Persuasive avatars: The effects of customizing a virtual salesperson׳s appearance on brand liking and purchase intentions

Author's Accepted Manuscript

Persuasive Avatars: The Effects of customizing a virtual salesperson's appearance on brand liking and purchase intentions Michael D. Hanus, Jesse Fox

www.elsevier.com/locate/ijhcs

PII: DOI: Reference:

S1071-5819(15)00121-4 http://dx.doi.org/10.1016/j.ijhcs.2015.07.004 YIJHC1973

To appear in:

Int. J. Human-Computer Studies

Received date: 16 October 2014 Revised date: 11 May 2015 Accepted date: 21 July 2015 Cite this article as: Michael D. Hanus, Jesse Fox, Persuasive Avatars: The Effects of customizing a virtual salesperson's appearance on brand liking and purchase intentions, Int. J. Human-Computer Studies, http://dx.doi.org/10.1016/j. ijhcs.2015.07.004 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Running Head: CUSTOMIZING A VIRTUAL SALESPERSON

1

Persuasive Avatars: The Effects of Customizing a Virtual Salesperson’s Appearance on Brand Liking and Purchase Intentions

Michael D. Hanus*, Jesse Fox The Ohio State University

*

Corresponding Author: Michael Hanus Present Address: 808 Riverview Dr, #6A, Columbus OH 43202 Phone: (859) 312-9520 [email protected]

The Ohio State University Derby Hall DB Building 025 154 North Oval Mall United States, Columbus, OH 43210 [email protected] [email protected] Abstract Transformed social interaction (TSI) suggests that virtual environments have unique advantages over traditional forms of interpersonal communication. Recent research has demonstrated that a persuasive speaker can use these advantages to create a more persuasive message. However, little has been done to establish how a receiver of a persuasive message in a virtual, interactive environment might also use these advantages. Giving individuals the ability to customize a persuasive source may empower them and affect their perceptions of the persuasive message. In a 2 (forewarned or not) x 2 (customize source or watch customization) experiment, participants were forewarned or not forewarned that they would be hearing a persuasive message about buying a new energy drink. Participants were then allowed to customize (or watch someone customize) the appearance of the speaker’s avatar before the sales pitch. Participants who had the

CUSTOMIZING A VIRTUAL SALESPERSON

2

opportunity customize the speaker’s appearance liked the product more and had higher purchase intentions than those who did not have control, regardless of forewarning. Findings suggest a new application for TSI, and that both users and sources may benefit when users have the power to customize the appearance of a virtual salesperson.

Highlights • Virtual environments (VEs) offer new opportunities for persuasion • Message sources and receivers can utilize affordances of interactive media • Allowed users in VE to customize avatar of a persuasive speaker before speech • Customizers liked product and had higher purchase intention than non-customizers

Keywords Persuasion, avatars, intrinsic motivation, customization, brand attitudes, virtual environments

1. Introduction As digital technology continues to grow in popularity, persuaders adapt their messaging tactics to take advantage of these new media. Advertisements are present across social networks, websites, and video games, while millions of users can be contacted quickly and cheaply via email. Virtual environments (VEs) offer another opportunity for those with the intent to persuade to find potential targets for their messages, as persuaders can closely interact with their consumers while tightly controlling the interactive experience in ways that maximize the message’s effect. VEs also allow unique opportunities for consumers to engage with advertising. By offering the ability to interact with the environment, the advertisement, and the source making the sales pitch, interactive media may change the way individuals process advertisements by increasing the persuasive outcomes sought by marketers. Having control over an

CUSTOMIZING A VIRTUAL SALESPERSON

3

advertisement might empower individuals and make them more engaged with the product, like it more, and be more likely to purchase it later. For example, the Lay’s “Do Us A Flavor” contest allows users to suggest a new chip flavor and to create packaging for a bag that would hold that new chip. Users from all over the country suggested new flavors, and Lay’s created and distributed the top three flavors. The winner was determined by consumer vote, and the result was a consumer created and approved product. Advertisers may be hesitant to grant total control of their product or message to consumers as Lay’s did, however it is plausible that customization of some aspects of the product, such as a brand logo, virtual salesperson, mascot, or website could be opened up to consumers. In this sense, the user might be able to influence some aspects of the persuasive source while keeping the core message intact, and this experience may affect their perception of the message. This study will test the effects of how customizing a virtual salesperson’s avatar will affect an individual’s perceptions of a persuasive speech given by that avatar. If people were given the power to actively influence and control their experience with a sales pitch it might change how they perceive that message. Research on persuasion using digital technology shows that messages can be efficiently tailored for individual recipients, increasing persuasive outcomes (Lairson, Newmark, Rakowski, Tiro, & Vernon, 2004; Noar, Benac, & Harris, 2007; Rimer & Kreuter, 2006). Tailoring has also been shown to be effective in virtual worlds (Authors, 2009; 2014; Schmiel & Suggs, 2014). Yet virtual worlds also create the opportunity for individuals to customize their experience and interact with the message in novel ways. One understudied area is the way in which users—that is, the targets of the persuasive message—may be persuaded by modifying aspects of the persuasive source using virtual affordances. Transformed social interaction suggests that virtual technologies enable

CUSTOMIZING A VIRTUAL SALESPERSON

4

communication in virtual environments to be modified in ways that are not feasible in the physical world (TSI; Bailenson, Beall, Loomis, Blascovich, & Turk, 2004). It is possible that affordances unique to virtual environments could be applied to increase the persuasive effects of an advertising message. Virtual worlds typically come with the ability to create one’s avatar in order to interact with the environment or other people in the world. Options for customizing these avatars vary across the virtual worlds, and users are often faced with customization systems when entering a virtual world. Because of this familiarity and the rich ability to customize, we propose that virtual worlds are an excellent space with which to test the effects of having control over a digital salesperson. By allowing participants to customize the appearance of a human salesperson’s avatar we can examine the affordances unique to virtual worlds and examine their effects on advertising reception and attitudes towards a persuasive salesperson making a pitch.

2. Virtual Environments and Avatars A virtual environment is a digital space in which a user’s movements are tracked and his or her surroundings rendered, or digitally composed and displayed to the senses, in accordance with those movements. For example, in a computer game, a user’s joystick motions can be tracked and his or her character will move forward, rendering a new environment. Or, Xbox Kinect players can physically swing their arms remote, and the screen will show a bowling ball rolling down the lane. The goal of a virtual environment is to replace the cues of the real world environment with digital ones. Thus, the affordances of VEs can be used to dramatically change the nature of communication beyond the boundaries of the physical world.

