and Wang, 2011). Next to that, the need for conformity is a consumer characteristic that comes back in literature about both aforementioned variables. Therefore, it is a moderator in this research. Taking all this variables together, this leads to the following research question:

“How does the valence of an OCR influence the attitude and the purchase intention of a consumer towards a product that is either social visible or not social visible, and to what extend is this relationship moderated by the need for conformity of a consumer?”

1.5 Theoretical and social relevance

10

1.6 Structure of the thesis

11

2. Theoretical framework

The theoretical framework gives more background information from relevant theories in current literature about the subjects mentioned in the introduction. All aforementioned variables will be discussed followed by relevant hypotheses. First, an explanation about the independent variables of this research is given, which are the valence of OCRs and the social visibility of a product. After that, the moderator ‘need for conformity’ will be elaborated on. Lastly, the relationship between the two dependent variables is explained. An overview of all hypotheses and the conceptual framework conclude this chapter.

2.1 Independent variable: The valence of an OCR

The valence of an OCR depends on the satisfaction of a consumer with the brand or product (Kietzmann and Canhoto, 2013). As a result, a satisfied consumer writes a positive OCR, while a dissatisfied consumer writes a negative OCR. This is due to the disconfirmation a consumer experiences (Oliver, 1985). His expectation-disconfirmation theory explains that when a consumer experiences a product better than expected, a consumer is positively disconfirmed. When a product is experienced by a consumer worse than expected, a consumer is negatively disconfirmed. A consumer is indifferent about the actual performance if its expectations of a product are confirmed. The satisfaction of a consumer is strongly related to disconfirmation (Niedrich et al., 2005). A positively confirmed consumer is most likely to write about its experience (Oliver, 1985; Donovan, 1999). Kietzmann and Canhoto (2013) built their research on this statement and their findings show that consumers are more likely to write about positively disconfirmed experiences if the experience was a face-to-face interaction. However, if the experience was an electronic or personal-but-distant (for example, a telephone call) interaction, consumers are more likely to write about negative disconfirmed expectations. Additionally, East et al. (2007) mention two studies that confirm aforementioned theories. The first study investigated comments on television programs, and found out that positive film judgments occurred twice as much as negative film judgments. The second research studied 97 users of a health and fitness resort. Of the 97 people, 94 respondents gave positive comments, while only 62 respondents gave negative comments. All the people who gave negative comments also gave positive ones. Thus, it can be expected that there are more positive than negative reviews online.

12 negativity effect also plays a role in eWOM. Park and Lee (2009) state that negative eWOM has a greater effect than positive eWOM even though consumers could be more likely to share positive experiences. An argument supporting this statement is that negative publicity is attracting more attention than positive publicity (Ahluwalia et al., 2000). Next to that, negative information is easier to recall (Purnawirawan et al., 2013), and stories about negative experiences tend to be more vivid (Donovan, 1999). East et al. (2007) state that there are more positive reviews available online, which results in the fact that negative reviews are examined more thorough than positive reviews. Therefore, negative reviews have a greater impact on consumer behavior. Negative reviews also have a large influence in sales elasticity compared to positive reviews (Floyd et al., 2014). The loss aversion principle could offer an explanation for this effect: a loss has a greater influence on perception and decision making than the same amount of gain, because the value function is steeper for losses than gains (Floyd et al., 2014). On the other hand, others find that the effect of positive information is greater than negative information. For example, Menon and Johar (1997) state that people recall positive experiences easier than negative experiences, while Chevalier and Mayzlin (2006) mention that the increase of sales is affected more by positive reviews than negative ones. Senecal and Nantel (2004) also conclude that products are selected twice as much when the products were recommended by others.

This paragraph with theories from current literature shows that valence is an important aspect of an OCR, which has an influence on the behavior of a consumer. Negative valence expects to have a negative influence on the attitude and the purchase intention of a consumer, while positive valence expects to have a positive influence on the attitude and the purchase intention of a consumer. Also, negative influence is expected to have a greater effect on the attitude and the purchase intention of a consumer than the positive valence. This leads to the following hypotheses:

H1a: Negative (positive) valence has a negative (positive) influence on the attitude (purchase intention) of a consumer towards a product

H1b: The negative influence of a negatively valenced OCR on the attitude (purchase intention) of a consumer towards a product is relatively larger than the positive influence of a positively valenced OCR

2.2 Independent variable: The social visibility of a product

13 and clothes, while private products are washing machines, refrigerators and beds. The attitude of a consumer may change depending on the social visibility of a product. The definition of MacDonald and Nail (2005) of private attitudes is “attitudes that are consciously recognizable, controllable, and

that the attitude holder believes are not directly accessible to anyone other than him or herself”,

while public attitudes are “verbal or non-verbal expressions related to an attitude domain that are

made with the belief that one or more other people are able to learn of that expression and attribute it to the attitude holder”. The findings of their research show that people often have different public

and private attitudes.

This difference in attitudes can lead to different behavior, since the attitude of a consumer affects its behavior (Ghen and Liu, 2004), thus purchase intention (Ajzen, 1991). A reason for the differences between private and public consumption could be social motives (Wood, 2000; Berger and Rand, 2008). People like to be perceived positively by others and therefore present themselves in ways that encourage that (Berger, 2014). Therefore, Wakefield and Inman (2003) state that consumers are most sensitive to the opinion of other consumers of the product, when the product is publicly consumed. Griskevicius et al. (2010) explain that status motives can make people choose a different product. Their findings state that consumers desired less luxurious green products when they were shopping in public, but luxurious non-green products when they were shopping in private even though both product categories were equally priced. The acquisition of products to show off your status is called conspicuous consumption. The ability to communicate that status message is only possible if the product is consumed in public (Hoyer et al., 2013). Schlosser (2009) did a research about the effect of computer-mediated versus face-to-face communication on conformity. The findings of the research show that persons in a face-to-face situation adjust their public attitude more than persons who participated in a computer-mediated situation. This would mean that traditional WOM has more influence than eWOM in the case of the social visibility of the product.

