Mitigating Reactance to Personalized Advertising through
Advertisement Customization
Oskar op de Beke (11110767)
Master’s Thesis
Graduate School of Communication Master’s Programme Communication Science
Supervisor: Dr. Paul E. Ketelaar Date of Completion: June 22, 2016
Abstract
Research has shown that we experience reactance, a form of resistance, towards personalized advertising. This reactance results from a loss of perceived control. The purpose of this study was to determine whether customization as a method of increasing perceived control over personalized ads could be used to reduce reactance towards these ads. In addition, the study sought to uncover the role trait autonomy played in contributing to reactance and in the relationship between customization and reactance. Finally, the study attempted to uncover whether purchase intention could be increased as a result of lower reactance levels towards the personalized ad. In an online experiment (N = 126) comprised of a 2 (Customization: Customization vs. No customization) x 2 (Trait Autonomy: Low vs. High) between subjects’ mixed factorial design, participants were given the chance to either customize visual aspects of an ad that was personalized for them or not given the chance to customize the visual aspects of the personalized ad. The results displayed no effect of providing participants with the opportunity to customize the personalized ad on reactance. Nor was any indirect effect uncovered between customization, reactance and purchase intention. Only one the three dimensions of trait autonomy, authorship/self-congruence, was associated with higher reactance levels. No moderation effect of trait autonomy on the relationship between customization and reactance was found for all three dimensions of trait autonomy.
Imagine you are browsing online for a new television. You click on a television that you like, but you are not yet ready to make the purchase so you instead go to a news website. However, on the right hand side of that news website is an ad for the exact television you were looking at earlier. While you could perceive this as convenient (Vesanen, 2007), odds are you may feel a little uncomfortable (Tam & Ho, 2006; Tucker, 2012). It is possible that you are even annoyed that your behavior on one website is being used to provide targeted ads to you on another (Tucker, 2012). This targeting method, known as dynamic retargeting, is one of the most common forms of online ad personalization (Lambrecht & Tucker, 2013). Personalization involves the collection of information about the user which is later used to create and present a relevant ad (Bang & Wojdynski 2016; Murthi & Sarkar, 2003).
With the advent and growth of internet marketing, we have seen an exponential increase in the use of personal data to improve the targeting power of advertising (Aguirre, Mahr, Gewal, Ruyter & Wetzels, 2015). Retailers frequently use customer data specifically to influence online purchasing behavior (Hawkings, 2012). Big websites such as Google and Yahoo, partner with online retailers to collect and provide consumer data (Angwin, 2012). This data is in turn used to create personalized digital ads (Angwin, 2012). While this personalization of ads may lead to such things as greater ad relevance, privacy concerns cause a sense of discomfort for the consumer (Tam & Ho, 2006; Tucker, 2012). This situation is what is commonly referred to as the personalization-privacy paradox (Aguirre et al., 2015).
Personalization has shown to be an effective tool for improving the delivery of services to customers (Rust & Chung, 2006), improving the cognitive processing of ads (Tam & Ho, 2006), and even increasing clicks on websites (Ansari & Mela, 2003). Not all the benefits are held by the marketer. Personalization can also serve customers by providing them with products and
services that better match their preferences and needs (Vesanen, 2007). It is the juxtaposition of these benefits with privacy concern discomfort that results in the personalization-privacy paradox (Tucker, 2012). In essence, people are concerned about the lack of control over their personal data (Tucker, 2012). This privacy invasion may result in consumers attempting to avoid complying with the persuasive message and displaying several forms of resistance (Aguirre et al., 2015; Puzakova, Rocereto & Kwak, 2013; Tucker, 2012). They may avoid complying by lowering their click through intentions (Aguirre et al., 2015) or purchase behavior (Taylor, Davis & Jillapalli, 2009). One form of resistance in particular is known as reactance. Reactance is defined simply as a motivation to resist something, which has been aroused by a perceived threat to one’s behavioral freedom (Brehm, 1989). When an individual’s data is collected for the purpose of personalization, he or she feels a loss of control over the data, thus precipitating this motivational state (Puzakova et al., 2013).
Current literature regarding privacy concern and personalization provides us with several studies on ways to deal with reactance by providing increased privacy control (Taylor et al., 2009; Tucker, 2014). For example, Tucker (2011) investigated how people’s click through intentions improved for personalized ads when Facebook instigated new privacy controls, allowing its users to select which information Facebook shared with advertisers. This research is backed by fundamental behavioral psychology research that stipulates perceived control as an important factor in reducing reactance (Taylor, 1979). However, there remains the question, whether the perception of control must necessarily result from increased personal privacy controls. Privacy controls, as a method of reducing reactance, are not without negative side effects. Requesting consent when collecting personal information is too time consuming and inefficient (Interactive Advertising Bureaux Europe, 2011). Studies have found that it
significantly disrupts the online experience (Milne, Bahl & Rohm, 2008). Therefore, other methods of reducing reactance should also be considered.
Within the world of targeted ads, personalization is not the only method by which a message can be modified to better fit the tastes or preferences of the intended target customer. Customization, or when the end user has control over the media or marketing experience, is also a commonly used technique (Arora et al., 2008). Customization therefore takes on two forms. The first of these is the customizable media environment. Examples include yahoo.com which lets users select which features they wish to see on their home page. The second form is the customization of actual marketing material. For example, at shareacoke.com, Coca Cola allows customers the opportunity to put their name on a Coca Cola bottle. Nike.com allows its customers to design their own sneakers, selecting from a range of colors and designs. By definition, customization of advertising provides consumers with increased control over ads (Arora et al., 2008). Behavioral psychology would suggest that allowing consumers the chance to customize personalized ads can reduce their reactance (Tucker, 2012).
Current literature provides much information about the impact of privacy controls on reducing reactance (Taylor et al., 2009; Tucker, 2011; Tucker, 2014). However, there is a significant gap in research regarding how methods of increasing user control other than privacy controls can be used to reduce reactance. The aim of this research is to demonstrate that an alternative form of control (customization) over personalized ads may also be successful in mitigating the negative effects of personalized ads. In addition, past research on reactance has identified a positive association between reactance and certain personality variables such as assertiveness, possessiveness, and suspiciousness (Seemann, Buboltz, Thomas, Soper & Wilkinson, 2005). Evidence from past literature suggests there may be a negative relationship
between another basic personality variable, trait autonomy, and reactance (Knee & Zuckerman, 1998; Pavey & Sparks, 2010). However, the literature has yet to directly look into how trait autonomy, a personality variable directly tied to how we perceive and experience our own behaviors, is related to reactance (Weinstein, Pryzbylski & Ryan, 2012). Trait autonomy is characterized by three central components: claiming yourself as the origin of your behaviors, insusceptibility to external behavioral control, and taking an interest in your own behaviors (Weinstein et al., 2012). This study therefore also aims to further our understanding of the underlying mechanisms that contribute to reactance by exploring the relationship between trait autonomy and reactance, as well as the effect trait autonomy has on the relationship between customization and reactance. Finally, this study seeks to understand whether the mitigating effects of customization on reactance can improve a customer’s purchase intentions, an important marketing metric.
