1
CAN PERSONALISED ADVERTISING BECOME TOO PERSONAL?:
THE MODERATING ROLE OF PERCEIVED INTRUSIVENESS OF
FACEBOOK ADVERTISEMENTS ON THE RELATIONSHIP
BETWEEN PERSONALISATION OF ADVERTISEMENTS AND
ATTITUDES TOWARDS THE BRAND.
by
DOUWE-ALBERT BEERDA
University of Groningen
Faculty of Economics and Business
MSc Marketing Management
Januari 2020
Mr. Andreaestraat 22
9291 MA Kollum
(0511) 453366
(06) 30318380
d.a.g.beerda@student.rug.nl
Student number: 2974975
2
Abstract
Attitude towards the brand is a rather important variable in the marketing literature. Brands use advertising to strengthen a consumer‟s positive attitude towards the brand. Organisations are getting more knowledge in the form of data regarding their potential customers in a social media setting. With this data, organisations are able to create personalised advertisement, based on the consumers unique preferences and characteristics. This subsequently leads to more relevant advertisement and a more positive attitude towards a brand. However, there are also potential negative effects of personalisation in the form advertisement intrusiveness. This master‟s thesis presents the results of an online survey (N = 209), in which the role of
perceived advertisement personalisation, perceived advertisement relevance and perceived advertisement intrusiveness were studied on the attitude towards the advertised brand. The independent variable consisted of actual personalisation by exposing respondents to an experimental manipulation, which divided the respondents into a low (n = 106) and high (n = 103) personalised condition. Also, the respondents‟ perceived personalisation was measured from these two different advertisement conditions. Firstly, results show that the stronger one‟s perception of the personalisation of an advertisement, the more relevant an advertisement is perceived by a consumer. Also, the results indicated that the more relevant an ad is perceived, the more positive one‟s attitude towards the advertised brand is. A mediation analysis was executed and a full mediation effect of perceived advertisement relevance on the relationship between perceived advertisement personalisation and attitude towards the brand was found. Lastly, perceived advertisement intrusiveness appeared to be a mediator. The more intrusive an advertisement is perceived, the less positive the relationship between perceived
advertisement relevance and the attitude towards the advertised brand is. Keywords:
4
1. Introduction
Over the recent years the usage of social media platforms by consumers has grown
tremendously. Facebook for example, currently the biggest social media platform, had 517.75 million users in 2010. In 2018 this number has grown to 2.38 billion (Ortiz-Ospina, 2019). The growth of social media users subsequently led to social media becoming an interesting platform for advertisers to reach their audience. Subsequently, generally social media spending as percentage of the total marketing budget has grown from being only 3.6% in 2009 to 11.4% in 2019 (CMO survey, 2019). In addition, social media spending is expected to rise by 73% in the upcoming 5 years (CMO survey, 2019).
Social media platforms offer advertisers the opportunity to target their audience specifically, which cannot be done to such extent by traditional channels for advertising, like print ads or television commercials. This is possible since a social media platform like
Facebook stores and uses enormous amounts of personal data of their users, which advertisers consequently use for their targeting efforts (Kelly, Kerr & Drennan, 2010). This precise way of targeting often leads to extreme personalised advertisements which are based on social media users‟ previous website browsing, demographics, personal preferences and past purchasing information (Li, 2016).
Consumers may typically respond in two different ways when they are exposed to a highly personalised ad. Firstly, targeted people can perceive these ads as appealing since they are relevant to them and fit into their interests (Anand & Shachar, 2009; Lambrecht & Tucker, 2013). Thus, the increased perceived advertisement personalisation may lead to a more
positive response towards the brand through one‟s perceived relevance. Positive attitudes towards the brand typically result in stronger purchase intentions among receivers of these personalised advertisements (Lee & Cranage, 2011), and, (indirectly) to more purchase behaviours (Ajzen, 1982).
Secondly , personalised advertisements can also be perceived as intrusive (Stone, 2010; Tucker, 2011). The consumers that perceive personal advertisements as intrusive, will resist the appeal of the ad because of a mechanism called „reactance‟ (White, Zahay,
5 to show negative responses towards the brand (White et al., 2008), which may result in a negative attitude towards the brand.
The potential negative and the positive reactions that personalised advertisements can evoke are dependent on the way receivers perceive the advertisement, which consequently leads to the personalisation paradox. In the current literature little research has been done in regard to this personalisation paradox. There is extensive literature on how personalised advertisements can lead to positive responses by consumers and preferable outcomes for companies in online contexts such as mobile advertising (Lee, Lee & Yang, 2017), direct e-mailing (Baek & Morimoto, 2012) and advertisement personalisation on websites (Ho & Bodoff, 2014). Furthermore, research has been done in regard to the negative effects of advertisement personalisation (Tucker, 2011; White et al., 2008). Moreover, some studies have researched the personalisation paradox, like van Doorn and Hoekstra (2013) and Lee and Cranage (2011). However, these papers researched this paradox in a website context. Little research has been done in regard to the role of perceived intrusiveness and the potential negative effect it could have on the relationship between personalised advertisements and the attitude that consumers have towards a brand in a social media context. Yet, especially in a social media context, this could be very different from the already researched website context. Since personalised advertisements may be perceived as more intrusive in a social media context. Because the environment in which advertisements are shown in a social media environment is considered to be a personal space by the receivers (Kelly et al., 2010). This personal space is designed and controlled by the receiver (Kelly et al., 2010). Therefore, receivers may perceive these personalised advertisements as more intrusive and subsequently these advertisement will seem more irrelevant in a social media context compared to a
website context (De Keyzer, Dens & De Pelsmacker, 2015).
Since there is a lack of research in regard to this paradox, especially in combination with a social media setting, this study aims to examine the processes underlying this
personalisation paradox in a Facebook setting. This leads to the following research question: To what extent does perceived advertisement intrusiveness affect the relationship between advertisement personalisation and attitude towards the brand in a social media context?
6 targeting audiences, Facebook will be used as the social media research context in this study. The research proceeds as follows. First, the literature regarding the personalisation paradox will be elaborated and the hypotheses will be presented. Secondly, the methodology will be described. Thirdly, the gathered results will be presented. And finally, the results, limitations and future research will be discussed in the discussion and conclusion.
2. Literature review
2.1 Attitude towards the brand
The effect of advertisement personalisation has been researched on different outcome variables. For example, on advertisement avoidance, advertisement attention and purchase intentions (Malheiros, Jennett, Patel, Brostoff & Sasse, 2012 ;van Doorn and Hoekstra, 2013). However, little research has been done to the effect of advertisement personalisation on attitude towards the brand. Yet, this could be useful since consumers commonly come into contact with brands and form attitudes towards them through the brand its advertisements (Lee, Lee, Yang, 2017). Therefore, it is important research what kind effect personalised advertisements have on the attitudes towards a brand. This study therefore focuses on attitude towards brand as dependent variable in relation advertisement personalisation.
Attitude towards the brand has thoroughly been researched and is deemed as a rather important variable in the marketing literature. For example, attitude towards the brand has a positive effect on the perceived quality of services and products (Dodds, Monroe & Grewal, 1991). Moreover, attitude towards the brand impacts purchase intentions positively (Dodds et al., 1991).