CUSTOMIZING A VIRTUAL SALESPERSON

5

Common elements of virtual environments are avatars and agents. The term avatar refers to any representation of a user in a virtual environment. These representations can vary in their form and detail, or whether they are enacted in real time or not. Avatars are controlled by humans and should be distinguished from agents, which are digital representations controlled by computer algorithms (Authors, 2014). Typically, avatars are used to represent people in Internet chat, video games, virtual reality, and other mediated contexts. In 3D environments, avatars provide an essential, functional representation with which the user can enact virtual behaviors such as navigating virtual space or engaging virtual objects or other avatars. A recent metaanalysis revealed that when virtual interactants are believed to be avatars, they are more persuasive than those believed to be agents (Authors, 2014). At this stage, further research needs to investigate how avatars or agents can be further manipulated to promote persuasive outcomes.

2.1 Transformed Social Interaction According to Bailenson et al. (2004), virtual technologies have advantages over traditional forms of interpersonal communication in three ways: sensory capabilities, situational context, and self-representation. TSI gives users the opportunity to modify and augment their senses, gaining more information than they would be able to in regular, face-to-face interactions (Bailenson & Beall, 2006). For example, perceptions could be enhanced in a virtual classroom, allowing teachers to see students’ names, grades, or other measures of comprehension or attention as they look at each student’s avatar. A salesperson could track the nonverbal behaviors of customers, identifying changes in heart rate or breathing frequency indicate interest or excitement in the pitch, and change the message based on this feedback during the pitch. Thus,

CUSTOMIZING A VIRTUAL SALESPERSON

6

the messengers’ senses are enhanced as they are able to detect, assess, and react to information and feedback from multiple sources at once. TSI also allows for the manipulation of the context surrounding the situation (Bailenson & Beall, 2006). If a user missed an important conversational point, he or she could rewind and replay the conversation. Buyers having a hard time focusing in a virtual car dealership could pause the tour and step away. Alternatively, the virtual dealership itself could disappear, and buyers could step in to a virtual model and explore its features on their own time, without other distractions. Altering elements of the context may improve desired persuasive outcomes. Finally, TSI allows for the manipulation of one’s self-representation (Bailenson & Beall, 2006), and as such users may choose avatars that are accurate or inaccurate representations of themselves (Vasalou, Joinson, Bänziger, Goldie, & Pitt, 2008). Individuals can change their avatar’s height, weight, or hair color in a virtual environment. They can make changes drastically beyond possibility in the real world, such as inhabiting the body of a narwhal. Individuals could also manipulate their appearance based on situational contexts or change them over time. For example, a salesperson could alter his or her emotional expressions depending on the course of a conversation. The ability to alter one’s self-representation maximizes the opportunity for creating the most persuasive version of the self possible. In this study, we seek to expand the domain of TSI. Although TSI research has thus far focused on modifying one’s own self-representation, the same affordances could enable users to modify aspects of others’ representations in the virtual world. Given that VEs enable multiple viewpoints even in the same virtual space, John could customize the features of Kim’s representation so that it appeared a certain way to John, although Kim would still see the representation she had chosen for herself. This flexibility could maximize desired outcomes.

CUSTOMIZING A VIRTUAL SALESPERSON

7

Perhaps a child learns more from a teacher represented by a teddy bear, or a socially anxious person may feel most comfortable inhabiting a virtual world where other people are always smiling. Customization of a product has been shown to be an effective means of increasing engagement with an advertisement (Bright & Daugherty, 2012), and the ability to customize the representation of another person could be very useful in interpersonal persuasive contexts, such as a sales pitch. Customization of the persuasive source could benefit both the user and the salesperson: users get some control over the interaction and may customize the salesperson in a way that makes the users feel more comfortable, thus increasing the likelihood of a sale.

2.2 Persuasion in Virtual Environments Because of their immersiveness and unique affordances, virtual environments are ideal for many types of persuasive messages (Askoy, Bloom, Lurie, & Cooil, 2006). Studies conducted in advertising contexts have found that interacting with virtual objects and virtual salespeople has a strong positive effect on persuasive outcomes. Li, Daugherty, and Biocca (2001, 2002) demonstrated that interacting with three-dimensional representations of products caused users to rate the experience as more natural and engaging, and users felt more present than those who saw and could not interact with two-dimensional representations. Interacting with three-dimensional objects also increases product attitude accessibility (Schlosser, 2003), and being able to interact with an object generates more mental imagery of product use, which increases purchase intentions (Lee, Li, & Edwards, 2012). Previous research indicates that the appearance of a persuasive source can make a difference in online environments (e.g., Bente, Baptist, & Leuschner, 2012; Qiu & Benbasat, 2010). Recent scholarship has focused on the use of virtual humans as persuasive agents. When

CUSTOMIZING A VIRTUAL SALESPERSON

8

virtual representations resemble the human form, users often treat them similarly to how they treat real people (Bailenson, Blascovich, Beall, & Loomis, 2003). Thus, virtual salespeople with human forms may be more effective than other common types of interaction in digital sales contexts, such as using a text-based chat (Askoy et al., 2006; Holzwarth, Janiszewski, & Neumann, 2006). Consumers on e-commerce websites seek interpersonal connection, and the inclusion of recommendation agents, or virtual human salespersons, have been shown create perceptions of interpersonal interaction (Holzwarth et al., 2006; Puzakova, Rocereto, & Kwak, 2013). Recommendation agents appear human to better increase similarity and identification with the consumer (Holzwarth et al., 2006). Consumers who interacted with a recommendation agent were more satisfied with the retailer and had more favorable attitudes and higher purchase intentions than those who did not have an agent (Askoy et al., 2006). Moreover, users like the recommendation agents more if the site matches the agent with the user’s ethnicity (Qiu & Benbasat, 2010), and users have more brand liking when persuasive avatars look like them (Ahn & Bailenson, 2011). At this time, however, it is uncertain how giving the user control over the source’s appearance will affect persuasive outcomes. In contrast to tailoring or targeted marketing, where a persuasive source customizes a message designed specifically for an individual (Kalyanaraman & Sundar, 2006; Sundar & Marathe, 2010), users also benefit when they have the ability to customize their virtual experience (Kalyanaraman & Sundar, 2006). Customization allows an individual user to be the source of the content (Sundar, 2008; Sundar & Nass, 2001). Many websites today offer the ability for users to influence their experiences by changing the website layout or selecting the most pertinent information. Instead of more traditional mass advertising where a single message is sent to lots of individuals, digital technology allows for users to actively shape their own

