Online behavioral advertising - the influence of personalization, consumer characteristics, brand familiarity, brand trust, and transparency on consumers’ attitudes, intentions and behaviors

Universidade do Minho Escola de Economia e Gestão Menna-Allah Mohamed Mostafa Online Behavioral Advertising – The Influence of Personalization, Consumer Characteristics, Brand Familiarity, Brand Trust, and Transparency on Consumers’ Attitudes, Intentions and Behaviors Online Behavioral Advertising – The Influence of Personalization, Consumer Characteristics, Brand Familiarity, Brand Trust, and Transparency on Consumers’ Attitudes, Intentions and Behaviors Menna-Allah Mohamed Mostafa U Minho | 2024 June 2024 Universidade do Minho Escola de Economia e Gestão Menna-Allah Mohamed Mostafa Online Behavioral Advertising – The Influence of Personalization, Consumer Characteristics, Brand Familiarity, Brand Trust, and Transparency on Consumers’ Attitudes, Intentions and Behaviors Doctoral Thesis In Business Administration Conducted under the supervision of: Professora Ana Maria Soares June 2024 ii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Atribuição-NãoComercial-SemDerivações CC BY-NCND h tt p s :/ / crea t ive c o mm on s .or g / lic en s es / by - nc - n d/ 4.0/ iii ACKNOWLEDGMENT I would like to thank God, The Great Lord, without you I would have achieved nothing in my life. I would like to thank my dear supervisor, Professora Ana Maria Soares, you have been the best motivator, supporter, and friend. Your patience and consideration were endless. Thank you for being the person you are. I will forever cherish our relationship. I would also like to thank my Mom Amani Mostafa, for everything you have done for me over the past years, you are the strongest person I know. I would like to thank my family, Uncle Khalid Khalil, Motherin-law Mariam Hussein and My Husband Ahmed Moheib. Your love and support were what lifted me up during my darkest times. My Friends, Bassant Saeed, and Dina Yahia, how can I ever thank you enough! You made my PhD journey a very remarkable one. I am very thankful for you being in my life. I would like to extend my thanks to all my professors in the Ph.D. course, and all the team members of the School of Economics and Management, who were always supportive and readily available whenever I needed them. Finally, I would like to thank my Grandpa Mostafa Khalil for everything you ever did for me. You raised me up like the dad I never had and always wanted to see me a PhD holder, this is for you. May you rest in peace. Amen. iv STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. v RESUMO A tecnologia permite às empresas compreender melhor os seus clientes-alvo. Atualmente, os anunciantes podem visar os seus consumidores com base no conhecimento real de quem, quando, como e porque é mais provável que comprem um produto específico. As empresas utilizam a publicidade comportamental online (Online Behavioural Advertising - OBA) - ou seja, monitorizar o comportamento online dos consumidores para apresentar anúncios personalizados aos consumidores visados - para aumentar o retorno atual e futuro das empresas. No entanto, a OBA também leva a um aumento das preocupações com a privacidade, desconfiança e reação à compra do produto ou à utilização do serviço que a empresa oferece. A personalização é considerada a principal abordagem à OBA, uma vez que ajuda a tornar os anúncios mais apelativos do ponto de vista pessoal. No entanto, a investigação empírica mostra resultados contraditórios no que respeita aos efeitos da personalização e à forma como esta pode até causar reacções e atitudes negativas por parte dos consumidores devido ao aumento das preocupações com a privacidade. É proposta e testada empiricamente uma modelo concetual que engloba o papel da personalização, das características do consumidor, da familiaridade com a marca, da confiança na marca e da transparência nas atitudes e intenções dos consumidores e no comportamento protetor em relação às marcas. Além disso, o estudo explora a influência mediadora da perceção de intrusividade, da perceção de preocupação com a privacidade e da reactância nas atitudes e intenções dos consumidores e no comportamento de proteção, tudo isto com base nos fundamentos teóricos da Social Exchange Theory . Os resultados mostram que alguns dos construtos propostos, nomeadamente a personalização, as características do consumidor, a familiaridade com a marca e a confiança na marca, influenciam as atitudes dos consumidores em relação à marca, as atitudes em relação ao OBA, as intenções de clicar no anúncio, as intenções de comprar o produto anunciado e o comportamento de evitar o OBA. Além disso, os resultados confirmam que esta influência é afetada por uma série de elementos mediadores, nomeadamente a perceção da intrusividade pelos consumidores, a perceção das preocupações com a privacidade e a perceção da ameaça à escolha. Estes resultados têm implicações teóricas e práticas. No que respeita à teoria, o estudo aprofunda a compreensão do conceito de OBA e da sua eficácia. A investigação visou contribuir para os estudos académicos e proporcionar uma melhor compreensão da questão em estudo e da sua relação com a teoria das trocas sociais e a teoria da reação psicológica. Além disso, o estudo forneceu alguns conhecimentos importantes sobre as práticas de OBA para ajudar os profissionais de marketing a criar campanhas de OBA bem-sucedidas e a ultrapassar as preocupações associadas à OBA. PALAVRAS-CHAVE: Publicidade comportamental online, Personalização, Intrusividade percebida, evitamento de OBA, Social Exchange Theory. vi ABSTRACT Technology enables businesses to better understand their targeted customers. Nowadays, advertisers can target their consumers based on actual knowledge of who, when, how, and why they are more likely to buy a specific product. Businesses use Online Behavioral Advertising (OBA) – i.e. monitoring consumers’ online behavior in order to present individually tailored ads to targeted consumers – to increase business current and future returns, however, OBA also leads to increased privacy concerns, distrust, and reactance to buying the product or using the service the business offers. Personalization is considered the main approach to OBA since it helps make ads more personally appealing. However, empirical research shows mixed results regarding the effects of personalization and how it can even cause negative consumer responses and attitudes due to increased privacy concerns. A conceptual framework encompassing the role of personalization, consumer characteristics, brand familiarity, brand trust, and transparency on consumers’ attitudes and intentions and protective behavior on brands is proposed and empirically tested. Moreover, the study explores the mediating influence of perceived intrusiveness, perceived privacy concern, and reactance on consumers’ attitudes and intentions and protective behavior, all based on the theoretical foundations of the social exchange theory. The findings show that some of the proposed constructs, namely personalization, consumer characteristics, brand familiarity, brand trust influence consumers’ attitudes towards the brand, attitudes towards OBA, intentions to click the ad, intentions to purchase the advertised product, and OBA avoidance behavior. Also, the findings supported that this influence is affected by several mediator elements namely consumers’ perceived intrusiveness, perceived privacy concerns, and perceived threat to choice. These findings have implications for theory and practice. For theory, the study deepens the understanding of the OBA concept and its effectiveness. The research aimed to contribute to the academic scholarship and give a better understanding of the issue under investigation and its relation to the Social Exchange Theory and Psychological Reactance Theory. Also, the study provides some important insights for OBA practices to help marketing professionals create successful OBA campaigns and overcome concerns associated with OBA. KEYWORDS: Online Behavioral Advertising, Personalization, Perceived Intrusiveness, OBA Avoidance, Social Exchange Theory. vii ACRONYMS ATB – Attitude Towards the Brand ATO – Attitude towards OBA AVE – Average Variance Extracted BF – Brand Familiarity BT – Brand Trust CB-SEM – Covariance-based Structural Equation Modeling CC – Consumer Characteristics CMB – Common Method Bias F2 – Effect Size ICA – Intention to Click the Ad IPP – Intention to Purchase the Product OBA – Online Behavioral Advertising PB – Protective Behavior PER – Personalization PIN – Perceived Intrusiveness PLS-SEM – Partial Least Squares Structural Equation Modeling PRC – Privacy Concerns PRT - Psychological Reactance Theory Q2 – Predictive Relevance R2 – Coefficient Determinants REA – Reactance SEM – Structural Equation Modeling SET – Social Exchange Theory SPSS – Statistical Package for Social Science TA – Transparency Approach VIF – Variance Inflation Factor 2 1.1. Introduction This chapter presents the theme under study, the concept of online behavioral advertising (OBA) and the variables identified in the literature as affecting OBA, such as personalization, consumer characteristics, brand familiarity, brand trust, transparency, perceived intrusiveness, privacy concerns, reactance, consumer attitudes and intentions and protective behavior. The chapter includes the rationale and scope of the study deriving from the gap identified in extant literature. The research background supports the emergence of the gap and the development of the main research questions and objectives. Finally, this chapter briefly addresses the expected contributions of the thesis. 1.2. Research Background In today’s digital world, developments in technology have enabled businesses to better understand their target customers. Businesses are now able to identify customers with purchase intentions, what they are likely to purchase, repurchase or even recommend to others (Kumar and Gupta, 2016). Advertisers monitor online consumers’ activities, collect their personal data, and use sophisticated analysis mechanisms to target customers with individually tailored ads in order to maximize their customers’ base and accordingly maximize business returns. Collected data includes websites visited, content viewed (e.g. articles read and videos watched), in addition to search engine queries and online purchases. This is referred to as online behavioral advertising – OBA (Boerman et. al, 2017). Moreover, unexpected changes in the world, like the Covid-19 pandemic, resulted in people staying at home and/or working remotely from home, which resulted in massive growth in online traffic. The consequence of this massive growth was more opportunities for online businesses to engage in online digital marketing activities. The Interactive Advertising Bureau reported a total revenue from digital advertising of $209.7 billion between the year 2021-2022, while it was reported at a total of only $40.1 billion in the first half of the year 2017 (IAB, 2022). Advertisers as well as third party businesses (e.g. Google and Facebook) use consumers’ online activity history to infer consumers’ interests and demographics to tailor individually targeted ads based on these inferences that appear in the form of search results, social feeds, and ads (Mayer and Mitchell, 2012). Altaweel et al. (2012) stated that the 100 most popular sites collect more than 6,000 cookies, 83% of which are third-party cookies, and more than 350 collected by individual websites. Roesner et al. (2012) affirmed that several online trackers capture more than 20% of each user’s browsing behavior. Furthermore, Englehardt et al. (2015) claimed that advertisers can recreate 62-73% of each online user 3 browsing history. In 2022, the International Association of Privacy Professionals released a report stating that nowadays around 80% of users’ browsing histories are captured by an average of 177 different online trackers only within the first two hours of browsing the Internet (IAPP, 2022). The practice of OBA involves the collection, use and sharing of personal data. However, since consumers have little or no knowledge of the extent to which their online behavior is tracked or their personal information is being collected (Ur et al. 2012), it is said to raise privacy concerns and feeling of reactance (Chen and Stallaert, 2014). Reactance stems from Psychological Reactance Theory and can be defined as a state where consumers reject something forced on them by behaving opposite to intended (Tucker, 2014). Personalization is considered a main base for OBA. Increased personalization increases tailored ad relevance which in return increases product purchases or service adoption (Tucker, 2014), but paradoxically it also increases privacy concerns about being watched and how information is being collected, which may lead to a feeling of vulnerability and reactance. Tucker (2014) argued that although personalized advertisements can be twice as effective as impersonalized ads, consumers tend to experience reactance due to increased perceived privacy concerns and perceptions of intrusiveness. This calls for the need for understanding of the impact of personalization, consumer characteristics and privacy concerns on consumers’ attitudes, intentions and behavior (Aguirre et al., 2015). Also, how advertisers collect consumers’ data – either overt or covert – has an effect on the way consumers perceive personalized ads and eventually on the way they behave towards these ads. The way marketers collect consumer related data online is referred to as transparency approach (Boerman et al., 2017). Transparency is believed to increase privacy concerns and feeling of intrusiveness, if consumers are presented by a personalized ad that contains identifiable personal information that has been collected without their consent (Milne et al., 2008). This calls for a better understanding of the effect of transparency on OBA (Boerman et al., 2017). Brand familiarity can be defined as consumers’ knowledge about a brand, developed through previous interactions with that brand, it is also considered as a mental shortcut used by consumers to make a purchasing decision where brand familiarity influences consumers’ trust (Phelan et al., 2016). Kumar and Gupta (2016) claimed that trust is a mental state that motivates consumers to accept a risky situation in trade for a positive outcome from that situation. This can be explained in the light of the Social Exchange Theory (SET). SET posits that concerned parties in a social relationship consider the cost of engaging into the relationship and try to maximize their gains and minimize their costs from the exchanges in that relationship (Emerson, 1976), therefore, online consumers prefer to engage in 4 transactions with people they trust due to the lack of face-to-face communication in the online environment (Chou and Hsu, 2016) Thus, although ad personalization causes favorable outcomes, it can also cause unfavorable consumer responses. A personalized ad with matching consumer preferences, can cause increased feelings of loss of choice and manipulation (Tucker, 2012). Also, since personalization is clear evidence that a user’s behavior was tracked and analyzed, it raises users’ privacy concerns and feelings of intrusiveness (Goldfarb and Tucker, 2011). It is therefore not surprising that a large majority of online consumers tend to avoid OBA and/or develop negative purchase intentions towards brands using OBA (Guild, 2013). Referring to what was previously mentioned, previous literature is not unanimous to whether a personalized message is always more effective than a standardized one (Noar et al., 2009). There also appears a dearth of studies addressing how different consumer characteristics affect consumers’ personal privacy concerns, feelings of intrusiveness and threats to choices (Boerman et al., 2017). All the previous interwind with how much the advertised brand is trusted, familiar to consumers’ memory and recall, the transparency approach used by the advertiser to collect consumers’ information and how these factors eventually affect consumers’ attitudes and behaviors towards the ad message. Hence existing literature lacks comprehensive studies on the effect of personalization, different consumer characteristics, brand-related factors and transparency. This gap highlights the need for further studies illuminating the usage of OBA and its effects. 1.3. Research Problem, Questions, Objectives and Contribution The research in hand aims to explore the role of some specific variables on consumers’ behavioral outcomes of OBA. The research problem can be summarized in the question of “What are the factors that influence consumers' attitudes, intentions, and behaviors in the online environment towards an online behaviorally targeted advertisement? And what factors can mediate such effect?”. Specifically, this research aims to explore the role of personalization, consumer characteristics, brand familiarity, brand trust, and transparency on attitudes toward OBA, attitudes toward the brand, intention to click on the ad, intention to purchase the advertised product and finally on OBA avoidance. Moreover, the study also explores the mediating influence of perceived intrusiveness, perceived privacy concerns, and reactance on consumers’ attitudes, intentions and behavioral outcomes. Thus, this research aims to answer the following questions: 5 1. How do personalization, consumer characteristics, brand familiarity, brand trust, and transparency influence consumers’ attitudes, intentions and OBA avoidance in the online environment? 2. How do perceived intrusiveness, privacy concerns, and reactance mediate such influence? The finding of this study is expected to advance the knowledge of OBA on two sides. First, it aims at contributing to the academic scholarship, to give a better understanding of the issue under investigation – OBA and other variables – and its relation to the Social Exchange theory and Psychological Reactance Theory. Our research adds to the body of knowledge on how these theories could be used to describe consumers’ attitudes and behaviors toward OBA. Also, this study deepens the understanding of the OBA concept and its effectiveness. Thus, this research provides useful information for future scholars to build on in their coming research. Second, this research has important practical implications for brands, marketing, and communication managers. This research provides some important insights into the online behavioral advertising practices. These insights are expected to help marketing professionals in creating successful OBA campaigns and overcome concerns and risks associated with online behavioral advertising practices. 1.4. Thesis Structure The overall structure of this dissertation takes the form of five chapters. After the present General Introduction chapter, Chapter two presents the Literature Review that supports the theoretical arguments of the thesis. This chapter addresses the primary constructs of the study, based on the previous theoretical background, the end of the chapter presented the research model and the research hypotheses. Chapter three introduces the research methodology part. After presenting the research design, the data collection methods of the empirical study (quantitative method) are discussed. Chapter four presents the results and a discussion about the main findings. Finally, chapter five discusses the research's main conclusions and points out the current study's limitations and thoughts that may assist in future research. The thesis outline is presented in the following figure (Figure 1). 6 Figure 1: Thesis Structure 7 LITERATURE REVIEW 8 2.1. Introduction The Literature Review chapter consists of three main sections. The first section discusses one of the primary concepts of the current study, Online Behavioral Advertising (OBA). It starts with exploring the general concept of online behavioral adverting, broadening the concept by addressing different OBA definitions, techniques and components as stated in past research. The second section introduces the theoretical framework of the study addressing the Social Exchange Theory and the Psychological Reactance Theory. The third section concentrates on the other primary concepts of the study, which were found to have a direct/indirect effect on OBA in previous literature. Definitions for the concepts of Personalization, Consumer Characteristics, Brand Familiarity, Brand Trust, Transparency, Perceived Intrusiveness, Privacy Concerns, Reactance, Consumer Attitudes and Intensions, and Protective Behaviors are presented as stated in previous literature and then the links between those concepts and the concept of OBA are discussed. Finally, the fourth section is devoted to developing and presenting the theoretical framework of the research, the literature review chapter then ends with the developed hypotheses of the current study. 2.2. The Concept of OBA Online behavioral advertising (OBA) can be simply defined as “the practice of tracking an individual’s online activities in order to deliver advertising tailored to the consumer’s interests” (Federal Trade Commission, 2009). Since OBA has become mandatory for many businesses to survive in the online market nowadays, most of the large portals (ex. Amazon), search sites (ex. Google AdSense), and social networking sites (ex. Facebook) have created their own OBA services (Kim and Huh, 2017). Moreover, almost all free web services rely on monetizing personal information through OBA (Mikians et al., 2012). Online behavioral advertising is often referred to as “Retargeted Advertising” because it has some unique characteristics that differ it from any other type of targeted ads (Bourial, 2015). Goldfarb (2014) stated that OBA refers only to advertising that is based on consumers past online behaviors and interactions with websites or search engines, while a regular targeted advertisement can be based on general consumer characteristics such as demographics (age, gender, education or employment status) or location. The availability of Web Analytics and Big Data mining techniques has dramatically enhanced the level of individualized consumer targeting, these techniques enabled advertising firms to predict what their 9 current and potential future customers are interested in buying based on their past online search behaviors (Kim and Huh, 2017). Advertisers, aggregators and ad-networks can collect consumer information over time while consumers’ browse the web and build consumers’ profiles based on their overall activity. These profiles can then be used to show each individual consumer advertising messages matching their interests more closely; instead of what they might be searching for, which eventually would lead to higher click-through rates and in turn, higher revenues for the advertiser (Carrascosa et al., 2015). While OBA has proved its importance as a technique to increase the effectiveness of online advertising, it was also claimed that OBA raised some consumers’ problems including but not limited to privacy concerns, perceived threat to choice and perceived intrusiveness (Carrascosa et al., 2015). Kim and Huh (2017) claimed that OBA can be seen as double-edged sword, with a possibility of having either favorable or unfavorable impacts on consumer responses (Bleier and Eisenbeiss 2015). In the next three sections, different OBA definitions from previous literature are discussed, the techniques of how OBA works, and the types of data collected are mentioned. 