7 Apart from these characteristics of an attitude , Eagly and Chaiken (1973) define an attitude as not only an internal state, as Mitchell and Olsen (1981) state, but also as an
enduring state. In this definition of an attitude, Eagly and Chaiken (1973) state that an attitude endures for at least a short period of time and that such an attitude can direct and energize behaviour. Finally, Spears and Singh (2004) add that an attitude toward the brand is an unidimensional summary evaluation of the brand. This is in line with papers from Zanna and Rempel (2008) and Machleit, Allen and Madden (1993) which are treating attitude as a summary evaluation in order to distinguish it from the evaluation which also includes beliefs, feelings, behaviours and expressions of attitudes (Giner-Sorolla 1999, Spears & Singh, 2004). They do not include these factors in their definition of attitude towards the brand since for example feelings cannot be compared to attitudes towards a brand. This is because feelings toward a brand are transitory, contrary to attitudes that are more enduring. Therefore in the context of this study we use the definition of Spears and Singh (2004): Attitude towards the brand is a relatively enduring, unidimensional summary evaluation of the brand that most likely is able to energize behaviour. This definition will be used since it only has an unidimensional approach towards the concept. By using this unidimensional definition, a more precise relationship on only attitudes can be researched in this study.
2.2 The effect of perceived advertisement personalisation and perceived advertisement relevance on attitude towards the brand
Brands use advertising to strengthen a consumer‟s positive attitude towards the brand. An effective way to do this is in a social media setting, is through the use of advertisement personalisation (Montgomery & Smith, 2009). Advertisement personalisation can be defined as delivering advertisements that contain individualized information, based on the unique preferences of the receiver (Arora, et al., 2008; Li, 2016; Shanahan, Tran & Taylor, 2019). Advertisement personalisation has become an interesting topic, due to organisations getting increasingly more knowledge in the form of data regarding their potential customers in a social media setting (Li, 2016). With this data, organisations are able to create personalised advertisement based on the receivers unique preferences and characteristics.
8 be described as perceived advertisement relevance. Perceived advertisement relevance in this study is defined as that what is being displayed in the advertisement is related to the recipient personal needs and values (Jung, 2017; Celsi & Olsen, 1988). Whether the recipient actually is perceiving an advertisement as relevant is not always obvious (Li, 2016). This is due to a difference in actual personalisation, created by the advertiser, and perceived advertisement personalisation, which is personalisation actually felt by the recipient. Therefore, an
advertisement can be personalised by the advertiser. However it could still be possible that the recipient does not perceive the ad as personalised and vice versa (Simonson, 2005: Li, 2016). In this study we are therefore using the construct of perceived advertisement personalisation, in contrast to other studies that only use advertisement personalisation. Since in this construct a recipient needs to recognize the personalisation that is visible in the advertisement before any favourable effect of personalisation to take place (Kramer, 2007). When the recipient of the ad actually perceives the ad as personalised, then the ad is able to increase the appeal of an advertisement (Kramer, 2007). This is because the receiver is more likely to suggest that there are similarities between the receiver self and what is displayed in the advertisement (Anand & Shachar, 2009; Malheiros et al., 2012). Consequently, the receiver of the ad tends to perceive these personalised advertisements to be more relevant (Noar, Harrington & Aldrich, 2009). Which leads to the following hypothesis.
H1. The stronger one‟s perception of the personalisation of an advertisement, the more
relevant an advertisement is perceived.
According to Noar et al., 2009, Perceived advertisement relevance is able to positively influence the brand in multiple ways. According to Jung (2017) perceived advertisement relevance is known to favourably affect attention towards the advertisement (Pechmann and Stewart, 1990). Furthermore, an advertisement that is perceived as highly relevant, is able to generate positive attitudes towards the advertisement ( Trampe, Stapel, Siero & Mulder, 2010). This leads to believe that higher levels of perceived advertisement relevance could also lead to more positive attitudes towards a brand. However, little research has been done in regard to this relationship.
9 consumers about the brand, can be advertisements. Brands are able to elicit positive attitudes among consumers by exposing them to advertisement that are perceived as relevant by consumers (Debevec & Iyer, 1988). Because of this, it is expected that the attitude towards the brand will be more positive when the advertisements put out by the brand are perceived as relevant by the recipient of the advertisement. Since stronger perceptions of advertisement relevance lead to more positive images, and hence more positive attitudes of the brand, we expect that this will result in more positive attitudes towards the brand. Subsequently the following hypothesis was generated.
H2. The more relevant an ad is perceived, the more positive one‟s attitude will be towards the
advertised brand.
Furthermore, Noar et al. (2009), identified perceived advertisement relevance as a primary mediator between advertisement personalisation and positive personalisation effects. However, this mediating role of perceived relevance has not been confirmed yet in combination with perceived advertisement personalisation, as independent variable, and attitude towards the brand, as dependent variable, in a social media setting. However, since we expect that attitude towards the brand is positively influenced by perceived advertisement relevance and we expect perceived advertisement relevance to be positively influenced by perceived advertisement personalisation, the following hypothesis was generated.
H3. Perceived advertisement relevance mediates the relationship between perceived
advertisement personalisation and attitude towards the advertised brand.
2.3 The moderating role of perceived advertisement intrusiveness on the relationship between perceived advertisement relevance and attitude towards the brand
10 Intrusiveness in psychology literature is defined as „‟an creation of an imbalance between closeness and autonomy‟‟ (Lavy, Mikulincer, Shaver & Gillath, 2009, p.990). Closeness refers to the degree of interdependence and relatedness between two parties. Furthermore, autonomy is the degree to which personal identity can be preserved (lavy et al,. 2009). In the advertising literature perceived advertisement intrusiveness is defined as a psychological reaction to advertisements that interfere with a consumers on going cognitive processes (Edwards, Li & Lee, 2002).
When personal information and preferences are used in advertising, then this can lead to recipients perceiving threats to their freedom due to increased perceived advertisement intrusiveness (Brehm, 1966). Consequently, this could lead to negative attitudes and behaviours towards the advertised brand . This is the case when receivers of these advertisement turn into a state of psychological reactance (Brehm, 1966). Consequently, recipients will attempt to retain their restricted freedom (Aguirre, Mahr, Grewal, de Ruyter & Wetzels, 2015). Subsequently this could lead to changes in behaviour and a negative attitude towards an object, in this case the brand (Brehm, 1966).
In a Facebook setting, this effect of perceived advertisement intrusiveness caused by a higher degree of personalised advertising, could even be more present compared to
personalised advertising in a website context. Because, the website in which advertisements are shown are not controlled by the recipient of the advertisement (De Keyzer et al., 2015). And since the Facebook timelines of the consumers, on which the brand puts their
advertisement, is designed and controlled by the receiver this could lead to different outcomes (De Keyzer et al., 2015). This is mainly due to the fact that these Facebook timelines are considered as personal space by the users of Facebook (De Keyzer et al., 2015). Therefore, these advertisements are affected by the Psychological ownership theory. Psychological ownership is a state that lets a person have a sense of ownership over external objects, in this case the Facebook timeline (Pierce, Kostova and Dirks, 2001; Aguirre et al., 2015). In this personal space of Facebook, receivers may perceive these personalised and relevant advertisements as more intrusive. This subsequently leads to perceived relevant
11
H4. There is a weaker positive relationship between perceived advertisement relevance and
the attitude towards a brand, if the receiver of the ad perceives a higher degree of intrusiveness caused by the ad.
