Thesis

Theses in Economics and Business The urge to splurge: Understanding the effects of media cues on impulse buying urges and behavior Anne Moes

The urge to splurge Understanding the effects of media cues on impulse buying urges and behavior

Published by University of Groningen, Groningen, The Netherlands. Printed by Ipskamp Printing. Graphic design and illustrations Camilo Garcia A. © 2025 Anne Moes. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system of any nature, or transmitted in any form or by any means, electronic, mechanical, now known or hereafter invented, including photocopying or recording without prior permission of the publisher. Deze publicatie maakt deel uit van het project I want it, I buy it met dossiernummer 023.011.008 dat mede is gefinancierd door de Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO).

The urge to splurge Understanding the effects of media cues on impulse buying urges and behavior PhD thesis to obtain the degree of PhD at the University of Groningen on the authority of the Rector Magnificus Prof. J.M.A. Scherpen and in accordance with the decision by the College of Deans. This thesis will be defended in public on Thursday 20 February 2025 at 12.45 hours born on 16 July 1987 by Anne Moes

Supervisors Prof. B.M. Fennis Prof. M.L. Fransen Co-supervisor Dr. T. Verhagen Assessment Committee Prof. B. Briers Prof. P. Kerkhof Prof. K. Van Ittersum

Content Chapter 1: General Introduction Chapter 2: In-store interactive advertising screens: the effect of interactivity on impulse buying explained by self-agency Chapter 3: A good reason to buy: Justification drives the effect of advertising frames on impulsive socially responsible buying Chapter 4: Take a deep breath: Exploring the role of mindfulness in impulse buying Chapter 5: General Discussion References Summary in Dutch (samenvatting) Acknowledgements (dankwoord) 7 17 39 63 85 99 119 125

6 The Urge to Splurge

7 Chapter 1: General introduction

8 The Urge to Splurge

9 Chapter 1 We probably all recognize the feeling of suddenly wanting to buy an item. Whether it is a comfortable pair of jeans, a delicious lemon meringue tart, or a stylish leather belt, the impulse to obtain such items can be strong. Impulse buying can be defined as unplanned behavior, accompanied with a strong sudden urge to buy (e.g., Rook, 1987; Beatty & Ferrell, 1998; Amos et al., 2014). Precisely how much is spent on impulse purchases is unclear. However, there seems to be scientific consensus that in capitalist societies such behavior is frequent. For example, Iyer et al. (2020) state that consumers impulsively spend up to $450 a month on such purchases (also see Kacen et al., 2012). Due to its magnitude, impulse buying carries significant consequences for various stakeholders. Therefore, it is interesting to know what affects these purchases. Academic consensus maintains that impulse buying can be affected by media cues (e.g., Adelaar et al., 2010; Baker Qureshi, Murtaza, & Kazi, 2019; Sharifi et al., 2023). At least three contemporary societal developments have brought forward new types of media cues that could be relevant to modernday impulse buying, but due to their novelty, these have been understudied. First, due to the continuous merging of the physical and digital world, we see an increase in the amount of so-called phygital cues, which are also seen in stores (e.g., Colombo, 2016). Phygital cues are technologies that bridge the digital and physical world, such as interactive screens. A second contemporary societal development is the increasing number of companies that have incorporated corporate social responsibility (CSR) initiatives, which have led to a rising number of CSR advertisement cues that promote a company’s charitable commitments (e.g., Coleman, Royne, & Pounders, 2019). The third development is the growing attention to mindfulness and mindful consumption (i.e., tempering excessive consumption), which has resulted in an increase of mediated mindfulness cues aimed at guiding people to become more mindful (e.g., Sheth, Sethia, & Srinivas, 2011). In the present dissertation we substantiate and measure how these three contemporary media cues (phygital cues, CSR advertisement cues, mediated mindfulness cues) may influence impulse buying urges and behavior. Urges refer to the sudden, persistent desire for an item and the tendency to buy, while impulse buying behavior refers to the actual act of buying impulsively. Moreover, in this dissertation we will also focus on how media cues affect impulse buying, by studying the self-inference processes preceding the urge to buy on impulse. Self-inference processes refer to the internal mechanisms individuals employ to make judgments about themselves (Olson & Hafer, 2013) and can, arguably, affect impulse buying. Self-inference processes relevant to impulse buying are not yet fully understood, while they hold the potential to offer important insights into the phenomenon of impulse buying (Pham et al., 2017).

10 The Urge to Splurge The central research question of this dissertation is the following: • To what extent do media cues (phygital, CSR advertising, and mediated mindfulness cues) affect consumers’ impulse buying urges and behaviors, and what is the role of self-inference processes in these potential effects? Figure 1 provides an overview of this dissertation by presenting the three contemporary societal developments and the media cues arising from them, which, arguably, affect impulse buying. Moreover, it highlights the self-inference processes that may play a role in the expected effects of media cues on impulse buying. In the remainder of this introduction, we will first shortly elaborate on the construct of impulse buying. Then we will further explain Figure 1 by discussing the three contemporary media cues that may affect impulse buying and addressing the three self-inferences processes relevant to the topic of this dissertation. At the end of this introduction, we will address the societal relevance of this research for various stakeholders, provide an outline of the empirical chapters, and describe the intended theoretical contribution of this dissertation. Figure 1: Overview of empirical chapters Impulse buying Rook (1987) argued that impulse buying is unplanned buying behavior, involving quick decision-making, and it is driven by an instantaneous desire for the relevant item. Desire is an important part of the definition, since it very clearly distinguishes unplanned purchases from impulse purchases (e.g., Muruganantham & Bhakat, 2013). This dissertation follows the idea that impulse buying is unplanned behavior, accompanied with a strong sudden urge to buy (e.g., Rook, 1987; Beatty & Ferrell, 1998; Amos et al., 2014). In addition, academics generally agree that impulse buying can be classified as hedonic (Barley & NancarChapter 2 Chapter 3 Societal development Merging of physical and digital world Increase of CSR initiatives Mindful consumption Phygital cues Interactive screens Advertising cues Self vs other- benefit frames Mediated minfulness cues Mf.instruction video Self-agency Self-justification Selfpresentation bias Media Cues Self-inference process Impulse buying urges and/or behavior Chapter 4