2.2.1. OBA Definitions There are many definitions of OBA, which can be also referred to as online profiling, behavioral targeting (Bennett, 2010) and retargeted advertising (Bourial, 2015). The work of Boerman et al. (2017) cited many definitions of OBA including the definition McDonald and Cranor (2010:2) that defined OBA as “ the practice of collecting data about an individual’s online activities for use in selecting which advertisement to display ”, Smit et al. (2014:15) definition of OBA as “ adjusting advertisements to previous online surfing behavior ”, and Ham and Nelson (2016:690) definition as “ a technology-driven advertising personalization method that enables advertisers to deliver highly relevant ad messages to individuals ”. Table (1) below sums up definitions from previous literature. Although OBA definitions are not consistent across previous literature, Boerman et al. (2017) noted that these definitions share two significant aspects: (1) Monitoring of consumers’ online behavior and (2) the use of the collected data to generate personally targeted ads. Thus, Boerman et al. (2017:364) introduced a new definition to OBA as: “The practice of monitoring people’s online behavior and using the collected information to show people individually targeted advertisements”. 10 Table 1: Online Behavioral Advertising Definitions. OBA Definition Author/Year The practice of tracking an individual’s online activities in order to deliver advertising tailored to the consumer’s interests. (Federal Trade Commission, 2009) The practice of collecting data about an individual’s online activities for use in selecting which advertisement to display. McDonald and Cranor (2010:2) Adjusting advertisements to previous online surfing behavior. Smit et al. (2014:15) A technology-driven advertising personalization method that enables advertisers to deliver highly relevant ad messages to individuals. Ham and Nelson (2016:690) The practice of monitoring people’s online behavior and using the collected information to show people individually targeted advertisements. Boerman et al. (2017:364) Source: Own Elaboration 2.2.2. OBA Techniques OBA involves tracking customers’ online activities and associating these activities with a specific computer or device (Berger, 2010). Firms who employ this type of advertising can use multiple approaches to collect consumers’ information, including cookie-based approach, spyware-based approach, deep packet inspection-based approach (Berger, 2010), first party model approach and browser fingerprinting approach (Lerner et al., 2016). A description of these approaches is provided in (table 2). Using the previously mentioned techniques, Englehardt et al. (2015) claimed that companies have the ability to reconstruct 62-73% of users’ browsing history. Although behavioral advertisers do not directly collect personally identifying data like a consumer’s name or physical address; Zhonghao et al. (2016) reported that 78% of sites contained trackers that attempted to transfer unsafe personal data, these data make it possible to identify the real consumer’s identity in the offline world (Barbaro et al., 2006). 11 Table 2:Online Behavioral Advertising Techniques and Approaches Technique Name How it works Cookie-based approach A cookie is a named piece of data that a visited website requests the user’s web browser to accept and keep saved. Then, each time the consumer uses the web browser to visit the cookie owner website, the consumer’s web browser sends the content of the cookie back to the owner website with the saved consumer online behavior information (Berger, 2010). Spyware-based approach Refers to willingly or unwillingly, installing a tracking software directly to the consumer’s computer to track and record the consumer data and behavior on the Internet. The installation can happen through the Internet Service Provider (ISP) or even a website the consumer visits (Berger, 2010). Deep Packet Inspection-based approach (DPI) Using this approach, the consumer’s ISP installs powerful hardware devices to examine all of the traffic going in and out of the consumers’ computers, which will be later used to compile information and sell this information to marketers to create additional revenues. DPI is practically impossible to stop as it does not require any software to be installed on the consumer’s computer, nor would an anti-spyware program be able to easily detect it (Berger, 2010). First Party Model approach Refers to targeting based on data collected at and by a single website. First Party targeting does not pose a threat to consumer privacy as it does not involve the sharing of data with third parties or across multiple websites (Berger, 2010). Browser Fingerprinting approach Refers to a stateless tracking technique that uses consumer’s device configuration information unprotected by the browser through JavaScript APIS and HTTP headers. Browser 18 2.8. Transparency Transparency refers to the way marketers collect consumer related data online, for the purpose of presenting consumers with personalized ads associated with their collected data. Transparency approach includes both overt and convert data collection techniques (Boerman et al., 2017). An overt data collection approach means informing consumers that their data is being collected. The fundamental assumption here is that after firms have informed consumers about their data collection, the continued use of the services grants a consent to further data collection by the firm in the future (Sundar and Marathe, 2010). While a covert data collection approach occurs when firms does not inform users about their data being collected, instead marketers discreetly collect and store consumer data while consumers surf the Web (Montgomery and Smith, 2009). Advocates of covert data collection techniques claim they are more beneficial to the firms, as they enable them to collect unbiased data, and gain a deeper understanding of their customers (Verhoef et al. 2010). On the other hand, it also benefits consumers, by not disturbing the flow of their online surfing experience (Milne et al., 2008). Although Internet users generally know that firms are collecting their information, being exposed to an ad message that contains recognizable personal information signals customers that their information has been collected, without their consent. This results in negative consumers' reactions and increased consumers' discomfort, leading to lower response rates to the ad message (Milne et al., 2008). This calls for a better understanding of the effect of transparency on OBA (Boerman et al., 2017). Firms seek to gain consumer trust through cues that show their transparency and reliability (Kim and Kim 2011). Alternatively, privacy laws also require companies to be transparent about their data collection practices (EU Data Protection Directive, 1995; EU General Data Protection Regulation, 2016). These efforts can be realized in the inclusion of privacy statements on a marketer's website. A privacy statement is a document that reveals which personal data was collected through the website, as well as how and why (Milne et al., 2008). However, privacy statements are not often read due to the complexity of their language and their length, customers often either agree or ignore them (McDonald and Cranor, 2010; Milne et al., 2006). Additionally, in order to improve transparency practices and increase consumers’ acceptance to OBA, the Digital Advertising Alliance (DAA, 2011) developed a set of icons to be shown on advertisements tailored by its members. These icons (e.g. the standard icon and an “asterisk man” icon) were developed to explain OBA mechanisms and to provide opt out options to consumers (Samat et al., 2017). However, the work of Leon et al. (2012) showed how participants misunderstood or rarely recognized disclosure 19 icons, 53% of their participants remembered OBA disclosure icons and only 12% remembered seeing a tagline “AdChoices”. Moreover, the tagline was incorrectly understood to intend selling advertising space instead of being a link to pages where they can make choices about OBA. Van Noort et al. (2013) suggested to complement the standard OBA icon with a label stating “ This ad is based on your surfing behavior ” in order to increase OBA awareness. 2.9. Perceived Intrusiveness Li et al. (2002:39) defined intrusiveness as “a psychological reaction to ads that interfere with a consumer’s ongoing cognitive processing” . Perceived intrusiveness is related to the perceived value of the ad message (Ying et al., 2009). This can be explained in the light of the Social Exchange theory, when consumers weigh the benefits of an ad message more than the threats it possesses, they perceive it as less intrusive (Rony, 2018). However, consumers’ perceptions of OBA appear to be mixed; while some see retargeted ads as beneficial (McDonald and Cranor 2010), others find them intrusive and creepy (Smit et al., 2014). McDonald and Cranor (2010) reported that some consumers actually change their online behavior when they know their data are being collected. Intrusiveness was addressed in previous literature in relation to OBA. Li et al. (2002) argued that ad messages with high level of personalization affects consumers’ cognitive abilities and can induce intrusive feelings towards the ad, Van Doorn and Hoekstra (2013) discussed the negative consumer attitudes associated with feeling of intrusiveness, generated by ads based on consumers past online behaviors. Zhao et al. (2017) claimed that ad content, brand familiarity, and brand love could lead consumers to experience the ad as less intrusive. 2.10. Privacy Concerns Privacy has a multidimensional meaning, that depends on the context in which it is defined and used. Westin (1967:7) defined privacy as “ the ability of the individuals to control the terms under which personal information is acquired and used ”. In marketing literature, a privacy concern is defined as “ the degree to which a consumer is worried about the potential invasion of the right to prevent the disclosure of personal information to others ” (Baek and Morimoto, 2012:63). Rognehaugh (1999:125) defined information privacy as “ users’ rights to keep information about themselves from being disclosed to others [marketers and other unknown people] ”. In the online environment, information privacy became a critical topic because of the collection, storing, and use of massive amounts of personal data (Okazaki et al., 2009). OBA tends to generate high levels 20 of privacy concerns, compared to other forms of online advertising. When a retargeted ad message contains extensive consumer related unique information, it tends to activate concerns over privacy, resulting in a negative consumer reaction towards OBA (Jai et al., 2013). When consumers become concerned about their personal data, they tend to remove their information (Sheehan and Hoy, 1999), or they even might provide incomplete and inaccurate data about themselves, leading to inefficient targeting and wasted advertising efforts (Awad and Krishnan, 2006). Empirical research addressed privacy concerns and provided evidence for its negative influence on consumer attitudes and intentions towards OBA. Baek and Morimoto (2012) stated how privacy concerns increased advertising skepticism and eventually ad avoidance. Van Doorn and Hoekstra (2013) showed a decrease in consumers purchase intentions as a result of high privacy concerns. Bleier and Eisebeiss (2015) discussed how consumers with high privacy concerns showed lower tolerance towards personalized ad messages. Aguirre et al. (2015) reported huge declines in click through rates after consumers realized their data was covertly collected to deliver personalized banners, while Smit et al. (2014) reported that consumers with less privacy concerns showed a more positive attitude toward a personalized ad message. Also, Phelan et al. (2016) claimed that some factors such as trust and brand love can influence consumers’ privacy concerns levels. In a recent study by Samat et al. (2017:303), the result showed privacy concerns as a main reason for avoiding OBA, the authors concluded a wide range of negative opinions towards targeted ads, and reported the second most common reason observed from a hypothetical scenario was “ because it feels like an invasion of my privacy. ” However, despite such empirical evidences about privacy concerns, consumers’ actual behavior is, in many cases, not consistent with or even opposite to their attitude. Contradictory to what might seem logical, Acquisti et al. (2013) found that people who claim high privacy concerns agree to disclose their personal information for small intensives such as free services or coupons in what is known as ‘privacy paradox’ (Boerman et al., 2017). 2.10.1. Privacy Paradox Prior research, such as the work of Dhar and Varshney (2011) and Sutanto et al. (2013), used the term ‘personalization-privacy paradox’ to describe the contradictory feelings consumers experience between their desire for personalization and their increased privacy concerns and need to protect their personal information. Smith et al. (2011:993) described the phenomenon as “ despite reported high privacy concerns, consumers still readily submit their personal information in a number of circumstances ”. The 21 concept of privacy paradox stems from the idea that individuals do nothing to protect their private information, though they are aware that the online environment possess a threat to their privacy (Norberg et al., 2007). The Privacy Calculus Model, is built upon a theory called “Calculus of Behavior”, developed by Laufer and Wolfe (1977). The theory suggests that consumers’ personal information disclosure behavior is influenced by their perceptions about the benefits they expect from that disclosure behavior. Therefore, consumers are only expected to share their personal information online, if they perceive the benefit of sharing such information - equal to or greater than - the risk of the disclosure behavior (Culnan and Bies, 2003). The Privacy Calculus Model is also rooted to the Social Exchange Theory (Schumann et al., 2014), where people evaluate social exchanges in terms of costs and benefits, and are only expected to join the exchange when benefits exceed the costs (Schumann et al., 2014). 2.11. Reactance The concept of reactance stems from the Psychological Reactance Theory (PRT) (Brehm, 1966; Brehm and Brehm, 1981). From a consumer research perspective, Reactance “ describes a strategy that consumers use to avoid complying with a persuasion attempt but does not directly characterize the cause of discomfort associated with a privacy invasion ” (Tucker 2012:2). Boerman et al. (2017:367) argued that “ Highly personalized ads lead people to perceive a loss of choice, control, or ownership, and thus cause negative feelings and responses ”. Thus, although consumers might find usefulness in a personalized ad message, the highly persuasive message might also trigger feelings of reactance, which leads to unfavorable attitudes and negative behavioral outcomes towards the ad message or the advertised brand (Bleier, & Eisenbeiss, 2015; Van Noort et al., 2014). Previous consumer research found that highly personalized ads lead consumers to feel loss of their perceived choice and increased feelings of being manipulated (Tucker, 2012). White et al. (2008), claimed that higher levels of personalization with no justification to why all these consumer information has been used, is considered way less effective and generates more reactance feelings, than when a justification is provided. Leon et al. (2012) stated that if users felt loss of control over their data through a targeted ad message, they will retaliate by avoiding the exact products shown to them in the targeted advertisement. Tucker (2014) found a link between introducing privacy controls and how they help to reduce reactance towards personalized advertisements. 22 2.12. Consumer Attitudes and Intentions Attitudes can be defined as general reactions – usually both cognitive and affective – towards a stimulus (Haugtvedt and Kasmer, 2008). Rony (2018) claimed that attitude is a person’s logical and emotional evaluation of something that is formed after exposure to different situations. Intentions, on the other hand; can be defined as “ the motivational factors that influence a behavior; they are indications of how hard people are willing to try, and how much of an effort they are planning to exert to perform the behavior ” (Ajzen, 1991:181). Attitudes can be either positive or negative, they are formed and changed regularly under certain circumstances (Haugtvedt and Kasmer, 2008). Past research recognized attitude-behavior consistency in many studies; thus, it can be said that attitudes can predict behaviors (Crano and Prislin, 2006). Ajzen (2008) claimed that intentions were also a direct cause of human behavioral outcome. Previous research shows varied consumers’ attitudes and intentions toward OBA, Lee et al. (2015), discussed how OBA triggers affective responses such as privacy concerns and reactance, which eventually affects the behavioral outcomes of targeted consumers (e.g., lower intentions to click the ad). The scholars furthermore, discussed how other factors such as personalization, trust, level of privacy concerns, and transparency moderates the effect of OBA on consumers’ behavior and purchase intentions. Other several studies also established how the level of personalization in OBA influences click-through intentions and click-through rates. Van Doorn and Hoekstra (2013), discussed how higher levels of personalization in an ad message, had led to an increased feelings of intrusiveness and thus negatively affected the purchase intentions. Tucker (2014) studied Facebook ads, and showed how a Facebook ad based on a person’s interest generated higher click-through rates, than those of an ad based on a person’s personal characteristics. Bleier and Eisenbeiss (2015) found that banner ads with high level of personalization (e.g. showing items from consumers’ shopping cart) increased click-through rates compared to ads with a lower level of personalization but only within the boundaries of a trusted retailer or brand. Furthermore, several studies established how transparency and consumer awareness of OBA guide consumers’ attitudes and intentions towards online behavioral ads. Aguirre et al. (2015) discussed how overt data collection mechanisms increased click-through rates compared to covert data collection mechanisms. The scholars also discussed how the inclusion of an OBA icon positively increased trust in the advertiser, and the advertised brand and eventually increased the advertising outcomes. 23 Consumers’ attitudes and understanding of OBA has as well been addressed in a number of studies, for example, multiple studies discussed how people with lower privacy concerns levels tend to be more accepting towards OBA (Miyazaki, 2008; Baek and Morimoto, 2012; Smit et al., 2014). Lambrecht and Tucker (2013) showed how OBA generated better behavioral outcomes and purchase intentions when consumer had narrowly construed preferences. Leon et al. (2013) argued that consumers only want to share certain amount of personal information but are uncomfortable sharing the rest. Similarly, Yao et al. (2017) showed that consumers cared about the types of personal information collected more than they were concerned with who was collecting it. Plane et al. (2017) found users were more concerned if an ad was targeted based on demographic information, such as age, gender, or race, than based on interests. 2.13. Protective Behavior Smit et al. (2014) stated that protective behavior depends mainly on varying consumer characteristics where consumers with more privacy concerns practice increased protective behavior to protect their online privacy. However, the work of Leon et al. (2012) shows that people have problems protecting their online privacy as they do not understand or know how to use the available protective tools (e.g. tools that block access to advertising websites) to opt out OBA. Only few consumers delete cookies and search history in an attempt to protect their online information (McDonald and Cranor 2010). Empirical research also mentioned other factors that can explain OBA acceptance and avoidance. Baek and Morimoto (2012) discussed how privacy concerns and ad irritation leads to avoidance of OBA due to increased ad skepticism. The scholars also discussed how perceived personalization can lead to less ad avoidance. Jai et al. (2013) explained how transparency about consumers’ data collection techniques and reasons influences OBA acceptance and avoidance, where informing consumers their data was being shared with third-party websites led to lower repurchase intentions compared to sharing their data internally within the same website. In addition, Schumann et al. (2014) findings showed that consumer believed that receiving a free web service is an acceptable trade-off to their collected information. 