CUSTOMIZING A VIRTUAL SALESPERSON

9

experiences. For example, Amazon.com allows users to customize the site and make it their own by uploading images, creating wish lists, writing reviews, and tagging items to show up in future searches. The ability to customize has been shown to increase feelings of involvement, relevance, interactivity, and novelty, leading to more positive evaluations of a website (Kalyanaraman & Sundar, 2006). It is possible that the benefits shown with websites also apply to interpersonal contexts of communicating with a virtual salesperson. In a virtual interaction, users should be more persuaded by an avatar or agent that delivers a persuasive message than just seeing the message (Askoy et al., 2006; Holzwarth, et al., 2006). TSI would suggest that salespeople could further use techniques unique to virtual worlds to give a more effective pitch. TSI also suggests that virtual worlds could enable users to customize a virtual salesperson, although this facet of TSI remains understudied. Recommender agents that have been customized by designers, rather than users, have been shown to increase liking and persuasive outcomes (Ahn & Bailenson, 2011; Qiu & Benbasat, 2010), but little is known about shifting this power to the user. We believe that the ability to for the user to customize the source will gratify one’s need for autonomy, which ultimately increases how much one likes a product and how likely one is to buy it.

3. Intrinsic Motivation and Autonomy Self-determination theory posits that humans need to fulfill three basic psychological needs: autonomy, competence, and relatedness (Deci & Ryan, 2000). Autonomy is fulfilled when one believes that one’s actions come as a result of one’s internal core of desires, rather than due to an external pressure (Ryan & Deci, 2000). Interactive media have been demonstrated to be effective at fulfilling these needs (Przybylski, Rigby, & Ryan, 2010; Ryan, Rigby, & Przybylski,

CUSTOMIZING A VIRTUAL SALESPERSON

10

2006; Sheldon & Filak, 2008). Interactive media offer the ability for individuals to have active control over their experiences and make choices (Liu & Shrum, 2002), which should successfully fulfill the need for autonomy. Interacting in a virtual world has been shown to satisfy needs for autonomy (Jung, 2011; Partala, 2011) and increase intrinsic motivation (Verhagen, Feldberg, van den Hoof, Meents & Merikivi, 2012). We suggest that these findings are due to the interactive, choice-driven nature of virtual worlds that allows users to customize their own experiences. The satisfaction of autonomy via customization should lead to positive experiences; customization has been shown to increase enjoyment of video games (Schmierbach, Limperos, & Woolley, 2012; Teng, 2010; Trepte & Reinecke, 2010). Self-determination theory also states that events that satisfy autonomy needs result in higher intrinsic motivation (Ryan & Deci, 2000). When considering a behavior, one attributes the reason for performing that behavior to either intrinsic (internal) or extrinsic (external) forces (Deci & Ryan, 2000; Ryan & Deci, 2000). For example, students may study because of an internal satisfaction that comes from learning, or they may study because their parents pay them for good grades. A customer might purchase a product because that customer chose to do so, or because of influence from an outside source or pressure. Intrinsic motivation is the key to engagement; making one more intrinsically motivated to do a behavior should lead to more enjoyment and more internal drive to continue with the behavior (Ryan & Deci, 2000). Because of this, increasing intrinsic motivation has been shown to lead to positive persuasive outcomes (Pelletier & Sharp, 2008). By giving users control over a virtual salesperson, users should experience more autonomy and have more motivation to continue with the task which should lead to a more enjoyable and engaging experience. This satisfaction of needs may make the persuasive message

CUSTOMIZING A VIRTUAL SALESPERSON

11

more appealing, despite what the user chooses to do with the speaker’s avatar. If control and motivation are the driving forces behind taking control over a persuasive speaker’s avatar then we would expect to see individuals who have more preference for the brand and report higher purchase intentions. It is possible that this occurs even for individuals who use customization as a means to not be persuaded, perhaps by making the salesperson’s avatar look silly or nonthreatening. The act of making choices and exercising control over the persuasive source should increase motivation in any case, making it possible that need fulfillment actually overrides, or reduces, one’s initial inclination to not be persuaded.

4. Overview TSI suggests virtual environments offer opportunities to interact with and experience messages in new ways. Self-determination theory suggests that the ability to customize a speaker’s avatar fulfills the core psychological need for autonomy. Need satisfaction boosts intrinsic motivation, which should lead to a more positive experience and desire to continue with the experience (Ryan & Deci, 2000). Fulfillment of autonomy may shed additional light on the mechanism behind the effectiveness of customization. We hypothesize that these positive feelings will carry over onto the reception of the sales pitch and create positive attitudes for the salesperson, product, and brand. H1: Those who customize a salesperson’s avatar will like the brand more than those who do not. H2: Those who customize a salesperson’s avatar will report higher purchase intentions than those who watch the creation of the salesperson’s avatar. H3: Intrinsic motivation will mediate the relationship between customizing a

CUSTOMIZING A VIRTUAL SALESPERSON

12

salesperson’s avatar and brand liking. H4: Intrinsic motivation will mediate the relationship between customizing a salesperson’s avatar and purchase intentions. In addition, there may be some important additional variables that affect the extent to which customization benefits persuasive outcomes. Forewarning occurs when one is told a persuasive message is coming before that individual encounters the message, and has been proven to be a way to induce resistance to persuasion through means of giving time to generate more resistance (Petty & Cacioppo, 1977, 1979). When a forewarned individual is told a persuasive message is coming, that individual has time to generate counterarguments, or thoughts against the advocated position of the persuasive message. Then, when the individual hears the persuasive message, the counterarguments are already in place and are actively used to resist the message (Benoit, 1998; Petty & Cacioppo, 1977, 1979). It is possible that forewarned individuals use the time watching or customizing the avatar to generate counterarguments about the message that will come, which would reduce persuasive outcomes. We suggest forewarning negatively affects persuasive outcomes only for individuals who do not customize the salesperson’s avatar. Individuals who have the ability to customize will be too engaged and intrinsically motivated to effectively counterargue, thus negating forewarning’s effects. H5: Forewarned individuals who do not customize the salesperson’s avatar will have lower brand attitudes than those who are not forewarned. H6: Forewarned individuals who do not customize the salesperson’s avatar will have lower purchase intentions than those who are not forewarned.