11 Chapter 1 row, 1998); therefore, consumers experience impulse buying urges in particular with hedonic products, such as fashionable sweaters or delicious chocolates (e.g., Gültekin & Özer, 2012). Furthermore, there is academic consensus on the fact that impulse buying can be affected by external stimuli, such as media cues (e.g., Stern, 1962; Rook, 1987; Rook & Gardner, 1993; Dutta & Mandal, 2018). In the next section, we will elaborate on the three contemporary media cues central to this dissertation that may affect impulse buying. The possible effect of contemporary media cues on impulse buying Impulse buying is often affected by external stimuli, such as media cues (e.g., Rook & Gardner, 1993). In response to contemporary societal developments, at least three new media cues are now more commonly used in society: phygital cues, corporate socially responsibility (CSR) advertising cues, and mediated mindfulness cues. In this section, we will discuss 1) The societal developments from where the studied contemporary media cues arise; 2) How these cues potentially affect impulse buying; and 3) The potential role of self-inference processes in these effects (also see Figure 1). Furthermore, this paragraph outlines the three main objectives of this dissertation. Phygital cues The continuous merging of the physical and digital world is a first contemporary societal development relevant to impulse buying. Until recently, characteristics of online stores, such as scrolling through what appears to be an unlimited selection of products at one’s own pace (e.g., Abdelsalam et al., 2020), seemed to be reserved for online retailers only. However, due to the continuous merging of the digital and physical world, it is now possible to transfer these benefits of online stores to physical ones through so-called phygital cues. Phygital cues are technologies that bridge the digital and physical world, such as the placement of interactive screens in brick-and-mortar stores. The implementation of interactive screens in stores is expanding (Wang, 2021). The literature shows that interactivity levels of these interactive screens are related to persuasive outcomes, like impulse buying, in online contexts (e.g., Kim & LaRose, 2004; Huang, 2016; Yim et al., 2017; Hu and Wise, 2021). However, it is unclear whether this interactivity also affects impulse buying in a physical surrounding. Therefore, we will first study the effect of interactive screens in physical stores on impulse buying urges. The literature suggests that consumers are easier to persuade through high (versus low) interactive content, since they feel freer to make their own choices. This is in line with the Reactance Theory (Brehm, 1966), which states that people are easier to persuade when they feel enabled to make their own choices (e.g., through interactive content; see Hu and Wise, 2021) than when they feel threatened in this freedom (e.g., by non-interactive ads; see Edwards et al., 2002). Therefore, we argue that the self-inference process “self-agency” functions as a mediator in the suggested effect of interactive screens on impulse urges. This means that we expect that interactive screens affect self-agency, and self-agency affects impulse buying urges. Self-agency refers to the feeling that people shape their own actions and influence their outcomes; therefore, it is linked to freedom

12 The Urge to Splurge of choice (Barlas and Obhi, 2013). We will introduce the term self-agency paradox which entails that individuals who feel they are in charge may actually be more vulnerable to persuasion (Kang and Sundar, 2016) and, consequently, could be more susceptible to impulse buying. In other words, increased perceptions of being in charge may actually foster less deliberative, more impulsive purchasing. The first goal of this dissertation is, therefore, the following: Goal 1: To examine the effect of interactive screens on consumers’ impulse buying urges and gain insight into the role of self-agency in this effect. CSR advertising cues A second contemporary societal development, which is potentially relevant to consumers’ impulse buying behavior, is the increasing number of companies that have started corporate social responsibility initiatives. Likewise, consumers also increasingly expect companies to act socially responsible. Consequently, the number of advertisements promoting a company’s corporate social responsibility (CSR) activities has grown (Coleman, Royne, & Pounders, 2019). It is unclear whether such advertising cues can foster impulsive sustainable purchases. Therefore, it is interesting to study the effect of two types of advertising cues on impulse purchases from socially responsible companies: 1) self-benefit frames, emphasizing the benefits for the consumer (e.g., “These chocolates will make you happier”) or 2) other-benefit frames, emphasizing the benefits for the social or environmental cause that the company supports (e.g., “These chocolates empower cacao farmers”). There is a growing body of literature that acknowledges that these two types of advertising cues prompt different responses (e.g., Ryoo, Sung, & Chechelnytska, 2020). However, it is unclear which advertising cue elicit more impulse purchases (e.g., Yadav, 2016; Jäger & Weber, 2020). Therefore, we will study the effect of self-benefit versus other benefit advertising frames on impulse buying urges and behavior. The literature suggests that impulse buying behavior often clashes with personal goals, such as eating healthy or saving. When feeling an impulsive urge to buy, people may experience a conflict between desire (vice – to buy) and reflection on long-term consequences (virtue – pursuing personal goals) (e.g., Hoch & Loewenstein, 1991; Lades, 2014; Fenton-O’Creevy, Dibb, & Furnham, 2018). This conflict can result in feelings of cognitive dissonance (Festinger, 1957), which refers to the uncomfortable feeling that one experiences when attitudes, believes, and behavior do not align. Self-justification potentially alleviates cognitive dissonance and, therefore, helps consumers to feel good about their impulse buying (Okada, 2005). Self-justification is the search for reasons or excuses for one’s own actions (Park & Hill, 2018). We will argue that the self-inference process of self-justification possibly functions as a mediator in the suggested influence of CSR advertising frames on impulse buying, meaning that we expect that advertising cues affect self-justification, which in turn, affect impulse buying. Therefore, the second goal of this dissertation is the following: Goal 2: To examine the effect of advertising frames from socially responsible companies on consumers’ impulse buying urges and behavior and gain insight into the role of self-justification in this effect.