Figure 1
12
3. Method
3.1 Research design
To analyse the relationships between perceived advertisement personalisation, perceived advertisement relevance, perceived advertisement intrusiveness and attitude towards the brand a one-way between-subjects experimental design was employed. The independent
manipulation variable, perceived advertisement personalisation, was manipulated with two levels: a high and a low personalised advertisement. After being exposed to the experimental manipulation, the respondents answered questions in relation to the mediator variable
„perceived advertisement relevance‟ and the moderator variable „perceived advertisement intrusiveness‟. The dependent variable of the study was the attitude towards the brand. All these measures were continuous variables.
In this study, actual personalisation was manipulated and perceived advertisement personalisation was measured with two statements, as used by Li (2016). These two statements were: (1) „‟The ad seemed to be designed specifically for me‟‟, and (2) „‟The advertisement targeted me as an unique individual‟‟. The statements were measured on a seven-point Likert scale with 1 being “strongly disagree” and 7 being “strongly agree”. The two statements that measured the perceived advertisement personalisation were averaged (α = .74). A dummy-coded variable was created for actual personalisation, with 0 being the low personalised condition and 1 being the high personalized condition.
Furthermore, we controlled for the influence of socio-demographics (age and income) to see whether they would have an effect on the tested relationships. Age was important since generally spoken younger people are on the platform of Facebook (Clement, 2019). Income was taken into account out of practical reasons, since people with lower incomes could face problems with identifying with the person in the proposed scenario. Since travelling can seem as rather expensive by people with lower incomes, which subsequently might cause these people to generally be more reluctant towards such advertisements. Finally, the frequency of Facebook usage was included as a covariate. Because it could be that the more someone uses Facebook, the more likely it is that this person has become tolerant or „numb‟ for personalised advertising, so the less likely that „personalisation‟ will be perceived at all.
13 The target population for this research were Dutch people who use Facebook. The research was conducted in the Netherlands. Hence, a Dutch speaking population was targeted and only people who were fluent in Dutch could participate in the study. Furthermore, since this research focused on a Facebook context, people who did not own a Facebook account were also excluded from the study.
Since the main preconditions to participate in this study were to be Dutch and a user of Facebook, convenient sampling was used. By sending a Dutch survey invitation through the platform Facebook, individuals with both preconditions could easily be reached. Apart from that, digital questionnaires are a convenient way of data collection when limited resources and fast response times are required (Ilieva, Baron & Healey, 2002). The data was gathered
between the 27th of November 2019 and the 12th of December 2019. Seven days after the initial distribution a reminder was send out.
14 Table 1
Descriptive statistics of the sample.
Frequency Gender Male 37.8% Female 60.3% Other 0.5% Unknown 1.4% N Mean (SD) Age 209 28 (11.30) Educationa 207 15.94 (1.25) Incomeb 171 12.11 (1.90)
Frequency of Facebook usage 209 4.48 (1.45)
Note.
a
Education was measured categorically. The mean of 15.94 implies that the respondents‟ average education was between secondary vocational education and higher professional education.
b
Income was also measured categorically. The mean of 12.11 implies that the respondents‟ average income was between €10.000 - €19.999 and €20.000 - €29.999 per year.
3.3 Materials
15 were either shown an advertisement for a trip to Bali (high personalisation) ( n = 103) or a weekend holiday to Amsterdam (low personalisation) (n = 106).
A low personalised advertisement was chosen instead of a „none personalised‟ or generic advertisement in regard to the measurement of the control group. This is because the present research aimed to investigate the difference between the extent of personalisation (instead of the difference between personalised and generic advertisements). The destination Bali was chosen to represent the high personalised group, since the person in the scenario is planning to go to Bali for a trip and shows interest in this by liking and following the Facebook page „‟BaliTrips‟‟. In this case the Bali advertisement is likely to be perceived as highly personalised. The Amsterdam advertisement was shown to the low personalised group, since a weekend holiday to Amsterdam is still likely to be regarded as travelling, which the person in the scenario likes to do, but it cannot be compared to a trip to Bali in terms of type of holiday.
Furthermore, a fictitious brand was used in this experiment to avoid that there would be any confounds of previous attitudes towards an actual brand. Both of the advertisements used in the experiment looked similar to ensure treatment equivalence. This was done to avoid potential appearance related effects (Lee & Cranage, 2011). Only the background picture and the text were slightly adjusted to fit either a weekend holiday to Amsterdam or a 7-day trip to Bali (see Appendix B).
3.4 Procedure
The survey started with whether respondents had a Facebook account and how frequently they used it. When respondents indicated that they did not own a Facebook account, they were thanked and excluded from further participation (n = 18). After these questions, the
respondents had to read the scenario in which they were asked to identify themselves with the person in the scenario. Next, the respondents were randomly assigned to the one of the two experimental condition: They either received the highly personalised advertisement of a fictitious brand regarding a trip to Bali (n = 103) or they received the low personalised advertisement regarding a weekend holiday to Amsterdam (n = 106). After that, the respondents received the questions regarding the manipulation check, perceived
16 advertisement intrusiveness was random. Finally, the dependent variable, attitude towards the brand, was measured. There was intentionally chosen for showing the questions regarding perceived advertisement relevance and perceived advertisement intrusiveness after the advertisement and perceived advertisement personalisation questions (manipulation check) were shown and before the questions regarding attitude towards the brand were shown. Since this gave respondents more time to form an attitude towards the brand which they never saw before.
3.5 Measures
The measures for the constructs perceived advertisement relevance, perceived advertisement intrusiveness and attitude towards the brand were all adopted from already existing literature to ensure the reliability and the validity of the constructs. The items used for the construct perceived advertisement relevance were adopted from Laczniac and Muehling, (1993). The items used for perceived advertisement intrusiveness were adapted from (Edwards et al., 2002; Mooradian 1996; van Doorn & Hoekstra, 2013). Both perceived advertisement
relevance and perceived advertisement intrusiveness were based on a seven-point Likert scale, ranging from “strongly disagree” to “strongly agree”. The items used for the construct attitude towards the brand were adopted from (Spears & Singh, 2004). Furthermore, this construct was measured on a seven-point semantic bipolar scale. All of the items used in this study can be seen in Appendix A.
Since the data for this research was gathered in the Netherlands, all of the items were translated through the use of back translation. This was done to make the items clear and understandable for the Dutch population. This is important to eventually acquire the most valid and reliable data. Due to linguistically differences between the English and Dutch language, the items were marginally modified during the translation process. The translation of the items can be seen in Appendix A.
3.6 Manipulation check
A manipulation check was included to see whether respondents were actually perceiving the highly personalised advertisement as more personalised than the low personalised
17 successful manipulation is important since the goal of an experimental research is to draw conclusions about the cause-and-effect relationship between two variables. Also, the researcher can then conclude that the participants have perceived and interpreted the
manipulation correctly. Moreover, conclusions on the relationship between the independent and dependent variables can be drawn more accurately. Furthermore, an unsuccessful
manipulation (i.e. when the participants did not perceive the high personalisation as high) is a potential threat to the internal validity of a study. In addition to that, within the research field of marketing, Li (2016) concluded that an actual personalised advertisement can be viewed as not personalised and vice versa. So, it is important to incorporate a manipulation check into the research design to be able to conclude that respondents actually perceived an high personalised advertisement as more personal than a low personalised advertisement.