CUSTOMIZING A VIRTUAL SALESPERSON

13

5. Method 5.1 Sample Participants (37 men and 57 women) were recruited from a large Midwestern university. Participants received course credit for their participation. Participants (n = 14) that indicated they did not believe the salesperson was controlled by a real human were excluded from analysis. The final sample (N = 80) consisted of 28 men and 52 women, ranged in age from 18 to 51 (M = 20.57, SD = 4.47) and reported their race/ethnicity as: Caucasian/ European-American/White (n = 47); Black/African/African-American (n = 17); Asian/Asian-American (n = 8); Latino/Latina/Hispanic (n = 1); multiracial (n = 4); and other (n = 2).

5.2 Procedure Upon arrival at the laboratory, participants were assigned to one of four conditions: Customization, No Forewarning (n = 22); Customization Forewarning (n = 18); No Customization, No Forewarning (n = 22); and No Customization, Forewarning (n = 18). Participants in all conditions were told that they would be interacting through a virtual world equipped with voice chat with Alex, a representative from Ternio (a fictional company), in a test of the “effectiveness of using virtual worlds to communicate with companies.” Voice chat was selected to ensure that participants felt like they were talking with a real, responsive person and not a bot. These measures were necessary to bolster the feeling that participants were interacting with an avatar rather than an agent, because variations in these perceptions can cause undesired variance in measured outcomes (Authors, 2014).

CUSTOMIZING A VIRTUAL SALESPERSON

14

In the Forewarning conditions, participants were told that Alex was going to try to persuade them to support the company’s bid for a contract with the university’s vending services. In the No Forewarning conditions, Alex’s persuasive intent was not revealed. Thus, participants were made aware or not aware of Alex’s attempt to persuade before the avatar customization step. In the Customization conditions, participants were given seven minutes to use the virtual world’s interface to modify Alex’s avatar. They had complete control over how Alex would appear to them in the virtual world; participants could change Alex’s clothes, face, hair style, and hair color. All participants in the Customization completed the task, and made a number of changes to the avatar. In the No Customization conditions, participants observed the video from another person’s avatar customization process. All of the Customization videos were recorded. Each participant in the No Customization condition was randomly assigned a video from another same sex participant. Thus, participants in the No Customization condition had no control over how Alex would appear. After this step, participants then entered the virtual world, which was set up as a casual room. To improve the cover story, the researcher made a call to Alex, made introductions, and then handed the headset to the participant. Then, the avatar they had customized (or watched being customized) that represented Alex walked into the room and sat down as Alex spoke with the participant. A second research assistant playing the role of Alex in a separate room then followed a script featuring information about the product, a fictional energy drink. Some interactive questions were included in the script so that the dialogue seemed more natural and so that Alex’s agency (i.e., that Alex was a real person, not a computer or a recording) was reinforced. After the interaction with Alex, participants completed survey measures on a separate computer and were subsequently debriefed.

CUSTOMIZING A VIRTUAL SALESPERSON

15

5.3 Measures 5.3.1 Intrinsic motivation Participants completed the Intrinsic Motivation Inventory (Ryan, Koestner, & Deci, 1991), which contains 22 items with statements such as, “I felt like it was my choice to do the task,” “I think I am pretty good at this task,” “I would describe the task as enjoyable,” and “I felt tense during the task” (Cronbach’s α = .82). Items were measured on a 7-point Likert scale from 1 (strongly disagree) to 7 (strongly agree).

5.3.2 Brand liking and purchase intention All items were measured on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). Brand liking was measured with four items that included statements such as “I like Ternio energy drinks” and “I would want to hear more about Ternio energy drinks” (Cronbach’s α = .82). Purchase intention was measured with three statements, such as “I would buy Ternio energy drinks” and “I was convinced by the speech to buy Ternio energy drinks” (Cronbach’s α = .82).

5.3.3 Perception of agency One item was included as a manipulation check where participants were asked to rate their perceptions of whether Alex was human on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). Because previous research has shown dramatic differences in effects based on perceived agency (Authors, 2014), participants who did not agree that Alex was human were excluded from analysis.

CUSTOMIZING A VIRTUAL SALESPERSON

16

6. Results Due to the imbalance in participant sex, a series of preliminary analyses were conducted to ensure that sex was not a confounding factor. There were no differences in assignment to the customization condition; males and females were equally likely to be assigned to the Customization and No Customization conditions, χ2(1, N = 80) = .88, p = .348. There were no differences in assignment to the forewarning condition; males and females were equally likely to be assigned to forewarning and non-forewarning conditions, χ2(1, N = 80) = 1.28, p = .258. Independent samples t-tests found that there were no significant differences between participant sex on brand liking, t(78) = 1.02, p = .313, purchase intention, t(78) = 1.78, p = .078, or intrinsic motivation, t(78) = .10, p = .924. H1 predicted that those who customize an avatar will have higher brand liking than those who do not. To address H1, a 2x2 MANOVA was run. There was a significant main effect for avatar customization on brand liking, F(1, 76) = 6.42, p = .01, partial η2 = .08, supporting H1. Those who customized the source (M = 5.32, SD = .69) expressed greater brand liking than those who did not (M = 4,74, SD = 1.26). H2 predicted that those who customize an avatar will have higher purchase intention than those who do not. Tests on H2 show that the effect of avatar customization on purchase intentions only bordered on significance, F(1, 76) = 3.28, p = .07, partial η2 = .04. Those who customized the source (M = 5.54, SD = .77) demonstrated greater purchase intentions than those who did not (M = 5.12, SD = 1.21), although this difference was not significant. We predicted interactions between forewarning and avatar customization such that those who were forewarned but did not customize would have the lowest brand liking (H5) and purchase intention (H6). There were no main effects for forewarning on brand liking, F(1,