13 Chapter 1 Mediated mindfulness cues A third societal development which is relevant to impulse buying behavior, is the attention for meaning-orientated consumption, such as mindful consumption (e.g., Sheth et al., 2011; Jain et al., 2023). When consuming mindfully, people not only take care of themselves, but also support the environment. This often results in tempering excessive consumption (Sheth et al., 2011). Even though impulse purchases can help retailers to thrive (Mehra et al., 2017) and provide consumers with short-term positive emotions (e.g., Verplanken & Sato, 2011), the negative consequences of such purchases should not be overlooked. Impulse buying often conflicts with consumers’ personal goals. As a result, the brief euphoria of buying on impulse can quickly be replaced by feelings of regret (e.g., Skelton & Allwood, 2017), guilt, or shame (e.g., Yi & Baumgartner, 2011). Many impulsively bought items are, perhaps therefore, thrown out within a short period of time after purchase or are stored somewhere in the back of a closet, attic, or garage (Boersma, 2020). This behavior not only harms consumers but also has serious negative environmental implications due to the waste of resources and transportation (Boersma, 2020). As a response to this dark side of consumption, both society and science are increasingly paying attention to mindful consumption (e.g., Sheth et al., 2011; Jain et al., 2023). In addition, the number of people who practice mindfulness in Western societies has grown substantially (Arthington, 2016). A growing body of research suggests that mindfulness seems promising in reducing impulse buying (e.g., Rosenberg, 2004; Bahl et al., 2016) by showing a negative correlation between trait mindfulness and impulse buying trait (e.g., Dhandra, 2020). However, the effect of mindfulness on impulse buying is not clear. Some studies discuss that mindfulness could perhaps even lower the threshold for giving in to one’s desire (to buy) (Friese & Hoffman, 2016). Therefore, we will study the influence of mediated mindfulness cues, such as online guided mindfulness meditations, on impulse buying urges and behavior. Although mindfulness is often referred to as a seemingly promising phenomenon when it comes to decreasing impulse buying, we will also address the possible risk of self-presentation bias in these studies on mindfulness and impulse buying. Self-presentation bias is the difference between what people say they do when asked about their behavior (in self-reported data) and how they actually act. We will approach self-presentation bias as a possible distorting self-inference process that can cause misleading outcomes in research on mindfulness and impulse buying. Therefore, the third and final goal of this dissertation can be formulated as follows: Goal 3: To examine the effect of mediated mindfulness instructions on consumers’ impulse buying urges and behavior and critically discuss the role of self-presentation bias in research on mindfulness and impulse buying. In sum, societal developments have introduced new commonly used media cues, which may affect impulse buying urges and behavior. We will study the effects on impulse buying of three media cues that are most relevant to the current times, namely phygital cues, CSR advertising cues, and mediated mindfulness cues. In the final section of this introduction, we will address the societal and scientific relevance of this dissertation, along with the outline of the empirical chapters.

14 The Urge to Splurge Contributions and outline empirical chapters In the first empirical chapter, we study the impact of interactive screens in store windows (phygital cues) on impulse buying urges. We argue that high interactivity levels lead to stronger impulse buying urges than low interactivity levels through self-agency. This chapter is particularly relevant for traditional retailers. Due to strong online competition from large retail chains, traditional physical retailers are increasingly dependent on impulse purchases (Mehra et al., 2017). Arguably, phygital cues, such as interactive screens, could help traditional physical retailers enhance impulse buying and strengthen their competitive position toward (online) competitors. The second empirical chapter studies the influence of self-benefit versus other benefit frames (CSR advertising cues) on impulse buying urges and behavior. We argue that other-benefit frames lead to more impulse buying compared to self-benefit frames and that self-justification explains this possible effect. This chapter is highly relevant to socially responsible entrepreneurs. Although growing, the market share of responsibly made products is still relatively small (d’Astous & Legendre, 2009; Walk-Morris, 2023). More clarity on which advertising cues socially responsible companies can use to increase impulse purchases can help such companies -and the causes they support- to thrive. Finally, the third empirical chapter addresses the effect of mindfulness instruction videos (mediated mindfulness cues) on impulse buying urges and behaviors. We elaborate on mindfulness as a seemingly promising phenomenon when it comes to decreasing impulse buying. However, we will also discuss why mindfulness could potentially increase impulse buying and address the possible risk of self-presentation bias in existing studies on mindfulness and impulse buying. We argue that self-presentation bias can explain the contradictory results in some previous studies and underscore the importance of using measurements that are less sensitive to self-presentation bias when studying mindfulness and impulse buying. This chapter is particularly relevant to consumers and the environment, since it might provide insights into how to consume less on impulse. Moreover, this dissertation aims to contribute to the marketing and consumer psychology literature, and more explicitly to the literature on impulse buying behavior, in at least three ways. First, we offer insights into the effects of promising contemporary media cues, which has not been previously studied in relation to impulse buying. Second, we extend the existing knowledge of the psychological processes that might explain impulse purchases. The literature indicates that both self-agency (e.g., Sundar, 2008) and self-justification (e.g., Okada, 2005) might function as such an explaining mechanism; however, this is the first research to actually study them as mediators in the effect of a stimulus on impulse buying. Third, we contribute by critically discussing the use of self-reported data in studies on impulse buying and by measuring actual impulse buying behavior in two of the three empirical chapters. Taken as a whole, this dissertation studies the effect of three media cues on impulse buying and aims to examine the role of three self-inference processes in these possible effects (see Figure 1 for a visual overview of each chapter). We aim not only to extend the literature on impulse buying but also to provide relevant insights for society.