The research question this study aims to answer is therefore, to what extent increased perceived control through the customization of personalized ads can serve to reduce reactance towards the ads and indirectly increase purchase intention. In addition, how does trait autonomy relate to reactant motivations and to what extent does trait autonomy moderate the relationship between customization and reactance?
Theoretical Section Personalization
The simplest definition of personalization is, “making something identifiable as belonging to a particular person” (“Personalize,” n.d.). Bang & Wojdynski (2016) have defined personalized advertising as advertising that “incorporates information about the individual, such as demographic information, personally identifiable information and shopping related
information” (p. 868). As a marketing strategy, personalization is aimed at providing the individual customer with relevant content (Tam & Ho, 2006). The strategy is intended to reduce the amount of effort a customer has to expend in order to meet his/her needs (Aguirre et al., 2015; Montgomery & Smith, 2009). The rapid development of advanced online data collection has made personalization a common practice for online retailers such as Amazon.com (Aguirre et al., 2015).
Personalization, according to the Federal Trade Commission, can be categorized by the level of anonymity represented in the information collected from the user (Privacy Online: Fair Information, 2000). At the most basic level, the information collected is practically anonymous (browser version, language, or local time). The next categorical level includes information that is personally unidentifiable such as age, date of birth, or gender. In addition, modern technology allows firms to track online behaviors (Goldfarb & Tucker, 2011). Firms can track a user’s movements online and use that information to target ads (Aguirre et al., 2015). Because online behavior on its own cannot be used to identify a person, this form of personalization can also be included in the second category. The third category, representing the highest level of personalization, includes information that can be used to identify or locate a person such as an email address, name, postal address, credit card number or even a social security number (Privacy Online: Fair Information, 2000).
Because personalization is an important marketing tool that involves the collection of personal data, often without an individual even noticing, it is important to address possible forms of resistance (Aguirre et al., 2015; Lambrecht & Tucker, 2013). One form of resistance emanates from the restriction of behavioral control over personal data that personalization places on users.
Thus, the following section discusses a form of resistance users might have towards this perceived restriction of behavioral freedom.
Reactance
According to self-determination theory (SDT), central to human motivations is our innate need for autonomy, relatedness, and competence (Deci & Ryan, 2000). Autonomy relates to the ability to maintain behavioral control (Weinstein et al., 2012). Because it is an innate need, we, as human beings, become resistant when our autonomy is threatened. This form of individual resistance was identified by Brehm (1989) in his works regarding reactance theory.
Reactance theory states that people acquire reactant motivations when they feel they are experiencing a threat to their behavioral control. In response to this threat, people then become motivated to act in a manner that will restore their autonomy (Brehm, 1989). In advertisement personalization, autonomy is threatened because we lose behavioral control over our personal data. We are no longer able to control everything that happens to our data. As a result, we become reactant or motivated to act in order to retain the threatened control (Fitzimons & Lehmann, 2004). Reactance therefore represents a cognitive manner by which consumers become motivated to, and subsequently resist, personalized advertising (Aguirre et al., 2015).
However, in order to fully capture why reactance is such a powerful state, it is necessary to address the reasons why personal data mean so much to a person. Based on psychological ownership theory, as humans we can form strong cognitive or affective attachments to external objects resulting in a sense of ownership over those objects (Avey, Avolio, Crossley & Luthans, 2009; Pierce, Kostova, & Dirks, 2001). This attachment causes us to create a sense of ownership over our personal data. Ownership can be defined as the “act, state, or right of possessing something,” (“Ownership,” n.d.) and therefore requires us to maintain control over our personal
data. Any violation of that sense of ownership can result in significant negative emotional consequences (Aguirre et al., 2015). Hence, we are prone to act towards maintaining ownership over our personal data. Consumers therefore become reactant towards personalization because they feel a loss of ownership or freedom to control their personal data (Puzakova et al., 2013).
Past research has demonstrated that consumers experience reactance when they receive personalized messages (White, Zahay, Thorbjørnsen, & Shavitt, 2008). Further research has demonstrated that introducing privacy controls can serve to reduce the levels of reactance that consumers feel towards these messages (Tucker, 2014). Other studies have shown that just by increasing the perception of information control, the negative effects of privacy concerns on behavioral intentions to resist an ad can be mitigated (Taylor et al., 2009). Research has also demonstrated that perceived control may play an important role in mitigating reactant motivations as well as in increasing intentions to use personalized web services (Tucker, 2012; van Velsen et al., 2015). The following section, therefore, delineates how another form of control, customization, can serve to reduce reactance levels by increasing levels of perceived control over the personalized ad.
Reactance and Customization
Behavioral control can be defined as the “availability of a response which may directly influence or modify the objective characteristics of an event” (Averill, 1973, p. 293). Our perception of this concept can be referred to as perceived control. Simply by being given choices in our environment we have increased levels of perceived control (Veitch & Gifford, 1996). Fundamental behavioral psychology research indicates that perceived control has many benefits in general, such as increased environmental satisfaction and improved job performance (Greenberger, Strasser, Cummings & Dunham, 1989). Of great significance is the notable effect
perceived control has on the impact of a stressor. Research conducted using elderly patients in a nursing home demonstrated that providing increased environmental control, such as caring for a plant, resulted in positive health benefits (Rodin & Langer, 1977). In addition, Thompson and Prottas (2006) demonstrated that increased perceived control can reduce the negative impact of a stressor in both familial and occupational environments.
In the context of personalized advertising, the threat to our freedom of behavioral control, caused by firms collecting and using our personal information, is a stressor. Based on the above discussion of perceived control, it is plausible that increasing one’s perceived behavioral control in the context of personalized advertising can reduce the negative impact of the stressor. When this stressor’s impact is reduced, the reactant motivations caused by the threat should in turn diminish. Thus, customization, or user control over the advertising material, may serve to reduce reactance to personalized advertising (Arora et al., 2008). It would achieve this reduction through the mitigating effect increased perceived control has on the stress caused by the threat to behavioral control.