CUSTOMIZING A VIRTUAL SALESPERSON

17

76) = .51, p = .47, partial η2 = .01, or purchase intentions, F(1, 76) = .57, p = .43, partial η2 = .01. There was also no interaction effect for brand liking, F(1, 76) = .65, p = .42, partial η2 = .01, or purchase intentions, F(1, 76) = .78, p = .38, partial η2 = .01, offering no support for H5 or H6. To address H3, a mediation analysis was run using PROCESS (Hayes, 2013). Avatar customization was entered as the independent variable, intrinsic motivation as the mediating variable, and brand liking as the dependent variable by generating 10,000 bootstrap confidence intervals in PROCESS model 4 (Figure 1). The amount of energy drinks a participant had during the week was not a significant covariate and was dropped from the model. Direct and indirect effects can be viewed in Table 1. Results show that avatar customization directly affects intrinsic motivation, where those in the customization condition have higher intrinsic motivation scores, a = .41; 95% CI [.09, .73], and that higher intrinsic motivation scores lead to higher brand liking regardless of condition, b = .76; 95% CI [.53, 1.06]. There is a direct effect of avatar customization on brand liking when holding intrinsic motivation constant, c = .58; 95% CI [.13, 1.03]. The data indicate that avatar customization affects brand liking indirectly through intrinsic motivation, ab = .32; 95% CI [.07, .63]. Relative to those in the non-customization condition, those that customized were more intrinsically motivated, which was associated with higher brand liking, supporting H3. To address H4, a second mediation analysis was run using PROCESS. Avatar customization was entered as the independent variable, intrinsic motivation as the mediating variable, and purchase intentions as the dependent variable in PROCESS model 4 (Figure 2). Number of energy drinks participants consumed per week was not a significant covariate in the model and was removed from analysis. Direct and indirect effects can be viewed in Table 2. Results show that avatar customization directly affects intrinsic motivation, where those in the

CUSTOMIZING A VIRTUAL SALESPERSON

18

customization condition have higher intrinsic motivation scores, a = .41; 95% CI [.09, .73], and that higher intrinsic motivation scores lead to greater purchase intention, regardless of condition, b = .59; 95% CI [.22, .96]. There is no direct effect of avatar customization on purchase intention when holding intrinsic motivation constant, c = .33, 95% CI [-.23, .88]. Avatar customization affected purchase intention indirectly through intrinsic motivation scores, ab = .24; 95% CI [.05, 60]. Relative to those in the non-customization condition, those that customized were more intrinsically motivated, which was associated with greater purchase intention, supportingH4.

7. Discussion The results confirm our primary hypotheses: individuals who customized a salesperson’s avatar had higher brand liking scores and purchase intentions than those who watched the customization of the salesperson’s avatar. The results suggest that customizing an avatar leads to increases in intrinsic motivation, and these feelings of autonomy, competence, and enjoyment then lead to increases in brand liking and purchase intention. These feelings carried over from the customization phase to listening to the sales pitch even though the avatar was removed from the participants’ control and given to the speaker. These results persisted across participants who varied in their level of drinking energy drinks during the week. Contrary to expectations, the results also show that our manipulation of forewarning had no effect on persuasive outcomes.

7.1 Implications Transformed social interaction was used as a framework to establish the unique capabilities of virtual environments to influence perception, manipulate the situation, and change self-representation. TSI suggests that the potential for virtual environments has only begun to be

CUSTOMIZING A VIRTUAL SALESPERSON

19

explored. Having participants customize a persuasive source takes advantage of interactive technology and virtual environments. Companies are beginning to turn to interactive media and highly customizable virtual environments as a way of reaching out to consumers. Allowing an individual to customize or interact with an established brand could be seen as a risk by opening the brand up to ridicule or allowing individuals to bypass a carefully constructed message. Yet our results suggest that companies would do well to allow customers to interact with a brand or other source material. Having customizable websites or brand logos may appeal to one’s intrinsic motivation and increase enjoyment and liking towards the brand. Our results suggest that intrinsic motivation is a driving force behind the persuasive benefits of customization. Examining other moderating or mediating variables might help elaborate additional theoretical mechanisms behind customization. For example, customizing a salesperson or brand website might also affect perceptions of source credibility, or perceived expertise and trustworthiness (Hovland & Weiss, 1951). It has been shown that highly credible sources have been found to be more persuasive than low credibility sources (Simons, Berkowitz, & Moyer, 1970), and recent work has demonstrated that websites are often evaluated in terms of credibility based on name alone, with specialized names leading to greater perceived credibility (Koh & Sundar, 2010). In addition to having a specialized name, the ability to interact with a website also increases perceptions of source credibility (Doney & Cannon, 1997; Sundar, 2008). Finally, the extent to which one identifies with a source has been shown to increase perceptions of that source’s credibility (Doney & Cannon, 1997). Avatar creation should strongly influence identification; in the creation process users are actively putting their own ideas and choices into the avatar, which is reflected on a screen. Avatar creation has been shown to increase identification (Trepte & Reinecke, 2010), and thus should affect perceptions of credibility when

CUSTOMIZING A VIRTUAL SALESPERSON

20

applied to a source’s avatar. Future research might examine the role of high or low credibility with a customizable source and the mechanisms driving customization. In addition to source credibility, efforts to determine the role of perceived interactivity in influencing brand attitudes and purchase intention ought to be investigated when the user is the source of a persuasive message. The perceived interactivity of a website has been shown to positively predict ratings of effectiveness, trust, and enjoyment and lead to more website loyalty (Cyr, Head, & Ivanov, 2009). This could also apply to avatar customization and interaction; the extent to which individuals believe they can interact with and control a salesperson’s avatar should mediate the effects of customization on brand attitudes and purchase intention. It should be noted that the positive effects on brand liking and purchase intention emerged through voice-to-voice communication between the salesperson and participant. To ensure that effects were attributable to the manipulation, participants in this study saw no brand logos or prototypes of the product itself; it was only described by the salesperson’s pitch. Incorporating visuals should make the product more memorable and future work might benefit from incorporating brand logos and looking at recall effects days or weeks after the initial introduction to the product. This represents the beginning of research on individuals’ ability to customize a persuasive source in virtual environments. Future research may investigate manipulating other facets of an avatar that have been shown to be persuasive, such as behavioral realism (Groom, Nass, Chen, Nielsen, Scarborough, & Robles, 2009) or its facial expressions or emotions (Krämer, Kopp, Becker-Asano, & Sommer, 2013). Manipulating a source does not have to be limited to creation of a speaker’s avatar, however. Giving participants the ability to customize a company website or logo would give additional insight on the effects of customization.