16 The Urge to Splurge

17 Chapter 2: In-store interactive advertising screens: the effect of interactivity on impulse buying explained by self-agency1 1 Adapted from: Moes, A., Fransen, M., Fennis, B., Verhagen, T. and van Vliet, H. (2022), «In-store interactive advertising screens: the effect of interactivity on impulse buying explained by self-agency», Journal of Research in Interactive Marketing, Vol. 16 No. 3, pp. 474-457. https://doi.org/10.1108/JRIM0097-2021-03

18 The Urge to Splurge

19 Chapter 2 We are steering the wheel of our own lives. We walk on the streets, where our will thrives. The more we feel that way, the less it is true. Interacting with screens, already will do. In this first empirical chapter we study the effect of interactive screens in store windows (as phygital cues) on impulse buying urges and to what extent self-agency functions as a mediator in this relationship. This chapter is a response to the contemporary societal development where the physical and digital world continuously merge.

20 The Urge to Splurge

21 Chapter 2 Abstract: Introduction Physical stores are increasingly dependent on impulse visits and the impulse purchases of passers-by. Interactive advertising screens in store windows could help retailers increase impulse-visit urges and impulse-buying urges. However, the effects of interactive screens in physical surroundings have not been studied before. Therefore, this study aimed to examine the effect of interactive screens on impulse urges and gain insight into the underlying mechanism that explains the possible effect. An interactive screen was placed in a store window. Using three field experiments, we studied the effect of interactivity-level (high vs low) on the impulse-visit and impulse-buying urges of passers-by, and the mediating role of self-agency in these effects. Highly interactive (compared to less interactive) advertising screens in store windows positively affect impulse-visit and impulse-buying urges through self-agency. Retailers can therefore use interactive advertising screens to increase the number of impulse purchases if feelings of self-agency are activated. This is the first study to examine the extent to which interactive screens in a store window enhance the impulse-visit and impulse-buying urges of passers-by and the mediating factor of these effects. By conducting three field experiments, we achieved a high external validity and managed to share very reliable results owing to the replication of the findings. The digital and physical worlds are increasingly merging in retail settings (Brynjolfsson et al., 2013). Typical online functionalities, such as interactive screens, have started entering physical store environments (e.g., Macy’s, Adidas, and Coca-Cola ; Pantano, 2016). The growing implementation of in-store interactive screens is, arguably, a logical result of 1) the habituation of consumers to shop and communicate through digital screens and other interactive in-store channels, 2) the increasingly active role that consumers play in the communication between brands and consumers, and 3) the fact that interactivity has become a crucial element in marketing practice (Wang, 2021). Remarkably, the effects of such interactive screens in physical storefronts on persuasive outcomes, such as impulse buying, are unclear. Physical stores depend more than ever on impulse visits and impulse purchases of passers-by, owing to the increased competition with online shops (Mehra et al., 2017). Therefore, retailers should use new marketing stimuli, such as interactive innovations (e.g., Berry et al., 2010), to enhance impulse visits and impulse buying (Iyer et al., 2020). Based on a qualitative study, Pantano (2016) suggests that compared to online stores, offering interactive content in shopping windows may improve the competitive positions of physical stores. Although research on interactivity and impulsivity is highly valuable for both researchers and practitioners, the effects of interactive content in storefronts on such visits and purchases have not yet been examined. Therefore, this study aims to answer the following two questions: Can interactive screens in store-windows trigger impulse visits and impulse-buying urges? Moreover, what mechanism could explain this effect if so? By ascertaining the above, this study makes a unique contribution to the interactive-marketing literature and responds to the 2https://www.ecommercenews.nl/winkels-proberen-interactieve-touchscreens/, retrieved 20 March 2019

22 The Urge to Splurge notion that more research on newly developed technologies in interactive marketing is necessary (Wang, 2021). Based on previous research on interactivity in online contexts, interactive screens could arguably increase impulse-visit urges and impulse-buying urges (e.g., Kim and LaRose, 2004; Huang, 2016; Yim et al., 2017; Hu and Wise, 2021). However, these studies state either that interactivity and impulsivity are related to each other or that interactivity can positively affect persuasive outcomes in general. They do not provide any insights into the causal effect of interactivity on impulsivity. This study, however, conducts three field experiments and can therefore draw valid conclusions on the effect of interactivity on impulse-buying urges. Moreover, the abovementioned studies were conducted in an (experimental) online setting. Online shopping motivations and behavior do not necessarily correspond with offline shopping motivations and behavior (Haridasan and Fernando, 2018); therefore, it is still unclear whether the results of online studies hold in physical surroundings. Furthermore, in-store shoppers (compared to online shoppers) are known to value interaction (Haridasan and Fernando, 2018). It is therefore remarkable that the effect of interactivity on impulse buying has not been studied in the context of physical shopping. Accordingly, the first aim of this study is to examine whether interactive screens in store windows positively affect impulse-visit urges and impulse-buying urges. This could provide relevant insights literature on interactive-marketing and for practice (Pantano, 2016). This study’s second aim is to gain insight into the underlying mechanisms that explain the possible effects of interactive screens on impulse-visit urges and impulse-buying urges. Although previous research on underlying mechanisms of impulse buying contributed significantly to the body of knowledge (e.g. Styvén et al., 2017), many studies assert that the antecedents, manifestations, underlying processes, and consequences of impulse buying, particularly as they pertain to interactive choice contexts, are still poorly understood (Pham et al., 2017). This study aims to clarify which psychological state, triggered by interactivity, enhances impulse visits and impulse purchases. The agency model of customization (AMC; Sundar, 2008) suggests that interactive techniques could possibly enhance self-agency, which could increase positive attitudes toward displayed content. We will explore, with three field experiments in an Amsterdam-based clothing store, the extent to which self-agency contributes to the effects of interactivity on impulse-visit urges and impulse-buying urges. Herewith, we contribute to the interactive-marketing literature since self-agency is not studied before as a possible mediator that could explain the positive effects of interactivity.