14 situation and influences judgments and decisions. The findings show that the interdependent self-concept only has an influence on the judgments of a person when those judgments are likely to be explained to other people, thus when it takes place in a public situation. Next to that, Lamberton et al. (2013) state that consumers use different criteria for evaluating their choices in a public situation than in a private situation. When the choice of a consumer is expected to stay private and does not have to be justified to others, consumers tend to follow their own opinion more. On the contrary, when the choice of a consumer is expected to be public, the influence of external factors becomes more important. Their findings also show that the confidence of a consumer concerning its choice is only affected when it is a choice in a public situation.

So, previous literature explains that the attitude of a consumer may be influenced by the social context and that attitude may influence the purchase intention of a consumer. One can define this as the social visibility, whether a product is publicly or privately consumed, of the product in this research. It is expected that the social visibility has a direct effect on the attitude and the purchase intention of a consumer. Next to that, this direct effect is expected to be positively moderated by the need for conformity. Also, when the opinion of others in an OCR has an influence on the attitude and purchase intention of the consumer, which is the valence in this research, the social visibility expects to affect that influence. All together, this leads to the following hypotheses:

H2a: A public (private) product has a positive (negative) influence on the attitude (purchase intention) of a consumer towards a product

H2b: The positive (negative) effect of a positive (negative) OCR on the attitude (purchase intention) of a consumer towards a product is increased if a product is publicly consumed vs. privately consumed

2.3 Moderator: The need for conformity

15 of the opinions of others were positive, the impressions of the consumers were also more positive. People are more likely to conform to a group when they identify themselves with the group or they value the group (Schlosser, 2009) and when there are less choices available (Curtis and Desforges, 2013). The research findings of Curtis and Desforges (2013) show that the responses of the participants of their research were more conform to those of the confederates when there were three instead of ten choices available. However, after ten available choices, the amount of the conformity did not change anymore.

On the other hand, Yoon et al. (2011) mention that previous research states that persons are also seeking for variety in their choices. This comes from the fact that one desires to be unique. Seeking for variety leads to counter-conformity (Barone and Jewell, 2012). The need for conformity and the need for uniqueness are connected to each other. When the need for conformity in a consumer is large, the need for uniqueness in a consumer is small and vice versa. Therefore, the degree of the need for conformity or the need for uniqueness of someone differs for each person. It depends on personal traits (Barone and Jewell, 2012). Sridhar and Srinivasan (2012) include both findings in their research. On one hand, it is mentioned that consumers experience conformity pressures from a group. This does not even have to be explicitly observed, writing a social norm can be enough. While on the other hand, consumers experience a uniqueness need, which is the opposite of the need for conformity.

The need for conformity is a relevant aspect concerning the attitude and thus the purchase intention of a consumer as can be read in this, and previous paragraphs of this chapter. It is a consumer characteristic that influences several variables, probably also the influence of the valence of an OCR and the social visibility of a product. Therefore, conformity is used as a moderator in this research. It is expected that the need for conformity positively moderates the effect of the valence of an OCR and the effect of the social visibility of a product on the attitude and the purchase intention of a consumer towards a product. Also, the need for conformity is expected to have a direct positive influence on the attitude and the purchase intention of a consumer towards a product. This leads to the following hypotheses:

H3a: The need for conformity positively influences the attitude (purchase intention) of a consumer towards a product

16

H3c: The positive (negative) effect of a public (private) product on the attitude (purchase intention) of a consumer towards a product is increased by the need for conformity

2.4 Dependent variables: the attitude and purchase intention of a consumer

The attitude of a consumer towards a brand is related to the purchase intention of a consumer (Ghen and Liu, 2004). This is explained by the theory of planned behavior. This theory indicates three determinants of purchase intention and one of them is attitude. The more favorable the attitude, the higher the purchase intention (Ajzen, 1991). Thus, besides the fact that the independent variables and the moderator tends to have an influence on the dependent variables, the attitude of a consumer also influences the purchase intention of a consumer. This leads to the following hypothesis:

17

2.5 Overview of hypotheses and conceptual framework

All aforementioned hypotheses are given in an overview below, which can be found in Table 1. Also, three conceptual frameworks are developed to give a visual presentation of this research, which can be seen in Figures 1, 2, and 3. The first framework explains the influence of the independent variables and the moderator on the dependent variable attitude. The second framework explains the influence of the independent variables and the moderator on the dependent variable purchase intention and the third framework states the relationship of the two dependent variables, namely the influence of the attitude on the purchase intention of a consumer.