CUSTOMIZING A VIRTUAL SALESPERSON

21

7.2 Limitations Previous research has demonstrated that forewarning effectively increases resistance to persuasion (Petty & Cacioppo, 1977, 1979), yet we found no differences between forewarning or not forewarning participants. Forewarning participants, particularly in the watch condition, should give them ample time to generate counterarguments and resistance to the persuasive message, but no such effects were found. This was surprising, in light of the strong support for reactance theory (Brehm, 1966; Brehm & Brehm, 1981; Wicklund, 1974), which states that individuals are strongly motivated to restore a freedom once it is perceived to be threatened. An attempt to persuade an individual should be perceived as a threat to one’s freedom to choose, and thus the individual would be motivated to restore that freedom. It is possible that the customization process allowed for an outlet of this reactance; individuals could restore freedom indirectly by making additional choices in the customization process without reacting against the source or the message. Unfortunately, reactance (and counterarguing) was measured indirectly by looking for decreases in brand liking and purchase intention. It is difficult to tell whether the lack of resistance to the persuasive message, enhanced by forewarning, is due to this ability to restore freedom indirectly or due to a lack of strength in the forewarning manipulation. Future research might directly measure counterarguments and angry affect (Dillard & Shen, 2005) before and after the sales pitch to establish the role of reactance in the customization process. Participants were initially asked to create an avatar that shared their sex. Alex was chosen as a gender-neutral name, but the confederate researcher who played the role of Alex was a male for all participants. The voice remaining male was done to maximize experimental control by keeping the confederate voice the same for all participants, though it is possible that females

CUSTOMIZING A VIRTUAL SALESPERSON

22

hearing a male voice may react to the message differently than males hearing a male voice. However, there were no sex differences in brand liking, purchase intention, or belief that Alex was a real representative of the company. Despite the precautions taken to ensure Alex appeared as a real representative, 14 participants still did not believe that Alex was human. Because perceptions of agency affect the degree to which people are influenced by virtual representations (Authors, 2014), these participants were excluded from this analysis. These participants may have thought this because so much of the experiment revolved around creating an avatar and looking at a screen, or perhaps this was a measuring artifact where participants thought the question asking if Alex was human indicated an experimental deception and did not want to seem fooled. We ran the primary analyses again including these participants and found that our results did not change. In future studies, manipulations of perceived agency may be an interesting angle, given most advertising contexts use computer-controlled agents rather than human-controlled avatars. From a practical standpoint, it may be argued that our experiment does not adequately reflect how current sites operate or wish to operate. A sales team committed to individual oneon-one online sales pitches could be costly and time consuming. Thus, future research should investigate whether manipulating computer-controlled agents has the same effect, as in some cases they may be perceived similarly to avatars (e.g., Shank, 2014), and using conversational agents would provide more flexibility and accessibility (Rossen & Lok, 2012). It is possible that our findings may be attributable as novelty effects (Tokunaga, 2013). The relatively new opportunities to customize a persuasive source may be exciting and interesting, driving engagement with the message. If the findings are limited to novelty effects, the positive effects of the message should diminish as these opportunities increase or as

CUSTOMIZING A VIRTUAL SALESPERSON

23

participants become used to the customization process. Our findings suggest that the ability to customize strongly influences core psychological needs (i.e., autonomy and competence), and it is this need satisfaction that drives positive ratings of the message. Because need satisfaction has been shown to be a long-term driving force with a number of positive benefits (Deci & Ryan, 2000; Ryan & Deci, 2000), it is plausible that the effects of customization will continue long past the experience being novel. Alternatively, our findings may be explained through the theoretical mechanism of cognitive dissonance (Festinger, 1954). Customization requires effort, and it is possible that the effort put forth may cause participants to feel more committed with the avatar and in turn like the avatar’s message more. We attempted to control for effort by giving participants in both tasks the same amount of time (seven minutes) with the watching or customization task. Still, we did not measure perceptions of commitment to the avatar or identification with the avatar, and this may be a relevant avenue for future research. Finally, participants were given the option to customize or watch the customization of an avatar without a true control condition where participants did not engage with an avatar at all and only received the pitch. It is possible that watching the creation of the avatar does have some persuasive benefits as viewers are more engaged. Future research should employ a true control condition where participants only receive the persuasive speech without engaging with an avatar.

7.3 Conclusion The results show that TSI can be an effective framework with which to test the effects of reactance and influence on motivation, brand attitudes, and purchase intentions. Despite using techniques traditionally shown to increase resistance to persuasion, having control over the

CUSTOMIZING A VIRTUAL SALESPERSON

24

creation of a persuasive speaker’s avatar results in liking the brand more and being more likely to purchase the product. These findings imply that allowing individuals to interact with persuasive sources can cause them to like the source and the product better. Advertising is only beginning to take advantage of the unique tools of interactive media, and research will continue to provide insights on how to design virtual environments to maximize both the consumer’s experience and the brand’s goals.

References Ahn, S. J., & Bailenson, J. N. (2011). Self-endorsing versus other-endorsing in virtual environments: The effect on brand preference. Journal of Advertising, 40, 93-106. doi:10.2753/JOA0091-3367400207 Aksoy, L., Bloom, P. N., Lurie, N. H., & Cooil, B. (2006). Should recommendation agents think like people? Journal of Service Research, 8, 297-315. doi:10.1177/1094670506286326 Authors. (2009). Authors. (2014). Bailenson, J. N., & Beall, A. C. (2006). Transformed social interaction: Exploring the digital plasticity of avatars. In R. Schroeder & A-S. Axelsson (Eds.), Avatars at work and play: Collaboration and interaction in shared virtual environments (pp. 1-16). New York: Springer-Verlag. Bailenson, J. N., Beall, A. C., Loomis, J., Blascovich, J., & Turk, M. (2004). Transformed social interaction: Decoupling representation from behavior and form in collaborative virtual environments. PRESENCE: Teleoperators & Virtual Environments, 13, 428-441. doi:10.1162/1054746041944803 Bailenson, J. N., Blascovich, J., Beall, A. C., & Loomis, J. M. (2003). Interpersonal distance in

CUSTOMIZING A VIRTUAL SALESPERSON

25

immersive virtual environments. Personality & Social Psychology Bulletin, 29, 819-833. doi:10.1177/0146167203029007002 Benoit, W. L. (1998). Forewarning and persuasion. In M. Allen, & R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 139-154). Cresskill, NJ: Hampton Press Inc. Bente, G., Baptist, O., & Leuschner, H. (2012). To buy or not to buy: Influence of seller photos and reputation on buyer trust and purchase behavior. International Journal of HumanComputer Studies, 70, 1-13. doi:10.1016/j.ijhcs.2011.08.005 Brehm, J. W. (1966). A theory of psychological reactance. New York: Academic Press. Brehm, S. S., & Brehm, J. W. (1981). Psychological reactance: A theory of freedom and control. New York: Academic Press. Bright, L. F., & Daugherty, T. (2012) Does customization impact advertising effectiveness? An exploratory study of consumer perceptions of advertising in customized online environments. Journal of Marketing Communications, 18(1), 19-37. doi: 10.1080/13527266.2011.620767 Cyr, D., Head, M., & Ivanov, A. (2009). Perceived interactivity leading to e-loyalty: Development of a model for cognitive-affective user responses. International Journal of Human-Computer Studies, 67, 850-869. doi:10.1016/j.ijhcs.2009.07.004 Deci, E. L., & Ryan, R. M. (1980). The empirical exploration of intrinsic motivational processes. Advances in Experimental Social Psychology, 13, 39–80. doi:10.1016/S00652601(08)60130-6 Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. New York: Plenum Press.

