Dennis Kaars
Student number: 5656397 Master’s Thesis
Graduate School of Communication
Master’s programme Communication Science
The effect of choice on YouTube pre-roll advertising avoidance
Supervisor: Mw. Dr. M.L. (Marieke) Fransen Date: 30-06-2017
1 ABSTRACT
The YouTube pre-roll is considered the benchmark in online video advertising, which is a multi-billion dollar industry. Pre-roll ads can be actively avoided (by skipping the ad), passively avoided (by not paying attention to the ad) or engaged with (by clicking the ad). This study investigated the effect of choice on YouTube pre-roll avoidance and –engagement. Although, previous studies uncovered intrusiveness, control and personal relevance as
antecedents of ad avoidance, this study is unique in studying their role as possible mediators of the effect of choice on ad- avoidance and –engagement. Results indicated that offering consumers choices may effectively decrease pre-roll skip rates and increase engagement. This effect was fully mediated by perceived personal relevance. No effect of choice on perceived intrusiveness or control was found. Implications of these results are discussed.
INTRODUCTION
In 2017, Google is projected to make $72.69 billion in ad revenues (Handley, 2017). This is about a third of advertisers’ total global digital ad expenses. A great deal of those ad revenues come from YouTube which was bought by Google in 2006 (Arrington, 2006). Exactly how much of Google’s ad revenues come from YouTube is unknown because revenue numbers for Google and YouTube are never reported separately, but it is safe to assume YouTube’s ad revenues run in the billions of dollars (Booton, 2017). YouTube is a video hosting service where users can view and upload videos. YouTube is localized in 75 countries and has over one billion users around the world with visitors watching one billion hours of video every day (Majhi, n.d.). Users that want to upload content to YouTube can create their own YouTube channel. They can import their videos into this channel and create their own playlists among other options (Majhi, n.d.). Other YouTube users can then view the videos on the channel. Owners of popular channels (10.000+ views) are able to monetize views by showing
2
advertisements on their channels (“How to earn money from your videos - YouTube Help,” n.d.). Different ad formats can be used on YouTube but the most popular format is currently a skippable ad format where viewers can skip the ad after five seconds. These skippable ads can appear before (pre-roll) or during content (mid-roll); pre-rolls are the most common and are considered the benchmark standard (Johnson, 2016; Katz, 2010). Channel owners get paid when a viewer watches at least thirty seconds of the advertisement, or the entire ad. They also get paid if a viewer engages with the ad. Skippable pre-roll ads on YouTube can come with Call-to-Action (CTA) overlays, for example a link that leads the viewer to the advertiser’s website. When, for instance, a viewer clicks this link, this is an engagement with the ad and the channel owner also gets paid (Johnson, 2016). Because of this business model, it is desirable for both YouTube channel owners and advertisers that ad viewers engage with, or entirely watch, an ad. Paradoxically, the skip function allows viewers to skip the ad after five seconds. A lot of viewers use this option; ads get skipped about 85 percent of the time while viewers only engage with an ad about 4% of the time (Johnson, 2016). Skipping an
advertisement is a form of ad avoidance. Previous research showed that empowering consumers by offering them choices can decrease this type of ad avoidance (Katz, 2010). Despite this, YouTube’s skippable pre-roll ads do not offer viewers the possibility to choose their own ad. First and foremost, this research paper explores if it is possible to decrease the ad avoidance of YouTube pre-roll advertisements by offering viewers a choice in which ad they want to watch. Secondly, while this research paper’s main focus is on ad avoidance, a measure for ad engagement is also incorporated. This is important because YouTube channel owners get paid when a viewer engages with ads on their channel. Doing such explorations is very relevant; if the ad skip rate could be dropped by just a couple of percent, or ad
engagement could be increased by a couple of percent, this would mean a huge difference to YouTube channel owners and advertisers in this multi-billion dollar industry, where billions
3
of advertised videos are being watched every month. For practitioners, insight into the effect of choice on ad skip rate and engagement is essential for developing the most effective ad experience. Previous research also identified perceived control (e.g. Freedman & Steinbruner, 1964), intrusiveness (e.g. E. Kim, Kim, Yeh, & Choi, 2011) and personal relevance (e.g. Hussain, 2014) as variables that can influence ad- avoidance and -engagement. This research paper proposes a model in which control, intrusiveness and relevance mediate the effect of choice on ad- avoidance and engagement. Although the influence of all of these variables on ad avoidance and (to a lesser extent) ad engagement has been studied before, they have never been tested as one single model. As a matter of fact, the impact of choice on a skippable online video advertising format has never been tested before. This research project adds to the existing literature by examining the relationship between avoidance of- and engagement with- ads that use this fairly new, increasingly popular advertising format and choice, intrusiveness, control and personal relevance. In sum, this paper focuses on the effect of choice on the avoidance of, and engagement with, skippable pre-roll advertising on YouTube. This leads to the following research questions:
RQ1a: What is the effect of choice on the avoidance of skippable pre-roll advertising on YouTube?