23 Chapter 2 Theoretical Background and Hypotheses Development Impulse visits and impulse buying Impulse purchases can be defined as unplanned purchases accompanied by a sudden and strong urge to buy (Rook, 1987; Amos, 2014). Building on this definition, we define impulse visits as the unplanned entry into a physical store accompanied by a sudden strong urge to do so. A meta-analysis on impulse buying (Amos, 2014) shows that antecedents of impulse buying can be categorized into dispositional factors (e.g., impulse-buying trait), sociodemographic factors (e.g., age), and situational factors (e.g., retail environment). Additionally, research shows that internal perceptions (e.g., time pressure), product characteristics (e.g., hedonic products), and product’s promotion factors (e.g., discounts) affect impulse buying (see Khan et al., 2015). More specially, for the impulse buying of fashion items, Lee and Johnson (2010) stress the effect that store layout has on consumers’ impulse buying. Compared to impulse buying, impulsive store visits are studied less often. However, it is known that store visits and impulse buying are often triggered by the same antecedents. For example, both store visits (Beak et al., 2020) and impulse buying (Sun and Yazdanifard, 2015) are stimulated by sensory experiences. It can be concluded that various antecedents affect both impulse buying and store visits. However, studies have not yet identified the effect of in-store interactive advertising screens on impulse-visit urges and impulse-buying urges. Interactive advertising screens can be classified under user-machine or user-message interactions, depending on the screen’s features. In user-machine interaction, the emphasis lies on human interaction with computers. User-message interaction entails the possibility of users modifying a message (Liu and Shrum, 2002). Based on different interactivity types, Liu and Shrum (2002, p. 54) propose the following definition of interactivity: “The degree to which two or more communication parties can act on each other, on the communication medium, and on the messages and the degree to which such influences are synchronized.” They also suggest that interactivity encompasses three dimensions: active control, twoway communication, and synchronicity (also see Liu, 2003). Even though mixed results on the effect of interactivity have been found (Wu, 2005), the notion that interactivity can affect persuasive outcomes for the better seems widely accepted (e.g., Chattaraman et al., 2014). For instance, studies demonstrate that interactivity positively affects constructs such as consumers’ perceptions of usefulness and enjoyment (Yim et al., 2017), attitudes (Hu and Wise, 2021), and online buying activity (Kim and LaRose, 2004). A meta-analysis of 63 studies on web interactivity underlines this notion and shows that it is positively correlated with attitudes and desirable behavioral intentions of consumers (Yang and Shen, 2018). Additionally, Kakalejčík et al. (2020) show that positive experiences between companies and consumers with respect to online interactions can enhance consumers’ (re)visit behavior and purchase behavior. Correspondingly, Effect of interactivity on impulse urges

24 The Urge to Splurge Bressolles et al. (2007) argue that website interactivity benefits feelings of gratification and could therefore induce buying impulses. They found that interactivity and impulse-buying buying urges are positively related. Additionally, through a survey, Huang (2016) showed that online interactive activity (browsing) and impulse buying are related. It is interesting to study the extent to which the positive effects of interactivity in online settings will hold in offline settings, as online consumers differ from offline consumers. The former seek variety (e.g., Donthu and Garcia, 1999) and convenience (Monsuwé et al., 2004) more often than the latter. Additionally, some consumers prefer to interact in an online retail setting over interacting in an offline retail setting (Becker and Pizzutti, 2017). Moreover, some consumers can experience more risk when shopping online than when shopping offline (see Kim et al., 2020). These differences could imply that the positive effects of online interactivity do not necessarily apply to in-store interactivity. However, this study examines the effect of in-store interactive digital screens, which are in functionality and use comparable with interactive digital screens that are used for online shopping, such as laptop-screens. Therefore, we also expect a positive effect of interactivity as a functionality of digital screens in physical shop windows on impulsive consumer behavior. Impulsive consumer behavior may manifest itself in various ways. In general, one may argue that such behavior manifests itself in terms of experiential shopping motives and behavior (where the experience itself is the goal of the behavior), as well as materialistic motives and behavior (where the acquired product is the goal of the behavior, see e.g., Carter and Gilovish, 2012). In the current context such experiential behavior translates into impulse store visits while materialistic behavior translates into impulse buying. There is no a priori reason to suspect that interactivity affects both types of impulsive consumer behavior differently (see Prediger et al., 2019). However, considering marketing implications, a separate examination of the effect of interactivity on both types is interesting. A visit does not necessarily result in a purchase (Vukadin et al., 2016). Likewise, with the possibility of online purchases store visits are no longer essential to impulse buying. We, therefore, explicitly distinguish impulse visits from impulse purchases and propose the following two hypotheses: We further study the extent to which the positive effects of interactivity can be explained by self-agency. Self-agency refers to the feeling that one shapes her/his own actions and, therefore, is frequently intertwined with freedom of choice (Barlas and Obhi, 2013). According to Reactance Theory (Brehm, 1966), people are easier to persuade when they feel free to make their own choices (e.g., through interactive Self-agency as a mediator in the effect of interactivity on impulse urges H1a: High interactivity leads to stronger impulse-visit urges than low interactivity. H1b: High interactivity leads to stronger impulse-buying urges than low interactivity.