2.13. Hypothesis Development In the past two decades, the effect of personalization has been widely examined in many prior studies across different domains such as marketing, mass communication, information technology, and public health (Li and Kalyanaraman, 2013). However, the effectiveness of personalized messages compared to standardized ones remains ambiguous and unclear in the existing literature (Noar et al., 2009). The 24 effect of personalization was different and sometimes even opposite in prior research, where some research confirmed the effectiveness of personalization (e.g., Beam and Kosicki, 2014; Ha and Janda, 2014), while other studies showed there was no significant difference between personalization and standardization (e.g., Li, 2016; Porter and Whitcomb, 2003). Personalization has been treated as a double-edged sword. Researchers, such as, Li (2016) advised researchers to further investigate why the effect of personalization looked so unpredictable, random and unstable. This calls for the first hypothesis of the study: ➢ H1: Personalization has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Some previous studies showed that consumer characteristics had an effect on perceptions, attitudes and behavior towards OBA, researchers pointed what is called “chilling effect”, where online users change their online behavior when they know their data is being collected (McDonald and Cranor, 2010). However, the available studies are scarce, and most previous studies did not study the condition of implementing appropriate online personalization strategies, altered based on consumer characteristics (Smith et al., 2011). This calls for the second hypothesis of the study: ➢ H2: Considering consumer characteristics has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Previous literature suggests that brand familiarity affects consumer attitudes and behaviors, it also acts as a shortcut to online purchase decision making for online consumers. Familiarity also has an effect online consumers’ trust in the advertiser (Shoenberger and Thorson, 2014). This supports the third hypothesis of the study: ➢ H3: Brand Familiarity has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Previous research pointed out how understanding the dynamics of consumer trust was crucial in the online environment, due to the lack of face-to-face transactions, thus suggesting that online consumers 25 are more likely to engage with firms or brands they trust (Rony, 2018). In fact, previous research on trust mostly focused on institutional trust, this trust tends to be created based on the reputation of the company or the brand (Kehr et al., 2015). Kumar and Pansari (2016) claimed that when consumers trust the advertising firm or the brand, they will be more willing to share their private information, which then can be used for creating an effective online advertising campaign, that would eventually generate better revenues for the firm. This supports the fourth hypothesis of the study: ➢ H4: Brand trust has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Previous literature showed that online consumers appreciated honesty, and wanted to be informed about the collection, usage, and sharing of their personal data (Turow et al. 2009). Data collection approaches, in addition to, implementing transparency mechanisms were found to affect feelings of vulnerability, which in turn affects consumers’ attitudes and behaviors towards OBA (Aguirre et al. 2015). Accordingly, when advertisers include transparency mechanisms into an ad message, consumers may experience more trust or become more willing to accept vulnerability, so that they would accept to click a highly personalized advertisement despite their privacy concerns and feelings of perceived intrusiveness (Boerman et al., 2017). This supports the fifth hypothesis of the study: ➢ H5: Being transparent about data collection mechanisms has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Previous research highlights that the level of personalization of the online ad causes favorable as well as unfavorable consumer attitudes and behaviors towards the ad (Tucker 2012b). Tucker (2014) claims personalization causes feeling of intrusiveness to increase, leading to negative attitudes and behaviors from consumers in response to the personalized advertisement. Another study by Hühn et al. (2017), supported the notion that those feelings of intrusiveness generate ad annoyance and avoidance behavior among users in the online environment and mobile media. This supports the sixth hypothesis of the study: ➢ H6: Perceived intrusiveness mediates the influence of i) personalization, ii) consumer characteristics, iii) brand familiarity iv) brand trust and v) transparency approach on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ 26 intention to on click the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Over the past decade, literature has shown how consumers have become more concerned about their online privacy, especially towards OBA practices. Research highlighted how privacy concerns play an important role in consumer acceptance of OBA, where the level of personalization affects consumerrelated factors such as privacy concerns which accordingly affect attitudes and behaviors (Bleier and Eisenbeiss, 2015). Also, previous research findings suggest that since OBA is based on tracking consumers’ online behavior and collecting consumers’ information, therefore, the negative impact of privacy concerns would be even higher in the context of OBA leading to negative attitudes and behavioral responses to the retargeted advertisement. This calls for the seventh hypothesis of the study: ➢ H7: Privacy concerns mediate the influence of i) personalization, ii) consumer characteristics, iii) brand familiarity iv) brand trust and v) transparency approach on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Previous literature discussed how highly personalized ads can generate feelings of loss of choice and control among consumers, and eventually cause negative attitudes and responses to the online retargeted advertisement. Thus, it can be expected that the level of online ad personalization influences consumer-related factors, such as feelings of reactance (Bleier and Eisenbeiss, 2015), which in turn influences OBA outcomes, such as purchase intentions and click-through rates (Aguirre et al., 2015). Therefore, more research is needed to disclose the role reactance plays in influencing consumers’ attitudes, intentions and behaviors towards OBA context. This calls for the eighth hypothesis of the study: ➢ H8: Reactance mediates the influence of i) personalization, ii) consumer characteristics, iii) brand familiarity iv) brand trust and v) transparency approach on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. 27 2.14. Research Model Boerman et al. (2017) claimed that OBA research lacks a clear understanding of the concept due to the interdisciplinary nature of the field and the number of interested parties including advertisers, consumers, computer scientists, and policymakers. However, OBA research has studied various independent, mediating, moderating, and outcome variables. Personalization proved to have a great impact on consumers’ responses to OBA. However, it is not well understood up to this date what levels of personalization consumers find acceptable and what they consider creepy and too invasive (Boerman et al. 2017). Moreover, prior research recommends it is important to know who is being targeted, as there seem to be individual differences in responses to OBA. Therefore, the purpose of this study is to explore the role of personalization, consumer characteristics, brand familiarity, brand trust, and transparency on consumers’ attitudes towards the brand, consumers’ attitudes towards OBA, consumers’ intention to click on the Ad, consumers’ intentions to purchase the product and consumers’ protective behavior. Moreover, the study explores the mediating influence of perceived intrusiveness, perceived privacy concerns, and reactance on consumers’ attitudes and intentions and protective behavior. 34 3.6. Research Approach Research approaches are classified into two main approaches; the deductive approach where the researcher develops a theory, formulates hypotheses and then designs a conceptual framework to test the hypotheses using data, and the inductive approach , in which data is collected and analyzed to develop theory. In the deductive approach previous literature is used to identify theories and to build the theoretical framework that will be tested using data (Saunders et al., 2009). Sekaran and Bougie (2016) claimed that deductive reasoning works from the more general to the more specific, where researchers start with a general theory that is then narrowed down into specific hypotheses that can be tested, these hypotheses are then further narrowed down when specific data is collected and analyzed. Saunders et al. (2009) stated that deduction is the dominant research approach in natural sciences, and are more often used in causal and quantitative studies where deduction owes more to positivism research philosophy. An important characteristic of deductive research is generalization. It is essential to select samples of adequate numerical size in order to be able to form a statistical generalization about regularities in human social behavior (Saunders et al., 2009). The research in hand follows a deductive approach, where previous literature is used to generate hypotheses based on the theoretical foundations of the Social Exchange Theory and the Psychological Reactance Theory. Hypotheses are then tested using quantitative research method. The next section, section (3.7) entails the research methodology chosen in this study, the data collection technique and the instruments used to collect the data. 3.7. Research Methodology The term methods refer to techniques and procedures used to obtain and analyze data. This includes research instruments such as questionnaires, observation and interviews as well as both statistical quantitative and nonstatistical qualitative analysis techniques. Whereas the term methodology refers to the theory of how research should be undertaken (Saunders et al., 2009). The next sections (3.7.1, 3.7.2, and 3.7.3) discusses the research method choice, data collection technique, the research instruments and the sampling design. 35 3.7.1. Research Method Choice (Quantitative Method) The study adopts a mono-method choice, which is the quantitative method. Mono methods can be defined as the use of a single data collection technique and corresponding analysis procedures as a way to answer the research questions. Quantitative methods are used to define any data collection technique (such as a questionnaire) or data analysis procedure (such as statistics) that uses or produces numerical data (Saunders et al., 2009). This choice is widely used within business and management research (Curran and Blackburn 2001). Cronbach and Webb (1975) argued that quantitative methods use correlational designs to reduce error, bias, and other noise that keeps one from clearly understanding social facts. Saunders et al. (2009) claimed that using a quantitative method such as a questionnaire to collect data and test hypotheses provide a more efficient data collection technique from a large sample because each respondent will respond to the same question. The main data collection instrument used in the study is a survey using a questionnaire. The data is considered primary data collected by the questionnaire. Previous literature was thoroughly investigated to identify measurement scales for the questions included in the questionnaire. The following section (3.7.2) presents the data collection method of this study, followed by a description of the research instrument used to collect the data. 3.7.2. Data Collection Technique (Experiment + Survey) An experimental study followed by a survey was conducted to test the hypotheses and answer the research questions stated in the previous section (3.2). The key approach of positivist researchers is the experiment, and it is usually associated with a hypothetico‐deductive approach to research, where the researcher manipulates the independent variable to study the effect of this manipulation on the dependent variable (Sekaran and Bougie, 2016). Experiments therefore are used to answer ‘how’ and ‘why’ questions (Saunders et al., 2009). Experimental research is appropriate for this study because “ experiments are well-suited to studying causal relationships ” (Shadish et al., 2002:7). 36 Shadish et al. (2002:6) further argued that by using experiments: (1) We manipulate the presumed cause and observe an outcome afterward; (2) we see whether variation in the cause is related to variation in the effect; and (3) we use various methods during the experiment to reduce the plausibility of other explanations for the effect along with ancillary methods to explore the plausibility of those we cannot rule out. The study utilized scenario-based online experiments. Scenarios were chosen because similar previous studies showed that scenario-based experiments are common and appropriate in this area of research (e.g., Aguirre et al., 2015; Bleier & Eisenbeiss, 2015; Shoenberger & Thorson, 2014; White et al., 2008). Appendix (I) includes eight different versions of the experiment scenarios designed for the study. To design the experiment, eight scenarios featuring different combinations of familiarization, personalization, and transparency were created where respondents were asked to imagine that they are interested in weight-loss programs and they searched online about the topic. Following that, the next morning while reading the news on an online news website, they saw an ad about weight-loss medication. Appendix (I) includes the eight versions of the ads that were designed to fit each condition (Familiar – Non-Personalized – Non-Transparent, Familiar – Non-Personalized – Transparent, Familiar – Personalized – Non-Transparent, Familiar – Personalized – Transparent, Unfamiliar – Non-Personalized – Non-Transparent, Unfamiliar – Non-Personalized – Transparent, Unfamiliar – Personalized – NonTransparent, Unfamiliar – Personalized – Transparent). The research then adopts the survey strategy as a method to collect primary data. The survey strategy is usually associated with the deductive approach, it allows the collection of a large amount of data from a large population in a highly cost-effective way (Saunders et al., 2009). Surveys are often conducted by using a questionnaire that a respondent completes on his or her own, either on paper or via the computer (Sekaran and Bougie, 2016). The survey strategy enables the collection of quantitative data that would be analyzed quantitatively using descriptive and inferential statistics (Saunders et al., 2009). Appendix (II) includes the English language version of the survey instrument. The following subsection (3.7.3) introduces the instrument for data collection (online experiment and questionnaire), followed by the description of the experiment scenarios, questionnaire structure, and content. 3.7.3. Experimental design As mentioned before in section (3.7.2), the study used a scenario-based online experiment followed by a questionnaire to collect primary data. The experiment was replicated 8 times using a real brand (Lipo 37 6) and a fictitious brand (Miracle Cure), under different personalization and transparency settings. Weight loss was selected as the experiment topic as it is considered a sensitive yet important topic to consumers. A real familiar weight loss medication (LIPO 6) was selected for the real brand while a fictitious unfamiliar brand (Miracle Cure) was introduced to respondents. Experiment manipulations are included in Appendix (I). Replication helps to generalize research findings and experimental research should be replicated for external validity (Lynch, 1982). Also, since brand familiarity is one of the variables that need to be tested, using a real brand and a fictitious brand was a must to consider the differences in participants’ brand familiarity and brand trust responses. The research took place ONLINE using the GORILLA SC experiment builder. The experiment was then followed by a questionnaire on the same platform to gather experiment data. A simple link that directs respondents to the online experiment website was populated through email, social media sites (Facebook and Instagram), and friends. After participants clicked the experiment link, they were re-directed to the Gorilla SC website. Participants were asked to agree on a consent to participate in an online experiment, afterwards, they were randomly assigned one of the eight experiment scenarios. For those who were randomly assigned to one of the personalized scenarios; participants were asked to input their first names, these first names were used in creating the advertising message. After being exposed to the online scenario, participants are then re-directed to the next page to answer the online questionnaire which measured their view of ad personalization, familiarity with the advertised brand, trust in the advertised brand, privacy concerns, reactance, perceived intrusiveness, attitudes towards brand, attitudes towards OBA, intention to click the Ad, intentions to purchase the product and protective behavior. Participants’ demographic information was also collected in the questionnaire, including age, gender, income level, educational background, and employment status. 3.8. Questionnaire Development A questionnaire is a written set of questions to which respondents record their answers in a predetermined order, usually within rather closely defined alternatives (De Vaus, 2002). Using online questionnaires yields many advantages that include, gaining a deep understanding of the subject under investigation, wide geographical coverage and reach, and participants increased convenience. Also, the distribution of online questionnaires is fast and easy (Sekaran and Bougie, 2016). 38 The questionnaire was applied in English since the respondents are proficient in this language. Bryman (2003) claimed that the questionnaire design is more crucial than the questionnaire length in receiving a high response rate. Given that, the questions had to be understandable, with no use of ambiguous language also to limit the possibility of misunderstanding the questions, several questions were asked for each scale. Respondents’ anonymity and confidentiality was assured by the researcher through the phrase ‘ The information collected via the questionnaire is ANONYMOUS and used for academic purposes ONLY. ’, so that they won’t feel their privacy invaded. The researcher designed the questionnaire to be filled-in in about fifteen minutes maximum. The questionnaire encompasses several questions, the upcoming sections include further detail about these questions. The following sections discuss the questionnaire different section, questions and scales used to measure the variables. 3.8.1. Demographic Questions Demographic questions were used to gather respondents’ personal data. Demographic data gathered by the questionnaire include age, gender, income level, educational level and employment status. The only case when respondents were asked to enter their first name, it was only for ad personalization reasons but was not saved or recorded. Demographic questions appear at the end of this study questionnaire, yet researchers can add demographic questions at the beginning of the questionnaire. 3.8.2. Identification of OBA Ads Since the notion – OBA – is not very common among Internet users, the following definition of OBA was presented to ensure respondents understood the topic in hand before answering OBA related questions: ‘Online behavioral advertising (OBA) can be defined as: The practice of tracking individuals' online activities in order to deliver advertising tailored to the consumers' interests.’ (Federal Trade Commission, 2009) 3.8.3. The Questionnaire Main Sections The questionnaire included several sections, answers to each of these sections were mandatory. Respondents could not move to the next section without finishing the questions of the previous one, to guarantee the completion of the questionnaire and assure appropriate data for the analysis. As mentioned 39 before, Appendix (II) contains an English language version of the questionnaire. The next sections are classified into tables with variables name, the scale used to measure each variable, origin and description of the scale and the items in each scale. Personalization The questionnaire starts right after the respondent was subjected to the scenario of the experiment. The first question the respondent was asked about is the level of personalization of the ad. Respondents indicate on a single item seven-point scale whether they agree or disagree with the question ‘ This ad is directed to me personally ’ adapted from De Keyzer, Dens, and De Pelsmacker (2015). Another aspect to personalization - which is how consumers view the importance of the amount of personalized information an ad contains – is measured using a five-item seven-point Likert type scale (1Strongly Disagree, 7-Strongly Agree) as shown the table below. Table 3: Personalization (Informativeness) Scale Items Variable Source: Ozcelik, and Varnali (2019) Personalization (Informativeness) 1. Customized online ad gives me quick and easy access to large volumes of information. 2. Information obtained from the customized online ad is useful. 3. I learned a lot from customized online ads. 4. I think information obtained from customized online ads is helpful. 5. Customized online ad makes acquiring information inexpensive. 40 Brand Familiarity Next, respondents use a single bipolar scale to indicate whether they were familiar or unfamiliar, with the brand. This answers the question ‘ Regarding the brand, I am: ’ Table 4: Brand Familiarity Scale Items Variable Source: Baker, Hutchinson, Moore and Nedungai (1986) Brand familiarity A unidimensional construct that is directly related to the amount of time that has been spent processing information about the brand, regardless of the type or content of the processing that was involved. The respondents indicated on a single bipolar scale whether they were familiar or unfamiliar. Consumer Attitudes Consumer attitudes can be classified into Consumer Attitudes towards OBA and Consumer Attitudes towards the brand . To measure these constructs, scales sourced from the literature were used as shown in the tables below. As a measurement scale, seven-point scale were used. Table 5: Consumer Attitudes towards OBA Scale Items Variable Source: MacKenzie and Lutz (1983) Attitude Toward OBA (General and Specific) 1. 1= bad vs. 7 = good, 2. 1 = unpleasant vs. 7 = pleasant, 3. 1 = unfavorable vs. 7 = favorable. 41 Table 6: Consumer Attitudes towards the Brand Scale Items Variable Source: Mitchell and Olson (1981) Attitude Toward the Brand 1. 1= bad vs. 7 = good, 2. 1 = unpleasant vs. 7 = pleasant, 3. 1 = dislike very much vs. 7 = like very much, 4. 1 = poor quality vs. 7 = high quality. Consumer Intentions Consumer intentions include Intentions to Click on the Ad and Intentions to Purchase the Advertised Product . Three-item seven-point scale used to measure intention toward the OBA. Table 7: Consumer Intentions Scale Items Variable Source: MacKenzie, Lutz, and Belch (1983). Intention to Click on the Ad Intention to Purchase the Product . 1. 1= unlikely vs. 7 = likely, 2. 1 = improbable vs. 7 = probable, 3. 1 = impossible vs. 7 = possible. Brand Trust Brand trust was measured with a four-item seven-point Likert type scale (1-Strongly Disagree, 7-Strongly Agree). The scales sourced from the literature were used as shown in the table below. Table 8: Brand Trust Scale Items Variable Source: Chaudhuri and Holbrook (2001). Brand Trust 1. I trust XYZ 2. I rely on XYZ. 42 3. XYZ is an honest brand. 4. The XYZ brand is safe. Perceived Intrusiveness Perceived intrusiveness was measured using Li, Edwards and Lee (2002) Seven-item, seven-point Likert type scale (1-Strongly Disagree, 7-Strongly Agree) as shown in the table below. Table 9: Perceived Intrusiveness Scale Items Variable Source: Li, Edwards, and Lee (2002). Perceived Intrusiveness 1. Distracting. 2. Disturbing. 3. Forced. 4. Interfering. 5. Intrusive. 6. Invasive. 7. Obtrusive. Privacy Concerns Privacy concerns were measured using Baek and Morimoto (2012) scale of six items, seven-point Likert type scale (1-Strongly Disagree, 7-Strongly Agree) as shown the table below. Table 10: Privacy Concerns Scale Items Variable Source: Baek and Morimoto (2012) Privacy Concerns 1. I feel uncomfortable when information is shared without permission. 43 2. I am concerned about misuse of personal information. 3. It bothers me to receive too much advertising material of no interest. 4. I feel fear that information may not be safe while stored. 5. I believe that personal information is often misused. 6. I think companies share information without permission. Reactance Reactance was captured using Gardner and Leshner (2016) scale of perceived threat to choice. The scale was adapted because it contained very easy to understand straightforward questions. The scale questions are shown in the table below. Table 11: Reactance Scale Items Variable Source: Gardner and Leshner (2016) Reactance (Perceived Threat to Choice) 1. The message threatened my freedom to choose. 2. The message tried to make a decision for me. 3. The message tried to manipulate me. 4. The message tried to persuade me. 50 model is reflective. Indicators are considered reflective if they establish a representative set of all possible items within the construct (Diamantopoulos and Winklhofer, 2001). Reliability and validity are the common tests applied to assess the propriety of the measurement model (outer model). (Hair et al., 2019; Hair Jr et al., 2017). If the measurement model meets all the required criteria, then the researcher needs to evaluate the structural model which will be elaborated in section (3.13.7) (Hair et al., 2019). 