Overview of the hypotheses

H1a: Negative (positive) valence has a negative (positive) influence on the attitude (purchase intention) of a consumer towards a product

H1b: The negative influence of a negatively valenced OCR on the attitude (purchase intention) of a consumer towards a product is relatively larger than the positive influence of a positively valenced OCR

H2a: A public (private) product has a positive (negative) influence on the attitude (purchase intention) of a consumer towards a product

H2b: The positive (negative) effect of a positive (negative) OCR on the attitude (purchase intention) of a consumer towards a product is increased if a product is publicly consumed vs. privately consumed

H3a: The need for conformity positively influences the attitude ( purchase intention) of a consumer towards a product H3b: The positive (negative) effect of a positive (negative) OCR on the attitude (purchase intention) of a consumer towards a product is increased by the need for conformity

H3c: The positive (negative) effect of a public (private) product on the attitude (purchase intention) of a consumer towards a product is increased by the need for conformity

H4: The positive attitude of a consumer positively influences the purchase intention of a consumer

18

Figure 1. Conceptual framework dependent variable attitude

Figure 2. Conceptual framework dependent variable purchase intention

19

3. Methodology

The previous chapters conceptualized all variables used in this research, concluded by the hypotheses and conceptual frameworks. The research continues by testing these hypotheses. The methodology explains the operationalization of the research starting by elaborating on the research design followed by explaining the respondents that participated in the research, the procedure of the research, and the measurements of the variables. After that, the constructs of the research are validated with a factor analysis and reliability analysis. Also, manipulation checks are performed in order to check whether the manipulations of the research were successful. This chapter concludes by describing a plan of analysis and the check of the VIF scores.

3.1 Research design and procedure

In order to answer the research question ‘How does the valence of an OCR influence the attitude and

the purchase intention of a consumer towards a product that is either social visible or not social visible, and to what extend is this relationship moderated by the need for conformity of a consumer?’,

20

Experimental design

Positive valence Neutral valence Negative valence

Privately consumed (washing machine) Condition 1 Condition 2 Condition 3

Publicly consumed (smartphone) Condition 4 Condition 5 Condition 6

Table 2. Overview conditions

In previous studies concerning consumer decision making, a between-participants design is often used, more often than a within-participants design where all respondents of the research are assigned to all conditions of the research. An argument for this type of research is that it prevents from practice effects, which means that if a respondent is assigned to all conditions, they can get experienced with the research during their participation. Next to that, the participants can become exhausted or bored if they have to respond to all conditions of the research. Also, carry-over effects can occur, which means that the effect of one condition can affect another condition due to comparison (Bordens and Abbott, 2014).

For every condition, the minimum amount of respondents has to be 30 (n≥30). Therefore, the amount of respondents for this research should at least be 180.

21 visibility. Next to that, control questions are included to check whether the participants find the products high involvement products and if they find the products hedonic or utilitarian. The last question is only used in further research if the manipulation of the social visibility was unsuccessful. At the end, general questions about the demographics of a participant are asked, for example the participant’s age, gender, and level of education. These control questions are necessary to give the descriptives of the sample. The link to find the questionnaire will be spread via social media as Facebook and Twitter, and through the e-mail addresses available from the network from the author. Hopefully, a snowball effect will occur in order to reach the amount of respondents that are needed fast.

Examples of the texts of an positive, neutral, and negative OCR of one of the conditions (in this case, an OCR about a private product (washing machine), but the word washing machine can be replaced by smartphone for the OCRs with a public product) are given below:

As can be seen, in case of a positive OCR, the consumer is very pleased with its purchase and a few advantages are summed up. A consumer is content with its purchase in case of a neutral OCR, in which also a few benefits of the product are mentioned. A negative OCR states a few disadvantages of the product and the consumer mentions that he is unsatisfied with its purchase. An overview of all OCRs of the six conditions can be found in Appendix 1 in both English and Dutch.

“My old washing machine broke down, so I bought a new one last week. I am very pleased with the purchase of my new washing machine. The machine is very easy to use, it does not make a lot of noise, and has a lot of functions. Also, it was not very expensive.”

“My old washing machine broke down, so I bought a new one last week. I am content with the purchase of my new washing machine. The machine cleans the clothes and clear instructions are included. Also, the price/quality ratio is ok.”

22

3.2 Participants

In total, 283 participants opened the questionnaire, but 78 persons did not finish it and were therefore excluded from the dataset. This makes a total of 205 participants that have filled out the questionnaire. Therefore, the sample size is sufficient, since it is more than 180. Also, there are 30 (or more) participants assigned to every condition. An overview of the exact amounts can be seen in Table 3. Slightly more women, namely 50.7%, (n=104) than men, 48.3% (n=99) participated in this research. Two participants did not indicate their gender, which explains the remaining 1%. The average age of the participants is 25.88 years. The most mentioned level of education is HBO (45.4%), followed by the University (22.9%), MBO (22%), and high school (9.3%). One participant did not indicate its highest level of education. The last part of the descriptive statistics contains the yearly gross income of the participants. The majority of the participants has a below modal income (46.8%), 10,2% of the participants has a modal income, and 20% of the participants has an above modal income. However, 23% of the participants did not indicate their income. Also, the participants had to indicate whether they find it useful to read OCRs in general on a seven point Likert scale. The mean of all participants is 5.18 with a standard deviation of 1.618. An overview of the descriptives of the participants per condition can be found in Table 3.

An overview of the descriptives of the participants per condition:

The descriptives of the participants per condition

Positive valence Neutral valence Negative valence

Privately consumed

n = 31

age = 33.52 (sd 16.318)

gender = 29% women, 67.7% men education = 9.7% high school, 12.9% MBO, 51.6% HBO, 25.8% University income = 41.9% beneath modal, 3.2% modal, 25.8% above modal, 22.6% don’t want to tell

use OCRs = 5.32 (sd 1.600)

conformity factor 1 = 4.492 (sd 1.433) conformity factor 2 = 4.121 (sd 1.178)

n = 37

age = 29.28 (sd 12.928)

gender = 59.5% women, 40.5% men education = 5.4% high school, 29.7% MBO, 43.2% HBO, 18.9% University income = 54.1% beneath modal, 10.8% modal, 18.9% above modal, 16.2% don’t want to tell use OCRs = 5.43 (sd 1.324) conformity factor 1 = 4.311 (sd 1.267) conformity factor 2 = 4.060 (sd 1.242) n =41 age = 37.70 (sd 17.037)

gender = 48.8% women, 51.2% men education = 12.2% high school, 19.5% MBO, 41.5% HBO, 26.8% University income = 46.3% beneath modal, 12.2% modal, 19.5% above modal, 22% don’t want to tell use OCRs = 5.46 (sd 1.690) conformity factor 1 = 4.500 (sd 1.396) conformity factor 2 = 4.439 (sd 1.223) Publicly consumed n = 30 age = 29.62 (sd 12.520)

gender = 53.3% women, 46.7% men education = 3.3% high school, 30%

n = 34

age = 34.29 (sd 15.190)

gender = 61.8% women, 35.3% men education = 14.7% high school, 23.5%

n = 32

age = 32.18 (sd 13.372)

23

MBO, 30% HBO, 36.7% University income = 43.3% beneath modal, 10% modal, 26.7% above modal, 16.7% don’t want to tell

use OCRs = 4.93 (sd 1.680)

conformity factor 1 = 4.375 (sd 1.298) conformity factor 2 = 4.167 (sd 1.053)

MBO, 50% HBO, 11.8% University income = 44.1% beneath modal, 14.7% modal, 14.7% above modal, 26.5% don’t want to tell

use OCRs = 5.38 (sd 1.596)

conformity factor 1 = 4.493 (sd 1.142) conformity factor 2 = 4.287 (sd 1.192)

MBO, 56.3% HBO, 18.8% University income = 50% beneath modal, 9.4% modal, 15.6% above modal, 25% don’t want to tell

use OCRs = 4.41 (sd 1.663)

conformity factor 1 = 4.492 (sd 1.222) conformity factor 2 = 4.648 (sd 1.475)

Table 3. Overview descriptives of the participants per condition

To make sure that the participants are randomly assigned between the six conditions of the questionnaire, a Pearson Chi-Square test is conducted including the variables age, gender, education, and income. The results of the test, which can be found in Table 4, show that all significance levels are higher than 0.05, which means that there is no statistically significant relationship. Based on these results it can be concluded that the randomization was done successfully.

Results Pearson Chi-Square test

Variables Chi-Square value df Sig.

Age 191.143 210 .820

Gender 8.619 5 .125

Education 15.160 15 .440

Income 6.481 15 .970

Table 4. Results Pearson Chi-Square test

3.3 Operationalization

The variables are measured with the use of the questionnaire. The online review that the respondents will see is about either a private product, or a public product. Next to that, the respondents will get a review with a positive, neutral, or negative valence. The respondents will be randomly assigned to one of the six conditions. An overview of the operationalization can be found in Table 5.

24 Next to that, the dependent variable, purchase intention of the consumer is measured. The purchase intention is measured by the four items in the scale of Li et al. (2002) measured on a seven point Likert scale. These dimensions are unlikely versus likely, improbable versus probable, uncertain versus certain, and not definitely and definitely. The scale showed a Cronbach’s alpha of 0.90.

The need for conformity is measured using a measurement scale of Mehrabian and Stefl (1995). The scale consists of eleven items and are measured on a nine point Likert scale. Examples of items are ‘I don’t give in to others easily’, ‘I tend to follow family tradition in making political decisions’, and ‘ I tend to rely on others when I have to make an important decision quickly’. The scaled showed a Cronbach’s alpha of 0.77.

The manipulation check is measured by asking the participants their opinion about the valence and the social visibility. The control variables of the questionnaire are whether the product is utilitarian or hedonic, the product involvement, and the use of OCRs in general by the consumer. All these variables are measured on a seven point Likert scale.

Lastly, the background variables are age, gender, the level of education, and the yearly income. The questions are measured on different scales. Age is measured on a continuous scale, gender is measured on a dichotomous scale, and both the level of education and the yearly income are measured on a categorical scale.

3.4 Construct validation

3.4.1 Factor analysis

A factor analysis, a principal component analysis with Varimax rotation, is performed in order to evaluate whether the attitude of the consumer, the purchase intention of the consumer, and the need for conformity are indeed unidimensional. In order to make sure that a factor analysis is appropriate, the Kaiser-Meyer-Olkin Measure of Sampling Adequacy should be over 0.5. Next to that, Bartlett’s test of sphericity should be significant, the communalities should be over 0.4 and the factor loadings should be over 0.5 (Malhotra, 2010).1 All outcomes can be found in Table 5.

1

25 For the attitude of the consumer, all items belong to one factor (Eigenvalue=4.323). The KMO measure is 0.905, the Bartlett’s test of sphericity is significant (0.000), and all communalities are over 0.4. Also, all factor loadings are over 0.5.

For the purchase intention of the consumer, all items also belong to one factor (Eigenvalue=3.345). The KMO measure is 0.812, Bartlett’s test of sphericity is significant (0.000), and all communalities are over 0.4. Also, all factor loadings are over 0.5.

For the need for conformity, three factors were found to which the items belong. The KMO measure is 0.702, Bartlett’s test of sphericity is significant (0.000). The communalities of items 1,2 and 5 are the only ones that are under 0.4. All rotated factor loadings are over 0.5. According to the factor analysis, items 5, 6, 8 and 10 belong to the first factor (Eigenvalue=2.850). The second factor consists of items 2, 7, 9 and 11 (Eigenvalue=1.759). Items 1, 3 and 4 belong to the third factor (Eigenvalue=1.042). The reliability analysis will investigate whether items should be from further analysis and which factors are going to be used for further analysis.