CUSTOMIZING A VIRTUAL SALESPERSON

26

Deci, E. L., & Ryan, R. M. (2000). The ‘‘what’’ and ‘‘why’’ of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11, 227–268. doi: 10.1207/S15327965PLI1104_01 Doney, P. M., & Cannon, J. P. (1997). An examination of the nature of trust in buyer-seller relationships. The Journal of Marketing, 61, 35-51. Retrieved from http://www.jstor.org/stable/1251829 Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140. doi: 10.1177/001872675400700202 Groom, V., Nass, C., Chen, T., Nielsen, A., Scarborough, J. K., & Robles, E. (2009). Evaluating the effects of behavioral realism in embodied agents. International Journal of HumanComputer Studies, 67, 842-849. doi:10.1016/j.ijhcs.2009.07.001. Holzwarth, M., Janiszewski, C., & Neumann, M. M. (2006). The influence of avatars on online consumer shopping behavior. Journal of Marketing, 70, 19-36. doi: 10.1509/jmkg.70.4.19 Hovland, C. I. & Weiss, W. (1951). The influence of source credibility on communication effectiveness. Public Opinion Quarterly, 15(4), 635-650. doi: 10.1086/266350 Jeong, H. J., & Lee, M. (2013). Effects of recommendation systems on consumer inferences of website motives and attitudes towards a website. International Journal of Advertising, 32(4), 539-558. doi: 10.2501/ija-32-4-539-558 Jung, Y. (2011). Understanding the role of sense of presence and perceived autonomy in users’ continued use of social virtual worlds. Journal of Computer-Mediated Communication, 16, 492-510. doi:10.111/j.1083-6101.2011.01540.x

CUSTOMIZING A VIRTUAL SALESPERSON

27

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, 110-132. doi:10.1111/j.1460-2466.2006.00006.x Koh, Y. J., & Sundar, S. S. (2010). Effects of specialization in computers, web sites, and web agents on e-commerce trust. International Journal of Human-Computer Studies, 68, 899912. doi: 10.1016/j.ijhcs.2010.08.002 Krämer, N., Kopp, S., Becker-Asano, C., & Sommer, N. (2013). Smile and the world will smile with you: The effects of a virtual agent’s smile on users’ evaluation and behavior. International Journal of Human-Computer Studies, 71, 335-349. doi: 10.1016/j.ijhcs.2012.09.006 Lairson, D. R., Newmark, G. R., Rakowski, W., Tiro, J. A., & Vernon, S. W. (2004). Development costs of a computer-generated tailored intervention. Evaluation and Program Planning, 27, 161-169. doi: 10.1016/j.evalprogplan.2004.01.004 Lee, K-Y., Li, H., & Edwards, S. M. (2012). The effect of 3-D product visualisation on the strength of brand attitude. International Journal of Advertising, 31, 377-396. doi: 10.2501/ija-31-377-396 Li, H., Daugherty, T., & Biocca, F. (2001). Characteristics of virtual experience in electronic commerce: A protocol analysis. Journal of Interactive Marketing, 15(3), 13-30. doi: 10.1002/dir.101 Li, H., Daugherty, T., & Biocca, F. (2002). Impact of 3-D advertising on product knowledge, brand attitude, and purchase intention: The mediating role of presence. Journal of Advertising, 31(3), 43-57. doi: 10.1080/00913367.2002.10673675

CUSTOMIZING A VIRTUAL SALESPERSON

28

Liu, Y., & Shrum, L. J. (2002). What is interactivity and is it always such a good thing? Implications of definition, person, and situation for the influence of interactivity on advertising effectiveness. Journal of Advertising, 31(4), 53-64. doi: 10.1080/00913367.2002.1067368 Noar, S. M., Benac, C. N., & Harris, M. S. (2007). Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychological Bulletin, 133, 673693. doi: 10.1037/0033-2909.133.4.673 Partala, T. (2011). Psychological needs and virtual worlds: Case Second Life. International Journal of Human-Computer Studies, 69, 787-800. doi: 10.1016/j.ijhcs.2011.07.004 Petty, R. E., & Cacioppo, J. T. (1977). Forewarning, cognitive responding, and resistance to persuasion. Journal of Personality and Social Psychology, 35, 645-655. doi:10.1037/0022-3514.35.9.645 Petty, R. E., & Cacioppo, J. T. (1979). Effects of forewarning of persuasive intent and involvement on cognitive responses and persuasion. Personality & Social Psychology Bulletin, 5, 173-176. doi: 10.1177/014616727900500209 Przybylski, A. K., Rigby, C. S., & Ryan, R. M. (2010). A motivational model of video game engagement. Review of General Psychology, 14(2), 154-166. doi: 10.1037/a0019440 Puzakova, M., Rocereto, J. F., & Kwak, H. (2013). Ads are watching me – a view from the interplay between anthropomorphism and customization. International Journal of Advertising, 32, 513-538. doi: 10.2501/ija-32-4-513-538 Qiu, L., & Benbasat, I. (2010). A study of demographic embodiments of product recommendation agents in electronic commerce. International Journal of Human-