3.12.1. Internal Consistency Reliability Reliability tests are used to identify the extent to which an instrument is consistent and stable in measuring a concept. The first part of a reliability assessment is the internal consistency of the measuring instrument, which is a measure of the homogeneity of the items of a certain construct. While testing a reflective model, all the measurement items should be highly correlated and interchangeable to make sure they measure the construct from all aspects (Sekaran & Bougie, 2016). Cronbach’s alpha is the most popular test for internal consistency. At a value of at least 0.7 the construct is considered reliable (Nunnally J, 1978). The second part of a reliability assessment is a composite reliability test, which takes into consideration the distinct outer loadings of the indicators and thus it provides relatively higher reliability estimates. A composite reliability value ranging from 0.6 to 0.9 is considered acceptable reliability (Diamantopoulos et al., 2012). The true value of reliability lies between the lower bound represented by Cronbach’s alpha and the upper bound represented by the composite reliability (Hair Jr et al., 2017; Sarstedt et al., 2021). The current study will apply both tests to assess the reliability of the measuring instrument. 3.12.2. Indicator Reliability Indicator reliability is the second step in ensuring the reliability of the reflective model. It examines how much variance of each indicator is explained by its construct. Indicator reliability is measured by computing the square of the indicator loading, which is the bivariate correlation between the indicator and its construct, since reflective indicators relate to the construct through loadings. Indicator loadings of 0.7 and above are the recommended threshold, as this value indicates that the construct explains more than 50% of the indicator’s variance, hence, achieving acceptable indicator reliability (Sarstedt et al., 2021). However, in social science research, weaker indicator loadings (<0.7) are often obtained, especially in the case of using newly developed scales (Hulland, 1999). Indicators with very low loadings (<0.4) should 51 be removed from the measurement model. While, loadings lying between 0.4 and 0.7, are not removed unless their deletion will improve the value of internal consistency reliability or convergent validity and will not affect the content validity. Otherwise, these indicators are retained (Hair et al., 2022). 3.12.3. Validity Validity assesses the certainty of the measuring instrument; it ensures how accurately an item measures a particular variable. Construct validity examines how well the measuring instrument fits the theoretical foundation of the construct (Sekaran & Bougie, 2016). Convergent and discriminant validity are applied to test construct validity. Convergent Validity indicates the extent to which a construct comes together to explain the variance of its indicators, and is evaluated by assessing the correlation of the construct with an alternative measure of the same concept. The average variance (AVE) was used to evaluate convergent validity. The minimum accepted value for AVE to achieve convergent validity is 0.50, and above (Hair et al., 2022). Discriminant validity is a test applied to ensure that a construct is truly distinct from other constructs in the model empirically. Discriminant validity was assessed using the Fornell-Larcker criterion which compares the square root of each construct’s AVE to its correlation with other constructs. For discriminant validity to be achieved, the square root value of the construct’s AVE should be greater than its correlations with all other constructs (Hair Jr et al., 2017; Fornell-Larcker, 1981) 3.13. Structural Model Assessment Structural model assessment is the next step in evaluating PLS-SEM results after the measurement model has been assessed and satisfactory results were achieved. The standard assessment procedures to be considered when analyzing the structural model are (a) assessing the structural model for collinearity issues, (b) assessing the significance and relevance of the model relationships, (c) assessing the level of coefficient determination ( R2 ), (d) assessing the effect size ( f 2 ) and (e) assessing the predictive relevance ( Q2 ) (Benitez et al., 2020; Hair et al., 2011, 2019; Hair Jr et al., 2017). 3.13.1. Collinearity Assessment Collinearity refers to the existence of high correlation between two independent variables in regression analysis. It is assessed by calculating the Variance Inflation Factor (VIF) values for each set of predictor constructs in every sub-part of the model. Determining the multicollinearity problems is considered the 52 first step in evaluating the structural model in PLS-SEM (Hair et al., 2020). VIF values greater than 5 indicate critical levels of collinearity. In such cases, collinearity must be treated either by removing certain constructs or merging several predictors into one construct. Low levels of collinearity must be achieved to ensure that every independent variable possesses a unique percentage of the explained variance of the dependent variable (Hair Jr et al., 2010, 2017). 3.13.2. Structural Model Path Coefficients Examining the size and significance of the path coefficients is the next step after solving the critical collinearity problems. Path coefficients are the estimates obtained to verify the hypothesized relationships between the variables in the structural model. The path coefficients values range from -1 to +1 and rarely be -1 or +1; values closer to +1 indicate a strong positive relationship, while those closer to -1 indicate a strong negative relationship. The more the value gets closer to 0, the weaker the relationship is. However, the true significance of the path coefficient is determined according to its standard error. A path coefficient standard error is obtained by applying the bootstrapping technique that computes the t values and p values that truly indicate the significance of the relationships (Hair Jr et al., 2017). Bootstrapping is a resampling technique that estimates the significance of the hypothesized relationships by drawing a sample from the original sample with replacements. The number of observations of each bootstrap sample is equal to the number of observations of the original sample. Prior studies recommended the use of 5,000 bootstrap samples (Henseler et al., 2009; Hair Jr et al., 2017). The t values computed from the bootstrapping represent the significance of the path coefficients, while the p is the probability of erroneously rejecting a true null hypothesis (probability of error). The predetermined critical t values for a one-tailed test are 1.28, 1.65, 2.33 for significance levels 10%, 5%, and 1% respectively. For two-tailed tests, critical t values are 1.65, 1.96, 2.57 for significance levels 10%, 5%, and 1% respectively. Therefore, when the computed t value exceeds these critical values, according to the chosen significance level (the most common is 5%), the hypothesized relationship is validated to be statistically significant. The choice of the significance level and type of test (one or two tails) depends on the field of study and the study’s objective (Hair Jr et al., 2017). 3.13.3. Coefficient of Determination ( R2 Value) The most used metric to assess structural model prediction is the coefficient of determination ( R2 value), which is calculated as the square correlation between the dependent variable’s actual and predicted 53 values. This coefficient measures the accumulated effects of independent variables on the dependent variable. R2 value evaluates the predictive accuracy of the structural model as it determines the amount of the dependent variable’s variance explained by all independent variables (Hair et al., 2019). R2 is measured in a range from 0 to 1, a higher value of R2 indicates higher predictive accuracy denoting that the tested independent variables are indeed causing most of the variance of the dependent variable (Hair Jr et al., 2017). R 2 values of 0.75, 0.50, and 0.25 are considered substantial, moderate, and weak, respectively. However, these values are recommended for marketing research (Hair Jr et al., 2010). In other disciplines, an R 2 value starting from 0.10 is considered acceptable (Falk and Miller, 1992; Raithel et al., 2012). According to Chin (1998), R 2 values 0.67, 0.33 and 0.19 are considered substantial, moderate, and weak respectively (Purwanto and Sudargini, 2021). Moreover, Cohen (1988) suggested R 2 values for endogenous latent variables as follows: 0.26 (substantial), 0.13 (moderate), 0.02 (weak). Furthermore, too high values of R 2 might indicate a model over fit (Sarstedt et al., 2021). 3.13.4. Effect Size ( f 2 ) The effect size ( f 2 ) is considered a second measurement of the significance of the hypothesized relationship in addition to the statistical significance provided by the R2 . f 2 estimates the predictive ability of each construct in the model, it measures the change in the R2 value when a specific independent variable is removed from the structural model, thereby assess whether the removed variable had a substantiative effect on the dependent variable (Hair Jr et al., 2017). An effect is considered small, medium, or large at f 2 values of 0.02, 0.15, and 0.35, respectively (Cohen, 1998). In most cases the f 2 is similar to the size of the path coefficients (Sarstedt et al., 2021). 3.13.5. Predictive Relevance ( Q2 ) Another way to assess the model predictive accuracy is by calculating its predictive relevance ( Q2 value). The predictive relevance ( Q2 ) measures if the path model has an adequate predictive power on the “out of sample” population. Meaning, the model can predict data not used in the estimation (Hair Jr et al., 2017). This approach is based on the blindfolding procedure that eliminates single points in the data matrix, assigns the removed points with the mean and estimates the model parameters (Rigdon, 2014). Q2 value is the difference between the original and the predicted data points. The procedure is repeated to replace each data point and re-estimate the model repeats to omit each data point and re-estimate the model. 54 Q2 values greater than zero for a specific endogenous variable indicates that the path model has a predictive relevance for that dependent variable, while less than zero indicates a lack of predictive relevance, moreover, values within 0.25 and 0.50 indicates medium and large predictive relevance of PLS-SEM model (Hair Jr et al., 2017). 3.14. Mediation Analysis Mediation happens when a variable - referred to as a mediator variable - intervenes between two other related variables. In particular, the change in the exogenous construct (independent variable) causes a change in the mediator variable, in turn, results in a change in the endogenous construct (dependent variable) in the PLS path model (Hair Jr et al., 2017). MacKinnon et al. (2007), claimed that mediation analysis assumes that a sequence of relationships in which an independent variable affects the mediating variable, which then affects the dependent variable. Mediating effects can be either direct or indirect. Direct effects are the relationships that links two constructs with a single arrow, while indirect effects are a sequence of two or more relationships represented by multiple arrows, they include a series of relationships with at least one intervening construct involved in the relationship (Sarstedt et al., 2020). Mediation analysis reveals whether the mediation has a direct effect or indirect effect and tests the significance level of the mediator(s) (Hair Jr et al., 2017). Hair Jr et al. (2017) describe two types of non-mediation: 1. Direct-only non-mediation, where the direct effect is significant but not the indirect effect. 2. No-effect non-mediation, where neither the direct nor indirect effect are significant. In addition, they identify three types of mediation: 1. Complementary mediation, where the indirect effect and the direct effect both are significant and point in the same direction. 2. Competitive mediation, where the indirect effect and the direct effect both are significant and point in opposite directions. And, 3. Indirect-only mediation, where the indirect effect is significant but not the direct effect. The current study includes three mediator variables, between exogenous and endogenous constructs; this type of analysis is referred to as multiple mediation analysis (Sarstedt et al., 2020). 3.14.1. Testing Mediating Effects Many prior studies relied on the 1982 Sobel test, to test the significance of mediating effects. Sobel test compares the direct relationship between the independent variable and the dependent variable with the indirect relationship between the independent variable and dependent variable that includes the 55 mediation construct (Helm et al., 2010). However, Sobel test assumes a normal data distribution which is not consistent with the nonparametric PLS-SEM method. Also, the parametric assumptions of the Sobel test usually do not regard the indirect effect between constructs. Besides, Sobel test requires unstandardized path coefficients as input for the test statistic and lacks statistical power, especially when applied to small sample sizes (Hair Jr et al., 2017; Sattler et al., 2010). For these reasons, the current research has dismissed the Sobel test for evaluating mediation analysis. Since the current study applies PLS-SEM, therefore, the mediation analysis is conducted using bootstrapping test, to test the direct and indirect effects of the mediation. Indirect Effects provide an overview of results, including standard errors, bootstrap mean values, t values, and p values. In addition, it shows the confidence interval as derived from the BCa method (Nitzl et al., 2016). 56 DATA ANALYSIS AND FINDINGS 57 4.1. Introduction The main findings of the empirical investigation are presented in this chapter. This chapter first introduces the experiment manipulation check, followed by the main findings and results after running the data analysis, including validity and reliability test, descriptive analysis, and structural equational modelling (SEM). The findings are reported using SPSS 26 and partial least squared-structural equation modelling (PLS-SEM) using SmartPLS4. SPSS is used in the descriptive analysis while PLS is used to test the research hypotheses, as indicated in the research model. 4.2. Experiment Manipulation Check A manipulation check as defined by Hoewe (2017:1) is " a test used to determine the effectiveness of a manipulation in the experimental design ." If the manipulation check is successful (prompt an expected difference between or among experimental conditions), the researcher can satisfactorily conclude participants correctly understand, interpret, or respond to the stimulus and draw more accurate conclusions about the relationship between the independent and dependent variables (Hoewe, 2017) The current study has three manipulation scenarios: Personalization (Personalized vs. Non-Personalized), Brand Familiarity (Familiar vs. Unfamiliar), and Transparency Approach (Transparent vs. NonTransparent). During the experiment participants are subjected to a scenario containing the stimulus materials, followed by a survey to check the manipulations. An independent samples t-test using SPSS26 was used in order to determine whether the manipulation check was successful or not. In the current study, participants’ first name was used to manipulate personalization where they were asked to enter their names before they were subject to the stimulus screen. In the personalized condition participants were greeted using their first names in the ad, while in the non-personalized condition, they were greeted using the phrase “Hello There,”. Personalization manipulation was checked using a oneitem scale: “This ad is directed to me personally,” adapted from De Keyzer, Dens, and De Pelsmacker (2015), and a 7-point Likert-type scale (1-Strongly Disagree, 7Strongly Agree) was used to measure their responses. An independent sample t-test was conducted to check the manipulation. Participants in the personalized condition reported higher agreement (M = 5.619, SD = 1.411) than those in the nonpersonalized condition (M = 2.511, SD = 1.422), t (470) = 24.113, p≤0.001. The results show statistically significant difference between the two scenarios; therefore, the manipulation was successful. 58 Table 13: Personalized/Non-Personalized Group Statistics N Mean Std. Deviation Std. Error Mean Non-Personalized 266 2.5113 1.42281 0.08724 Personalized 221 5.6199 1.41111 0.09492 Table 14: Personalized/Non-Personalized Independent Samples Test Levene's Test for Equality of Variances 95% Confidence Interval of the Difference F Sig. t df Sig. (2tailed) Mean Differe nce Std. Error Differe nce Lower Upper Equal variances assumed 0.074 0.786 -24.094 485 0.000 -3.108 0.129 -3.362 -2.855 Equal variances not assumed -24.113 470 0.000 -3.108 0.128 -3.361 -2.855 Brand Familiarity was manipulated using an image as a stimulus in the ad, the image either showed a familiar brand or an unfamiliar one. Brand Familiarity manipulation was checked using a one-item scale: “Regarding the brand, I am:” adapted from Baker, Hutchinson, Moore and Nedungai (1986), and a 7point Likert-type scale (1-Unfamiliar, 7Familiar) was used to measure their responses. An independent sample t-test was conducted to check the manipulation. Participants in the familiar condition reported higher agreement (M = 5.876, SD = 0.914) than those in the unfamiliar condition (M = 2.250, SD = 0.989), t (485) = 42.010, p≤0.001. The results show statistically significant difference between the two scenarios; therefore, the manipulation was successful. Table 15:Familiar/Unfamiliar Group Statistics N Mean Std. Deviation Std. Error Mean Unfamiliar 244 2.2500 0.98914 0.06332 Familiar 243 5.8765 0.91429 0.05865 59 Table 16: Familiar/Unfamiliar Independent Samples Test Levene's Test for Equality of Variances 95% Confidence Interval of the Difference F Sig. t df Sig. (2tailed) Mean Differen ce Std. Error Differenc e Lower Upper Equal variances assumed 4.425 0.036 -42.010 485 0.000 -3.626 0.086 -3.796 -3.456 Equal variances not assumed -42.016 482 0.000 -3.626 0.086 -3.796 -3.456 Regarding Transparency, some participants were randomly subjected to a transparent experiment displaying a link with “Why am I seeing this ad?” question, while others were subjected to a nontransparent one lacking the same question. The survey did not have an explicit question to test the manipulation for transparency, however, multiple related questions were tried as proxies for explicit transparency question (such as “I think companies share information without permission”, “The advertised brand is an honest brand”). Yet, all tested proxy questions yielded statistically insignificant results (manipulation unsuccessful), this suggests that transparency manipulation is either overlooked/misunderstood by the respondents or actually has no bearing on their responses. This result is actually not new as mentioned earlier in the Literature Review chapter. For the first two scenarios since the manipulation checks were successful, it could be assumed that participants could identify the different (non)personalized/(un)familiar scenarios, and thus their answers to the survey questions were based on that identification. As for transparency, the experiment number referring to whether the experiment was transparent or not was recorded to test the effect of transparency on user responses. Survey data examination was carried out using SPSS26 and SmartPLS4, data examinations are covered in the next sections. 4.3. Data Examination The data collected from questionnaires were first entered into the SPSS 26 software to obtain descriptive statistics of the responses. Then, statistical methods, using SPSS 26, were employed to verify that the data is free of errors before analysis to ensure reliable and valid data, to ensure the final results are 66 4.5.3. Discriminant Validity Fornell-Larcker (1981) is used to assess the discriminant validity. For discriminant validity to be achieved, the square root value of the construct’s AVE should be greater than its correlations with all other constructs (Hair Jr et al., 2017). Results support (table 20), the discriminant validity of all variables. The correlation coefficients in brackets denote the strength of relationships between the constructs, reflecting their discriminant validity. Notably, constructs like "Attitudes towards the Brand" and "Attitude towards OBA" exhibit high correlation coefficients of 0.914 and 0.954, respectively, implying potential overlap or shared variance, which raises concerns regarding their discriminant validity. Similarly, "Intention to Click on the Ad" and "Intentions to Purchase the Product" reveal high coefficients of 0.961 with themselves, suggesting robust internal consistency. Additionally, constructs such as "Protective Behavior," "Personalization," "Perceived Intrusiveness," "Privacy Concerns," and "Reactance" also exhibit high coefficients with themselves, indicating internal reliability. While most constructs demonstrate strong internal consistency, the substantial correlations between certain constructs, like "Attitudes towards the Brand" and "Attitude towards OBA," hint at potential overlaps in their conceptual domains, warranting further examination of their discriminant validity. 67 Table 20: Correlation Coefficient between the Research Variables - Fornell-Larcker ATB ATO BF BT CC ICA IPP PB PER PIN PRC REA TA Attitude Towards the Brand 0.914 Attitude Towards OBA 0.854 0.954 Brand Familiarity 0.514 0.435 1.000 Brand Trust 0.733 0.627 0.468 0.965 Consumer Characteristics 0.706 0.721 0.320 0.541 0.933 Intention to Click on the Ad 0.873 0.858 0.428 0.628 0.740 0.961 Intention to Purchase the Product 0.857 0.851 0.433 0.627 0.724 0.850 0.961 Protective Behavior 0.865 0.849 0.434 0.631 0.706 0.840 0.849 0.957 Personalization 0.670 0.577 0.291 0.503 0.524 0.604 0.583 0.600 0.962 Perceived Intrusiveness 0.821 0.656 0.457 0.629 0.509 0.667 0.629 0.661 0.550 0.946 Privacy Concerns 0.787 0.616 0.397 0.587 0.506 0.670 0.623 0.629 0.526 0.731 0.948 Reactance 0.823 0.632 0.384 0.583 0.532 0.673 0.664 0.681 0.588 0.767 0.732 0.949 Transparency Approach 0.048 0.026 0.108 0.066 -0.032 0.044 0.041 0.054 0.018 0.064 0.051 0.010 1.000 68 4.5.4. Convergent Validity and Composite Reliability This section represents the convergent validity and the composite reliability of the research variables. For the convergent validity, it is measured through AVE. AVE should be extracted when the constructs of 0.5 or greater (Hair et al., 2022). From the analysis, it is noticed that the AVE of all the variables is greater than 0.5. For the reliability, it is measured through Cronbach's Alpha and Composite Reliability (Hair et al., 2017; Sarstedt et al., 2021). A satisfactory degree of reliability is indicated by alpha of at least 0.7 (Nunnally, 1978), while composite reliability coefficients ranging from 0.6 to 0.9 are considered acceptable reliability (Diamantopoulos et al., 2012). From these results, it is noticed that all the values of Cronbach’s Alpha and Composite Reliability values are greater than 0.7. The results of AVE, Cronbach alpha, and Composite Reliability of research constructs are shown in (table 21). Table 21: Convergent Validity and Composite Reliability Cronbach's alpha Composite reliability Average variance extracted (AVE) Attitude Towards the Brand 0.934 0.953 0.835 Attitude Towards OBA 0.980 0.984 0.910 Brand Trust 0.975 0.982 0.931 Consumer Characteristics 0.853 0.931 0.870 Intention to Click on the Ad 0.959 0.973 0.924 Intention to Purchase the Product 0.959 0.973 0.924 Protective Behavior 0.988 0.990 0.915 Personalization 0.984 0.987 0.925 Perceived Intrusiveness 0.980 0.983 0.894 Privacy Concerns 0.977 0.982 0.898 Reactance 0.963 0.973 0.900 4.6. Structural Model Assessment Structural model assessment is the next step in evaluating PLS-SEM results after the measurement model has been assessed and satisfactory results were achieved. The standard assessment procedures to be considered when analyzing the structural model are (a) assessing the structural model for collinearity issues, (b) assessing the significance and relevance of the model relationships, (c) assessing the level of coefficient determination ( R2 ), (d) assessing the effect size 69 ( f 2 ) and (e) assessing the predictive relevance ( Q2 ) (Benitez et al., 2020; Hair et al., 2011, 2019; Hair Jr et al., 2017). 