To check whether the manipulation of the personalisation of the advertisement was successful, an independent samples t-test was executed with level of personalisation (low and high) being the grouping variable and perceived advertisement personalisation being the outcome variable. Before carrying out the independent samples t-test, the assumptions were checked. The assumptions were: independency of cases, normality and homogeneity of
variances. A test of normality (p < .001) and Levene‟s test (p < .05) both showed significance, which indicates that the assumptions of normality and homogeneity could not be assumed. Despite that, the independent samples t-test was carried out since the Central Limit Theorem states that “the sampling distribution of the mean for any population, given an adequate sample size, will approximate a standard normal distribution” (Aberson, Berger, Healy, Kyle & Romero, 2000) and the sample size of this study (N = 209) was adequate (VanVoorhis & Morgan, 2007). Also, the sample sizes of the groups were approximately equal (low
18
3.7 Plan of analysis
IBM SPSS version 25 was used to perform the statistical tests that were needed to answer the hypotheses. Firstly, the dataset was cleaned. With the results a two-step analysis was
executed, which is recommended by Henseler, Ringle and Sinkovics (2009). Step one entails examining the reliability and validity of all of the items in each construct by means of an exploratory factor analysis. A Kaiser-Meyer-Olkin measure of sampling adequacy was used to examine the appropriateness of the factor analysis. Values below 0.50 indicate that the factor analysis might not be appropriate (Malhotra, 2010). Values between 0.50 and 1.00 indicate that the factor analysis is appropriate to execute. Also, Bartlett‟s test of sphericity was used to examine whether the variables were uncorrelated (Malhotra, 2010). The value of this test should be significant (p < 0.05) in order to perform a factor analysis. Moreover, the factor loadings were analysed to check whether items that should be measuring each construct correlated enough with each other. After the factor analysis, a reliability analysis using
Cronbach‟s Alpha was conducted to check if an averaged-variable could be computed. The minimum for computing an averaged-variable is a Cronbach‟s Alpha of .60 (Malhotra, 2010). If the factor loadings showed satisfactory loadings (> .50) (Hair, Anderson, Tatham and Black (1998) and the items formed a reliable construct, then the items would be taken together to compute the theoretical construct.
In the second step of the two-step analysis, analyses were carried out to answer the hypotheses one to four. To test hypothesis 1, a linear regression analysis was carried out with the continuous variable „‟perceived advertisement personalisation‟‟, that has been used as manipulation check, as the independent variable and perceived advertisement relevance as the outcome variable. In order to conduct a linear regression analysis, the assumptions for this analysis should be met. In the results section, the assumptions are reported and checked. After that, the analysis was carried out and potential confounding variables were checked by adding them as predictors in a multiple linear regression. To validate the results of the linear
19 To test hypothesis 2, a linear regression analysis was performed with perceived
advertisement relevance as the independent variable and attitude towards the brand as the dependent variable. First, an assumption check was done. Results of this check are reported in the results section. Secondly, the linear regression analysis was performed. Lastly, potential confounding variables were checked again by adding them as predictors in an multiple linear regression.
To test hypothesis 3, Hayes‟ PROCESS version 3.4 (Hayes, 2012) Mediation Test model 4 using the 95% confidence interval from 5000 bootstrapped samples was performed to test whether the relationship between perceived advertisement personalisation (i.e. continuous variable) and attitude towards the advertised brand is mediated by perceived advertisement relevance. As done before, an assumption check was performed before carrying out the analysis. Results of the assumption check can be found in the results section. After that, the mediation test was executed. To validate the results, the mediation test was repeated with condition (low and high) as the predicting variable. Then, possible covariates were checked by adding them one by one as covariates in the mediation test (model 4) with perceived advertisement personalisation being the independent variable and brand attitude being the outcome variable.
Lastly, to test hypothesis 4, Hayes‟ PROCESS version 3.4 (Hayes, 2012) Moderation Test model 1 using the 95% confidence interval from 5000 bootstrapped samples was
performed to test whether the relationship between perceived advertisement relevance and attitude towards the brand was moderated by the perceived advertisement intrusiveness. An assumption check was done, of which results can be found in the results section. The
20
4. Results
4.1 Reliability and validity
A factor analysis was used to reduce and summarize the data (Malhotra, 2010). The factor analysis was used to test convergent and discriminant validity and reliability. First of all, the Kaiser-Meyer-Olkin measure of sampling adequacy was executed to check if it was
appropriate to combine the two perceived advertisement personalisation (i.e. continuous variable) items, the ten perceived advertisement relevance items, the ten perceived
advertisement intrusiveness items and the five attitude towards the brand items into factors. All of the constructs showed satisfactory values in regard to the Kaiser-Meyer-Olkin Test and Bartlett‟s Test. The lowest outcome with the items regarding the perceived advertisement personalisation (i.e. continuous variable) showing the lowest KMO-outcome of 0.5 and a significant Bartlett‟s Test (see Appendix C). Since this is still regarded as acceptable, the factor analysis for the perceived advertisement personalisation was suitable, just like for the other constructs. Following, all of the correlations were checked to determine whether there was too little correlation (< 0.30) between the items, or too much (> 0.90) which could have resulted in multicollinearity. No correlation coefficients below the threshold of .30 or above the threshold of .90 were found, so there were no multicollinearity issues. After that, the factor analysis was carried out. According to Hair et al. (1998) items are considered
practically significant for their overarching construct when they surpass a threshold regarding the factor loading of .50. Table 2 shows the results of the factor analysis, including the
Cronbach‟s Alpha for each construct. All of the construct items showed satisfactory values (> .50). To determine how many factors should be created, the scree-plot, eigenvalues and the total variance explained were observed (see Appendix C). According to Kootstra (2004) eigenvalues should be above 1.00 and the total variance explained should be about higher than 70%. Although the total variance explained of the construct „perceived advertisement
21 Table 2
Factor analysis
Construct and item wording Loadings Cronbach‟
s alpha
Perceived advertisement personalisation .74 PAP_1: The ad seemed to be designed specifically for me. .89
PAP_2: The advertisement targeted me as a unique individual. .89
Perceived advertisement relevance .95
PAR_1: I felt that it might be important to me. .79 PAR_2: I felt that it could be Meaningful to me. .85 PAR_3: I felt that it could be created just for me.
PAR_4: I felt that it could be worth remembering. PAR_5: I felt that it could be of value to me. PAR_6: I felt that it could be relevant to my needs. PAR_7: I felt that it could be useful to me.
PAR_8: I felt that it could be worth paying attention. PAR_9: I felt that it could be interesting to me.
PAR_10: I felt that it could be likely to give me new ideas.
.51 .88 .92 .91 .93 .89 .90 .70
Perceived advertisement intrusiveness .92
PAI_1: I think this offer is disturbing. .81
PAI_2: I think this offer is alarming. .82
PAI_3: I think this offer is obtrusive. .82
PAI_4: I think this offer is irritating. PAI_5: I think this offer is annoying. PAI_6: I think this offer is uncomfortable.
PAI_7: I think it is uncomfortable that personal information is used in this offer.