RQ1b: What is the role of intrusiveness, control and personal relevance in mediating the effect of choice on the avoidance of skippable pre-roll advertising on YouTube?
RQ2a: What is the effect of choice on the engagement with skippable pre-roll advertising on YouTube?
RQ2b: What is the role of intrusiveness, control and personal relevance in mediating the effect of choice on the engagement with skippable pre-roll advertising on
4
THEORETICAL FRAMEWORK
Ad avoidance & choice
Advertising, like the YouTube pre-roll ads, are persuasive attempts to convince a viewer to buy a product, use a service or like a brand (among many other reasons for advertising). Pre-rolls are similar to regular televised commercials where people are enticed to buy the latest smartphone, switch energy supplier or get an anti-dandruff shampoo for instance. Advertisers often try to tell consumers what to do when promoting their products or services. But
generally, people do not like to be told what to think or do (Brehm, 1966). When people are told what to think or do they experience a loss of freedom. This leads to resistance. (Ad) avoidance is a strategy people use to exert resistance (Fransen, Verlegh, Kirmani, & Smit, 2015). Speck & Elliott (1997) identified skipping-, ignoring- and eliminating ads as ways to avoid ads in traditional media. Skipping an ad is for instance skipping a TV-channel an ad is broadcasted on. Eliminating an ad is zapping to another channel when an ad starts running. Ads can be ignored by simply not paying attention. Fransen et. al (2015) categorized these different types of avoidance as physical, mechanical and cognitive avoidance. In the case of YouTube pre-roll advertising, physical avoidance is for example walking away from the device or closing YouTube. Examples of mechanical avoidance are: skipping the ad or using an ad blocker. Cognitive avoidance is not paying attention to the pre-roll thus not storing any information. The counterpart of avoidance is approach where attention is being paid to the pre-roll, or for instance when a viewer engages with a pre-roll. Tang, Zhang & Wu (2013) differentiate between active- or passive approach- and avoidance strategies. Waiting for an ad to go away is a passive avoidance strategy in their definition, while getting rid of the ad would be a very active avoidance strategy. This research paper uses an adaption from work by Tang & Zhang (2013) to define ad avoidance as a variable with an active and a passive component. Active avoidance is defined in this research as skipping the pre-roll. Passive avoidance is
5
defined as letting the pre-roll play but not paying attention. Tang, Zhang & Wu (2013) define clicking an ad, bookmarking an ad or visiting a brand’s website via the ad as increasingly active approach strategies. This paper’s definition of ad engagement includes intentions to click, bookmark or visiting a brand’s website via the ad. In sum, this paper divides ad
avoidance into (1) active avoidance and (2) passive avoidance. Ad engagement (3) is defined as an approach strategy.