25 Chapter 2 content, see Hu and Wise, 2021) than when they feel threatened in this freedom (e.g., by non-interactive ads; see Edwards et al., 2002). Correspondingly, Hu and Wise (2021) showed that the interactive elements of playable ads (indirectly) reduced perceived freedom threat, and therefore decreased consumers’ resistance to the ads. Self-agency has been found to (unconsciously) strengthen a sense of authorship (e.g. Aarts et al., 2009), empowerment, and mastery (e.g. Chiu et al., 2013) and may therefore be expected to represent the opposite of freedom threat. In line with Hu & Wise, we will examine to what extent interactivity enhances feelings of self-agency and, subsequently, if self-agency affects impulse buying. Customization is known to evoke feelings of self-agency (Sundar, 2008). The possibility of customization allows consumers to actively to adjust the content on, for example, an interface. Therefore, this option makes users part of the communication process rather than just the receivers, which, in turn, enhances feelings of self-agency (Sundar, 2008). Without implying that customization and interaction are the same, it could be argued that the option to change the communication process actively (e.g., changing the interface) plays an important role in both constructs. Furthermore, although it has not provided proof, the AMC (Sundar, 2008) proposes that interactivity techniques may enhance agency (also see Sundar and Marathe, 2010, p. 304). This agency model also posits that self-agency results in positive attitudes toward the displayed content (Sundar, 2008). Attitudes toward products are known to correlate with impulse-buying behavior (Chen, 2008), indicating that self-agency may lead to higher impulse-visit urges and impulse-buying urges. Moreover, the literature reveals that people who experience a high sense of self-agency perceive messages as more important and are easier to persuade, to visit and buy impulsively for example, than people who experience a low sense of self-agency (Kang and Sundar, 2016), corresponding with Reactance Theory. Similar as the reasoning for hypothesis 1a and 1b, there are no a priori reasons to suspect that interactivity affects both types of impulsive consumer behaviors through different processes (see e.g., Prediger et al., 2019). Based on the preceding, the following hypotheses are formulated (see Figure 2 for the conceptual model): Figure 2: Conceptual model hypotheses. H2a: Self-agency mediates the effect of interactivity (high versus low) on impulse-visit urges. H2b: Self-agency mediates the effect of interactivity (high versus low) on impulse-buying urges. Interactivity Self-agency Impulse-visit urge Impulse-buying urge + + + H2a H1a H2b H1b + +

26 The Urge to Splurge We conducted a series of three field experiments (Experiments 1, 2a, and 2b) to test the hypotheses. All experiments have a one-factor between-subject design with ‘level of interactivity’ (low versus high) as the independent variable and ‘impulse-visit urge’ and ‘impulse-buying urge’ as dependent variables. We tested the hypotheses at a significance level of .05. The experiments have a total sample size of 436 consumers. The required sample size was calculated a priori with G*Power in the F tests family. The expected effect size (f2) we used to calculate the sample size was .15 since previous research found an average effect size of .17 when measuring the effect of marketing stimuli on impulse-buying behavior (Iyer et al., 2020). As there is just one predictor in our studies (interactivity), a total sample size of at least 89 participants per study was needed according to G*Power tool, a requirement that we have met in all experiments. The first experiment is both confirmative and explorative in nature since some additional exploratory variables were examined next to the hypotheses testing (see Experiment 1, ‘Measurements’). Both the second experiment and third experiment are solely confirmative in nature and were performed to check whether the results of Experiment 1 could be replicated. Replication is often undervalued, although of great importance for gaining confidence in the found results (McEwan et al., 2018). The data for Experiment 1 were collected in one week, the data for Experiments 2a and 2b were collected in two weeks each. All three experiments had an approximately equal number of participants per day, except Sundays, when the shop was closed. All experiments included a manipulation check. Below we describe the method, results, and discussion for each experiment. Overview of Studies An interactive screen (42 inches) was placed in the window of a women’s clothing store located in a well-visited, high-end shopping street in Amsterdam, Netherlands. The content and functionalities of the screen were specifically designed for the three experiments. In line with Liu and Shrum’s (2002) conceptualization, screen interactivity was manipulated through active control, two-way communication, and synchronicity. In both conditions, participants had to press play to start interacting with the screen. In the high-interactivity condition, participants could interact with the screen by swiping through the store’s collection and zooming in on the displayed items (active control). Users could also like products by pressing on a ‘thumbs up’ and interact with the screen by leaving a comment. Both were followed by a short reaction displayed on the screen (two-way communication). Additionally, in the high-interactivity condition, the screen responded immediately to the user’s actions (synchronicity). In the low-interactivity condition, participants had fewer interaction options. After pressing play, they could only reverse the order in which items were shown instead of swiping through them. They could not zoom-in, like, or rate any products or the screen itself, and the response time of the screen on the user’s action was delayed by a full second. The two conditions differed in the interactive features only. All other elements, such as content and size, were identical (see image A). Method. Stimulus material. Experiment 1