3.4.2 Reliability analysis

A reliability analysis is conducted in order to test the reliability of the constructs that measure the variables the attitude of the consumer, the purchase intention of the consumer, and the need for conformity. A reliability analysis measures the internal consistency of the marker items, the Cronbach’s alpha, of the constructs. The Cronbach’s alpha has to be higher than 0.6, because that is the widely accepted threshold in academic research (Malhotra, 2010). After measuring the Cronbach’s alpha, the ‘Cronbach’s alpha if item deleted’ has to be checked in order to verify whether the alpha would be higher if an item is deleted. An overview of all alphas can be found in Table 5.

26 When taking both the factor analysis and the reliability analysis into consideration, all items are used for the concepts attitude and purchase intention. They are valid constructs. For the need for conformity, only the first two factors are used. This means that items 1, 3 and 4 are deleted.

An overview of the operationalization and construct validation:

Overview operationalization construct validation

Concept Items Commu

nalities Factor loadings KMO measure Bartlett’s test of sphericity Source α this study Attitude 1: Unappealing/appealing 2: Bad/good 3: Unpleasant/pleasant 4: Unfavorable/favorable 5: Unlikable/likable .867 .829 .915 .876 .835 .931 .911 .956 .936 .914 .905 .000 Spears and Singh (2004) Composite reliability = .94 AVE = .77 .961 Purchase intention 1: Unlikely/likely 2: Improbable/probable 3: Uncertain/certain 4: Not definitely/definitely .860 .866 .872 .746 .928 .931 .934 .864 .812 .000 Li et al. (2002) α = .90 .934 Need for conformity Factor 1 Need for conformity Factor 2

5: Basically, my friends are the ones who decide what we do together 6: A charismatic and eloquent speaker can easily influence and change my ideas

8: If someone is very persuasive, I tend to change my opinion and go along with them

10: I tend to rely on others when I have to make an important decision quickly

2: I would be the last one to change my opinion in a heated argument on a controversial topic

7: I am more independent than

27 Need for conformity Factor 3 (not used in this study) conforming in my ways

9: I don’t give in to others easily 11: I prefer to make my own way in life rather than find a group I can follow

1: I often rely on, and act upon, the opinion of others

3: Generally, I would rather give in and go along for the sake of peace than struggle to have my way

4: I tend to follow family tradition in making political decisions

.545 .589 .382 .634 .585 .628 .729 .522 .764 .752 .579

Table 5. Overview operationalization and construct validation

3.5 Manipulation checks

Manipulation checks are performed in order to check whether the respondents did perceive the reviews as positive, neutral or negative and whether they perceived the products as a private or public product. Also, a control variable is checked, namely whether the products are indeed perceived as high involvement products. This is all done with the use a One-Way ANOVA test. The means of the conditions are tested on significant differences (p<0.05).

For the valence of the OCR, the respondents had to rate the question ‘I find this online consumer

review..’ on a seven point Likert scale with 1 as negative and 7 as positive, which leads to the means

28 excluded from further analysis. Continuing with the manipulation check, the neutral conditions differ significantly from the negative conditions (p<0.05). Also, the means of the positive and the negative conditions are compared and they are significant (p<0.05). This means that this manipulation was successful.

Manipulation check of the valence of the conditions

Mean Std. Deviation Differs from

Positive conditions 4.92 1.345 Negative

Neutral conditions 4.70 1.356 Negative

Negative conditions 2.34 1.145 Positive and neutral

Overall F

Overall Significance

87.080 .000

Table 6. ANOVA test valence

For the social visibility of the product, the respondents had to rate the question ‘I use this type of

product mostly in a..’ on a seven point Likert scale with 1 as private situation and 7 as public

situation. An independent t-test is performed that compares ratings on the products of the conditions about the smartphones with the ratings on the products of the conditions about the washing machines. The mean of the smartphones (3.59) is higher than the mean of the washing machines (1.83), see Table 7. Also, the significance is lower than 0.05. This means that this manipulation was successful.

Manipulation check of the social visibility of the products

Mean Std. Deviation Differs from

Smartphones 3.59 1.546 Washing machines

Washing machines 1.83 1.211 Smartphones

Overall t

Overall Significance

5.101 .000

Table 7. Independent t-test social visibility

Next to the manipulations, the control variable involvement is checked, see Table 8. For the involvement of the product, the respondents had to rate the question ‘When I would purchase this

29 strongly disagree and 7 as strongly agree. In this research, a mean score higher than 5 is perceived as high involved. A one-sample t-test is performed with a test value of 5 that compares the involvement for the conditions with the smartphones and the conditions with the washing machines. The mean of the smartphones (6.05) is a little bit lower than the mean of the washing machines (6.28). Also, the significance of both tests is lower than 0.05 (p<0.05). This means that both products are perceived as high involvement products.

Manipulation check of the control variable involvement

Mean Std. Deviation t Sig.

Involvement smartphones 6.05 1.210 9.138 .000

Involvement washing machines 6.28 1.252 9.998 .000

Table 8. Independent t-test control variable involvement

3.6 Plan of analysis

Regression analyses are performed to measure the effects of the independent variables and the moderators on the dependent variables, starting with multiple linear regressions. The first model is the base model, which contains the main effects of the independent variables on the dependent variables and the interaction effect of the independent variables on the dependent variables. In the second model, the main effects of the first moderator is added, which is factor 1 of the need for conformity. Also, the interaction effects of the first moderator with the independent variables on the dependent variables is added. The third model is the primary model extended by the main effects and the interaction effects of the second moderator, which is factor 2 of the need for conformity. The final model contains all main effects and interactions effects of previous models. After the multiple linear regressions, a bivariate regression is performed to measure the effect of the attitude on the purchase intention of a consumer. The R² adjusted states the total variance explained by the models and the significance has to be under 0.05. The betas explain the direction of the effect and also has to be significant.