CUSTOMIZING A VIRTUAL SALESPERSON

29

Computer Studies, 68, 669-688. doi: 10.1016/j.ijhcs.2010.05.005 Rimer, B. K., & Kreuter, M. W. (2006). Advancing tailored health communication: A persuasion and message effects perspective. Journal of Communication, 56, 184-201. doi: 10.111/j.1460-2466.2006.00289.x Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68-78. doi: 10.1037/0003-066X.55.1.68 Ryan, R. M., Rigby, C. S., & Przybylski, A. (2006). The motivational pull of video games: A self-determination theory approach. Motivation and Emotion, 30, 347-363. doi: 10.1007/s11031-006-9051-8 Schlosser, A. E. (2003). Experiencing products in the virtual world: The role of goal and imagery in influencing attitudes versus purchase intentions. Journal of Consumer Research, 30, 184-198. doi: 10.1086/376807 Schmeil, A., & Suggs, S. S. (2014). An empirical comparison of variations of a virtual representation of an individual’s health. In A. Spagnolli, L. Chittaro, & L. Gamberini (Eds.), Persuasive technology: Persuasive, motivating, empowering videogames (pp. 201-223). Switzerland: Springer. Schmierbach, M., Limperos, A. M., & Woolley, J. K. (2012). Feeling the need for (personalized) speed: How natural controls and customization contribute to enjoyment of a racing game through enhanced immersion. Cyberpsychology, Behavior, & Social Networking, 15, 364-369. doi: 10.1089/cyber.2012.0025 Shank, D. B. (2014). Impressions of goodness and power for computer agents: Computer identity weakens power impressions. International Journal of Human-Computer Studies,

CUSTOMIZING A VIRTUAL SALESPERSON

30

72, 747-756. doi: 10.1016/j.ijhcs.2014.05.002 Sheldon, K. M., & Filak, V. (2008). Manipulating autonomy, competence, and relatedness support in a game-learning context: New evidence that all three needs matter. British Journal of Social Psychology, 47, 267-283. doi:10.1348/014466607X23879 Simons, H. W., Berkowitz, N. N., & Moyer, R. J. (1970). Similarity, credibility, and attitude change: A review and a theory. Psychological Bulletin, 73, 1-16. doi:10.1037/h0028429 Sundar, S. S. (2008). The MAIN model: A heuristic approach to understanding technology effects on credibility. In J. Metzger & A. J. Flanagin (Eds.), The John D. and Catherine T. MacArthur foundation series on digital media and learning (pp.73-100). Cambridge: MA: MIT Press. Sundar, S. S., & Marathe, S. S. (2010). Personalization versus customization: The importance of agency, privacy, and power usage. Human Communication Research, 36, 298-322. doi: 10.1111/j.1468-2958.2010.01377.x Sundar, S. S., & Nass, C. (2001). Conceptualizing sources in online news. Journal of Communication, 51(1), 52-72. doi: 10.1111/j.1460-2466.2001.tb02872.x Teng, C. I. (2010). Customization, immersion satisfaction, and online gamer loyalty. Computers in Human Behavior, 26, 1547-1554. doi: 10.1016/j.chb.2010.05.02 Tokunaga, R. S. (2013). Engagement with novel virtual environments: The role of perceived novelty and flow in the development of the deficient self-regulation of Internet use and media habits. Human Communication Research, 39, 365-393. doi: 10.111/hcre.12008 Trepte, S., & Reinecke, L. (2010). Avatar creation and video game enjoyment: Effects of lifesatisfaction, game competitiveness, and identification with the avatar. Journal of Media Psychology, 22(4), 171-184. doi: 10.1027/1864-1105/a000022

CUSTOMIZING A VIRTUAL SALESPERSON

31

Vasalou, A., Joinson, A., Bänziger, T., Goldie, P., & Pitt, J. (2008). Avatars in social media: Balancing accuracy, playfulness, and embodied messages. International Journal of Human-Computer Studies, 66, 801-811. doi: 10.1016/j.ijhcs.2008.08.002 Verhagen, T., Feldberg, F., van den Hooff, B., Meents, S., & Merikivi, J. (2012). Understanding users’ motivations to engage in virtual worlds: A multipurpose model and empirical testing. Computers in Human Behavior, 28, 484-495. doi: 10.1016/j.chb.2011.10.020 Wicklund, R. A. (1974). Freedom and reactance. Lawrence Erlbaum. Yee, N., & Bailenson, J. N. (2007). The Proteus effect: Self transformations in virtual reality. Human Communication Research, 33, 271-290. doi:10.1111/j.1468-2958.2007.00299.x Yee, N., Bailenson, J. N., & Duchenaut, N. (2009). The Proteus effect: Implications of transformed digital self-representation on online and offline behavior. Communication Research, 36, 285-312. doi: 10.1177/0093650208330254

CUSTOMIZING A VIRTUAL SALESPERSON

32

Figures

Figure 1. Path coefficients for simple mediation analysis of avatar customization on brand liking through intrinsic motivation. ab = .32, SE = .14, 95% CI [.63, 07]. Note: Dotted line denotes the effect of avatar customization on final brand liking when intrinsic motivation is not included as a mediator. b,c, and c’ are unstandardized logistic regression coefficients, a is an unstandardized OLS regression coefficient. * denotes p < .05.

Figure 2. Path coefficients for simple mediation analysis of avatar customization on purchase intention through intrinsic motivation. ab = .24, SE = .13, 95% CI [.60, 05]. Note: Dotted line denotes the effect of avatar customization on final brand liking when intrinsic motivation is not included as a mediator. b,c, and c’ are unstandardized logistic regression coefficients, a is an unstandardized OLS regression coefficient. * denotes p < .05.

CUSTOMIZING A VIRTUAL SALESPERSON

33

Tables: Table 1 Model for Effect of Avatar Customization on Brand Liking Mediated by Intrinsic Motivation Intrinsic Motivation

Brand Liking

Manip.

Coeff .41

SE .16

p .01

LLCI .73

ULCI Coeff .09 .28

SE .20

IMI

--

--

--

--

--

.76

.13

Constant

4.75

.25

< .0005

4.24

5.25

2.13

.70

p .15

LLCI ULCI -.67 .11

< .0005 .49 .003

.75

1.02 3.52

R2 = .075

R2 = .36

F(1, 78) = 6.34, p = .01

F(2, 77) = 21.41, p < .0005

Table 2 Model for Effect of Avatar Customization on Purchase Intention Mediated by Intrinsic Motivation Intrinsic Motivation

Purchase Intention

Manip.

Coeff .41

SE .16

p .01

LLCI .73

ULCI Coeff .09 .09

SE .27

p .75

LLCI ULCI -.63 .46

IMI

--

--

--

--

--

.59

.18

.002

.22

.96

Constant

4.75

.25

< .0005

4.24

5.25

2.91

.97

.003

.98

4.85

R2 = .075

R2 = .13

F(1, 78) = 6.34, p = .01

F(2, 77) = 5.90, p .004