27 Chapter 2 Image A: impression of the conditions. Image 1: Low-interactivity condition after the participant pressed play, where levels of control, two-way communication, and synchronicity were kept low. Image 2: High-interactivity condition after the participant pressed play, where levels of control, two-way communication, and synchronicity were higher. Participants. All participants (N = 102) were adults (Mage = 48.42, SD = 19.96) and passers-by of the store where the interactive screen was located. Only women were approached to participate since the store in question only sells women’s clothing and therefore mostly sells to women. The participants fitted the natural group that could possibly be induced to buy impulsively at the store. Procedure. The condition the participants were exposed to depended on the day they were asked to participate. The conditions were altered every day and week (also between the three experiments) to counterbalance the possible effect that ‘day of the week’ could have on the dependent variables. We hired student assistants to collect the data in the shopping street where the store was located. They attended a short training in advance, where they were trained on how to approach passers-by and what procedure to follow during data collection. After reading and signing an informed consent that, among other things, stated that participation was voluntary, participants were asked to use the screen in the store window. Since Experiment 1 was a partially explorative study, the participants were either asked to interact with the screen for one minute or for as long as they pleased, so it could be determined if a set interaction time would lead to different results than with no set time. Next, the participants filled in the questionnaire to measure the dependent and mediating variables. Afterward, they could voluntarily leave their email addresses to participate in a raffle to win a voucher worth 50 euros for the store in question. Last, all participants were debriefed and thanked for their participation.

28 The Urge to Splurge Measurements. All constructs were measured using established measurement instruments from previous literature, on seven-point Likert scales (1 = totally disagree and 7 = totally agree). Manipulation check. More interactive options do not necessarily mean higher perceptions of interactivity (Voorveld et al., 2011). We further measured active control, twoway communication, and synchronicity (Liu and Shrum, 2002), and overall perceived interactivity to ascertain whether the high-interactivity condition resulted in more feelings of interactivity than the low-interactivity condition.3 See Table 1 for the manipulation-check measurements for each experiment. 3 Since the Cronbach’s alpha of the two-way communication scale and the synchronicity scale appeared to be insufficient in the first experiment, we proceeded the manipulation check analyses in Experiment 1 only with the active-control construct (Cronbach’s alpha = .77). The low Cronbach’s alpha for ‘two-way communication’ and ‘synchronicity’ were likely caused by the fact that many participants overlooked some of the items that measured two-way communication and synchronicity. Therefore, we made it mandatory in Experiments 2 and 3 to answer all manipulation-check questions. Table 1: Measurements manipulation checks Experiments 1, 2a, and 2b. Active control Two-way communication Synchronicity Overall perceived interactivity Construct Items Cronbach’s Alpha 1/2a/2b Based on scales used by: I felt that I had control over the screen; While interacting with the screen, I could browse at my own pace; While interacting with the screen, I had absolutely no control over what I could do with the screen.*^ It is impossible to share my opinion via the screen*^; The screen makes me feel it wants to listen to its visitors; The screen gives visitors the opportunity to talk back. When I clicked on the screen, I felt I was getting an instantaneous response; The screen was very slow in responding to my requests.*^ On a scale from 1 (not at all) to 7 (very), how interactive did you find the screen? .77/.75/.82 Liu and Shrum (2002). Liu and Shrum (2002). Liu and Shrum (2002). .52/.81/.70 .37/.75/.83 - - * Item was recoded in the analysis ^ Item was deleted in at least one of the studies to increase the Cronbach’s Alpha.

29 Chapter 2 Dependent variables. Impulse-buying urge was measured using a threeitem scale (explained variance = 76.26%, Cronbach’s alpha = .84). Two of the items were based on a scale used by Sultan et al. (2012), namely, ‘I feel a strong urge to buy (one of) the products that the screen displayed’ and ‘I want (one of the) products that the screen displayed.’ The third item was added validate the definition of impulse buying, where unplanned needs play an important role: ‘I did not intend to buy (one of) the displayed products, but I now have the desire to do so.’ We could not find an existing scale that measures impulse-visit urges. Therefore, we adjusted the scale we used to measure impulse-buying urges as follows: ‘I feel a strong urge to enter this store,’ ‘I want to visit this store,’ ‘I did not intend to visit this store, but now I do feel the desire to do so’ (explained variance = 80.24%, Cronbach’s alpha = .88). Mediator. Self-agency was measured using the items ‘It felt like I was responsible for what happened on the screen’ and ‘It felt like I was the one influencing what happened on the screen,’ based on a scale by Ruth et al. (2002) (explained variance = 90.01%, Cronbach’s alpha = .96). See Table 2 for an overview of all key measurements that were used per experiment. Table 2: Measurements key-constructs Experiments 1, 2a, and 2b. Impulsebuying urge Impulsevisit urge Feelings of self-agency Construct Items Explained variance 1/2a/2b Cronbach’s Alpha 1/2a/2b Based on scales used by: I feel a strong urge to buy (one of the) products that the screen displayed; I want (one of the) products that the screen displayed; I did not intend to buy (one of the) displayed products, but now I do feel the desire to do so. I feel a strong urge to enter this store; I want to visit this store; I did not intent to visit this store, but now I feel the desire to do so. It felt like I was responsible for what happened on the screen; It felt like I was the one influencing what happened on the screen. 76.26%/ 70.23%/ 74.01% .84/.79/.82 Sultan et al., (2012). Sultan et al., (2012). Ruth et al., (2002). 80.24%/ 73.40%/ 73.62% 88/.82/.82 90.01%/ 95.32%/ 87.79% .96/.91/.86 * Item was recoded in the analysis