4.6.1. Multicollinearity Assessment Verification of multicollinearity between independent variables in the model is conducted; the correlation between two or more variables causes difficulty in determining variance explained by independent variables and creating a regression model (Hair et al., 2017). Redundant dependent variable info presented. The inner VIF shows that all VIFs are less than 5, which means that there is no problem in multicollinearity. The results of inner VIF are shown in (table 22). Table 22: Inner VIF Attitudes towards the Brand Attitude towards OBA Intention to Click on the Ad. Intentions to Purchase the Product Protective Behavior Perceived Intrusivenes s privacy concerns Reactance Brand Familiarity 1.372 1.372 1.372 1.372 1.372 1.306 1.306 1.306 Brand Trust 2.063 2.063 2.063 2.063 2.063 1.774 1.774 1.774 Consumer Characteristics 1.703 1.703 1.703 1.703 1.703 1.626 1.626 1.626 Personalization 1.755 1.755 1.755 1.755 1.755 1.524 1.524 1.524 Perceived Intrusiveness 3.192 3.192 3.192 3.192 3.192 Privacy Concerns 2.663 2.663 2.663 2.663 2.663 Reactance 3.122 3.122 3.122 3.122 3.122 Transparency Approach 1.025 1.025 1.025 1.025 1.025 1.020 1.020 1.020 4.6.2. Structural Model Path Coefficient (Testing Research Hypotheses) PLS analysis is done to test the hypotheses through making structural equation modeling (SEM). The results of analysis are shown in (table 26). From the table, it could be observed that: ➢ H1: Personalization has a positive impact on the dependent variables. This hypothesis is divided into five sub-hypotheses: For the first sub-hypothesis, H1a, “Personalization has a positive impact on Consumers’ Attitudes Towards the Brand”, the results shows that there is a positive significant impact of Personalization on Consumers’ attitudes towards the brand (p≤0.001) and the estimate equals to 0.111 (more than 0). Based on the previous results, the first sub-hypothesis is supported. 70 For the second sub-hypothesis, H1b, “Personalization has a positive impact on Consumers’ Attitudes Towards OBA”, SEM shows that there is a positive significant effect of Personalization on Consumers’ attitudes towards OBA (p≤0.001) and the estimate equals to 0.104. Based on the previous results, the second sub-hypothesis is supported. For the third sub-hypothesis, H1c, “Personalization has a positive impact on Consumers’ Intention to Click on the Ad”, SEM indicated a positive significant influence of Personalization on Consumers’ intention to click on the Ad (p≤0.001) and the estimate equals to 0.114. Based on the previous results, the third sub-hypothesis is supported. For the fourth sub-hypothesis, H1d, “Personalization has a positive impact on Consumers’ Intentions to Purchase the Product”, SEM shows that Personalization has a positive significant impact on Consumers’ intentions to purchase the product (p≤0.001) and the estimate equals to 0.103. Based on the previous results, the fourth sub-hypothesis is supported. For the fifth sub-hypothesis, H1e, “Personalization has a positive impact on Consumers’ Protective Behavior”, SEM shows that there is a positive significant effect of Personalization on Consumers’ protective behavior (p≤0.001) and the estimate equals to 0.123. Based on the previous results, the fifth sub-hypothesis is supported. From the above results, it is concluded that the first hypothesis is supported. ➢ H2: Considering consumer characteristics has a positive impact on the dependent variables. This hypothesis is divided into five sub-hypotheses: For the first sub-hypothesis, H2a, “Considering consumer characteristics has a positive impact on Consumers’ Attitudes Towards the Brand”, the analysis shows that Understanding consumer characteristics has a positive significant effect on Consumers’ attitudes towards the brand (p≤0.001) and the estimate equals to 0.224. Based on the previous results, the first sub-hypothesis is supported. For the second sub-hypothesis, H2b, “Considering consumer characteristics has a positive impact on Consumers’ Attitudes Towards OBA”, SEM shows that there is a positive significant influence of Understanding consumer characteristics on Consumers’ attitudes towards OBA (p≤0.001) and the estimate equals to 0.416. Based on the previous results, the second sub-hypothesis is supported. For the third sub-hypothesis, H2c, “Considering consumer characteristics has a positive impact on Consumers’ Intention to Click on the Ad”, SEM indicated a positive significant influence of 71 Understanding consumer characteristics on Consumers’ intention to click the Ad (p≤0.001), Based on the previous results, the third sub-hypothesis is supported. For the fourth sub-hypothesis, H2d, “Considering consumer characteristics has a positive impact on Consumers’ Intentions to Purchase the Product”, SEM shows that there is a positive significant effect of Understanding consumer characteristics on Consumers’ intention to purchase the product (p≤0.001). Accordingly, the fourth sub-hypothesis is supported. For the fifth sub-hypothesis, H2e, “Considering consumer characteristics has a positive impact on Consumers’ Protective Behavior”, the analysis proved a positive significant impact of Understanding consumer characteristics on Consumers’ protective behavior (p≤0.001) and the estimate equals to 0.369. Accordingly, the fifth sub-hypothesis is supported. From the above results, it is concluded that the second hypothesis is supported. ➢ H3: Brand Familiarity has a positive impact on the dependent variables. For the first sub-hypothesis, H3a, “Brand Familiarity has a positive impact on Consumers’ Attitudes Towards the Brand”, the results show that there is a positive significant impact of Brand Familiarity on Consumers’ attitudes towards the brand (p≤0.001) and the estimate equal to 0.087. Based on the previous results, the first sub-hypothesis is supported. For the second sub-hypothesis, H3b, “Brand Familiarity has a positive impact on Consumers’ Attitudes Towards OBA”, the results indicated a positive significant influence of Brand Familiarity on Consumers’ attitudes towards OBA as P-value is 0.011 and the estimate equals to 0.081. Based on the previous results, the second sub-hypothesis is supported. For the third sub-hypothesis, H3c, “Brand Familiarity has a positive impact on Consumers’ Intention to Click on the Ad”, SEM proved a positive significant impact of Brand Familiarity on Consumers’ intention to click on the Ad as P-value is 0.027 and the estimate equals to 0.063. Based on the previous results, the third sub-hypothesis is supported. The analysis of the fourth sub-hypothesis, H3d, “Brand Familiarity has a positive impact on Consumers’ Intentions to Purchase the Product”, proved a positive significant impact of Brand Familiarity on Consumers’ intentions to purchase the product as P-value is 0.003 and the estimate equals to 0.087. Based on the previous results, the fourth sub-hypothesis is supported. For the fifth sub-hypothesis, H3e, “Brand Familiarity has a positive impact on Consumers’ Protective Behavior”, the results show that there is a positive significant impact of Brand Familiarity 72 on Consumers’ protective behavior as P-value is 0.011 with an estimate equal to 0.075. Accordingly, the fifth sub-hypothesis is supported. From the above results, the third hypothesis is supported. ➢ H4: Brand trust has a positive impact on the dependent variables. This hypothesis is divided into five sub-hypotheses: For the first sub-hypothesis, H4a, “Brand trust has a positive impact on Consumers’ Attitudes Towards the Brand”, the results show that there is a positive significant impact of Brand trust on Consumers’ attitudes towards the brand (p≤0.001) and the estimate equal to 0.147. Based on the previous results, the first sub-hypothesis is supported. For the second sub-hypothesis, H4b, “Brand trust has a positive impact on Consumers’ Attitudes Towards OBA”, SEM proved a positive significant influence of Brand trust on Consumers’ attitudes towards OBA, as the P-value equals to 0.001 and the estimate equals to 0.124. Based on the previous results, the second sub-hypothesis is supported. For the third sub-hypothesis, H4c, “Brand trust has a positive impact on Consumers’ Intention to Click on the Ad.”, SEM shows a positive significant impact of Brand trust on Consumers’ intention to click on the Ad., as the P-value equals to 0.009 and the estimate equals to 0.089. Accordingly, the third sub-hypothesis is supported. For the fourth sub-hypothesis, H4d, “Brand trust has a positive impact on Consumers’ Intentions to Purchase the Product”, SEM shows that Brand trust has a positive significant influence on intentions to purchase the product (p≤0.001) and the estimate is 0.128. Based on the previous results, the fourth sub-hypothesis is supported. For the fifth sub-hypothesis, H4e, “Brand trust has a positive impact on Consumers’ Protective Behavior”, SEM indicated a positive significant impact of Brand trust on Consumers’ protective behavior, as the P-value equals to 0.002 and the estimate is 0.121. Based on this result, the fifth subhypothesis is supported. From all the above results, the fourth hypothesis is proved to be supported. ➢ H5: Being transparent about data collection mechanisms has a positive impact on the dependent variables. This hypothesis is divided into five sub-hypotheses: 73 For the first sub-hypothesis, H5a, “Being transparent about data collection mechanisms has a positive impact on Consumers’ Attitudes Towards the Brand”, the results show an insignificant impact of Being transparent about data collection mechanisms on Consumers’ attitudes towards the brand, as the P-value equals to 0.276 (greater than 0.05). Based on the previous results, the first subhypothesis is not supported. For the second sub-hypothesis, H5b, “Being transparent about data collection mechanisms has a positive impact on Consumers’ Attitudes Towards OBA”, SEM proves that Being transparent about data collection mechanisms has an insignificant influence on Consumers’ attitudes towards OBA, as P-value equals to 0.433. Accordingly, the second sub-hypothesis is not supported. For the third sub-hypothesis, H5c, “Being transparent about data collection mechanisms has a positive impact on Consumers’ Intention to Click on the Ad.”, SEM indicated an insignificant effect of Being transparent about data collection mechanisms on intention to click on the Ad., as the Pvalue equals to 0.138. Based on the previous results, the third sub-hypothesis is not supported. For the fourth sub-hypothesis, H5d, “Being transparent about data collection mechanisms has a positive impact on Consumers’ Intentions to Purchase the Product”, the results show that Being transparent about data collection mechanisms has an insignificant effect on Consumers’ intentions to purchase the product, as P-value is 0.156. Accordingly, the fourth sub-hypothesis is not supported. For the fifth sub-hypothesis, “Being transparent about data collection mechanisms has a positive impact on Consumers’ Protective Behavior”, SEM proves an insignificant influence of Being transparent on protective behavior, as P-value is 0.086. Based on the previous results, the fifth sub-hypothesis is not supported. From all the above results, the fifth hypothesis is not supported. 74 Figure 3: Measurement Model 75 Figure 4: Structural Model 82 For the mediation role of Perceived Intrusiveness between independent variables and Intentions to Purchase the Product, results can be interpreted as follows: Regarding the mediating role of PIN in the relationship between PER and IPP, the first mediation assessment reveals an insignificant indirect effect of PER on IPP through PIN (H6id: β= 0.009, t= 0.668, p=0.025). Therefore, H6id has no mediation effect, suggesting that Perceived Intrusiveness does not mediate the relationship between Personalization and consumers’ Intentions to Purchase the Product. Regarding the mediating role of PIN in the relationship between CC and IPP, the first mediation assessment reveals an insignificant indirect effect of CC on IPP through PIN (H6iid: β= 0.005, t= 0.645, p=0.260). Therefore, H6iid has no mediation effect, suggesting that Consumers’ perceptions of Intrusiveness do not play a role in influencing their Intentions to Purchase the Product. Regarding the mediating role of PIN in the relationship between BF and IPP, the first mediation assessment reveals an insignificant indirect effect of BF on IPP through PIN (H6iiid: β= 0.006, t= 0.650, p=0.258). Therefore, H6iiid has no mediation effect, suggesting that Perceived Intrusiveness has no role in explaining how Brand Familiarity shapes consumers’ Intentions to Purchase the Product. Regarding the mediating role of PIN in the relationship between BT and IPP, the first mediation assessment reveals an insignificant indirect effect of BT on IPP through PIN (H6ivd: β= 0.012, t= 0.670, p=0.252). Therefore, H6ivd has no mediation effect, suggesting that Perceived Intrusiveness has no role in explaining how Brand Trust shapes consumers’ Intentions to Purchase the Product. Finally, Regarding the mediating role of PIN in the relationship between TA and IPP, the first mediation assessment reveals an insignificant indirect effect of TA on IPP through PIN (H6vd: β= 0.002, t= 0.327, p=0.372). Therefore, H6vd has no mediation effect, suggesting that Perceived Intrusiveness does not influence the relationship between Transparency Approach and consumers’ Intentions to Purchase the Product. For the mediation role of Perceived Intrusiveness between independent variables and Protective Behavior, results can be interpreted as follows: Regarding the mediating role of PIN in the relationship between PER and PB, the first mediation assessment reveals a significant indirect effect of PER on PB through PIN (H6ie: β= 0.026, t= 0.013, p=0.020). The direct effect of PER in the presence of the mediator on PB is significant (β=0.123, t= 3.829, p≤0.001). The total effect of PER on PB is also significant (β=0.226, t= 7.251, p≤0.001). This shows a complementary partial mediating role of PIN in the relationship between PER and PB (Hair et 83 al., 2021; Zhao et al., 2010). Therefore, H6ie is partially supported. This suggests that Personalization's impact on consumers’ Protective Behavior can explained through the effect of Perceived Intrusiveness. Regarding the mediating role of PIN in the relationship between CC and PB, the first mediation assessment reveals a significant indirect effect of CC on PB through PIN (H6iie: β= 0.014, t= 1.823, p=0.034). The direct effect of CC in the presence of the mediator on PB is significant (β=0.369, t= 12.823, p≤0.001). The total effect of CC on PB is also significant (β=0.426, t= 14.295, p≤0.001). This shows a complementary partial mediating role of PIN in the relationship between CC and PB (Hair et al., 2021; Zhao et al., 2010). Therefore, H6iie is partially supported. The finding suggest that Consumers’ perceptions of Intrusiveness play a role in shaping their Protective Behavior. Regarding the mediating role of PIN in the relationship between BF and PB, the first mediation assessment reveals a significant indirect effect of BF on PB through PIN (H6iiie: β= 0.018, t= 1.910, p=0.028). The direct effect of BF in the presence of the mediator on PB is significant (β=0.075, t= 2.306, p=0.011). The total effect of BF on PB is also significant (β=0.121, t= 3.300, p≤0.001). This shows a complementary partial mediating role of PIN in the relationship between BF and PB (Hair et al., 2021; Zhao et al., 2010). Therefore, H6iiie is partially supported. Thus, Perceived Intrusiveness plays a significant role in explaining how Brand Familiarity shapes consumers’ Protective Behavior. Regarding the mediating role of PIN in the relationship between BT and PB, the first mediation assessment reveals a significant indirect effect of BT on PB through PIN (H6ive: β= 0.036, t= 2.098, p=0.018). The direct effect of BT in the presence of the mediator on PB is significant (β=0.121, t= 2.891, p=0.002). The total effect of BT on PB is also significant (β=0.228, t= 5.380, p≤0.001). This shows a complementary partial mediating role of PIN in the relationship between BT and PB (Hair et al., 2021; Zhao et al., 2010). Therefore, H6ive is partially supported. Thus, while Brand Trust directly impacts consumers’ Protective Behavior, part of this relationship is mediated by Perceived Intrusiveness. In contrast, Regarding the mediating role of PIN in the relationship between TA and PB, the first mediation assessment reveals an insignificant indirect effect of TA on PB through PIN (H6ve: β= 0.004, t= 0.595, p=0.276). Therefore, H6ve has no mediation effect, suggesting that Perceived Intrusiveness does not influence the relationship between Transparency Approach and consumers’ Protective Behavior. For the mediation role of Privacy Concerns between independent variables and Attitudes towards the Brand, it could be observed that: 84 Regarding the mediating role of PRC in the relationship between PER and ATB, the first mediation assessment reveals a significant indirect effect of PER on ATB through PRC (H7ia: β= 0.041, t= 4.595, p≤0.001). The direct effect of PER in the presence of the mediator on ATB is significant (β=0.111, t= 6.348, p≤0.001). The total effect of PER on ATB is also significant (β=0.282, t= 11.691, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between PER and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H7ia is partially supported. This suggests that Personalization's impact on Attitudes towards the Brand can explained through the effect of Privacy Concerns in this model. Regarding the mediating role of PRC in the relationship between CC and ATB, the first mediation assessment reveals a significant indirect effect of CC on ATB through PRC (H7iia: β= 0.029, t= 3.738, p≤0.001). The direct effect of CC in the presence of the mediator on ATB is significant (β=0.224, t= 11.688, p≤0.001). The total effect of CC on ATB is also significant (β=0.322, t= 12.727, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between CC and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H7iia is partially supported. The finding suggest that consumer perceptions of Privacy play a role in shaping consumers’ Attitudes towards the Brand. Regarding the mediating role of PRC in the relationship between BF and ATB, the first mediation assessment reveals a significant indirect effect of BF on ATB through PRC (H7iiia: β= 0.021, t= 2.838, p=0.002). The direct effect of BF in the presence of the mediator on ATB is significant (β=0.087, t= 4.134, p≤0.001). The total effect of BF on ATB is also significant (β=0.169, t= 5.303, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between BF and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H7iiia is partially supported. Thus, Privacy Concerns plays a significant role in explaining how Brand Familiarity shapes consumers’ Attitudes towards the Brand. Regarding the mediating role of PRC in the relationship between BT and ATB, the first mediation assessment reveals a significant indirect effect of BT on ATB through PRC (H7iva: β= 0.053, t= 4.811, p≤0.001). The direct effect of BT in the presence of the mediator on ATB is significant (β=0.147, t= 5.775, p≤0.001). The total effect of BT on ATB is also significant (β=0.337, t= 9.419, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between BT and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H7iva is partially supported. The findings highlight the pivotal role of Privacy Concerns in shaping consumers’ Attitudes towards the brand based on Brand Trust. Finally, Regarding the mediating role of PRC in the relationship between TA and ATB, the first mediation assessment reveals an insignificant indirect effect of TA on ATB through PRC (H7va: β= 0.006, t= 0.513, 85 p=0.304). Therefore, H6va has no mediation effect, suggesting that Privacy concerns does not mediate the relationship between Transparency Approach and consumers' Attitudes towards the Brand. For the mediation role of Privacy Concerns between independent variables and Attitudes towards OBA, results can be interpreted as follows: Regarding the mediating role of PRC in the relationship between PER and ATO, the first mediation assessment reveals a significant indirect effect of PER on ATO through PRC (H7ib: β= 0.019, t= 1.682, p=0.046). The direct effect of PER in the presence of the mediator on ATO is significant (β=0.104, t= 3.295, p≤0.001). The total effect of PER on ATO is also significant (β=0.186, t= 5.974, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between PER and ATO (Hair et al., 2021; Zhao et al., 2010). Therefore, H7ib is partially supported. This suggests that Personalization's impact on Attitudes towards OBA can explained through the effect of Privacy Concerns. Regarding the mediating role of PRC in the relationship between CC and ATO, the first mediation assessment reveals an insignificant indirect effect of CC on ATO through PRC (H7iib: β= 0.014, t= 0.009, p=0.056). Therefore, H7iib has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Consumer Characteristics and Attitudes towards OBA. Regarding the mediating role of PRC in the relationship between BF and ATO, the first mediation assessment reveals an insignificant indirect effect of BF on ATO through PRC (H7iiib: β= 0.010, t= 1.457, p=0.073). Therefore, H7iiib has no mediation effect, suggesting that Privacy Concerns has no significant role in explaining how Brand Familiarity shapes consumers’ attitudes towards OBA. Regarding the mediating role of PRC in the relationship between BT and ATO, the first mediation assessment reveals an insignificant indirect effect of BT on ATO through PRC (H7ivb: β= 0.025, t= 1.650, p=0.050). Therefore, H7ivb has no mediation effect, suggesting that Privacy Concerns has no significant role in explaining how Brand Trust shapes consumers’ Attitudes towards OBA. Finally, Regarding the mediating role of PRC in the relationship between TA and ATO, the first mediation assessment reveals an insignificant indirect effect of TA on ATO through PRC (H7vb: β= 0.003, t= 0.439, p=0.330). Therefore, H7vb has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Transparency Approach and consumers' Attitudes towards OBA. For the mediation role of Privacy Concerns between independent variables and Intentions to Click on the Ad, results can be interpreted as follows: 86 Regarding the mediating role of PRC in the relationship between PER and ICA, the first mediation assessment reveals a significant indirect effect of PER on ICA through PRC (H7ic: β= 0.041, t= 3.312, p≤0.001). The direct effect of PER in the presence of the mediator on ICA is significant (β=0.114, t= 3.994, p≤0.001). The total effect of PER on ICA is also significant (β=0.215, t= 7.060, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between PER and ICA (Hair et al., 2021; Zhao et al., 2010). Therefore, H7ic is