Research on reactance suggests that reactant motivations can occur unconsciously (Chartrand, Dalton & Fitzimons, 2007). In addition, research on unconscious goal pursuit proposes that these motivations can be satisfied, and therefore reduced, through goal satiation tasks (Chartrand et al., 2008; Laran, Dalton, & Andrade, 2011). Moreover, research shows that certain unconscious resistance motivations can be satiated without a person’s awareness (Laran et al., 2011). In addition to this, goal satiation tasks do not need to be directly related to the source of the motivation. For example, Laran et al. (2011) discovered that an unconscious motivation to resist a slogan could be satiated by picturing a parental resistance situation. This evidence suggests that unconscious reactant motivations can be satisfied through a goal satiation
task that is not directly related to the origin of the motivation -- the origin of this motivation being lack of behavioral control over personal data. Customization, in giving the user a way of satiating their motivation for control, may therefore serve to reduce reactance to personalized ads.
Considering the evidence regarding the effect of perceived control on a stressor, as well as the possibility for unconscious motivations to be satisfied through an unrelated satiation task, the question remains whether increased perceived control through customization can serve to reduce reactance levels. Therefore, the following hypothesis was developed.
H1: The presence of control through customization over the personalized ad will lead to a lower level of reactance towards the ad compared to the absence of control.
Trait Autonomy and Reactance
Conceptually, autonomy exists in two forms. The first form is a state of being, for example the state of having behavioral control over one’s personal data (Weinstein et al., 2012). The second form is an aspect of one’s personality (Weinstein et al., 2012). This can be referred to as trait autonomy. Individuals who are high in trait autonomy are high in authorship/self-congruence, meaning they are willing to endorse their behaviors as originating from themselves and being self-congruent. They are also not susceptible to control and take an interest in their own actions. In other words, all or most of their behaviors are self-initiated and they are aware of this fact (Weinstein et al., 2012). Self-determination theory (SDT) provides context for the relationship between autonomy and reactance. SDT proposes that we have an innate need for autonomy (Deci & Ryan, 2000). When we have reduced behavioral control over our personal data, we experience reduced state autonomy, because actions regarding our data were regulated
by an external force. Based on the SDT and reactance theory, this infringement upon our state autonomy may then become central to our motivations to restore our sense of behavioral control. These motivations will likely manifest themselves as reactance. However, past research suggests that high levels of trait autonomy may make up for this diminished level of state autonomy. For example, people with high levels of trait autonomy perceive information to be less freedom threatening (Pavey & Sparks, 2010). This research suggests that people with high levels of autonomy may perceive the loss of control over their personal data as less of an encroachment on their behavioral freedom. As a result, their reactant motivations may be mitigated. In addition, individuals with higher levels of autonomy exhibit fewer defensive behaviors and coping strategies (Knee & Zuckerman, 1998). Reactant behaviors, such as ignoring or avoiding an ad, are defensive reactions to the loss of control caused by personalization as well as a form of coping with this loss of control. The evidence presented above suggests that those with higher levels of trait autonomy may demonstrate lower levels of reactance.
H2: Participants with high levels of trait autonomy will demonstrate lower levels of reactance towards the personalized ad compared to participants with low levels of trait autonomy.
Furthermore, research has discovered that trait autonomy is significantly correlated with greater feelings of perceived control (Newcomb, Huba & Bentler, 1986). When an individual already experiences a high level of perceived control, it is unlikely that this person will be strongly affected by a manipulation intended to increase his or her perceived control. For this reason, customization of personalized advertising should not significantly increase the perception of control for individuals with high autonomy and therefore will likely have no impact on their levels of reactance. However, individuals with low levels of autonomy should
experience a significant change in perceived control as a result of being allowed to customize a personalized ad and therefore have reduced levels of reactance.
H3: For participants with low levels of trait autonomy, the presence of control through customization over the personalized ad will lead to lower levels of reactance towards the ad compared to the absence of control. However, for participants with high levels of trait autonomy, the presence of control through customization will not produce any significant difference in levels of reactance compared to the absence of control.
Reactance and Purchase Intention
Reactance is characterized by a motivation to restore behavioral control (Brehm, 1989). When experiencing reactance, consumers can either ignore the persuasive message or perform a behavior that intentionally contradicts the persuasive attempt (Fitzsimons & Lehmann, 2004). Ads often serve the purpose of simply getting a brand into the consumers’ consideration set (Belch & Belch, 2015, p. 119). A consumer’s consideration set is a set of the brands or products that a consumer would consider purchasing for a given purchase goal (Shapiro, Macinnis & Heckler, 1997). It is from the consideration set that consumers develop a purchase intention, or a predisposition to buy a certain brand or product (Belch & Belch, 2015, p. 124). Because it is the goal of reactant motivations to perform a behavior contradictory to the intention of the control reducing threat, in this case the personalized ad, it is plausible that consumers may deny entry of the brand or product displayed in the personalized ad into their consideration set. As such, it is plausible that by reducing reactance, consumers will be more likely to include the brand in their consideration set, increasing their purchase intention.
H4: The presence of control through customization over the personalized ad will lead to increased purchase intention by lowering reactance levels to the ad.
See Figure 1 for a visual depiction of the proposed model.
Methods
The experiment was comprised of a 2 (Customization: Customization vs. No customization) x 2 (Trait Autonomy: Low vs. High) between subjects’ mixed factorial design with customization as a true independent variable and trait autonomy as a quasi-independent
variable. Participants were exposed to the same personalized ad and the same scenario after which they were randomly sorted into either the Customization or the No-customization condition. Those in the Customization condition were able to visually customize the personalized ad, while those in the No-customization condition were not.
Participants
N = 126 (Male = 53, Female = 73) of internet users in the U.S. were recruited to participate in this study. Three participants were excluded as outliers, because they had scores two or more standard deviations above the mean on at least two measures. All participants fell in the 18-32 age bracket (M = 24.03, SD = 2.77), which makes up the generation that spends the most time on the Internet per day (“Average duration”, n.d.). Furthermore, around 31% of online shoppers in the U.S fell in this age range between the years 2012 and 2016 (“Age distribution”, n.d.). Therefore, this is an ideal sample with which to study the effects of online personalized advertising. The sample was selected using a combination of a convenience sampling method and a snowballing method facilitated by the use of social networking sites. This method ensured that the sample consisted of internet users as well as people within the correct age range. A one-way ANOVA indicated that the Customization and No-customization groups did not differ in age, F (1, 121) = 0.28, p = 599, p2 = .000, 95% CI [-.63, 1.33]. Chi-squared tests indicated that
the groups did not differ in gender, χ2 (1, N = 126) = 1.46, p = .925, nor in education level, χ2 (1, N = 123) = .22, p = .636. Thus, the groups were successfully randomized.