30

VIF scores

Model Variables Old VIF scores

1 Valence 2.146 SocialVisibility 1.854 ValenceXSocialVisibility 2.889 2 Valence 13.387 SocialVisibility 13.811 ValenceXSocialVisibility 2.894 ConformFactor1 3.535 ValenceXConformFactor1 13.134 SocialVisibilityXConformFactor1 14.444 3 Valence 15.624 SocialVisibility 15.621 ValenceXSocialVisibility 2.971 ConformFactor2 2.542 ValenceXConformFactor2 13.908 SocialVisibilityXConformFactor2 14.437 4 Valence 26.779 SocialVisibility 29.141 ValenceXSocialVisibility 2.982 ConformFactor1 3.744 ValenceXConformFactor1 13.922 SocialVisibilityXConformFactor1 15.656 ConformFactor2 2.676 ValenceXConformFactor2 14.547 SocialVisibilityXConformFactor2 15.987

Table 9. Overview of the VIF scores

31 the two. Since the VIF scores of the base model are not too high, a regression will be performed to understand the independent-dependent variables relationship. However, the effect of the interaction terms with the moderators Conformity Factor 1 and Conformity Factor 2 had to be measured in a different way. The problem in this study is that the independent variables are dummy variables and the moderators are categorical variables. A simple and elegant solution is to divide the moderators into two categories as well based on a median-split. Values that are higher than the median are coded 1 and values that are under the median are coded 0. In this way, all variables are categorical. After that, an ANOVA with two-way interactions is performed on the interaction terms with the moderators to see the results. The downside of this method is that the variability in the moderators is lost. However, the findings can at least be interpreted as there is no multicollinearity.2

32

4. Results

The 205 questionnaires that are filled out by the respondents form the collected data of this research. This data is analyzed with the use of the Statistical Package for Social Science (SPSS) 22. The results of this data analysis are discussed in this chapter. First, the descriptives of all conditions are stated for first insights. Secondly, the results of the regression analyses and the ANOVAs are given followed by an overview of the acceptation or rejection of the hypotheses.

4.1 Descriptives of all conditions

The mean of the attitude and the purchase intention of the consumer are both measured on a seven point Likert scale. Two ANOVAs are performed to calculate the means of the dependent variables per condition. Both analyses are significant with F=30.817 and Sig.=.000 (attitude) and F=15.234 and Sig.=.000 (purchase intention). An overview of the outcomes of this test can be found in Table 10.

Overview of the means of the concepts per condition

Positive valence Neutral valence Negative valence

Privately consumed (n=31) (n=37) (n=41) Attitude 4.723 (sd 1.293) 4.768 (sd 1.267) 2.356 (sd 1.005) Purchase intention 3.919 (sd 1.096) 4.108 (sd 1.262) 2.244 (sd 1.126) Publicly consumed (n=30) (n=34) (n=32) Attitude 4.633 (sd 1.594) 4.100 (sd .992) 2.650 (sd .806) Purchase intention 3.408 (sd 1.401) 3.150 (sd 1.358) 2.313 (sd .925)

Table 10. Overview of means of the concepts per condition

When taking the manipulation check of the valence of the OCRs and the means of the concepts into consideration, it can be concluded that the participants perceived the positive and the neutral conditions almost similar. The manipulation check was not significant and also the behavior concerning the attitude and purchase intention of the consumer does not differ much. Therefore, the neutral conditions are not used in further analyses.

4.2 Regression analyses

33 11. This is the only model for which the VIF scores are not too high. Because of the high VIF scores in the other models, this research continues with two-way ANOVAs to interpret the results as there is no multicollinearity. The moderators are divided into two categories as well, based on a median-split. Four ANOVAs are performed following the same models as would have been used for regression analyses. The first model is the base model, which contains the main effects of the independent variables on the dependent variables and the interaction effect of the independent variables on the dependent variables. In the second model, the main effects of the first moderator is added. The third model is the primary model extended by the second moderator, which is factor 2 of the need for conformity. The final model contains all main effects and interactions effects of previous models. All outcomes of the ANOVAs can be found in Tables 12 and 13. To measure the influence of the attitude on the purchase intention of a consumer a regression analysis is used again of which the outcomes can be found in Table 14. Lastly, Table 15 gives an overview of the acceptation and rejection of all hypotheses.

4.2.1 The main effects of the independent variables

The regression analyses measured the main effects of the independent variables on the dependent variables. Dummy variables are used with 1 as positive and 1 as washing machines. Both analyses are significant with R² adjusted=.455 and Sig.=.000 (attitude) and R² adjusted=.271 and Sig.=.000 (purchase intention). As can be seen in Table 11, a positive OCR has a significant and positive effect on the attitude and the purchase intention of a consumer. Thus a negative OCR has a significant and negative effect on the attitude and purchase intention of a consumer. The parameters of the social visibility are both not significant, which means that the social visibility has no main effects on the dependent variables. As can be seen in Tables 12 and 13, the outcomes of the ANOVAs show the same results.

4.2.2 The interaction effect of the social visibility of a product

34

Results regression analyses base model

Attitude Purchase intention

Variables Base model Base model

(Constant) 2.650(.211)*** 2.313(.202)***

Valence (0=negative, 1=positive) 1.983(.303)*** 1.096(.291)***

SocialVisibility (0=smartphone, 1=washing machine) -.294(.281)n.s. -.069(.270)n.s.