partially supported. This suggests that Personalization's impact on Intentions to Click on the Ad can be explained through the effect of Privacy Concerns. Regarding the mediating role of PRC in the relationship between CC and ICA, the first mediation assessment reveals a significant indirect effect of CC on ICA through PRC (H7iic: β= 0.028, t= 2.901, p=0.002). The direct effect of CC in the presence of the mediator on ICA is significant (β=0.420, t= 15.161, p≤0.001). The total effect of CC on ICA is also significant (β=0.481, t= 17.078, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between CC and ICA (Hair et al., 2021; Zhao et al., 2010). Therefore, H7iic is partially supported. The finding suggest that consumers' perceptions of Privacy Concerns play a role in shaping their Intentions to Click on the Ad. Regarding the mediating role of PRC in the relationship between BF and ICA, the first mediation assessment reveals a significant indirect effect of BF on ICA through PRC (H7iiic: β= 0.021, t= 2.342, p=0.010). The direct effect of BF in the presence of the mediator on ICA is significant (β=0.063, t= 1.928, p=0.027). The total effect of BF on ICA is also significant (β=0.112, t= 3.209, p=0.001). This shows a complementary partial mediating role of PRC in the relationship between BF and ICA (Hair et al., 2021; Zhao et al., 2010). Therefore, H7iiic is partially supported. Thus, Privacy Concerns plays a significant role in explaining how Brand Familiarity shapes consumers' Intentions to Click on the Ad. Regarding the mediating role of PRC in the relationship between BT and ICA, the first mediation assessment reveals a significant indirect effect of BT on ICA through PRC (H7ivc: β= 0.052, t= 3.251, p=0.001). The direct effect of BT in the presence of the mediator on ICA is significant (β=0.089, t= 2.348, p=0.009). The total effect of BT on ICA is also significant (β=0.205, t= 5.394, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between BT and ICA (Hair et al., 2021; Zhao et al., 2010). Therefore, H7ivc is partially supported. Thus, while Brand Trust directly impacts consumers’ Intentions to Click on the Ad, part of this relationship is mediated by Privacy Concerns. In contrast, Regarding the mediating role of PRC in the relationship between TA and ICA, the first mediation assessment reveals an insignificant indirect effect of TA on ICA through PRC (H7vc: β= 0.006, 87 t= 0.493, p=0.311). Therefore, H7vc has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Transparency Approach and consumers’ Intentions to Click on the Ad. For the mediation role of Privacy Concerns between independent variables and Intentions to Purchase the Product, results can be interpreted as follows: Regarding the mediating role of PRC in the relationship between PER and IPP, the first mediation assessment reveals a significant indirect effect of PER on IPP through PRC (H7id: β= 0.021, t= 1.777, p=0.038). The direct effect of PER in the presence of the mediator on IPP is significant (β=0.103, t= 3.368, p≤0.001). The total effect of PER on IPP is also significant (β=0.193, t= 6.478, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between PER and IPP (Hair et al., 2021; Zhao et al., 2010). Therefore, H7id is partially supported. This suggests that Personalization's impact on Intentions to Purchase the Product can be explained through the effect of Privacy Concerns. Regarding the mediating role of PRC in the relationship between CC and IPP, the first mediation assessment reveals a significant indirect effect of CC on IPP through PRC (H7iid: β= 0.015, t= 1.681, p=0.046). The direct effect of CC in the presence of the mediator on IPP is significant (β=0.414, t= 14.272, p≤0.001). The total effect of CC on IPP is also significant (β=0.467, t= 15.730, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between CC and IPP (Hair et al., 2021; Zhao et al., 2010). Therefore, H7iid is partially supported. This suggests that Consumers’ perceptions of Privacy Concerns impact their Intentions to Purchase the Product. Regarding the mediating role of PRC in the relationship between BF and IPP, the first mediation assessment reveals an insignificant indirect effect of BF on IPP through PIN (H7iiid: β= 0.011, t= 1.547, p=0.061). Therefore, H7iiid has no mediation effect, suggesting that Privacy Concerns has no role in explaining how Brand Familiarity shapes consumers’ Intentions to Purchase the Product. Regarding the mediating role of PRC in the relationship between BT and IPP, the first mediation assessment reveals a significant indirect effect of BT on IPP through PRC (H7ivd: β= 0.028, t= 1.791, p=0.037). The direct effect of BT in the presence of the mediator on IPP is significant (β=0.128, t= 3.309, p≤0.001). The total effect of BT on IPP is also significant (β=0.218, t= 5.665, p≤0.001). This shows a complementary partial mediating role of PRC in the relationship between BT and IPP (Hair et al., 2021; Zhao et al., 2010). Therefore, H7ivd is partially supported. This suggests that Brand Trust impacts consumers’ Intentions to Purchase the Product influenced by their Privacy Concerns. 88 Finally, Regarding the mediating role of PRC in the relationship between TA and IPP, the first mediation assessment reveals an insignificant indirect effect of TA on IPP through PRC (H7vd: β= 0.003, t= 0.457, p=0.324). Therefore, H7vd has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Transparency Approach and consumers’ Intentions to Purchase the Product. For the mediation role of Privacy Concerns between independent variables and Protective Behavior, results can be interpreted as follows: Regarding the mediating role of PRC in the relationship between PER and PB, the first mediation assessment reveals an insignificant indirect effect of PER on PB through PRC (H7ie: β= 0.015, t= 1.210, p=0.113). Therefore, H7ie has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Personalization and consumers’ Protective Behavior. Regarding the mediating role of PRC in the relationship between CC and PB, the first mediation assessment reveals an insignificant indirect effect of CC on PB through PRC (H7iie: β= 0.010, t= 1.196, p=0.116). Therefore, H7iie has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Consumer Characteristics and consumers’ Protective Behavior. Regarding the mediating role of PRC in the relationship between BF and PB, the first mediation assessment reveals an insignificant indirect effect of BF on PB through PRC (H7iiie: β= 0.007, t= 1.112, p=0.133). Therefore, H7iiie has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Brand Familiarity and consumers’ Protective Behavior. Regarding the mediating role of PRC in the relationship between BT and PB, the first mediation assessment reveals an insignificant indirect effect of BT on PB through PRC (H7ive: β= 0.019, t= 1.229, p=0.110). Therefore, H7ive has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Brand Trust and consumers’ Protective Behavior. Finally, Regarding the mediating role of PRC in the relationship between TA and PB, the first mediation assessment reveals an insignificant indirect effect of TA on PB through PRC (H7ve: β= 0.002, t= 0.393, p=0.347). Therefore, H7ve has no mediation effect, suggesting that Privacy Concerns does not influence the relationship between Transparency Approach and consumers’ Protective Behavior. For the mediation role of Reactance between independent variables and Attitudes towards the Brand, it could be observed that: Regarding the mediating role of REA in the relationship between PER and ATB, the first mediation assessment reveals a significant indirect effect of PER on ATB through REA (H8ia: β= 0.076, t= 5.759, 89 p≤0.001). The direct effect of PER in the presence of the mediator on ATB is significant (β=0.111, t= 6.348, p≤0.001). The total effect of PER on ATB is also significant (β=0.282, t= 11.691, p≤0.001). This shows a complementary partial mediating role of REA in the relationship between PER and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H8ia is partially supported. This suggests that Personalization's impact on Attitudes towards the Brand can explained through the effect of consumers’ feelings of Reactance. Regarding the mediating role of REA in the relationship between CC and ATB, the first mediation assessment reveals a significant indirect effect of CC on ATB through REA (H8iia: β= 0.042, t= 4.213, p≤0.001). The direct effect of CC in the presence of the mediator on ATB is significant (β=0.224, t= 11.688, p≤0.001). The total effect of CC on ATB is also significant (β=0.322, t= 12.727, p≤0.001). This shows a complementary partial mediating role of REA in the relationship between CC and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H8iia is partially supported. The finding suggest that consumer perceptions of Reactance play a role in shaping consumers’ Attitudes towards the Brand. Regarding the mediating role of REA in the relationship between BF and ATB, the first mediation assessment reveals a significant indirect effect of BF on ATB through PRC (H8iiia: β= 0.025, t= 2.353, p=0.009). The direct effect of BF in the presence of the mediator on ATB is significant (β=0.087, t= 4.134, p≤0.001). The total effect of BF on ATB is also significant (β=0.169, t= 5.303, p≤0.001). This shows a complementary partial mediating role of REA in the relationship between BF and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H8iiia is partially supported. Thus, Reactance plays a significant role in explaining how Brand Familiarity shapes consumers’ Attitudes towards the Brand. Regarding the mediating role of REA in the relationship between BT and ATB, the first mediation assessment reveals a significant indirect effect of BT on ATB through REA (H8iva: β= 0.064, t= 5.014, p≤0.001). The direct effect of BT in the presence of the mediator on ATB is significant (β=0.147, t= 5.775, p≤0.001). The total effect of BT on ATB is also significant (β=0.337, t= 9.419, p≤0.001). This shows a complementary partial mediating role of REA in the relationship between BT and ATB (Hair et al., 2021; Zhao et al., 2010). Therefore, H8iva is partially supported. The findings highlight the pivotal role of Reactance in shaping consumers’ perceptions of Brand Trust and their Attitudes towards the Brand. In contrast, Regarding the mediating role of REA in the relationship between TA and ATB, the first mediation assessment reveals an insignificant indirect effect of TA on ATB through REA (H8va: β= - 90 0.009, t= 0.598, p=0.275). Therefore, H8va has no mediation effect, suggesting that Reactance does not mediate the relationship between Transparency Approach and Attitudes towards the Brand. For the mediation role of Reactance between independent variables and Attitudes towards OBA, results can be interpreted as follows: Regarding the mediating role of REA in the relationship between PER and ATO, the first mediation assessment reveals an insignificant indirect effect of PER on ATO through REA (H8ib: β= 0.018, t= 1.061, p=0.144). Therefore, H8ib has no mediation effect, suggesting that Reactance does not influence the relationship between Personalization and Attitudes towards OBA. Regarding the mediating role of REA in the relationship between CC and ATO, the first mediation assessment reveals an insignificant indirect effect of CC on ATO through REA (H8iib: β= 0.010, t= 01.045, p=0.148). Therefore, H8iib has no mediation effect, suggesting that Reactance does not influence the relationship between Consumer Characteristics and Attitudes towards OBA. Regarding the mediating role of REA in the relationship between BF and ATO, the first mediation assessment reveals an insignificant indirect effect of BF on ATO through REA (H8iiib: β= 0.006, t= 0.911, p=0.181). Therefore, H8iiib has no mediation effect, suggesting that Reactance has no significant role in explaining how Brand Familiarity shapes consumers’ Attitudes towards OBA. Regarding the mediating role of REA in the relationship between BT and ATO, the first mediation assessment reveals an insignificant indirect effect of BT on ATO through REA (H8ivb: β= 0.015, t= 1.048, p=0.147). Therefore, H8ivb has no mediation effect, suggesting that Reactance has no significant role in explaining how Brand Trust shapes consumers’ Attitudes towards OBA. Finally, Regarding the mediating role of REA in the relationship between TA and ATO, the first mediation assessment reveals an insignificant indirect effect of TA on ATO through PRC (H8vb: β= -0.002, t= 0.411, p=0.340). Therefore, H8vb has no mediation effect, suggesting that Reactance does not influence the relationship between Transparency Approach and Attitudes towards OBA. For the mediation role of Reactance between independent variables and Intentions to Click on the Ad, results can be interpreted as follows: Regarding the mediating role of REA in the relationship between PER and ICA, the first mediation assessment reveals a significant indirect effect of PER on ICA through REA (H8ic: β= 0.034, t= 2.002, p=0.023). The direct effect of PER in the presence of the mediator on ICA is significant (β=0.114, t= 3.994, p≤0.001). The total effect of PER on ICA is also significant (β=0.215, t= 7.060, p≤0.001). This 91 shows a complementary partial mediating role of REA in the relationship between PER and ICA (Hair et al., 2021; Zhao et al., 2010). Therefore, H8ic is partially supported. This suggests that Personalization's impact on Intentions to Click on the Ad can be explained through the effect of Reactance. Regarding the mediating role of REA in the relationship between CC and ICA, the first mediation assessment reveals a significant indirect effect of CC on ICA through REA (H8iic: β= 0.019, t= 1.912, p=0.028). The direct effect of CC in the presence of the mediator on ICA is significant (β=0.420, t= 15.161, p≤0.001). The total effect of CC on ICA is also significant (β=0.481, t= 17.078, p≤0.001). This shows a complementary partial mediating role of REA in the relationship between CC and ICA (Hair et al., 2021; Zhao et al., 2010). Therefore, H8iic is partially supported. The finding suggest that consumer perceptions of Reactance play a role in shaping their Intentions to Click on the Ad. Regarding the mediating role of REA in the relationship between BF and ICA, the first mediation assessment reveals an insignificant indirect effect of BF on ICA through REA (H8iiic: β= 0.011, t= 1.497, p=0.067). Therefore, H8iiic has no mediation effect, suggesting that Reactance does not influence the relationship between Brand Familiarity and Intentions to Click on the Ad. Regarding the mediating role of REA in the relationship between BT and ICA, the first mediation assessment reveals a significant indirect effect of BT on ICA through REA (H8ivc: β= 0.029, t= 1.981, p=0.024). The direct effect of BT in the presence of the mediator on ICA is significant (β=0.089, t= 2.348, p=0.009). The total effect of BT on ICA is also significant (β=0.205, t= 5.394, p≤0.001). This shows a complementary partial mediating role of REA in the relationship between BT and ICA (Hair et al., 2021; Zhao et al., 2010). Therefore, H8ivc is partially supported. Thus, while Brand Trust directly impacts consumers’ Intentions to Click on the Ad, part of this relationship is mediated by Reactance. In contrast, Regarding the mediating role of REA in the relationship between TA and ICA, the first mediation assessment reveals an insignificant indirect effect of TA on ICA through REA (H8vc: β= -0.004, t= 0.548, p=0.292). Therefore, H8vc has no mediation effect, suggesting that Reactance does not influence the relationship between Transparency Approach and consumers’ Intentions to Click on the Ad. For the mediation role of Reactance between independent variables and Intentions to Purchase the Product, results can be interpreted as follows: Regarding the mediating role of REA in the relationship between PER and IPP, the first mediation assessment reveals a significant indirect effect of PER on IPP through REA (H8id: β= 0.061, t= 3.128, 98 Hypothesis β SD t p Hypotheses Decision H7ia: Personalization -> Privacy Concerns - > Attitude towards the Brand 0.041 0.009 4.595 0.000 Partially Supported H7ib Personalization -> Privacy Concerns - > Attitude towards OBA 0.019 0.012 1.682 0.046 Partially Supported H7ic Personalization -> Privacy Concerns - > Intention to Click on the Ad 0.041 0.012 3.312 0.000 Partially Supported H7id Personalization -> Privacy Concerns - > Intention to Purchase the Product 0.021 0.012 1.777 0.038 Partially Supported H7ie Personalization -> Privacy Concerns - > Protective Behavior 0.015 0.012 1.210 0.113 No Mediation H7iia: Consumer Characteristics -> Privacy Concerns -> Attitude towards the Brand 0.029 0.008 3.738 0.000 Partially Supported H7iib Consumer Characteristics -> Privacy Concerns -> Attitude towards OBA 0.014 0.009 1.587 0.056 No Mediation H7iic Consumer Characteristics -> Privacy Concerns -> Intention to Click on the Ad 0.028 0.010 2.901 0.002 Partially Supported H7iid Consumer Characteristics -> Privacy Concerns -> Intention to Purchase the Product 0.015 0.009 1.681 0.046 Partially Supported H7iie Consumer Characteristics -> Privacy Concerns -> Protective Behavior 0.010 0.009 1.196 0.116 No Mediation H7iiia: Brand Familiarity -> Privacy Concerns -> Attitude towards the Brand 0.021 0.007 2.838 0.002 Partially Supported H7iiib Brand Familiarity -> Privacy Concerns -> Attitude towards OBA 0.010 0.007 1.457 0.073 No Mediation H7iiic Brand Familiarity -> Privacy Concerns -> Intention to Click on the Ad 0.021 0.009 2.342 0.010 Partially Supported H7iiid Brand Familiarity -> Privacy Concerns -> Intention to Purchase the Product 0.011 0.007 1.547 0.061 No Mediation H7iiie Brand Familiarity -> Privacy Concerns -> Protective Behavior 0.007 0.007 1.112 0.133 No Mediation H7iva: Brand Trust -> Privacy Concerns -> Attitude towards the Brand 0.053 0.011 4.811 0.000 Partially Supported H7ivb Brand Trust -> Privacy Concerns -> Attitude towards OBA 0.025 0.015 1.650 0.050 No Mediation H7ivc Brand Trust -> Privacy Concerns -> Intention to Click on the Ad 0.052 0.016 3.251 0.001 Partially Supported 99 Hypothesis β SD t p Hypotheses Decision H7ivd Brand Trust -> Privacy Concerns -> Intention to Purchase the Product 0.028 0.015 1.791 0.037 Partially Supported H7ive Brand Trust -> Privacy Concerns -> Protective Behavior 0.019 0.015 1.229 0.110 No Mediation H7va: Transparency Approach -> Privacy Concerns -> Attitude towards the Brand 0.006 0.012 0.513 0.304 No Mediation H7vb Transparency Approach -> Privacy Concerns -> Attitude towards OBA 0.003 0.006 0.439 0.330 No Mediation H7vc Transparency Approach -> Privacy Concerns -> Intention to Click on the Ad 0.006 0.012 0.493 0.311 No Mediation H7vd Transparency Approach -> Privacy Concerns -> Intention to Purchase the Product 0.003 0.007 0.457 0.324 No Mediation H7ve Transparency Approach -> Privacy Concerns -> Protective Behavior 0.002 0.005 0.393 0.347 No Mediation H8ia: Personalization -> Reactance -> Attitude towards the Brand 0.076 0.013 5.759 0.000 Partially Supported H8ib Personalization -> Reactance -> Attitude towards OBA 0.018 0.017 1.061 0.144 No Mediation H8ic Personalization -> Reactance -> Intention to Click on the Ad 0.034 0.017 2.002 0.023 Partially Supported H8id Personalization -> Reactance -> Intention to Purchase the Product 0.060 0.019 3.128 0.001 Partially Supported H8ie Personalization -> Reactance -> Protective Behavior 0.062 0.018 3.406 0.000 Partially Supported H8iia: Consumer Characteristics -> Reactance -> Attitude towards the Brand 0.042 0.010 4.213 0.000 Partially Supported H8iib Consumer Characteristics -> Reactance -> Attitude towards OBA 0.010 0.010 1.045 0.148 No Mediation H8iic Consumer Characteristics -> Reactance -> Intention to Click on the Ad 0.019 0.010 1.912 0.028 Partially Supported H8iid Consumer Characteristics -> Reactance -> Intention to Purchase the Product 0.033 0.011 2.908 0.002 Partially Supported H8iie Consumer Characteristics -> Reactance -> Protective Behavior 0.034 0.011 3.052 0.001 Partially Supported H8iiia: Brand Familiarity -> Reactance -> Attitude towards the Brand 0.025 0.010 2.353 0.009 Partially Supported 100 Hypothesis β SD t p Hypotheses Decision H8iiib Brand Familiarity -> Reactance -> Attitude towards OBA 0.006 0.006 0.911 0.181 No Mediation H8iiic Brand Familiarity -> Reactance -> Intention to Click on the Ad 0.011 0.007 1.497 0.067 No Mediation H8iiid Brand Familiarity -> Reactance -> Intention to Purchase the Product 0.019 0.010 1.922 0.027 Partially Supported H8iiie Brand Familiarity -> Reactance -> Protective Behavior 0.020 0.010 1.962 0.025 Partially Supported H8iva: Brand Trust -> Reactance -> Attitude towards the Brand 0.064 0.013 5.014 0.000 Partially Supported H8ivb Brand Trust -> Reactance -> Attitude towards OBA 0.015 0.015 1.048 0.147 No Mediation H8ivc Brand Trust -> Reactance -> Intention to Click on the Ad 0.029 0.015 1.981 0.024 Partially Supported H8ivd Brand Trust -> Reactance -> Intention to Purchase the Product 0.050 0.016 3.056 0.001 Partially Supported H8ive Brand Trust -> Reactance -> Protective Behavior 0.052 0.016 3.268 0.001 Partially Supported H8va: Transparency Approach -> Reactance -> Attitude towards the Brand -0.009 0.015 0.598 0.275 No Mediation H8vb Transparency Approach -> Reactance -> Attitude towards OBA -0.002 0.005 0.411 0.340 No Mediation H8vc Transparency Approach -> Reactance -> Intention to Click on the Ad -0.004 0.008 0.548 0.292 No Mediation H8vd Transparency Approach -> Reactance -> Intention to Purchase the Product -0.007 0.013 0.574 0.283 No Mediation H8ve Transparency Approach -> Reactance -> Protective Behavior -0.008 0.013 0.589 0.278 No Mediation 101 DISCUSSION AND CONCLUSION 102 5.1. Introduction This chapter presents an overall analysis and discussion of the primary finding of the current research, in comparison with previously reviewed literature. Followed by the theoretical and practical contributions of the study. Then, research limitations and suggestions for future work are discussed. Finally, a short conclusion for the whole study is presented. 5.2. Discussion The purpose of the study was to explore the role of some specific variables on consumers’ behavioral outcomes of OBA. Specifically, the aim of this research was to explore the role of personalization, consumer characteristics, brand familiarity, brand trust, and transparency on consumers’ attitudes toward OBA, attitudes toward the brand, intention to click on the ad, intention to purchase the advertised product and finally on OBA avoidance. Moreover, the study also explored the mediating influence of perceived intrusiveness, perceived privacy concerns, and reactance on consumers’ attitudes, intentions and behavioral outcomes. Previous studies have addressed these previously mentioned variables, but not to all of them simultaneously. Also, some specific variables were not even studied before in relation to OBA but only to online advertising in general. In addition, we build upon the Social Exchange Theory and the Phycological Reactance Theory to explain the concept of OBA and its antecedents, The research problem can be summarized in the question of “What are the factors that influence consumers attitudes, intentions, and behaviors in the online environment towards an online behaviorally targeted advertisement? And what factors can mediate such effect?”. From this problem and after reviewing previous literature and selecting the appropriate variables, the research questions were developed as: 1. How do personalization, consumer characteristics, brand familiarity, brand trust, and transparency influence consumers’ attitudes, intentions and OBA avoidance in the online environment? 