PAI_8: The supplier knows a lot about me. PAI_9: This offer gives me an uneasy feeling. PAI_10: This offer gives me an unsafe feel.
22
Attitude towards brand .89
BA_1: Unappealing / Appealing .84
BA_2: Bad/Good .86
BA_3: Unpleasant / Pleasant BA_4: Unfavourable / Favourable BA_5: Unlikable / Likable
23
4.2 Influence of perceived advertisement personalisation on perceived advertisement relevance
To test hypothesis 1, a linear regression analysis was carried out with perceived advertisement personalisation (i.e. continuous variable) as the independent variable and perceived
advertisement relevance as the outcome variable. The assumptions for a linear regression analysis, linearity, normality, homoscedasticity and independency of cases, were checked. The scatterplot showed a strong positive linear relationship between the two variables. This was confirmed with a significant Pearson‟s correlation coefficient of .64. The assumptions of homoscedasticity and linearity were met, since the residuals in the scatterplot of standardised predicted values and standardised residuals were approximately normally distributed. The histogram of the standardised residuals also showed approximate normally distributed data. Thus, all assumptions were met.
Results indicated that there was a significant relationship between the two variables,
t(208) = 11.82, p < .001. The explained variance of the model was .40. The slope coefficient
for perceived advertisement personalisation was 0.55, which indicated that the stronger one‟s perception of the personalisation of an advertisement, the more relevant an advertisement is perceived. After that, potential confounding variables were checked by adding them as predictors in a multiple linear regression. The variables that were checked, were age,
frequency of Facebook usage and income. Multicollinearity was checked for these variables by observing Pearson‟s correlation coefficients. No correlation coefficients were found, therefore it could be assumed that there was no multicollinearity between the possible confounding variables. The multiple linear regression was performed and no significant confounding variables were found.
To validate these results and check for the potential causal relationship between advertisement personalisation (i.e. categorical variable) and perceived advertisement relevance, a one-way ANOVA was conducted with the condition (high vs. low
personalisation) as the independent variable and perceived advertisement relevance as the dependent variable. The assumptions for a one-way ANOVA, independency of cases, normality and homoscedasticity, were checked. To check the assumptions of
24 same. Since the sample sizes of the low personalisation group (n = 106) and the high
personalisation group (n = 103) were roughly the same, the assumption of homoscedasticity could be assumed. Results show that there was a significant difference in perceived
advertisement relevance between the low personalisation group and the high personalisation group, F(1, 208) = 29.37, p < .001. Potential covariates were checked by adding them as covariates in an ANCOVA-test. The extra assumptions, linearity in the relationship between the dependent variable and the covariate is linear and homogeneity of regression slopes, of an ANCOVA-test were checked. All assumptions were met. The variables that were checked, were age, frequency of Facebook usage and income. No significant covariates were found.
The ANOVA-test showed a significant difference in advertisement relevance between the low and high personalisation group. Therefore, these results confirmed a positive causal relationship between advertisement personalisation (i.e. categorical variable) and perceived advertisement relevance. Furthermore, the results from the regression analysis indicated that the stronger one‟s perception of the personalisation of an advertisement, the more relevant an
advertisement is perceived. This supported hypothesis 1.
4.3 Relationship between perceived advertisement relevance and attitude towards the brand
To test hypothesis 2, a linear regression analysis was performed with perceived advertisement relevance as the independent variable and attitude towards the brand as the dependent
variable. The assumptions for a linear regression analysis, linearity, normality,
homoscedasticity and independency of cases, were checked. The scatterplot showed a strong positive linear relationship between the two variables. This was confirmed with a significant Pearson‟s correlation coefficient of .41. The assumptions of homoscedasticity and linearity were met, since the residuals in the scatterplot of standardised predicted values and
standardised residuals were approximately normally distributed. The histogram of the standardised residuals also showed approximate normally distributed data. Thus, all assumptions were met.
Results indicated that there was a significant relationship between the two variables,
t(208) = 6.47, p < .001. The explained variance of the model was .17. The slope coefficient
25 adding them as predictors in an multiple linear regression. The variables that were checked, were age, frequency of Facebook usage and income. No significant confounding variables were found.
4.4 Perceived advertisement relevance as a mediator between perceived advertisement personalisation and attitude towards the brand
To test hypothesis 3, Hayes‟ PROCESS macro version 3.4 Mediation Test model 4 using the 95% confidence interval from 5000 bootstrapped samples was performed to test whether the relationship between perceived advertisement personalisation and attitude towards the brand is mediated by the perceived advertisement relevance (Hayes, 2012). Before carrying out the mediation test, the assumptions were checked. The assumptions were, continuous
measurements, normality, independence, linearity and homogeneity. All variables were measured on a continuous scale. Normality, linearity and homogeneity were checked by examining the expected and observed standardised residuals. The residuals in the scatterplot were randomly spread around y = 0 and no systematic deviations from the diagonal line could be perceived in the QQ-plot. Thus, all remaining assumptions were met.
After that, the mediation test was carried out. Table 4 provides an overview of the results of the mediation test. The left column of Table 4 shows that the perceived
advertisement relevance was higher when the respondents perceived the advertisement as more personalised (a = .55). The model explained 40% of variance in attitude towards the advertised brand. The right column of Table 4 shows that perceived advertisement relevance had a positive effect on attitude towards the advertised brand, which indicates that the higher the perceived advertisement relevance, the more positive one‟s attitude is towards the brand (b = .35). Hypothesis 3 states that the perceived advertisement relevance would mediate the relationship between perceived advertisement personalisation and attitude towards the brand. Since the direct effect of perceived advertisement personalisation on the attitude towards the brand did not remain significant after adding perceived advertisement relevance as a mediator (β = -.05, t(206) = -0.86, p = .39), whereas the indirect effect of the perceived advertisement relevance did show a significant effect , therefore it can be concluded that perceived
26 To validate these results, the mediation test was repeated with the categorical variable condition (low and high personalisation) as the predicting variable. Results show that the direct effect of the condition on the attitude towards the brand did not remain significant after adding perceived advertisement relevance as a mediator (β = -.25, t(206) = -1.80, p = .07), whereas the indirect effect of the perceived advertisement relevance did show a significant effect. Naturally, there were some slight variations in the results that were most likely caused by the different operationalization of the variable advertisement personalisation (categorical and continuous). However, in line with the mediation analysis reported above, the main findings did not change dramatically when using the continuous variable for personalisation instead of the categorical variable. Hence, both analyses showed a mediation effect, hereby supporting hypothesis 3: Perceived advertisement relevance mediates the relationship between perceived ad personalisation and attitude towards the advertised brand.