Personalization method. Both groups were exposed to the same personalized and non-customized ad for a new credit card from Commonwealth Bank (Figure 1). Commonwealth Bank was chosen because it is an Australian bank and therefore not likely to be familiar to the
sample. The personalization method was selected so as to meet both theoretical conceptualizations of personalization and to be comparable to methods used in past research (Bang & Wojdynski, 2016; Kobsa, 2007; Aguierre et al., 2015). Consistent with theory, the ad included specific pieces of personal data ranging
from information people readily disclose (age, username selected by the participant) to information people are less willing to disclose or generally not comfortable disclosing (city in which the participants live, zip or postal code, income range) (Bang & Wojdynski, 2016; Kobsa, 2007). This information was collected directly from the participant at the beginning of the experiment. In order to not compromise the participants’ privacy and sense of
comfort within the experiment, they were not asked to provide their actual income; rather they were asked to imagine having a certain income. In addition to this, past research has successfully utilized a scenario in order to inform participants that their online behavior was tracked, then
Figure 1. Personalized and non-customized
ad for both experimental groups.
Figure 2. Scenario
used to target them with an ad (Aguirre et al., 2015). Therefore, participants were also given a scenario indicating that their online behavior was being tracked. See Figures 1 and 2 for an example of a personalized ad as well as the accompanying scenario. While personalization is not manipulated in this study, this personalization method was pretested against a non-personalized ad that did not feature any personal details combined with a scenario that did not indicate that the participants’ online behaviors were being tracked. This was done in order to ensure the ad was sufficiently personalized in the eyes of the participants. The four items that make up the personalization scale are as follows: “This advertisement is directed to me personally,” “I recognize my personal situation in this advertisement,” “This advertisement takes into account the situation I am facing,” and “This advertisement takes into account my personal situation” (α = .92) (Aguirre et al., 2015). Results of this pretest indicated that the participants in the personalized ad condition (M = 5.76, SD = 1.23, n = 30) rated the ad as significantly more personalized than those in the non-personalized ad condition at a .01 alpha level (M = 3.33, SD = 1.28, n = 32), t (60) = 7.63, p2= .492, p < .001, 99% CI [1.79, 3.07].
Manipulation of customization. Both groups viewed a non-customized version of the personalized ad at the beginning of the experiment (See Figure 2). After seeing the same personalized, non-customized ad, the groups were randomly sorted into the Customization and No-customization group. Due to the lack of research specifically manipulating advertisement customization, the customization manipulation developed for this study is entirely novel. The Customization group was given three options to customize the personalized ad. In terms of online advertising, this could mean selecting different colors, shapes, graphics, as well as the order of message contents (Ko, Cho, & Roberts, 2005). Therefore, customization for this experiment involved the manipulation of visual aspects of the message. The participants could
choose one of four font colors (red, green, blue, purple), one of four fonts (Arial, Courier New, Comic Sans, and Georgia), as well as one of two font sizes for the phrase “Platinum Card” (font size 10 or 14). The No-customization group was also once again presented with the personalized ad. However, instead of being given customization options, they were simply told that the personalized ad would be modified. The ad was then modified using a random combination of the above three customization options. This customization method was pretested using a two-item measure of perceived control. The two-items are as follows: “Please indicate the extent to which you felt control over the Commonwealth credit card ad,” and “I felt control over the Commonwealth credit card ad” (r = .85, p < .001) (Arcury, Quandt, & Russell, 2002). Items were measured using a seven-point Likert scale (“None/Strongly disagree” to “A lot/Strongly agree”) Results of the pretest indicated that those in the Customization condition (M = 4.90, SD = 1.23, n = 30) rated their level of perceived control over the ad as significantly greater than those in the No-customization condition at a .01 alpha level (M = 3.13, SD =1.44, n = 32), t (60) = 5.20, p < .001, p2 = .311, 99% CI [1.09, 2.46].
Measures
Trait Autonomy. Participants completed a 15-item measure of trait autonomy developed by Weinstein et al. (2012) using theoretical and empirical analysis. The scale consists of three subscales based on the three central components of trait autonomy; authorship/self-congruence (α = .75), susceptibility to control (α = .77), and interest-taking (α = .88). Example items include “My decisions represent my most important values and feelings,” “I do things in order to avoid feeling badly about myself,” and “I often reflect on why I react the way I do”. Items were measured using a seven-point Likert scale (“strongly disagree” to “strongly agree”). The items under the susceptibility to control measure were reverse scored. For the purposes of maintaining
reliability and to uncover which component contributed most to reactance, each subscale was analyzed individually.
Purchase intention. To measure purchase intention, participants completed an adapted version of a four-item measure developed by Dodds, Monroe and Gewal (1991) using a seven-point Likert scale (“very low/strongly disagree” to “very high/strongly agree”) (α = .95). The items are as follows: “The likelihood of signing up for the Commonwealth credit card is,” “I would consider signing up for the Commonwealth credit card,” “The probability that I would consider signing up for the Commonwealth credit card is,” and “My willingness to sign up for the Commonwealth credit card is”. This scale was chosen based on the fact that it addresses cognitive (consideration), affective (willingness), and behavioral (likelihood of signing up) aspects of purchase intention. It was modified to say “signing up” rather than “purchasing” in order to make it more applicable to the experimental scenario of a credit card.
Reactance. Reactance was measured using an adapted version of a seven-item scale developed by White et al. (2008). A seven-point Likert scale was also used for this measure (“very unlikely” to “very likely”). Example items are as follows: “How likely are you to find the Commonwealth credit card ad interfering?” and “How likely are you to resist the Commonwealth credit card ad?” This scale was chosen because it addresses both the restriction of freedom (intrusive, unwelcome etc.) as well as the motivation to resist the message (resist, dismiss, ignore) that make up reactance. Furthermore, the scale is tailored specifically towards online advertising.
Manipulation check. Participants completed the same two-item measure of perceived control used in the pretest at the end of the experiment as a manipulation check.
Familiarity. Participants also completed a three-item measure of brand familiarity to ensure they were not familiar with the bank and therefore did not carry any preexisting attitudes towards the brand. The items are as follows: “Commonwealth Bank is familiar to me,” “I have heard of Commonwealth Bank,” “I do not know Commonwealth Bank” (α = .74) (Delgado-Ballester, Navarro & Sicilia, 2012). The third item was reverse scored. See Appendix for a complete list of items found in each measure.