ValenceXSocialVisibility .383(.415)n.s. .580(.398)n.s.

R² adjusted .455 .271

F (Sig.) 37.988*** 17.481***

n.s.=not significant, *p < .10, **p<.05, ***p<.01 Table 11. Regression analyses base model

4.2.3 The main effects of the moderators

The ANOVAs measured the main effects of the need for conformity factor 1 and factor 2 on the dependent variables. Model 2 measured the effects of conformity factor 1 and Model 3 measured the effects of conformity factor 2. The R² adjusted of Model 2 is .443 (attitude) and .266 (purchase intention) and the R² adjusted of Model 3 is .446 (attitude) and .271 (purchase intention). As can be seen in Tables 12 and 13, for both the attitude and the purchase intention of a consumer, the main effects of both moderators are not significant, which means that the need for conformity has no main effects on the dependent variables.

4.2.4 The interaction effect of the need for conformity

35

Results ANOVAs: DV Attitude

Model 1 Model 2 Model 3 Model 4

Variables F Sig. F Sig. F Sig. F Sig.

Valence 110.064 .000 106.974 .000 108.505 .000 105.436 .000 SocialVisibility .244 .622 .294 .588 .230 .633 .285 .595 ValenceXSocialVisibility .854 .357 .811 .370 .771 .382 .722 .397 ConformFactor1 .006 .940 .009 .924 ValenceXConformFactor1 .003 .954 .015 .903 SocialVisibilityXConformFactor1 .257 .613 .259 .611 ConformFactor2 .047 .828 .088 .767 ValenceXConformFactor2 .462 .498 .439 .509 SocialVisibilityXConformFactor2 .515 .474 .478 .491 R² adjusted .455 .443 .446 .434

Table 12. Results ANOVAs: DV Attitude

Results ANOVAs: DV Purchase intention

Model 1 Model 2 Model 3 Model 4

Variables F Sig. F Sig. F Sig. F Sig.

Valence 48.396 .000 46.416 .000 49.152 .000 47.013 .000 SocialVisibility 1.233 .269 1.322 .252 1.232 .269 1.288 .259 ValenceXSocialVisibility 2.117 .148 2.364 .127 1.863 .175 2.073 .152 ConformFactor1 1.759 .187 1.315 .254 ValenceXConformFactor1 .413 .522 .297 .587 SocialVisibilityXConformFactor1 .041 .841 .057 .812 ConformFactor2 .513 .475 .524 .470 ValenceXConformFactor2 .038 .846 .078 .781 SocialVisibilityXConformFactor2 2.656 .106 2.030 .157 R² adjusted .271 .266 .271 .263

36 4.2.5 The effect of the attitude on the purchase intention of the consumer

In previous paragraphs the effects of the independent variables and the moderators on the dependent variables are measured. Next to that, the attitude of a consumer might have an influence on the purchase intention of a consumer. A bivariate regression is performed in order to check that. The analysis is significant with R²=.468 and Sig.=.000. The attitude of a consumer towards a product has a significant (.000) positive (.571) effect on the purchase intention of a consumer.

Results regression analysis dependent variables: DV Purchase intention

Variable Model 1 (Constant) .921 (.202)*** Attitude .571 (.053)*** R² adjusted F (Sig.) .468 117.777*** n.s.=not significant, *p < .10, **p<.05, ***p<.01

37 4.2.6 Hypothesis testing

All aforementioned results lead to the acceptation or the rejection of the hypotheses of this research. Table 15 gives an overview of the acceptation and rejection of the hypotheses.

Hypotheses Accepted/Rejected

H1a: Negative (positive) valence has a negative (positive) influence on the attitude (purchase intention) of a consumer towards a product

Accepted

H1b: The negative influence of a negatively valenced OCR on the attitude (purchase intention) of a consumer towards a product is relatively larger than the positive influence of a positively valenced OCR

Not measured

H2a: A public (private) product has a positive (negative) influence on the attitude (purchase intention) of a consumer towards a product

H2b: The positive (negative) effect of a positive (negative) OCR on the attitude (purchase intention) of a consumer towards a product is increased if a product is publicly consumed vs. privately consumed

Rejected

Rejected

H3a: The need for conformity positively influences the attitude ( purchase intention) of a consumer towards a product

Rejected

H3b: The positive (negative) effect of a positive (negative) OCR on the attitude (purchase intention) of a consumer towards a product is increased by the need for conformity

H3c: The positive (negative) effect of a public (private) product on the attitude (purchase intention) of a consumer towards a product is increased by the need for conformity

Rejected

Rejected

H4: The positive attitude of a consumer positively influences the purchase intention of a consumer Accepted

38

5. Discussion, conclusion, implications and limitations

In this chapter, the research question will be answered based on the findings of this research. First, the findings are discussed. After that, the managerial implications are mentioned followed by the limitations of this study and possible directions for further research.

5.1 Discussion

The aim of this research was to investigate the effect of the valence of an OCR and the social visibility of a product on the attitude and the purchase intention of a consumer moderated by the need for conformity. This leads to the following research question, which will be discussed based on the findings of this research:

‘How does the valence of an OCR influence the attitude and the purchase intention of a consumer

towards a product that is either social visible or not social visible, and to what extend is this relationship moderated by the need for conformity of a consumer?’