27 Table 3
Mediation effect of perceived advertisement relevance on the relationship between perceived advertisement personalisation and attitude towards the brand
Variable Perceived advertisement relevance (M) Attitude towards the brand (Y)
β SE p 95% CI β SE p 95% CI Constant i1 2.01 .18 < .001 1.65; 2.37 i2 2.93 .21 < .001 2.52; 3.34 PARa (M) - - - - b .35 .06 < .001 0.23; 0.47 PAPb (X) a .55 .05 < .001 0.46; 0.64 c1 -.05 .05 .39 -0.15; 0.06 F(1, 207) = 139.60, p < .001, R2 = .40 F(2, 206) = 21.26, p < .001, R2 = .17 Direct effect c’ -.05 .05 .39 -0.15; 0.06 Indirect effect ab .19 .04 - 0.12; 0.27 Covariates FoFUc -.03 .05 .52 -0.13; 0.07 .10 .04 < .05 0.01; 0.19 PAId -.19 .05 < .001 -0.30; -0.08 -.30 .05 < .001 -0.40; -0.21 a
PAR = perceived advertisement relevance b
PAP = perceived advertisement personalisation c
FoFU = frequency of Facebook usage d
PAI = perceived advertisement intrusiveness
4.5 Perceived advertisement intrusiveness as moderator on the relationship between perceived advertisement relevance and attitude towards the brand
To test hypothesis 4, Hayes‟ PROCESS version 3.4 Moderation Test model 1 using the 95% confidence interval from 5000 bootstrapped samples was performed to test whether the relationship between perceived advertisement relevance and attitude towards the brand was moderated by perceived advertisement intrusiveness (Hayes, 2012). In this model, an interaction term is automatically created between the independent variable (perceived
advertisement relevance) and the moderation variable (perceived advertisement intrusiveness) (Aiken & West, 1991). Before carrying out the moderation test, the assumptions were
28 line could be perceived in the QQ-plot. Thus, all remaining assumptions were met. After that, the moderation test was carried out.
The results of the moderation test can be found in Table 4. The model overall explains 32% of variance in attitudes towards the brand. A main effect for perceived advertisement relevance was found: the higher the perceived advertisement relevance, the more positive one‟s attitude towards the advertised brand was (p < .001). Also, a negative significant effect for perceived advertisement intrusiveness was observed: the more intrusive one‟s perceived the advertisement, the less positive one‟s attitude is towards the advertised brand (p < .001).
Furthermore, the interaction effect contributed significantly to the explanation of the model (Table 4). The interaction term of perceived advertisement relevance and perceived advertisement intrusiveness also showed to have a significant effect (p < .05). After that, the details of the interaction effects were analysed by looking at the conditional effects of the focal predictor at values of the moderator. For those with low perceived advertisement intrusiveness (1 SD below the mean), perceived advertisement relevance has a stronger influence on the attitude towards the brand (β = .36, t(205) = 6.30, p <.001) than for those with high perceived advertisement intrusiveness (1 SD above the mean) (β = .17, t(205) = 2.67, p < .01). Further analyses also show that as perceived advertisement intrusiveness decreases, the relationship between relevance and attitude towards the brand becomes more positive. Since, at a score of approximately 5 on a seven-point Likert scale on perceived advertisement intrusiveness, the values of the perceived advertisement relevance will become β = .18, t(205) = 2.89, p < .001, whereas at the lowest perceived advertisement intrusiveness (a score of 1), the values of the perceived advertisement relevance will become β = .46, t(205) = 5.01, p <.001. Therefore, it was concluded that the more intrusive an advertisement is perceived, the less positive the relationship between perceived advertisement relevance and the attitude towards the brand is, which is supporting hypothesis 4.
29 Table 4
Moderation effect of perceived advertisement intrusiveness on the relationship between perceived advertisement relevance and attitude towards the advertised brand
Variable β t p 95% CI Constant 4.14 70.50 < .001 4.03; 4.26 F(3, 205) = 32.23, p < .001; R2 = .32 PARa 0.26 5.83 < .001 0.17; 0.35 PAIb -0.31 -6.51 < .001 -0.40; -0.21 PAR*PAI -0.07 -2.32 .022 -0.14; -0.01 ΔF(1, 204) = 5.68, p < .05; ΔR2 = .02 Covariates Condition -0.27 -2.20 .03 -0.52; -0.03 F(4, 204) = 25.84, p < .001; R2 = .34 Frequency of Facebook usage 0.09 2.34 .02 0.01; 0.17 F(4, 204) = 26.06, p < .001; R2 = .34 a
PAR = perceived advertisement relevance b
30
5. Discussion
This research contributes in four different ways to the current literature. First, this study explored the personalisation paradox in a social media setting. Second, it contributes as one of the first papers to do this as far as known, to the exploration of the effect of advertisement personalisation on the attitude towards the brand. Third, this study confirmed the causal effect of advertisement personalisation on perceived advertisement relevance. Finally, this research contributes to current literature by confirming that there is a mediating effect of perceived advertisement relevance between the relationship of advertisement personalisation on attitude towards the brand.
Next, the findings of the mediation and moderation analyses will be discussed. Furthermore, the theoretical contributions, limitations, future research and practical implications for advertisers in a social media setting will be presented.
Before the mediation effect of perceived advertisement relevance between the relationship of advertisement personalisation was checked, first the causal relationship
between advertisement personalisation (i.e. categorical variable) and perceived advertisement relevance, the linear relationship between perceived advertisement personalisation (i.e. continuous variable) and perceived advertisement relevance and the linear relationship between perceived advertisement relevance and attitude towards the brand.
Firstly, this study was able to contribute to the current literature by confirming, in the setting of social media, that there is a positive causal effect advertisement personalisation (i.e. categorical variable) on perceived advertisement. Secondly, the positive relationship between perceived advertisement personalisation (i.e. continuous variable) and perceived
advertisement relevance was confirmed. This find is essentially in line with what Petty et al., (2000) already found. Thirdly, a positive linear relationship between perceived advertisement relevance and attitude towards the brand was confirmed. This study is one of the first to confirm the positive direct effect of perceived advertisement relevance on the attitude towards the brand.
31 between perceived advertisement personalisation (i.e. continuous variable) and attitude
towards the brand. Therefore, this study confirmed as one of the first to do so, as far as known that the relationship between advertisement personalisation and attitude towards the brand is fully mediated perceived advertisement relevance. Overall, from these finds in regard to the mediation analysis can be concluded that personalisation in advertising indeed is able to elicit positive outcomes for the brand. With these findings in mind, it could be important to do more research in regard to these relationships. Since a limitation of this research is that the present study was only partially designed as an experimental study, and only the causal relationship between advertisement personalisation and perceived advertisement relevance could be confirmed. Other research could focus on examining all of these relationships with an experimental study design, in order to find out more about the causal relationships between these constructs (Shanahan, Tran & Taylor, 2019).
The moderating role of perceived advertisement intrusiveness was researched and there was found that the interaction effect of perceived advertisement intrusiveness was significantly and negatively influencing the relationship between perceived advertisement relevance and attitude towards the brand. Therefore, this study contributes to the current literature, as one of the first studies to do so in a social media setting, that when advertisement are perceived as intrusive, this feeling of intrusiveness will mitigate the positive effect of perceived advertisement relevance on attitude towards the brand. This however brings up the question what exactly causes the perceived advertisement intrusiveness. This was something that was not researched in this paper. It could be practically useful for advertisers to know what exactly causes advertisements to be perceived as highly intrusive. Therefore, in future research the exact drivers of perceived advertisement intrusiveness in personalised advertising could be further explored.