Procedure
Participants began the experiment by reading over a consent letter. In the consent letter, participants agreed that they had been sufficiently informed, that they retained the right to withdraw consent and halt participation at any point, that they were guaranteed anonymity, and were provided with contact information if they wished to know more about the study. Once they agreed to give their consent and participate, they moved on to the demographics section. In the demographics section, participants chose a username for themselves, provided their age, the city in which they live, their zip (postal) code, their gender, and highest level of completed education (Crissey, 2009). The demographics section occurred at the beginning of the experiment rather than at the end, because personal details from this section were needed for the personalized ad. Participants then completed the 15-item measure of trait autonomy. After this, both groups were told that they would be given a scenario and be shown an ad. To ensure that the participants spent a sufficient amount of time reading the scenario and looking at the ad, they were asked to read and look over both carefully as they would be asked questions regarding both later on. Both groups were then given the same scenario describing how they had recently posted their new job position on Facebook, what the income was for that job position, and that they had just canceled their current credit card and were actively looking for a new one (See Figure 2). They were then
asked to imagine that they came across the personalized ad on their Facebook timeline (Aguirre et al., 2015). Both groups witnessed the same personalized ad from Commonwealth Bank using the personal details they entered in the demographics section (See Figure 1). The lack of research involving advertisement customization required the
development of a novel customization process. Participants in the Customization group were then told that they would now have the opportunity to customize the ad. They were then given the three customization options (font, font color, and font size). Finally, they
were told that they would now have the opportunity to view their customized ad (See Figure 3),
after which they were shown their now-customized personalized ad (See Figure 4). Participants in the No-customization group, after seeing the personalized ad, were simply told that they
would now be shown the same ad, but with several of the visual aspects of the ad having been modified (See Figure 3). They were then shown the personalized ad with a random combination of the three customization options (i.e. red, comic sans, font size 10). See Figure 4 for an example of a customized (or modified) personalized ad. All participants then completed the measure of purchase intention followed by the measure of reactance. Finally, participants completed a manipulation check for perceived control as well as a measure of brand familiarity for Commonwealth Bank. To conclude the experiment, participants were thanked for their participation. Participants who wished to be debriefed were later sent a link to the experiment debrief.
Results
Manipulation Check and Familiarity
A one-way ANOVA was used to determine whether those in the Customization group experienced more perceived control over the personalized ad than the No-customization group. The ANOVA test indicated that those in the Customization condition (M = 3.25, SD = 1.60) reported having more perceived control over the personalized ad than those in the No-customization condition at a .01 alpha level (M = 2.51, SD = 1.27), F (1, 121) = 8.09, p = .005,
p2 = .063, 99% CI [.23, 1.26]. Therefore, the customization manipulation was successful in
increasing perceived control. In addition, participants were generally not very familiar with Commonwealth Bank (M = 1.35, SD = 1.42), with 1 being not familiar at all and 7 being very familiar.
Main Analysis
Hypothesis 1, that the presence of control through customization over the personalized ad will lead to a lower level of reactance towards the ad compared to the absence of control was not
Figure 4. Customized (or Modified)
supported. To test whether there was a significant difference in reactance scores between the Customization and No-customization groups, a one-way ANOVA was conducted. The one-way ANOVA indicated that those in the Customization condition (M = 5.39, SD = 1.11) did not differ significantly in reactance from the No-customization condition (M = 5.43, SD = .91), F (1, 121) = 0.05, p = .821, p2 = .000, 95% CI [-.32, .41].
Hypothesis 2, that participants with high levels of trait autonomy will demonstrate lower levels of reactance towards the personalized ad compared to participants with low levels of trait autonomy was not supported. To determine the relationship between trait autonomy and reactance, three separate bivariate correlation analyses were conducted for each of the three trait autonomy subscales. The first correlation analysis indicated that higher levels of authorship/self-congruence (Autonomy Subscale 1) was weakly associated with higher reactance scores at a .01 alpha level, r = .24, p = .007, N = 123. The other two correlation analyses indicated no significant association between levels of susceptibility to control (Autonomy Subscale 2) and reactance (p = .243), and interest taking (Autonomy Subscale 3) and reactance (p = .401).
Hypothesis 3 was also not supported. The hypothesis is as follows: for participants with low levels of trait autonomy, the presence of control through customization over the personalized ad will lead to lower levels of reactance towards the ad compared to the absence of control. However, for participants with high levels of trait autonomy, the presence of control through customization will not produce any significant difference in levels of reactance compared to the absence of control. Three individual simple slopes moderation analyses were conducted using Hayes (2013) PROCESS plug in for SPSS (Model 1) to determine whether the three trait autonomy subscales moderated the relationship between customization and reactance. A simple slopes analysis was conducted instead of performing a median split for the trait autonomy
subscales in order to avoid the reduction of predictor variability. The overall model with authorship/self-congruence (Autonomy Subscale 1) as the moderator variable approached significance, F (3, 119) = 2.50, p = .063, R2 = .06. However, this was driven almost entirely by authorship/self-congruence as a predictor, b = 0.35, t (119) = 1.82, p = .071. The interaction was not significant (p = .827). The simple slopes analyses with susceptibility to control (p = .296) and interest taking as moderators were not significant (p = .763).
Hypothesis 4, that the presence of control through customization over the personalized ad will lead to increased purchase intention by lowering reactance levels to the ad, was also not supported. A linear regression analysis using Hayes (2013) PROCESS plug in for SPSS (Model 4) was conducted to determine whether there existed an indirect relationship between customization and purchase intention through lower levels of reactance. The Hayes (2013) PROCESS plug-in tests for the prediction power of customization on reactance, reactance on purchase intention, and customization on purchase intention. The overall mediation model was nonsignificant (p = .821). Reactance scores did, however, significantly predict purchase intention at a .01 alpha level, b = -0.79, t (120) = -6.56, p < .001, 99% CI [-1.02, -.55]. See Figure 5 for an overview of the findings.
Conclusion
The aim of this study was to determine whether increased control over personalized ads through customization could be used to reduce reactance to those ads and in turn increase purchase intention. In addition, the study sought to understand how trait autonomy relates to
Figure 5. Research model and results.
reactance and how it impacts the relationship between customization and reactance. The analyses displayed no effect of providing participants with the opportunity to customize the personalized ad on reactance. Nor was any mediation effect uncovered between customization, reactance and purchase intention. One of three dimensions of trait autonomy, authorship/self-congruence, was associated with higher reactance levels, while the other two dimensions were not associated with reactance. No moderation effect of trait autonomy on the relationship between customization and reactance was found for all three subscales.