2. How do perceived intrusiveness, privacy concerns, and reactance mediate such influence? In order to address the research problem and answer the research questions, a quantitative methodology was adopted; where an online experiment and a structured questionnaire survey using Gorilla SC platforms were used as research instruments. The questionnaire was distributed through a non-probability sampling method (Convenience sampling), that resulted in 487 valid responses of online Internet users. The conducted 103 survey questions were based on the adoption of scales widely tested in previous literature, and thus the aim was to confirm to what extent the model presented in this study fitted the information presented in the literature. In this light, structural equation analysis was used using SmartPLS. The results obtained throughout this research will now be discussed in relation to the results of the existing literature, then theoretical and practical implications will be presented. H1: Personalization has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. This hypothesis is divided into five sub-hypotheses, where all the sub-hypotheses were supported. This finding goes in line with some previous literature, that suggest that different levels of personalization can influence consumer’s attitudes towards the ad and the brand, and intentions to click the personalized ad (Boerman et al., 2017). The results of work of Aguirre et al. (2015) points out that higher personalization leads to higher click-through intention. Bleier (2015) stated higher personalization influence OBA outcomes expressed in click through rates and purchase intentions, as it enhances online users’ perceived informativeness by receiving information relevant to their interests without going through loads of irrelevant information. In contrast, Van Doorn & Hoekstra (2013) found that a higher degree of personalization influence feeling of increased intrusiveness by online customers which in turn leads to OBA avoidance. H2: Considering consumer characteristics has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. This hypothesis is divided into five sub-hypotheses, where all the sub-hypotheses were supported. This finding goes in line with some previous literature, Turow et al. (2009) points out that perceptions of OBA and attitudes towards personalized ads depend on consumer characteristics such as age, where younger consumers are more accepting to OBA compared to older consumers. Also, the work of Smit et al. (2014) discussed how consumers’ characteristics such as age, education level and online experience influenced their attitudes and responses towards OBA, where the older female consumers with a low level of education and a low family income tended to avoid OBA. 104 H3: Brand Familiarity has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. This hypothesis is divided into five sub-hypotheses, where all the sub-hypotheses were supported. This finding goes in line with some previous literature, where the work of Dodds and Monroe (1985) reported that brand familiarity influenced consumers’ attitudes and perceptions about the product quality and eventually consumers’ intentions to purchase the product. Later, Dodds et al. (1991) in another study, reported that brand name had a positive effect on consumers’ purchase intention. Also, Shoenberger (2014) claimed that brand familiarity is a key heuristic used by consumers to make a purchasing decision. Finally, Zhao et al. (2017) claimed that brand familiarity led consumers to perceive a personalized ad as less intrusive and thus decreased OBA avoidance. H4: Brand trust has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. This hypothesis is divided into five sub-hypotheses, where all the sub-hypotheses were supported. This finding goes in line with some previous literature, where the work of McStay (2011) confirmed the trust is key determinant of consumers’ perceptions, attitudes and behaviors in the online environment. Kumar and Gupta (2016) stated that trust can be created through brand name. Also, a study by Earp and Baumer (2003) found that brand trust leads consumers to disclose their personal information online better than an unknown website with good privacy policy. Rony (2018) assumed that online consumers will only engage with firms or brands that they trust more. Bleier (2015) also found that increased click-through intentions rates to highly personalized ads when online customers trust the advertiser. Finally, Kumar and Pansari (2016) claimed that if online consumers trusted the advertising brand, then they will be more willing to share their private information. H5: Being transparent about data collection mechanisms has a positive impact on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. 105 This hypothesis is divided into five sub-hypotheses, where all the sub-hypotheses were not supported. Surprisingly, this finding also goes in line with some previous literature, where the work of Leon et al. (2012) showed how participants misunderstood or rarely recognized disclosure icons, and incorrectly understood the tagline to intend selling advertising space instead of being a link to pages where they can make choices about OBA. The same happened in the current study where participants either did not recognize or did not understand the question “Why am I seeing this ad?” associated with the transparent experiment they were subjected to. H6: Perceived intrusiveness mediates the influence of i) personalization, ii) consumer characteristics, iii) brand familiarity iv) brand trust and v) transparency approach on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. H7: Privacy concerns mediate the influence of i) personalization, ii) consumer characteristics, iii) brand familiarity iv) brand trust and v) transparency approach on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. H8: Reactance mediates the influence of i) personalization, ii) consumer characteristics, iii) brand familiarity iv) brand trust and v) transparency approach on a) Consumers’ attitudes towards the brand, b) Consumers’ attitudes towards OBA, c) Consumers’ intention to click on the Ad, d) Consumers’ intentions to purchase the product and e) Consumers’ protective behavior. Each one of the previously mentioned mediating hypotheses is divided into twenty-five sub-hypotheses, where all the sub-hypotheses had either a partial mediation role or a no mediation role at all. First, all of the three mediators namely, Perceived Intrusiveness, Privacy Concerns, and Reactance, had no mediation effect at all between Transparency Approach and all the dependent variables. This can be traced back to the fact that consumers did not recognize/ understand the transparency question associated to the random experiment they were subject to, which may in turn have affected their answers to the survey questions. This is contradicting with previous literature that claimed online consumers appreciated honesty, where implementing transparency mechanisms had an effect on consumers’ feelings of vulnerability, which in turn affected consumers’ trust, attitudes and behaviors towards OBA (Aguirre et al. 2015). Boerman et al. 106 (2017) also claimed that including transparency mechanisms into an ad message, lead consumers to accept to click a highly personalized advertisement despite their privacy concerns and feelings of perceived intrusiveness. Second, Perceived Intrusiveness partially mediated the effect of Personalization, Consumer Characteristics, Brand Familiarity and Brand Trust on Consumers’ Attitudes towards the Brand, Consumers’ Attitudes towards OBA, Consumers’ Intention to Click on the Ad, and Consumers’ protective behavior. However, there was no mediation effect of Perceived Intrusiveness between Personalization, Consumer Characteristics, Brand Familiarity, Brand Trust and Consumers’ Intentions to Purchase the Product. This goes in line with previous literature, where the work of Doorn & Hoekstra (2013) implied that personalization leads to better consumers’ attitudes and higher click-through intentions but also to higher levels of perceived intrusiveness, these feelings of intrusiveness interrupt the cognitive processing of online customers which may lead to lower click-through rates and intentions. Hühn et al. (2017), supported that those feelings of intrusiveness generate ad annoyance and avoidance behavior among users in the online environment. While, Zhao et al. (2017) claimed that brand familiarity, and brand trust could lead consumers to experience the ads as less intrusive. Third, Privacy Concerns had no mediation effect between Personalization, Consumer Characteristics, Brand Familiarity, Brand Trust and consumers’ Protective Behavior. Also, no mediation effect was found between Consumer Characteristics, Brand Trust, Brand Familiarity and Attitudes towards OBA. Finally, no mediation effect was found between Brand Familiarity and consumers’ Intentions to Purchase the Product. While, Privacy Concerns partially mediated the effect of the rest of the independent variables on the dependent variables. This goes in line with previous literature, Xu et al. (2011) claimed that personalized ads had significant effects on consumers’ privacy concerns when consumers were presented by an ad resembling their personal information or search history. Baek and Morimoto (2012) stated that when consumers’ feelings of privacy concerns increase, this leads to increased ad skepticism, which consequently leads to more avoidance of OBA. Additionally, Fortes et al. (2017) claimed that users’ perception of privacy issues affect their attitudes and intentions to click online ads. Finally, Reactance had no mediation effect between Personalization, Consumer Characteristics, Brand Familiarity, Brand Trust and consumers’ Attitudes towards OBA. Also, no mediation effect was found between Brand Familiarity and consumers’ Intentions to Click on the Ad. However, Reactance partially mediated the effect of the rest of the independent variables on the dependent variables. This goes in line with previous 107 literature, Bleier and Eisenbeiss (2015) claimed that although personalized ads are useful, they trigger feelings of reactance, which may lead to negative behavioral outcomes towards the ad message or the advertised brand, and lower purchase intentions and click-through rates (Aguirre et al., 2015). 5.3. Theoretical Implications This study deepens the understanding of OBA concept and its effectiveness. The research aimed to contribute to the academic scholarship and give better understanding to the issue under investigation – OBA and selected variables – and its relation to the Social Exchange Theory and Phycological Reactance Theory. The research used the Social Exchange Theory (SET) and the Phycological Reactance Theory to describe consumers’ attitudes and behavior towards OBA. SET suggests that people assess social exchanges in terms of costs and benefits. Accordingly, consumers would accept OBA if they weight the benefits arising from online ads greater than its perceived cost (Schumann et al., 2014). Also, the research studied OBA as an online marketing mechanism that is based on the collection of online personal data which could result in generating feelings of reactance among online consumers, and how these feelings can affect consumers' attitudes, intentions and behaviors accordingly. In addition to building on theories, the studied variables in this research namely personalization, consumer characteristics, brand familiarity, brand trust, and transparency approach, were not studied before combined together and also their collective influence on OBA avoidance behavior is novel to literature. Previous studies also used the mediating variables to test their effect on consumers’ attitudes, intentions to click on the ad, and purchase intentions but not on protective behavior against OBA. Thus, this research provides useful information for future scholars to build on in their coming research. 5.4. Practical Implications This research provides some important insights for the online behavioral advertising practices. These insights are discussed below to help marketing professional in creating successful OBA campaigns and overcome concerns associated with OBA. First, online ad message personalization proved to be an important aspect to OBA, where personalization and personalization informativeness showed positive effects on consumers’ attitudes towards the ad and intentions to purchase the advertised product. In addition to lower OBA avoidance behavior in the online environment. However, the current research also found that OBA works better for a familiar brand rather 114 Delgado-Ballester, E., Navarro, A., & Sicilia, M. (2012). Revitalising brands through communication messages: The role of brand familiarity. European Journal of Marketing, 46 (1/2), 31-51. Dhar, S., & Varshney, U. (2011). Challenges and business models for mobile location-based services and advertising. Communications of the ACM, 54 (5), 121-128. Diamantopoulos, A., & Winklhofer, H. M. (2001). Index construction with formative indicators: An alternative to scale development. Journal of marketing research , 38 (2), 269-277. Diamantopoulos, A., Sarstedt, M., Fuchs, C., Wilczynski, P., & Kaiser, S. (2012). Guidelines for choosing between multi-item and single-item scales for construct measurement: a predictive validity perspective. Journal of the Academy of Marketing Science , 40 , 434-449. Digital Advertising Alliance (2011). “Your Ad Choices”, http://www.youradchoices.com Dillard, J. P., & Pfau, M. (2002). The persuasion handbook: Developments in theory and practice . Sage Publications. Dinev, T., & Hart, P. (2006). An extended privacy calculus model for e-commerce transactions. Information Systems Research, 17 (1), 61-80. Dodds, W. B., & Monroe, K. B. (1985). The effect of brand and price information on subjective product evaluations. Advances in Consumer Research., 12 (1), 8590. Dodds, W., Monroe, K., & Grewal, D. (1991). Effects of price, brand, and store information on buyers' product evaluations. Journal of Marketing Research, 28 (3), 307-319. Dolin, C., Weinshel, B., Shan, S., Hahn, C. M., Choi, E., Mazurek, M. L., & Ur, B. (2018, April). Unpacking perceptions of data-driven inferences underlying online targeting and personalization. In Proceedings of the 2018 CHI conference on human factors in computing systems (pp. 1-12). Earp, J.B., & Baumer, D. (2003). Innovative web use to learn about consumer behavior and online privacy. Communications of the ACM , 46(4), 81-83. 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. Emerson, R. M. (1976). Social exchange theory. Annual Review of Sociology , 2 (1), 335-362. Englehardt, S., Reisman, D., Eubank, C., Zimmerman, P., Mayer, J., Narayanan, A., & Felten, E. W. (2015, May). Cookies that give you away: The surveillance implications of web tracking. In Proceedings of the 24th International Conference on World Wide Web (pp. 289-299). International World Wide Web Conferences Steering Committee. EU Data Protection Directive (1995), “Directive 95/46/EC of the European Parliament and of the Council of 24 October 1995 on the Protection of Individuals with Regard to the Processing of Personal Data and on the Free Movement of Such Data”, Official Journal of the European Communities, 38 (281), 31–50. EU General Data Protection Regulation (2016), “Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the Protection of Natural Persons with Regard to the Processing of Personal Data and on the Free Movement of such Data, and Repealing Directive 95/46/EC (General Data Protection Regulation)”, Official Journal of the European Union, 119 (1). 115 Falk, M., & Miller, A. G. (1992). Infrared spectrum of carbon dioxide in aqueous solution. Vibrational spectroscopy , 4 (1), 105-108. Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior research methods , 39 (2), 175-191. Federal Trade Commission. (2009). Self-regulatory principles for online behavioral advertising. Federal Trade Commission Staff Report, Washington, DC. Federal Trade Commission. 2009. “FTC Staff Revises Online Behavioral Advertising Principles.” Accessed December 29, 2023. http://www.ftc.gov/opa/2009/02/behavad.shtm/. Fitzsimons, G. J., & Lehmann, D. R. (2004). Reactance to recommendations: When unsolicited advice yields contrary responses. Marketing Science, 23(1), 82-94. Gardner, L., & Leshner, G. (2016). The role of narrative and other-referencing in attenuating psychological reactance to diabetes self-care messages. Health communication, 31 (6), 738-751. Gefen, D. (2000). E-commerce: The role of familiarity and trust. Omega, 28 (6), 725737. Gill, J. and Johnson, P. (2002). Research Methods for Managers (3rd ed.). London: Sage Publications. Goldfarb, A. (2014). What is different about online advertising? Review of Industrial Organization , 44 (2), 115129. Goldfarb, A., & Tucker, C. (2011). Online display advertising: Targeting and obtrusiveness. Marketing Science, 30 (3), 389-404. Guild, B. (2013). “2013 Choice stream Survey: Consumer Opinions on Online Advertising & Audience Targeting”. Accessed January 18, 2024), http://choicestream.com/2013 Staging/wpcontent/uploads/2013/10/2013-Survey.pdf. Ha, H. Y., & Janda, S. (2014). The effect of customized information on online purchase intentions. Internet Research, 24 (4), 496-519. Hair Jr, J.F., Black, W.C., Babin, B.J., & Anderson, R.E. (2010). Advanced diagnostics for multiple regression: A supplement to multivariate data analysis. Advanced Diagnostics for Multiple Regression: A Supplement to Multivariate Data Analysis . Hair Jr, J.F., Hult, G.T.M., Ringle, C.M., & Sarstedt, M. (2017). A primer on partial least squares structural equation modeling (PLS-SEM), 2nd ed . Sage publications. Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2022). A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM) . Springer International Publishing. Hair, J. F., Ringle, C. M., & Sarstedt, M. (2011). PLS-SEM: Indeed a silver bullet. Journal of Marketing theory and Practice , 19 (2), 139-152. Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. European business review , 31 (1), 2-24. Ham, C. D., & Nelson, M. R. (2016). The role of persuasion knowledge, assessment of benefit and harm, and third-person perception in coping with online behavioral advertising. Computers in Human Behavior , 62 , 689-702. 116 Ham, C. D. (2017). Exploring how consumers cope with online behavioral advertising. International Journal of Advertising, 36 (4), 632–658. Haugtvedt, C. P., & Kasmer, J. A. (2008). Attitude change and persuasion. In C. P. Haugtvedt, P. M. Herr, & F. P. Kardes (Eds.), Handbook of consumer psychology (pp. 419-435. New York, NY: Lawrence Erlbaum Associates. Heerwegh, D., Vanhove, T., Matthijs, K., & Loosveldt, G. (2005). The effect of personalization on response rates and data quality in web surveys. International Journal of Social Research Methodology , 8(2), 85-99. Helm, R., Mauroner, O., & Dowling, M. (2010). Innovation as mediator between entrepreneurial orientation and spin-off venture performance. International Journal of Entrepreneurship and Small Business, 11 (4), 472491. Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the academy of marketing science, 43 , 115-135. 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 (Vol. 20, pp. 277-319). Emerald Group Publishing Limited. Hoewe, J. (2017). Manipulation check. The international encyclopedia of communication research methods , 1-5. Hoyer, W., & Brown, S. (1990). Effects of brand awareness on choice for a common, repeatpurchase product. Journal of Consumer Research, 17 (2), 141-148. Huang, E. (2012). Online experiences and virtual goods purchase intention. Internet Research , 22(3) , 252274. Huang, E. (2012). Online experiences and virtual goods purchase intention. Internet Research, 22(3) , 252274. Hühn, A. E., Khan, V. J., Ketelaar, P., van't Riet, J., Konig, R., Rozendaal, E., ... & Markopoulos, P. (2017). Does location congruence matter? A field study on the effects of location-based advertising on perceived AD intrusiveness, relevance & value. Computers in Human Behavior . Hulland, J. (1999). Use of partial least squares (PLS) in strategic management research: A review of four recent studies. Strategic management journal , 20 (2), 195-204. Interactive Advertising Bureau (2022). “Internet Advertising Revenue Report: Full Year 2022”. Accessed April 12, 2024. https://www.iab.com/insights/internet-advertising-revenue-report-full-year-2022/ International Association of Privacy Professionals (2022). “80% of users' browser history is captured by online trackers, report finds” Accessed April 12, 2024. https://iapp.org/news/a/80-percent-of-users-browserhistory-is-captured-by-online-trackers-report/ Iqbal, U., Englehardt, S., & Shafiq, Z. (2021, May). Fingerprinting the fingerprinters: Learning to detect browser fingerprinting behaviors. In 2021 IEEE Symposium on Security and Privacy (SP) (pp. 1143-1161). IEEE. 117 Jai, T. M. C., Burns, L. D., & King, N. J. (2013). The effect of behavioral tracking practices on consumers’ shopping evaluations and repurchase intention toward trusted online retailers. Computers in Human Behavior, 29 (3), 901-909. Kalyanaraman, S., & Sundar, S. S. (2006). The psychological appeal of personalized content in web portals: Does customization affect attitudes and behavior?. Journal of Communication, 56 (1), 110-132. Kehr, F., Kowatsch, T., Wentzel, D., & Fleisch, E. (2015). Blissfully ignorant: the effects of general privacy concerns, general institutional trust, and affect in the privacy calculus. Information Systems Journal, 25 (6), 607-635. Kim, H., & Huh, J. (2017). Perceived relevance and privacy concern regarding online behavioral advertising (OBA) and their role in consumer responses. Journal of Current Issues & Research in Advertising, 38 (1), 92105. Kim, K., & Kim, J. (2011). Third-party privacy certification as an online advertising strategy: An investigation of the factors affecting the relationship between third-party certification and initial trust. Journal of Interactive Marketing, 25 (3), 145-158. Kissmetrics Blog (n.d.). Behind the scenes of behavioral advertising . Retrieved from https://blog.kissmetrics.com/behavioral-advertising/ , Accessed December 29, 2023 Komiak, S. X., & Benbasat, I. (2006). The effects of personalization and familiarity on trust and adoption of recommendation agents. MIS Quarterly , 30 (4), 941960. Kothari, C. R. (2004). Research methodology: Methods and techniques . New Age International. Kumar, V., & Gupta, S. (2016). Conceptualizing the evolution and future of advertising . Journal of advertising, 45 (3), 302-317. Kumar, V., & Pansari, A. (2016). Competitive advantage through engagement. Journal of marketing research, 53 (4), 497-514. Lambrecht, Anja and Catherine E. Tucker (2013), “When Does RetargetingWork? Information Specificity in Online Advertising”, Journal of Marketing Research, 50 (5), 561–76. Laufer, R. S., & Wolfe, M. (1977). Privacy as a concept and a social issue: A multidimensional developmental theory. Journal of Social Issues , 33 (3), 2242. Lee, S., Lee, Y., Lee, J. I., & Park, J. (2015). Personalized e-services: consumer privacy concern and information sharing. Social Behavior and Personality: an international journal, 43 (5), 729-740. Leon, P. G., Ur, B., Shay, R., Wang, Y., Balebako, R., & Cranor, L. F. (2012). “Why Johnny Can't Opt Out: A Usability Evaluation of Tools to Limit Online Behavioral Advertising,” In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems , 589-98. Leon, P. G., Ur, B., Wang, Y., Sleeper, M., Balebako, R., Shay, R., ... & Cranor, L. F. (2013, July). What matters to users? Factors that affect users' willingness to share information with online advertisers. In Proceedings of the ninth symposium on usable privacy and security (pp. 1-12). Lerner, A., Simpson, A.K., Kohno, T., & Roesner, F. (2016, August). Internet Jones and the Raiders of the Lost Trackers: An Archaeological Study of Web Tracking from 1996 to 2016. In USENIX Security Symposium . 