30 The Urge to Splurge In addition to the constructs displayed in Table 2, we measured several other variables in Experiment 1 for exploratory reasons, such as feelings of product ownership and numerous emotions.4 Results. The manipulation check showed that participants significantly scored higher (t = -4.90, p < .001) on active control in the high-interactivity condition (M = 4.34) than in the low-interactivity condition (M = 2.73). This was also the case for overall perceived interactivity (M = 3.67 versus M = 2.42, t = -4.14, p < .001). The manipulation succeeded. To test H1a and H1b (high interactivity leads to stronger impulse-visit urges and impulse-buying urges than low interactivity), we performed a oneway ANOVA. Results show that both direct effects are not significant (impulse-visit urges: F (1, 96) = .10, p = .754; impulse-buying urges: F (1, 96) = 2.51, p = .259). We also tested the proposed mediation hypotheses, such that self-agency mediates the effect of interactivity (high versus low) on impulse-visit urges and impulse-buying urges, using PROCESS’s model number 4 (Hayes, 2017). These analyses show that interactivity does not affect impulse-visit urges through self-agency (H2a, 95% CI: [-.01 to .76]). Therefore, hypotheses 1a, 1b, and 2a are rejected. We did find a positive effect of interactivity on self-agency (b = 1.22, SE = .28, 95% CI: [.66 to 1.78]) and of self-agency on impulse-buying urges (b = .31, SE = .10, CI: [.11 to .51]). Subsequently, we find a positive mediation effect of interactivity through self-agency on impulse-buying urges (b = .38, SE = .18, 95% CI: [.10 to .81], effect size PM = .15). Hypothesis 2b, self-agency mediates the effect of interactivity (high versus low) on impulse-buying urges, is therefore accepted. See Table 3 for the means of the outcome variables of each experiment. 4. The first author can be contacted for a full list of these explorative variables. We did not find any significant results of these explorative variables and will not mention these further in this study. Table 3: Means and standard deviations of self-agency, impulse-visit urges, and impulse-buying urges in low interactivity and high interactivity conditions per Experiment. * Significant difference between low and high interactivity within the same Experiment. Low interactivity High interactivity Condition SA IVU IBU Exp. 1: 1.88 (1.24)* Exp. 2a: 1.75 (1.15)* Exp. 2b: 2.06 (1.40)* Exp. 1: 3.05 (1.58) Exp. 2a: 2.89 (1.56) Exp. 2b: 2.94 (1.64) Exp. 1: 3.15 (1.49) Exp. 2a: 2.67 (1.35) Exp. 2b: 3.18 (1.44) Exp. 1: 2.28 (1.27) Exp. 2a: 2.34 (1.17) Exp. 2b: 2.88 (1.56) Exp. 1: 2.60 (1.48) Exp. 2a: 2.51 (1.37) Exp. 2b: 2.50 (1.35) Exp. 1: 3.02 (1.53)* Exp. 2a: 3.54 (1.82)* Exp. 2b: 3.44 (1.84)*

31 Chapter 2 Discussion. We did not find a direct effect of interactivity on impulse urges in Experiment 1. Nevertheless, we performed mediation analysis as a direct effect of X on Y is not required in modern mediation analysis: “Statistical mediation analysis has changed since the publication of Baron and Kenny (1986). The heyday of the causal steps ‘criteria to establish mediation’ approach is over. … Modern mediation analysis emphasizes an explicit estimation of the indirect effect, … and an acknowledgment that evidence of a statistically significant association between X and Y is not necessary to talk about and model intervening variable processes …” (Hayes, 2018, p.146). The mediation analysis shows an indirect effect of interactivity on impulse-buying urges through self-agency. There are no effects found on impulse-visit urges. Since Experiment 1 was partially explorative in nature, we replicate the study in Experiment 2a and Experiment 2b to confirm the results. For Experiment 2a and Experiment 2b, the same stimulus material was used, and largely the same procedure was followed as in Experiment 1. Experiment 1 participants were asked to interact with the screen either for one minute or for as long as they pleased. Since interaction time did not influence the results, we decided to ask all participants in Experiments 2a and 2b to interact with the screen for as long as they pleased to create a more natural situation for all participants. However, we set a minimal interaction time of 40 seconds, to provide the user with enough time to explore the condition. Participants. The samples of Experiments 2a (N = 153) and 2b (N = 190) were similar to those of Experiment 1. The mean ages in Experiment 2a and 2b were 47.69 (SD = 17.79) and 49.37 (SD = 19.12), respectively. All participants were again adult passers-by and female. Seven of the participants in Experiment 2b had already participated in Experiments 1 or 2a and were therefore excluded from further analyses (N = 183). Measurements. Manipulation check. We used the same manipulation check as in Experiment 1 (see Table 1). In Experiment 2a active control had a Cronbach’s alpha of .75, two-way communication of .81, and synchronicity of .75. In Experiment 2b active control had a Cronbach’s alpha of .82, two-way communication of .70, and synchronicity of .83. Dependent variables and mediators. As in Experiment 1, we measured impulse-buying urges (Experiment 2a: explained variance = 70.23%, Cronbach’s alpha = .79; Experiment 2b: explained variance = 74.01%, Cronbach’s alpha = .82), impulse-visit urges (Experiment 2a: explained variance = 73.40%, Cronbach’s alpha = .82; Experiment 2b: explained variance = 73.62%, Cronbach’s alpha = .82), and self-agency (Experiment 2a: ex- Method. Stimulus material & procedure. Experiment 2a & Experiment 2b

RkJQdWJsaXNoZXIy MjY0ODMw