Discussion
The lack of effect of customization on reactance indicates that no claim can be made stipulating that an alternative form of control, unrelated to the source of the reactant motivations, can be used to reduce reactance whether by reducing stress (Thompson & Prottas, 2006; Rodin & Langer, 1977) or satisfying unconscious motivations (Laran et al., 2011). With regard to reactance theory, this lack of effect indicates that no modification can be made to the theory regarding the importance of other sources of perceived control on reducing reactant motivations (Brehm, 1989).
However, the customization’s inability to reduce reactance may be attributed to a weak manipulation. On average, even those in the customization condition felt little control over the ad (M = 3.25, SD = 1.60), with 1 representing no control and 7 representing a lot of control, and a moderately high amount of reactance towards the ad (M = 5.39, SD = 1.11), with 1 representing very little reactance and 7 representing a great deal of reactance. The weak manipulation may have been the result of several important limitations. First, the three customization options (font, font color, font size) may have been too habitual for people who commonly use word processing
software and therefore lacked the novelty necessary to produce significant feelings of perceived control. This is supported by habit research stating that habitual behaviors require less self-control (Verplanken & Wood, 2006). As a result of needing less self-self-control participants may have felt less control over the ad, leading to lower feelings of perceived control and in turn greater reactance scores. In addition, habit research shows that habitual behaviors can reduce our sensitivity to changes in our environment (Verplanken & Wood, 2006). For this reason, it is plausible that participants were not sufficiently aware of the changes caused by the customization, contributing further to lower perceived control and greater reactance scores. The lack of interactivity involved in the customization process may have contributed further to the weak manipulation. This reasoning is supported by research showing that online interactivity leads to greater perceived control (Wu, 2005). Furthermore, interactivity is strongly associated with attitude (Chung & Zhao, 2004), such that the lack of interactivity in the customization process could have caused a negative impact on attitude towards the ad, reinforcing reactant intentions. This is consistent with findings showing that reactance is strongly associated with state anger (Quick & Stephenson, 2007). Finally, the weak manipulation was hindered by the fact that those in the No-customization group also felt control over the ad, leading to feelings of perceived control over the personalized ad that did not differ greatly from the Customization group (M = 3.25, SD = 1.60). This is evidenced by the fact that participants in the No-customization group on average felt a little control over the ad rather than none (M = 2.51, SD = 1.27). This finding can be explained by the fact that all participants were allowed to input the information used in the personalized ad themselves, providing them with some control over the actual content of the personalized ad. Empirical evidence regarding control over our environment
leading to greater perceived control supports the reasoning that all participants would have felt some perceived control over the personalized ad (Veitch & Gifford, 1996).
Hypothesis 2 expected to uncover a negative association between trait autonomy and reactance levels. However, the only association that was uncovered was a weak, though significant, positive association between authorship/self-congruence and reactance levels. The other two components of autonomy, susceptibility to control and interest taking, did not have any association with reactance. This finding is inconsistent with past research showing that people high in autonomy tend to perceive information to be less freedom threatening (Pavey & Sparks, 2010) and exhibit fewer defensive behaviors and coping strategies (Knee & Zuckerman, 1998). However, past literature does provide some support for the positive association between authorship/self-congruence and reactance levels. People with greater authorship/self-congruence are more aware of the source of their behaviors and therefore may be more aware when their behavioral control is restricted (Weinstein et al., 2012). In addition, when the ability to claim a behavior as self-originated is taken away, it could result in discomfort for an individual high in congruence (Weinstein et al., 2012). The association between authorship/congruence and reactance is consistent with research showing that people higher in self-awareness tend to be more reactant (Carver & Scheier, 1981). Also consistent, is research showing that reactance is associated with being more self-assured (Dowd, Wallbrown, Sanders & Yesenosky, 1994).
Hypothesis 3 expected to uncover a moderation effect such that people with higher levels of trait autonomy experienced no effect of customization on reactance while those with low levels did experience an effect. The lack of moderation effect in combination with the lack of main effect of customization on reactance indicates that no conclusion can be drawn regarding
whether the findings are consistent with the theorization that those who innately feel greater perceived control are not impacted by a manipulation designed to increase their perceived control (Newcomb et al., 1986). However, the finding from the Hypothesis 2 analysis, that reactance is associated with greater authorship/self-congruence, provides more support for a moderation effect, albeit a moderation effect specifically for the authorship/self-congruence component of trait autonomy. People who are higher in authorship/self-congruence are more aware of the source of their behavior. Therefore, when their behavioral control is restricted, they are not likely to reduce their reactance when they are given control over something unrelated. This is due to the fact that they are more consciously aware that this other form of control is not related to the source of their restricted behavioral control. Therefore, this finding provides support for the hypothesis that an effect of customization on reactance should occur for people with low authorship/self-congruence, but not for people high in this component.
Hypothesis 4 expected to discover an indirect positive effect of customization on purchase intention by reducing reactance towards the personalized ad. However, no mediation effect was uncovered. This lack of findings can also be explained by the weak customization manipulation. As reactance scores did predict scores on purchase intention, or in the case of this study, credit card sign-up intention. The fact that people with lower reactance scores demonstrated higher purchase intention provides supports for the theorization that consumers may refuse to include a brand or product in their consideration set as a way of expressing their reactant motivations (Belch & Belch, 2015, p. 119; Fitzsimons & Lehmann, 2004). This finding also is consistent with past research demonstrating lower purchase behavior as a form of resisting advertising (Taylor et al., 2009).
The findings from this study do not have any significant societal or consumer applications. However, the result indicating people with higher authorship/self-congruence tend to be more reactant to personalized ads suggests that marketers should use caution when targeting more autonomous people with personalized advertising. This result also provides a possible additional component to reactance theory. More specifically, this finding suggests that awareness of the loss of behavioral freedom is an important factor in determining reactance levels. In addition, customization may be one way of increasing user perceived control, though marketers need to ensure the customization method is novel and interactive.