5.1.1 The valence of an OCR

Consistent with current literature, the valence of an OCR has a significant influence on both the attitude and the purchase intention of a consumer. As expected, positive OCRs have a positive effect and negative OCRs have a negative effect. This is in line with the findings of the study of Purnawirawan et al. (2013) who state that people tend to follow the majority of the opinions. Whether the effect of the negative OCRs is larger than the effect of the positive OCRs is not measured, since the neutral conditions are excluded from further research and they would serve as a benchmark. It was expected that the influence of negative OCRs would be larger than the influence of positive OCRs following the studies of Price (1996), Gershoff et al. (2006), and Park and Lee (2009). Ahluwalia et al. (2000) support this by stating that negative information is more attractive than positive information. But, as mentioned earlier, this hypothesis could not be tested in this study anymore.

5.1.2 The social visibility of a product

39 would prefer themselves as a result of status motives. Also, Torelli (2006), Schlosser (2009), and Lamberton et al. (2013) states that people adjust their behavior in public situations.

5.1.3 The need for conformity

The need for conformity was expected to have both main effects and interaction effects. First, the need for conformity was expected to have a positive influence on the attitude and the purchase intention of a consumer. However, this hypothesis was rejected. Next to that, the need for conformity was expected to increase the effect of the valence of an OCR on the attitude and the purchase intention of a consumer. However, this hypothesis is rejected as well. These findings are contradictory with findings from other studies. For example, Asch (1951) mentions that people tend to conform when they want to fit in or when they assume that others have better information. Conformity is a reason that could encourage change in the attitude of someone (Wood, 2000). However, the results of this research does not confirm that.

5.1.4 The attitude and purchase intention

It was expected that a positive attitude has a positive influence on the purchase intention of a consumer. This expectation is supported by the findings of this research. This is in line with the theory of Ajzen (1991) who explained this behavior in his theory of planned behavior. Ghen and Liu (2004) built upon this theory and confirm that the attitude of a consumer is related to the purchase intention of a consumer.

5.2 Managerial implications

This research could be interesting for companies who want to know more about the impact of the valence of OCRs on their websites on the attitude and the purchase intention of their (potential) consumers towards their products. As the results of this research show, the valence of an OCR influences the attitude and the purchase intention, thus the buying behavior, of a consumer. Companies can respond to this knowledge by providing the opportunity on their website to write reviews about the products or services to benefit from the influence of positive reviews. However, the possibility that consumers write negative reviews is existing in that case as well. This might be a reason to not provide the opportunity to write reviews. However, it does not have to be a problem if companies finds a good way to deal with negative complaints.

5.3 Limitations and future research directions

40 First, only the text of a review was shown in an offline situation. The participants could for example not see a picture of the product, the features of the product, or the brand. In a normal situation, consumers would be able to browse through the website on which they are reading the review on. They would probably read several reviews, look at the features, and/or the brands that are available. This was not possible in this study. Therefore, the results of this research might not be comparable with a real life situation.

Next to that, the manipulation check of the valence of the OCRs showed that the positive and neutral OCRs were perceived almost similar, which resulted in the fact that the neutral conditions were excluded from the regression analyses. This might be due to the fact that the respondents only read one of the reviews. If they would have read all the reviews, the difference would probably be clear. Therefore, future research could perform the same experiment in which the respondents are assigned to all conditions instead of assigning the respondents to only one condition.

Concerning the social visibility of a product, current research is primarily about the difference in behavior of people in public or private situations. In this case however, the difference lies in public and private products. The respondents answered the questionnaire in their own environment without people around who would see their choices. This might be the reason why the effects of the social visibility of the product are not significant. Further research could perform a similar experiment that is done in a real life situation. Also, a smartphone is used in this research as a public product. As the results of the manipulation check showed, a smartphone was considered more as a public product than a washing machine. However, the means were not really high. It might be that people are so familiar with this product nowadays that they do not really consider this as a public product anymore. Further research could perform this experiment with a different public product.

The need for conformity also showed no significant effects in this research. The participants had to rate several statements concerning the need for conformity. However, there is a difference between how a person actually is and how a person desires to be. It might be possible that the participants rated the statements on how they desire to be instead of how they actually are.

41

6. References

Ahluwalia, R., Burnkrant, R.E., Unnava, H. (2000). “Consumer response to negative publicity: The moderating role of commitment”, Journal of Marketing Research, 27(2), 203-214

Ajzen, I. (1991). “The theory of planned behavior”, Organizational Behavior and Human decision

processes, 50, 179-211

Bansal, H.S., Voyer, P.A. (2000). “Word-of-mouth processes within a services purchase decision context”, Journal of service research, 3(2), 166-177

Barone, M.J., Jewell, R.D. (2012). “The category advertising norms and consumer counter-conformity

influence comparative advertising effectiveness”, Journal of Consumer Psychology, 22(4), 496-506

Berger, J. (2014). “Word of mouth and interpersonal communication: A review and directions for future research”, Journal of Consumer Psychology, 24(4), 586-607

Berger, J, & Rand L. (2008), “Shifting signals to help health: Using identity signaling to reduce risky health behaviors”, Journal of Consumer Research, 35, 509-518

Bordens, K.S., Abbott, B.B. (2014). “Research design and methods: A process approach”, New York,

NY : McGraw-Hill Education

Cialdini, R.B., Goldstein, N.J. (2004). “Social influence: Compliance and conformity”, Annual Review of

Psychology, 55, 591-621

Childers, T. L., Rao, A. R. (1992). “The influence of familial and peer-based reference groups on consumer decisions”, Journal of Consumer Research, 19, 198–211

Chevalier, J., Mayzlin, D. (2006). “The effect of word-of-mouth on sales: Online book review”, Journal

of Marketing Research, 43, 345-354

Curtis, D.A., Desforger, D.M. (2013). “Less is more: The level of choice affects conformity”, North