118 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. Li, C., & Kalyanaraman, S. (2013). “I, me, mine” or “Us, we, ours?”: The influence of cultural psychology on web-based customization. Media Psychology, 16 (3), 272-294. Li, C., & Liu, J. (2017). A name alone is not enough: A reexamination of web-based personalization effect. Computers in Human Behavior , 72 , 132-139. Li, H., Edwards, S. M., & Lee, J. H. (2002). Measuring the intrusiveness of advertisements: Scale development and validation. Journal of Advertising , 31 (2), 37-47. Li, W., & Huang, Z. (2016). The research of influence factors of online behavioral advertising avoidance. American Journal of Industrial and Business Management, 6 (09), 947. Liu, Z., Min, Q., Zhai, Q., & Smyth, R. (2016). Self-disclosure in Chinese micro-blogging: A social exchange theory perspective. Information & Management, 53 (1), 53-63. Luo, X. (2002). Trust production and privacy concerns on the Internet: A framework based on relationship marketing and social exchange theory. Industrial marketing management, 31 (2), 111-118. Lynch Jr., J. G. (1982). On the external validity of experiments in consumer research. Journal of Consumer Research , 9 (3), 225-239. MacKenzie, S. B., & Lutz, R. J. (1983). Testing competing theories of advertising effectiveness via structural equation models. In Proceedings of American Marketing Association Research Methods and Causal Modeling Conference, forthcoming. MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation analysis. Annu. Rev. Psychol., 58 , 593614. Martí-Parreño, J., Bermejo-Berros, J., & Aldás-Manzano, J. (2017). Product placement in video games: the effect of brand familiarity and repetition on consumers' memory. Journal of Interactive Marketing , 38 , 5563. Matore, E. M., & Khairani, A. Z. (2020). The Pattern of Skewness and Kurtosis Using Mean Score and Logit in Measuring Adversity Quotient (AQ) For Normality Testing. International Journal of Future Generation Communication and Networking, 13 (1), 688–702. Mayer, J. R., & Mitchell, J. C. (2012, May). Third-party web tracking: Policy and technology. In 2012 IEEE symposium on security and privacy (pp. 413-427). IEEE. McDonald, A., & Cranor, L. F. (2010). Beliefs and Behaviors: Internet Users’ Understanding of Behavioral Advertising. SSRN Scholarly Paper ID 1989092. Social Science Research Network, Rochester, NY. http://papers. ssrn. com/abstract, 1989092. McStay, A. (2011), The Mood of Information: A Critique of Online Behavioral Advertising, New York: Continuum. Mikians J., Gyarmati L., Erramilli V., and Laoutaris N. (2012). Detecting price and search discrimination on the Internet. In Proc. ACM HotNets . Milne, G. R., Bahl, S., & Rohm, A. (2008). Toward a framework for assessing covert marketing practices. Journal of Public Policy & Marketing, 27 (1), 57-62. 119 Milne, G. R., Culnan, M. J., & Greene, H. (2006). A longitudinal assessment of online privacy notice readability. Journal of Public Policy & Marketing, 25 (2), 238-249. 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. Miyazaki, A. D. (2008). Online privacy and the disclosure of cookie use: Effects on consumer trust and anticipated patronage. Journal of Public Policy & Marketing, 27 (1), 19-33. Montgomery, A. L., & Smith, M. D. (2009). Prospects for personalization on the Internet. Journal of Interactive Marketing , 23 (2), 130-137. Moon, Y. (2002). Personalization and personality: Some effects of customizing message style based on consumer personality. Journal of Consumer Psychology , 12(4), 313-326. Moorman, C., Gerald, Z., & Rohit, D. (1992). “Relationships Between Providers and Users of Market Research: The Dynamics of Trust”. Journal of Marketing, 29 (3), 314–28. Murthi, B. P. S., & Sarkar, S. (2003). The role of the management sciences in research on personalization. Management Science, 49 (10), 1344-1362. Nitzl, C., Roldan, J. L., & Cepeda, G. (2016). Mediation analysis in partial least squares path modeling: Helping researchers discuss more sophisticated models. Industrial management & data systems, 116 (9), 1849-1864. 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. 73-133). New York, NY: Routledge. Norberg, P. A., Horne, D. R., & Horne, D. A. (2007). The privacy paradox: Personal information disclosure intentions versus behaviors. Journal of Consumer Affairs , 41 (1), 100-126. Nunnally, Jum C. (1978). Psychometric Theory, 2nd ed. New York: McGraw-Hill Nuttin, J. M. (1985). Narcissism beyond Gestalt and awareness: The name letter effect. European Journal of Social Psychology , 15(3) , 353-361. Okazaki, S., Li, H., & Hirose, M. (2009). Consumer privacy concerns and preference for degree of regulatory control: A study of mobile advertising in Japan. Journal of Advertising, 38 (4), 63-77. Ozcelik, A. B., & Varnali, K. (2019). Effectiveness of online behavioral targeting: A psychological perspective. Electronic Commerce Research and Applications, 33 , 100819. Phelan, C., Lampe, C., & Resnick, P. (2016). It's creepy, but it doesn't bother me. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 5240-5251). ACM. Phelps, J. E., & Hoy, M. G. (1996). The Aad‐Ab‐PI relationship in children: The impact of brand familiarity and measurement timing. Psychology & Marketing , 13 (1), 77-105. Plane, A. C., Redmiles, E. M., Mazurek, M. L., & Tschantz, M. C. (2017). Exploring user perceptions of discrimination in online targeted advertising. In 26th USENIX Security Symposium (USENIX Security 17) (pp. 935-951). 120 Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of applied psychology , 88 (5), 879-903. Porter, S. R., & Whitcomb, M. E. (2003). The impact of contact type on web survey response rates. Public Opinion Quarterly , 67(4), 579-588. Purwanto, A., & Sudargini, Y. (2021). Exploring Factors Affecting the Purchase Intention of Halal Food Products: An Empirical Study on Student Consumers. International Journal of Social and Management Studies , 2 (4), 13-21. Raithel, S., Sarstedt, M., Scharf, S., & Schwaiger, M. (2012). On the value relevance of customer satisfaction. Multiple drivers and multiple markets. Journal of the academy of marketing science , 40 , 509-525. Reinartz, W., Haenlein, M., & Henseler, J. (2009). An empirical comparison of the efficacy of covariancebased and variance-based SEM. International Journal of research in Marketing , 26 (4), 332-344. Richardson, H. A., Simmering, M. J., & Sturman, M. C. (2009). A tale of three perspectives: Examining post hoc statistical techniques for detection and correction of common method variance. Organizational Research Methods , 12 (4), 762-800. Richins, M. L., & Bloch, P. H. (1986). After the new wears off: The temporal context of product involvement. Journal of Consumer Research , 13(2) , 280-285. Rigdon, E. E. (2014). Rethinking partial least squares path modeling: breaking chains and forging ahead. Long range planning , 47 (3), 161-167. Ringle, C. M., Sarstedt, M., & Straub, D. W. (2012). A critical look at the use of PLS-SEM in" MIS Quarterly". MIS quarterly , 36 (1), iii-xiv. Ringle, C., Da Silva, D., & Bido, D. (2015). Structural equation modeling with the SmartPLS. Bido, D., da Silva, D., & Ringle, C. (2014). Structural Equation Modeling with the Smartpls. Brazilian Journal Of Marketing , 13 (2), 56-73. Robson, C. (2002). Real world research: A resource for social scientists and practitioner-researchers . (No Title) (2nd ed.). Oxford: Blackwell. Roesner, F., Kohno, T., & Wetherall, D. (2012). Detecting and Defending Against {Third-Party} Tracking on the Web. In 9th USENIX Symposium on Networked Systems Design and Implementation (NSDI 12) (pp. 155168). Rognehaugh, R. 1999. The Health Information Technology Dictionary , Gaithersburg, MD: Aspen. Rony, M. N. (2018). Online Behavioral Advertising (OBA): The Influence of Reciprocity, Personalization, And Ad Content Type on Consumers’ attitude And Intention. Rousseau, D.M., Sitkin, S.B., Burt, R.S., & Camerer, C. (1998). Not so different after all: A cross-discipline view of trust. Academy of management review , 23(3), 393-404. Samat, S., Acquisti, A., & Babcock, L. (2017). Raise the curtains: The effect of awareness about targeting on consumer attitudes and purchase intentions. In Thirteenth Symposium on Usable Privacy and Security (SOUPS 2017) (pp. 299-319). 121 Sarstedt, M., & Mooi, E. (2014). A concise guide to market research:The Process, Data, and Methods using IBM SPSS Statistics (2nd ed) . Springer . Sarstedt, M., Hair Jr, J. F., Nitzl, C., Ringle, C. M., & Howard, M. C. (2020). Beyond a tandem analysis of SEM and PROCESS: Use of PLS-SEM for mediation analyses!. International Journal of Market Research, 62 (3), 288-299. Sarstedt, M., Ringle, C. M., & Hair, J. F. (2021). Partial least squares structural equation modeling. In Handbook of market research (pp. 587-632). Cham: Springer International Publishing. Sattler, H., Völckner, F., Riediger, C., & Ringle, C. M. (2010). The impact of brand extension success drivers on brand extension price premiums. International Journal of Research in marketing, 27 (4), 319-328. Saunders, M., Lewis, P., & Thornhill, A. (2009). Research methods for business students . Pearson education. Schumann, J. H., Von Wangenheim, F., & Groene, N. (2014). Targeted online advertising: Using reciprocity appeals to increase acceptance among users of free web services. Journal of Marketing, 78 (1), 59-75. Sekaran, U., & Bougie, R. (2016). Research methods for business: A skill building approach . John wiley & sons. Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasiexperimental designs for generalized causal inference (2nd ed.). Boston, MA: Houghton Mifflin Company. Shang, R. A., Chen, Y. C., & Liao, H. J. (2006). The value of participation in virtual consumer communities on brand loyalty. Internet Research, 16(4) , 398-418. Sheehan, K. B., & Hoy, M. G. (1999). Flaming, complaining, abstaining: How online users respond to privacy concerns. Journal of advertising, 28 (3), 37-51. Shoenberger, H. (2014). Risky business: Using heuristic-systematic processing theory to understand consumer data privacy concerns in an online behavioral advertising context (Unpublished doctoral dissertation). University of Missouri, Columbia, MO. Shoenberger, H., & Thorson, E. (2014). prediction of perceived online shopping benefits and risks from trust and knowledge of targeting. American Academy of Advertising. Conference. Proceedings (Online), 155-162. Smit, E. G., Van Noort, G., & Voorveld, H. A. (2014). Understanding online behavioural advertising: User knowledge, privacy concerns and online coping behaviour in Europe. Computers in Human Behavior, 32 , 15-22. Smith, A. D. (2006). Exploring service marketing aspects of e-personalization and its impact on online consumer behavior. Services Marketing Quarterly, 27 (2), 89-102. Smith, H., Dinev, T., & Xu, H. (2011). Information privacy research: An interdisciplinary review. MIS Quarterly, 35 (4), 989-1015. Statista [ONLINE, 2023]. (https://www.statista.com/statistics/1410056/egypt-digitalpopulation/#:~:text=As%20of%20January%202023%2C%20Egypt,those%20were%20using%20social%20me dia.). Accessed January, 20 2024. Sundar, S. S., & Marathe, S. S. (2010). Personalization versus customization: The importance of agency, privacy, and power usage. Human communication research, 36 (3), 298-322. 122 Sutanto, J., Palme, E., Tan, C., & Phang, C. (2013). Addressing the personalizationprivacy paradox: An empirical assessment from a field experiment on smartphone users. MIS Quarterly, 37 (4), 1141-1164. Tam, K. Y., & Ho, S. Y. (2006). Understanding the impact of web personalization on user information processing and decision outcomes. MIS quarterly , 865-890. Te’eni-Harari, T., Lehman-Wilzig, S. N., & Lampert, S. I. (2009). The importance of product involvement for predicting advertising effectiveness among young people. International Journal of Advertising, 28(2) , 203229. Tucker, C. E. (2014). Social networks, personalized advertising, and privacy controls. Journal of marketing research, 51 (5), 546-562. Tucker, Catherine E. (2012), “The Economics of Advertising and Privacy,” International Journal of Industrial Organization, 30 (3),326–9. Turow, J., King, J., Hoofnagle, C. J., Bleakley, A., & Hennessy, M. (2009). Americans reject tailored advertising and three activities that enable it. Available at SSRN 1478214 . Ur, B., Leon, P. G., Cranor, L. F., Shay, R., & Wang, Y. (2012, July). Smart, useful, scary, creepy: perceptions of online behavioral advertising. In proceedings of the eighth symposium on usable privacy and security (pp. 1-15). Urban, G.L., Amyx, C., & Lorenzon, A. (2009). Online trust: state of the art, new frontiers, and research potential. Journal of Interactive Marketing , 23(2), 179-190. Van Doorn, J., & Hoekstra, J. C. (2013). Customization of online advertising: The role of intrusiveness. Marketing Letters, 24 , 339-351. Van Noort, G., Smit, E. G., & Voorveld, H. A. (2013). The online behavioural advertising icon: two user studies. Advances in Advertising Research (Vol. IV) The Changing Roles of Advertising , 365-378. Verhoef, P. C., Reinartz, W. J., & Krafft, M. (2010). Customer engagement as a new perspective in customer management. Journal of service research, 13 (3), 247-252. Westin, A. F. (1967). Privacy and freedom . New York, NY: Atheneum. White, T., Zahay, D., Thorbjørnsen, H., & Shavitt, S. (2008). Getting too personal: Reactance to highly personalized email solicitations. Marketing Letters, 19(1), 39-50. Xu, H., Luo, X. R., Carroll, J. M., & Rosson, M. B. (2011). The personalization privacy paradox: An exploratory study of decision-making process for location-aware marketing. Decision support systems, 51 (1), 42-52. Yan, Z., Wang, T., Chen, Y., & Zhang, H. (2016). Knowledge sharing in online health communities: A social exchange theory perspective. Information & management, 53 (5), 643-653. Yao, Y., Lo Re, D., & Wang, Y. (2017, February). Folk models of online behavioral advertising. In Proceedings of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing (pp. 19571969). Yap, J. E., Beverland, M., & Bove, L. (2009). A conceptual framework of the causes and consequences of the privacy paradox. In The Australian and New Zealand Marketing Academy (ANZMAC) Conference 2009 Proceedings (pp. 1-9). Melbourne, VIC: Monash University. 123 Ying, L., Korneliussen, T., & Grønhaug, K. (2009). The effect of ad value, ad placement and ad execution on the perceived intrusiveness of web advertisements. International Journal of Advertising, 28 (4), 623-638. Yoon, C., & Rolland, E. (2012). Knowledge-sharing in virtual communities: familiarity, anonymity and selfdetermination theory. Behaviour & Information Technology , 31 (11), 1133-1143. Yu, J., & Cude, B. (2009). “Hello, Mrs. Sarah Jones! We recommend this product!” Consumers’ perceptions about personalized advertising: Comparisons across advertisements delivered via three different types of media. International Journal of Consumer Studies , 33(4) , 503-514. Yu, S., Hudders, L., & Cauberghe, V. (2017). Targeting the luxury consumer: A vice or virtue? A cross-cultural comparison of the effectiveness of behaviorally targeted ads. Journal of Fashion Marketing and Management, 21 (2), 187-205. Zaichkowsky, J. L. (1985). Measuring the involvement construct. Journal of Consumer Research, 12(3) , 341352. Zhao, X., Yang, J., Xie, T., & Wang, Z. (2017). Examining advertising intrusiveness on Instagram: Hedonic and utilitarian attributes of brand and sponsored content. American Academy of Advertising. Conference. Proceedings (Online), 243-255. Zhonghao Yu, Sam Macbeth, Konark Modi, and Josep M. Pujol. (2016). Tracking the Trackers. In Proceedings of the 25th International Conference on World Wide Web (pp. 121-132). Zuiderveen Borgesius, F. (2015). Improving privacy protection in the area of behavioural targeting. Available at SSRN 2654213. 130 APPENDIX II – QUESTIONNAIRE | ENGLISH 131 Consent to Participate in an Online Experiment This research is being conducted at The School of Economics and Management, University of Minho. The purpose of this research is to understand consumers’ views towards online behavioral advertising. If you agree to take part in this research, you will be redirected to a page that contains a scenario explaining the online advertisement and then asked to answer some questions about your opinion towards the ad. You will be also asked to answer some general usage and demographic questions. Please note that there are no right or wrong answers. The survey will take about 15 minutes to complete. The information collected via the questionnaire is ANONYMOUS and used for academic purposes ONLY. Thank you very much for your collaboration! I understand and agree to participate. 132 [Scenario]: Imagine you are interested in weight-loss programs and so, Last night you searched online for weight-loss programs and exercises. Next morning you opened a news website to read the morning news and you saw the following banner ad. Based on your understanding of the previous scenario, please indicate your opinion by selecting the appropriate option for each question: 1. This ad is directed to me personally [Personalization] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 2. Regarding the brand, I am: [Brand Familiarity] 1Unfamiliar 2 3 4 5 6 7Familiar 3. In general, my feeling toward advertising is: [Attitudes towards OBA] 1Bad 2 3 4 5 6 7Good 1Unpleasant 2 3 4 5 6 7Pleasant 1Unfavorable 2 3 4 5 6 7Favorable 133 The next two questions are based on your understanding of the following information: Online behavioral advertising (OBA) can be defined as: The practice of tracking individuals' online activities in order to deliver advertising tailored to the consumers' interests. 4. Do you remember seeing this type of ad in the last six months? a. Yes. b. No. 5. Approximately, how many Online Behavioral Advertising (OBA) do you see every day? a. None. b. One. c. Two. d. Three. e. Four. f. Five. g. More than five. 134 Based on your understanding of the previous scenario and OBA definition, please indicate your opinion by selecting the appropriate option for each question. 6. Thinking about the ad, I feel: [Attitude towards OBA] 1Bad 2 3 4 5 6 7Good 1Unpleasant 2 3 4 5 6 7Pleasant 1Unfavorable 2 3 4 5 6 7Favorable 7. Thinking about the brand, I feel: [Attitude towards the Brand] 1Bad 2 3 4 5 6 7Good 1Unpleasant 2 3 4 5 6 7Pleasant 1Dislike very much 2 3 4 5 6 7Like very much 1Poor Quality 2 3 4 5 6 7High Quality 8. My intention to click the ad is: [Intention to Click the Ad] 1Unlikely 2 3 4 5 6 7Likely 1Improbable 2 3 4 5 6 7Probable 1Impossible 2 3 4 5 6 7Possible 9. My intention to purchase the product is: [Intention to Purchase the Product] 1Unlikely 2 3 4 5 6 7Likely 1Improbable 2 3 4 5 6 7Probable 1Impossible 2 3 4 5 6 7Possible 135 10. I trust the advertised brand: [Brand Trust] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 11. I can rely on the advertised brand: [Brand Trust] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 12. The advertised brand is an honest brand: [Brand Trust] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 13. The advertised brand is safe: [Brand Trust] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 14. I think the ad was distracting: [Perceived Intrusiveness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 15. I think the ad was disturbing: [Perceived Intrusiveness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 16. I think the ad was forced: [Perceived Intrusiveness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 17. I think the ad was interfering: [Perceived Intrusiveness] 136 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 18. I think the ad was intrusive: [Perceived Intrusiveness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 19. I think the ad was invasive: [Perceived intrusiveness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 20. I think the ad was obtrusive: [Perceived Intrusiveness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree For the following questions, please indicate your opinion by selecting the appropriate option for each question: 21. Customized online ad gives me quick and easy access to large volumes of information. [Personalization - Informativeness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 22. Information obtained from the customized online ad is useful to me. [Personalization - Informativeness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 137 23. I learned a lot from customized online ads. [Personalization - Informativeness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 24. I think information obtained from customized online ads is helpful. [Personalization - Informativeness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 25. Customized online ad makes acquiring information inexpensive to me. [Personalization - Informativeness] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 26. I feel uncomfortable when information is shared without my permission [Privacy Concern] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 27. I am concerned about misuse of personal information [Privacy Concern] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 28. It bothers me to receive too much advertising material of no interest [Privacy concern] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 29. I feel fear that information may not be safe while stored [Privacy Concern] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 138 30. I believe that personal information is often misused [Privacy Concern] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 31. I think companies share information without permission [Privacy Concern] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 32. The message threatened my freedom to choose [Perceived threat to Choice, Reactance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 33. The message tried to make a decision for me [Perceived threat to Choice, Reactance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 34. The message tried to manipulate me [Perceived threat to Choice, Reactance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 35. The message tried to persuade me [Perceived threat to Choice, Reactance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 36. I deliberately ignore online behavioral ads when I’m surfing on the Internet. [Protective Behavior, OBA Avoidance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 139 37. I deliberately ignore online behavioral ads when I’m browsing on the shopping sites. [Protective Behavior, OBA Avoidance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 38. I deliberately ignore online behavioral ads when I open my mailbox. [Protective Behavior, OBA Avoidance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 39. Online behavioral ads are annoying. [Protective Behavior, OBA Avoidance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 40. Online behavioral ads make me feel disturbed. [Protective Behavior, OBA Avoidance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 41. Online behavioral ads are unappealing. [Protective Behavior, OBA Avoidance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree 42. If online behavioral ads pop out, I will close them. [Protective Behavior, OBA Avoidance] 1Strongly Disagree 2 3 4 5 6 7Strongly Agree