32
6. Conclusion
This study investigated the effect of advertisement personalisation on the attitude towards the brand, the mediating effect of perceived advertisement relevance on this relationship and the role of perceived advertisement intrusiveness as a moderator on the relationship between perceived advertisement relevance and attitude towards the brand. The aim of this master‟s thesis was to answer the following research question: To what extent does perceived
advertisement intrusiveness affect the relationship between advertisement personalisation and attitude towards the brand in a social media context? Results of this study showed that the stronger one‟s perception of the personalisation of an advertisement was, the more relevant an advertisement was perceived. Also, the more relevant an ad was perceived, the more positive one‟s attitude would be towards the advertised brand. The relationship between perceived advertisement personalisation and the attitude towards the brand was further investigated and showed to be fully mediated by the perceived advertisement relevance. Furthermore, evidence was found for a moderation effect: there was a weaker positive relationship between
33
7. References
Aberson, C. L., Berger, D. E., Healy, M. R., Kyle, D. J., & Romero, V. L. (2000). Evaluation of an interactive tutorial for teaching the central limit theorem. Teaching of Psychology,
27(4), 289-291.
Aguirre, E., Mahr, D., Grewal, D., de Ruyter, K., & Wetzels, M. (2015). Unraveling the personalization paradox: The effect of information collection and trust-building strategies on online advertisement effectiveness. Journal of retailing, 91(1), 34-49.
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Thousand Oaks, California: Sage.
Ajzen, I., Timko, C., & White, J. B. (1982). Self-monitoring and the attitude–behavior relation. Journal of personality and social psychology, 42(3), 426.
Anand, B. N., & Shachar, R. (2009). Targeted advertising as a signal. QME, 7(3), 237-266.
Arora, N., Dreze, X., Ghose, A., Hess, J. D., Iyengar, R., Jing, B., ... & Sajeesh, S. (2008). Putting one-to-one marketing to work: Personalization, customization, and choice. Marketing Letters, 19(3-4), 305.
Brehm, J. W. (1966). A theory of psychological reactance.
Clement, J. (2019, November 20). Global Facebook user age & gender distribution 2019. Retrieved November 15, 2019, from https://www.statista.com/statistics/376128/facebook-global-user-age-distribution/.
CMO survey (2019). CMO Survey Report: Highlights and Insights Report. Retrieved from
https://cmosurvey.org/wp-content/uploads/2019/02/The_CMO_Survey-Highlights-and_Insights_Report-Feb-2019-1.pdf
59-34 76.
Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. Journal of consumer research, 15(2), 210-224.
Chatterjee, P. (2008). Are Unclicked Ads Wasted? Enduring Effects of Banner and Pop-Up Ad Exposures on Brand Memory and Attitudes. Journal of electronic commerce Research, 9(1).
Debevec, K., & Iyer, E. (1988). Self‐referencing as a mediator of the effectiveness of sex‐role portrayals in advertising. Psychology & Marketing, 5(1), 71-84.
De Keyzer, F., Dens, N., & De Pelsmacker, P. (2015). Is this for me? How consumers respond to personalized advertising on social network sites. Journal of Interactive Advertising, 15(2), 124-134.
Dodds, W. B., Monroe, K. B., & Grewal, D. (1991). Effects of price, brand, and store information on buyers‟ product evaluations. Journal of marketing research, 28(3), 307-319.
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Harcourt Brace Jovanovich College Publishers.
Enberg, J. (2019, March 28). Global Digital Ad Spending 2019. Retrieved from https://www.emarketer.com/content/global-digital-ad-spending-2019
Edwards, S. M., Li, H., & Lee, J. H. (2002). Forced exposure and psychological reactance: Antecedents and consequences of the perceived intrusiveness of pop-up ads. Journal of Advertising, 31(3), 83-95.
35 Hair, J. F., Anderson, R. E., Tatham, R.L., & Black, W. C. (1998). Multivariate Data Analysis (5th ed). Upper Saddle River, New Jersey: Prentice Hall.
Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling.
Henseler, J., Ringle, C. M., & Sinkovics, R. R. (2009). The use of partial least squares path modeling in international marketing. In New challenges to international marketing (pp. 277-319). Emerald Group Publishing Limited.
Ho, S. Y., & Bodoff, D. (2014). The effects of Web personalization on user attitude and behavior: An integration of the elaboration likelihood model and consumer search theory. MIS
quarterly, 38(2).
Ilieva, J., Baron, S., & Healey, N. M. (2002). Online surveys in marketing research: Pros and cons. International Journal of Market Research, 44(3), 361.
Jung, A. R. (2017). The influence of perceived ad relevance on social media advertising: An empirical examination of a mediating role of privacy concern. Computers in Human
Behavior, 70, 303-309.
Kelly, L., Kerr, G., & Drennan, J. (2010). Avoidance of advertising in social networking sites: The teenage perspective. Journal of interactive advertising, 10(2), 16-27.
Kramer, T. (2007). The effect of measurement task transparency on preference construction and evaluations of personalized recommendations. Journal of Marketing Research, 44(2), 224-233.
Laczniak, R. N., & Muehling, D. D. (1993). Toward a better understanding of the role of advertising message involvement in ad processing. Psychology & Marketing, 10(4), 301-319.
36 Lavy, S., Mikulincer, M., Shaver, P. R., & Gillath, O. (2009). Intrusiveness in romantic relationships: A cross-cultural perspective on imbalances between proximity and autonomy. Journal of Social and Personal relationships, 26(6-7), 989-1008.
Lee, C. H., & Cranage, D. A. (2011). Personalisation–privacy paradox: The effects of personalisation and privacy assurance on customer responses to travel Web sites. Tourism
Management, 32(5), 987-994.
Lee, E. B., Lee, S. G., & Yang, C. G. (2017). The influences of advertisement attitude and brand attitude on purchase intention of smartphone advertising. Industrial Management &
Data Systems, 117(6), 1011-1036.
Li, C. (2016). When does web-based personalization really work? The distinction between actual personalization and perceived personalization. Computers in human behavior, 54, 25-33.
Machleit, K. A., Allen, C. T., & Madden, T. J. (1993). The mature brand and brand interest: An alternative consequence of ad-evoked affect. Journal of Marketing, 57(4), 72-82.
Malheiros, M., Jennett, C., Patel, S., Brostoff, S., & Sasse, M. A. (2012, May). Too close for comfort: A study of the effectiveness and acceptability of rich-media personalized
advertising. In Proceedings of the SIGCHI conference on human factors in computing systems (pp. 579-588). ACM.
Malhotra, N. K. (2010). Marketing research: An applied orientation (6th ed.). Boston, New York: Pearson.
Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitude?. Journal of marketing research, 18(3), 318-332.
37 Noar, S. M., Harrington, N. G., & Aldrich, R. S. (2009). The role of message tailoring in the development of persuasive health communication messages. In C. S. Beck (Ed.),
Communication yearbook 33 (pp. 73e133). New York, NY: Routledge.
Ortiz-Ospina, E. (2019, September 18). The rise of social media. Retrieved from https://ourworldindata.org/rise-of-social-media
Pechmann, C., & Stewart, D. W. (1990). The effects of comparative advertising on attention, memory, and purchase intentions. Journal of Consumer Research, 17(2), 180-191.
Perdue, B. C., & Summers, J. O. (1986). Checking the success of manipulations in marketing experiments. Journal of marketing research, 23(4), 317-326.
Petty, R. E., Wheeler, S. C., & Bizer, G. Y. (2000). Attitude functions and persuasion: An elaboration likelihood approach to matched versus mismatched messages. Why we evaluate: Functions of attitudes, 133-162.