When building off of this research in the future, there are some important recommendations to take into consideration. The first recommendation is to utilize a stronger, more novel and realistic customization method. More specifically, future research should utilize a method that involves a form of customization that people do not encounter in their everyday lives. For example, a novel method could come in the form of allowing participants to change the color of a pair of sneakers in the ad. In addition to this, the customization method should employ a more interactive design, due to interactivity’s close tie to increased perception of control (Wu, 2005). Using the sneaker example above, an interactive design suggests that the participants modify the color of the sneakers by clicking directly on the ad itself, rather than being given color options separately from the ad. In addition, interactivity suggests that the participants witness an immediate change in the ad rather than a change that is delayed by having to click to the next page. The limitations regarding the manipulation described earlier also present the need to reduce the perception of control over the ad experienced by the No-customization group. Therefore, future research should consider creating more mental distance between the collection of personal data and the personalized ad. This can be achieved possibly by collecting the
personal data at an earlier time, such as one week before the actual manipulation. The temporal distance will weaken the mental connection between the specific data that was collected and the personalized ad (Trope, Liberman & Wakslak, 2007). In addition to this, the personal data could be collected using a task users are made to perceive as unrelated to the personalized ad. For example, they could be told they are participating in two separate studies. In doing so, participants would be less likely to connect their input of personal information to the personalized ad they witness later on. Future research should make use of these recommendations in uncovering whether customization as a form of increasing user control can be used to reduce reactance to personalized ads and in turn increase purchase intention.
References
Aguirre, E., Mahr, D., Grewal, D., de Ruyter, K., & Wetzels, M. (2015). Unraveling the Personalization Paradox: The Effect of Information Collection and Trust-Building Strategies on Online Advertisement Effectiveness. Journal of Retailing, 91(1), 34-49. Angwin, J. (2012, June 17). Online tracking ramps up. The Wall Street Journal. Retrieved from
http://www.wsj.com/articles/SB10001424052702303836404577472491637833420 Ansari, A., & Mela, C. F. (2003). E-customization. Journal of marketing research, 40(2),
131-145.
Arora, N., Dreze, X., Ghose, A., Hess, J. D., Iyengar, R., Jing, B., Joshi, Y., Kumar, V., Lurie, N., Neslin, Sajeesh, S., Su, M., Syam, N., Thomas, J., & Zhang, Z. J. (2008). Putting
one-to-one marketing to work: Personalization, customization, and choice. Marketing Letters, 19(3-4), 305-321.
Averill, J. R. (1973). Personal control over aversive stimuli and its relationship to stress. Psychological bulletin, 80(4), 286.
Avey, J. B., Avolio, B. J., Crossley, C. D., & Luthans, F. (2009). Psychological ownership: Theoretical extensions, measurement and relation to work outcomes. Journal of Organizational Behavior, 30(2), 173-191.
Bang, H., & Wojdynski, B. W. (2016). Tracking users' visual attention and responses to personalized advertising based on task cognitive demand. Computers in Human Behavior, 55, 867-876.
Belch, G. E., & Belch, M. A. (2015). Perspectives on consumer behavior. In Advertising and
promotion: An integrated marketing communications perspective (10E Global ed., pp.
107-138). McGraw Hill Education.
Brehm, J. W. (1989). Psychological reactance: Theory and applications. Advances in consumer research, 16(1), 72-75.
Carver, C. S. & Scheier, M. F. (1981). Self-consciousness and reactance. Journal of Research in Personality, 15, 16-29.
Chartrand, T. L., Dalton, A. N., & Fitzsimons, G. J. (2007). Nonconscious relationship
reactance: When significant others prime opposing goals. Journal of Experimental Social Psychology, 43(5), 719-726.
Chartrand, T. L., Huber, J., Shiv, B., & Tanner, R. J. (2008). Nonconscious goals and consumer choice. Journal of Consumer Research, 35(2), 189-201.
Chung, H., & Zhao, X. (2004). Effects of perceived interactivity on web site preference and memory: Role of personal motivation. Journal of Computer‐Mediated
Communication, 10(1), 00-00.
Crissey, S. R. (2009). Educational Attainment in the United States: 2007. US Department of Commerce.
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(4), 227-268.
Dowd, E. T., Wallbrown, F., Sanders, D., & Yesenosky, J. M. (1994). Psychological reactance and its relationship to normal personality variables. Cognitive Therapy and
Research, 18(6), 601-612.
Fitzsimons, G. J., & Lehmann, D. R. (2004). Reactance to recommendations: When unsolicited advice yields contrary responses. Marketing Science,23(1), 82-94.
Goldfarb, A., & Tucker, C. (2011). Online display advertising: Targeting and obtrusiveness. Marketing Science, 30(3), 389-404.
Greenberger, D. B., Strasser, S., Cummings, L. L., & Dunham, R. B. (1989). The impact of personal control on performance and satisfaction. Organizational Behavior and Human
Decision Processes, 43(1), 29-51.
Hawkings, G. (2012). Will Big Data Kill All but the Biggest Retailers? Harvard Business
Review.
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A
regression-based approach. Guilford Press.
Interactive Advertising Bureaux. (2012, April). Internet ad revenues hit $31 billion in 2011, historic high up 22% over 2010 record-breaking numbers. Retrieved from
http://www.iab.com/news/internet-ad-revenues-hit-31-billion-2011-historic-high-22-2010-record-breaking-numbers/
Knee, C. R., & Zuckerman, M. (1998). A nondefensive personality: Autonomy and control as moderators of defensive coping and self-handicapping.Journal of Research in
Personality, 32(2), 115-130.
Knowles, E. S., & Linn, J. A. (2004). The importance of resistance to persuasion. In E. S.Knowles & J. A. Linn (Eds.), Resistance and Persuasion. Mahwah, New Jersey: Lawrence Erlbaum Associates.
http://www.communicationcache.com/uploads/1/0/8/8/10887248/resistance_and_persuasi on.pdf
Kobsa, A. (2007). Privacy-enhanced web personalization. In The adaptive web (pp. 628-670). Springer Berlin Heidelberg.
Lambrecht, A., & Tucker, C. (2013). When does retargeting work? Information specificity in online advertising. Journal of Marketing Research,50(5), 561-576.
Laran, J., Dalton, A. N., & Andrade, E. B. (2011). The curious case of behavioral backlash: Why brands produce priming effects and slogans produce reverse priming effects. Journal of Consumer Research, 37(6), 999-1014.
Milne, G. R., Bahl, S., & Rohm, A. (2008). Toward a framework for assessing covert marketing practices. Journal of Public Policy & Marketing, 27(1), 57-62.
Montgomery, A. L., & Smith, M. D. (2009). Prospects for Personalization on the Internet. Journal of Interactive Marketing, 23(2), 130-137.
Murthi, B. P. S., & Sarkar, S. (2003). The role of the management sciences in research on personalization. Management Science, 49(10), 1344-1362.
Newcomb, M. D., Huba, G. J., & Bentler, P. M. (1986). Life change events among adolescents: An empirical consideration of some methodological issues. The Journal of nervous and mental disease, 174(5), 280-289.