Pierce, J. L., Kostova, T., & Dirks, K. T. (2001). Toward a theory of psychological ownership in organizations. Academy of management review, 26(2), 298-310.
Shanahan, T., Tran, T. P., & Taylor, E. C. (2019). Getting to know you: Social media personalization as a means of enhancing brand loyalty and perceived quality. Journal of Retailing and Consumer Services, 47, 57-65.
Simonson, I. (2005). Determinants of customers‟ responses to customized offers: Conceptual framework and research propositions. Journal of marketing, 69(1), 32-45.
Spears, N., & Singh, S. N. (2004). Measuring attitude toward the brand and purchase intentions. Journal of current issues & research in advertising, 26(2), 53-66.
38 Trampe, D., Stapel, D. A., Siero, F. W., & Mulder, H. (2010). Beauty as a tool: The effect of model attractiveness, product relevance, and elaboration likelihood on advertising
effectiveness. Psychology & Marketing, 27(12), 1101-1121.
Tucker, C. (2011, May). Social networks, personalized advertising, and perceptions of privacy control. In Proceedings of the tenth workshop on the economics of information security
(WEIS).
Tye-Williams, S. (2018). Manipulation Check. In M. Allen (Ed.), The SAGE Encyclopedia of
Communication Research Methods (1st ed.) (pp. 901-903). Thousand Oaks, California: SAGE Publications.
Van Doorn, J., & Hoekstra, J. C. (2013). Customization of online advertising: The role of intrusiveness. Marketing Letters, 24(4), 339-351.
VanVoorhis, C. W., & Morgan, B. L. (2007). Understanding power and rules of thumb for
determining sample sizes. Tutorials in quantitative methods for psychology, 3(2), 43-50.
White, T. B., Zahay, D. L., Thorbjørnsen, H., & Shavitt, S. (2008). Getting too personal: Reactance to highly personalized email solicitations. Marketing Letters, 19(1), 39-50.
39
8. Appendices
Appendix A
Translations of the scenario and the items
Scenario
EN:
Imagine yourself as being someone who loves to travel, especially to Asia. You have already been to Indonesia, Thailand and Vietnam. Bali, an island part of Indonesia, is the next
destination you want to visit. Recently you were on Facebook and came across the Facebook page „‟BaliTrips‟‟. BaliTrips posts pictures and travel tips regarding Bali, named: Bali.nl. Since you are interested in travelling to Bali you decide to like and follow the page Bali.nl. The following day you receive the following advertisement on your Facebook timeline from the company travels-for-u. During the week you regularly see the same advertisement on your Facebook timeline.
You did not have any previous encounters with the company travels-for-u and therefore did not know about the existence of this company. Furthermore, the Facebook page „‟BaliTrips‟‟ and he company Travels-for-U are in no shape or form connected to eachother.
NL:
40 Je bent nog nooit eerder in aanraking gekomen met iets van het bedrijf Travels-for-U en je wist dus ook niet dat Travels-for-U bestond. Daarnaast zijn de Facebookpagina ''BaliTrips'' en het bedrijf Travels-for-U op geen enkele manier met elkaar verbonden.
Items
Perceived advertisement personalisation
PAP_1:
English: The ad seemed to be designed specifically for me. Dutch: De advertentie leek specifiek voor mij te zijn ontworpen.
PAP_2:
English: The advertisement targeted me as an unique individual. Dutch: De advertentie richtte zich op mij als een uniek persoon.
(Li, 2016)
Perceived advertisement relevance
Instruction:
English: When I saw the advertisement of Travels-for-U on Facebook…: Dutch: Toen ik de advertentie van Travels-for-U op Facebook zag…
PAR_1:
English: I felt that it might be important to me.
Dutch: Had ik het idee dat het belangrijk voor me kon zijn.
PAR_2:
English: I felt that it could be Meaningful to me.
Dutch: Had ik het idee dat het zinvol voor mij kon zijn.
PAR_3: I felt that it could be created just for me. Had ik het idee dat het alleen voor mij was gemaakt.
41 English: I felt that it could be worth remembering.
Dutch: Had ik het idee dat het de moeite waard was om het te onthouden.
PAR_5:
English: I felt that it could be of value to me.
Dutch: Had ik het idee dat het voor mij van waarde kon zijn.
PAR_6:
English: I felt that it could be relevant to my needs.
Dutch: Had ik het idee dat het relevant kon zijn voor mijn behoeften.
PAR_7:
English: I felt that it could be useful to me.
Dutch: Had ik het idee dat het voor mij nuttig kon zijn.
PAR_8:
English: I felt that it could be worth paying attention. Dutch: Had ik het idee dat het de aandacht waard was.
PAR_9:
English: I felt that it could be interesting to me.
Dutch: Had ik het idee dat het interessant voor me kon zijn.
PAR_10:
English: I felt that it could be likely to give me new ideas.
Dutch: Had ik het idee dat het me waarschijnlijk nieuwe ideeën kon geven.
Adopted from Laczniac & Muehling, (1993)
Perceived advertisement intrusiveness
PAI_1:
English: I think this offer is disturbing.
42 PAI_2:
English: I think this offer is alarming.
Dutch: Ik vind deze advertentie zorgwekkend.
PAI_3:
English: I think this offer is obtrusive. Dutch: Ik vind deze advertentie opdringerig.
PAI_4:
English: I think this offer is irritating. Dutch: Ik vind deze advertentie irritant.
PAI_5:
English: I think this offer is annoying. Dutch: Ik vind deze advertentie vervelend.
PAI_6:
English: I think this offer is uncomfortable. Dutch: Ik vind deze advertentie onaangenaam.
PAI_7: I think it is uncomfortable that personal information is used in this offer.
Ik vind het onaangenaam dat persoonlijke informatie in deze advertentie wordt gebruikt.
PAI_8:
English: The advertiser knows a lot about me. Dutch: De adverteerder weet veel over mij.
PAI_9: This offer gives me an uneasy feeling. Deze advertentie geeft me een ongerust gevoel.
PAI_10:
43 Dutch: Deze advertentie geeft me een onveilig gevoel.
(Edwards et al., 2002) (Mooradian 1996) (van Doorn & Hoekstra, 2013)
Attitudes towards the brand
Instruction:
English: Please describe your overall feelings about the brand described in the ad you just read.
Dutch: Beschrijf uw algemene gevoelens over het bedrijf dat werd beschreven in de advertentie die u zojuist hebt gelezen.
BA_1:
English: Unappealing / Appealing Dutch: Onaantrekkelijk / Aantrekkelijk
BA_2:
English: Bad/Good Dutch: Slecht / Goed
BA_3:
English: Unpleasant / Pleasant Dutch: Onaangenaam / Aangenaam
BA_4:
English: Unfavourable / Favorable Dutch: Ongunstig / gunstig
BA_5:
English: Unlikable / Likable
Dutch: Onsympathiek / Sympathiek
44
Appendix B
Experimental conditions
45 Appendix C
Table with KMO- and Bartlett’s-values, Eigenvalues and variance explained
Construct KMO Bartlett‟s
Test Eigenvalue Variance explained (%) Perceived advertisement personalisation .50 < .001 1.59 79.58