Ownership. (n.d.). In Oxford advanced learner's dictionary. Retrieved from http://www.oxforddictionaries.com/definition/english/ownership Pavey, J. L., & Sparks, P. (2010). Autonomy and reactions to health-risk
information. Psychology and Health, 25(7), 855-872.
Personalize. (n.d.). In Oxford dictionaries. Retrieved March 13, 2016, from
http://www.oxforddictionaries.com/definition/english/personalize?q=personalisation#pers onalize__20
Pierce, J. L., Kostova, T., & Dirks, K. T. (2001). Toward a theory of psychological ownership in organizations. Academy of management review, 26(2), 298-310.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of applied psychology, 88(5), 879.
Privacy online: Fair information practices in the electronic marketplace. (2000). Washington, D.C.: Federal Trade Commission (FTC).
Puzakova, M., Rocereto, J. F., & Kwak, H. (2013). Ads are watching me: A view from the interplay between anthropomorphism and customisation. International Journal of Advertising, 32(4), 513-538.
Quick, B. L., & Stephenson, M. T. (2007). Further evidence that psychological reactance can be modeled as a combination of anger and negative cognitions. Communication
Rodin, J., & Langer, E. J. (1977). Long-term effects of a control-relevant intervention with the institutionalized aged. Journal of personality and social psychology, 35(12), 897. Rust, R. T., & Chung, T. S. (2006). Marketing models of service and relationships. Marketing
Science, 25(6), 560-580.
Seemann, E. A., Buboltz, W. C., Thomas, A., Soper, B., & Wilkinson, L. (2005). Normal Personality Variables and Their Relationship to Psychological Reactance. Individual Differences Research, 3(2).
Shapiro, S., MacInnis, D. J., & Heckler, S. E. (1997). The effects of incidental ad exposure on the formation of consideration sets. Journal of consumer research, 24(1), 94-104. Tam, K. Y., & Ho, S. Y. (2006). Understanding the impact of web personalization on user
information processing and decision outcomes. Mis Quarterly, 865-890.
Taylor, S. E. (1979). Hospital patient behavior: Reactance, helplessness, or control?. Journal of Social Issues, 35(1), 156-184.
Taylor, D. G., Davis, D. F., & Jillapalli, R. (2009). Privacy concern and online personalization: The moderating effects of information control and compensation. Electronic Commerce Research, 9(3), 203-223.
Thompson, C. A., & Prottas, D. J. (2006). Relationships among organizational family support, job autonomy, perceived control, and employee well-being. Journal of occupational health psychology, 11(1), 100.
Trope, Y., Liberman, N., & Wakslak, C. (2007). Construal levels and psychological distance: Effects on representation, prediction, evaluation, and behavior. Journal of consumer psychology: the official journal of the Society for Consumer Psychology, 17(2), 83.
Tucker, C. E. (2014). Social networks, personalized advertising, and privacy controls. Journal of Marketing Research, 51(5), 546-562.
Tucker, C. E. (2012). The economics of advertising and privacy. International journal of Industrial organization, 30(3), 326-329.
van Velsen, L., van der Geest, T., van de Wijngaert, L., van den Berg, S., & Steehouder, M. (2015). Personalization has a Price, Controllability is the Currency: Predictors for the Intention to use Personalized eGovernment Websites. Journal of organizational computing and electronic commerce,25(1), 76-97.
Veitch, J. A., & Gifford, R. (1996). Choice, perceived control, and performance decrements in the physical environment. Journal of Environmental Psychology, 16(3), 269-276. Verbeke, W., & Vackier, I. (2004). Profile and effects of consumer involvement in fresh
meat. Meat Science, 67(1), 159-168.
Verplanken, B., & Wood, W. (2006). Interventions to break and create consumer habits. Journal of Public Policy & Marketing, 25(1), 90-103.
Vesanen, J. (2007). What is personalization? A conceptual framework. European Journal of Marketing, 41(5/6), 409-418.
Weinstein, N., Przybylski, A. K., & Ryan, R. M. (2012). The index of autonomous functioning: Development of a scale of human autonomy. Journal of Research in Personality, 46(4), 397-413.
White, T. B., Zahay, D. L., Thorbjørnsen, H., & Shavitt, S. (2008). Getting too personal: Reactance to highly personalized email solicitations. Marketing Letters, 19(1), 39-50. Wu, G. (2005). The mediating role of perceived interactivity in the effect of actual interactivity
Appendix
Pretest Measures:
Personalization (Aguirre et al., 2015)
Please indicate the extent to which you agree with the following statement: This advertisement is directed to me personally.
I recognize my personal situation in this advertisement.
This advertisement takes into account the situation I am facing. This advertisement takes into account my personal situation. (Likert; 1 = Strongly disagree, 7 = Strongly agree)
Perceived Control – (Also Manipulation Check)
Please indicate the extent to which you felt control over the Commonwealth credit card ad: None – A lot
(Likert; 1 – 7)
(Arcury, Quandt, & Russell, 2002).
Please indicate the extent to which you agree with the following statement: I felt control over the Commonwealth credit card ad:
(Likert; 1 = Strongly disagree, 7 = Strongly agree)
Measures:
Education Scale (Crissey, 2009) Highest level of completed education.
Elementary School High School Certificate Degree Bachelor’s Degree Master’s Degree Doctorate
Index of Autonomous Functioning (IAF)
Please indicate the extent to which you agree with the following statements: Authorship/Self-congruence
My decisions represent my most important values and feelings. I strongly identify with the things that I do.
My actions are congruent with who I really am.
My whole self stands behind the important decisions that I make. My decisions are steadily informed by things that I want or care about. Susceptibility to Control - **Reverse Scored**
I do things in order to avoid feeling badly about myself. I do a lot of things to avoid feeling ashamed.
I try and manipulate myself into doing certain things. I believe certain things so that others will like me. I often pressure myself.
Interest-Taking
I often reflect on why I react the way I do.
I am deeply curious when I react with fear or anxiety to events in my life. I am interested in understanding the reasons for my actions.
I am interested in why I act the way I do. I like to investigate my feelings.
(Likert; 1 = Strongly disagree, 7 = Strongly agree)
Purchase Intention (Dodds, Monroe & Grewal, 1991)
Please answer the following questions regarding the Commonwealth credit card:
The likelihood of signing up for the Commonwealth Credit Card is: (very low to very high)”
I would consider signing up for the Commonwealth Credit Card (strongly disagree to strongly agree)
The probability that I would consider signing up for the Commonwealth Credit Card is: (very low to very high)
My willingness to sign up for the Commonwealth Credit Card is: (very low to very high) (Likert; 1 – 7)
Reactance Measure (White et al., 2008)