According to Knowles & Riner (2007), there are three basic sources of resistance to persuasion. Resistance to the influence attempt (reactance), resistance to the proposal
(skepticism) and resistance to change (inertia). Skepticism and inertia may occur based on the content of the persuasive attempt while reactance occurs immediately and has nothing to do with content. Merely being exposed to an ad for a smartphone, for example, can already lead to reactance. In the actual ad, a smartphone brand might make claims about their latest product a viewer of this ad is skeptical about. When a viewer is encouraged by the ad to trade in their current phone for the advertising brand’s new model, this may lead to unwillingness to change. Edwards, Li & Lee (2002) concluded ad avoidance is a form of reactance; ads are avoided before effects based on the content of the ad (such as irritation) has a chance to set in. Reactance can occur when a person experiences a lack of choices or alternatives (Knowles & Riner, 2007). Providing people with choices may be successful in decreasing reactance (Fransen et al., 2015; Knowles & Riner, 2007). When you tell a child to brush its teeth this may cause reactance for instance. When you give the same child a choice by telling it that it can (1) brush its teeth or (2) put on its pajamas, reactance may be lower while there is still a desired outcome. The reasoning behind this is that when there is only one alternative all the reactance will be focused on this alternative; when there are multiple alternatives, reactance is spread amongst these alternatives (Knowles & Linn, 2004). In the case of this study, allowing viewers to choose which pre-roll they want to watch may decrease reactance. This could in
6
turn decrease ad avoidance. The first, immediate response a person may have when this person experiences reactance as a consequence of being exposed to a pre-roll ad, might be to actively avoid the ad by skipping it. This corresponds with behavioral avoidance since this person behaves by taking the action of skipping the ad. When a person has a choice, this person may be less inclined to skip the ad than when this person has no choice. This leads to the following hypotheses:
H1a: Participants in the no choice condition actively avoid ads more than participants in the choice condition
Reactance however may not always lead to active avoidance strategies. When a viewer decides not to skip the ad as an active reactance strategy this user might still use a passive reactance strategy. This includes strategies such as ignoring the ad or waiting for the ad to go away (Tang & Zhang, 2013). This corresponds with a cognitive avoidance strategy where no physical action is taken to avoid the ad but rather a form of mental action (e.g. not paying attention). People that are depleted of their self-control for instance, may have no motivation to physically click away the ad but may still want to avoid it (Janssen, Fransen, Wulff, & van Reijmersdal, 2016). Because, providing people with choices may be successful in decreasing ad avoidance, a person that has a choice may be less likely to passively avoid the ad than when this person has no choice. This leads to the following hypothesis:
H1b: Participants in the no choice condition passively avoid ads more than participants in the choice condition
Several researchers have acknowledged that there is a connection between avoidance and approach (Eroglu, Machleit, & Davis, 2003; Mollen, Rimal, Ruiter, & Kok, 2013; Tang & Zhang, 2013). At the same time avoidance is more closely related to resistance and approach is more closely related to persuasion. Knowles and Linn (2004) define strategies for
7
decreasing resistance such as ad avoidance, as omega strategies, as opposed to alpha strategies which attempt to persuade by using approach forces. In this paper, ad engagement is
considered an approach strategy. Although less avoidance does not necessarily leads to more approach this research tests the hypothesis that choice may increase ad engagement:
H1c: Participants in the choice condition intend to engage with ads more than participants in the no choice condition
Several factors may explain the hypothesized effect of choice on ad avoidance and ad
engagement. These factors are intrusiveness, control and personal relevance. In the following paragraphs, the roles of intrusiveness, control and personal relevance as possible mediators of the effect of choice on ad avoidance and ad engagement are discussed. Doing such an
exploration offers a deeper understanding of the way in which choice may influence ad- avoidance and –engagement.
Intrusiveness
The first variable that will be discussed that may explain a possible effect of choice on ad avoidance and ad engagement, is intrusiveness. Considering the fact that approximately 85 percent of all YouTube pre-roll commercials get skipped when the option is available, the conclusion can be drawn that people generally do not want to watch pre-roll advertising (Johnson, 2016). People visit YouTube with the goal to watch videos and not commercials. Ads like pre-rolls take away a user’s freedom to reach their goal directly. That is why ads are considered to be intrusive (Edwards et al., 2002; S. C. Kim & Yoon, 2015). Ha (1996) defines intrusiveness as the disruption of the flow of an editorial unit. A pre-roll ad on YouTube arguably disrupts the flow of the experience of watching a video since viewers can not directly watch the video they want. C. Li & Meeds (2005) found that viewers experienced situations where they perceived to have no control when they had to watch a video ad, as
8
more intrusive than situations where they perceived to have control (i.e. where they could click away pop-up ads). Putting viewers in control by offering them the option to choose the pre-roll ad they want to watch may therefore decrease the perceived intrusiveness of YouTube pre-roll advertising.
H2a: Participants in the no choice condition perceive ads as more intrusive than participants in the choice condition
Intrusiveness is also defined as the interruption of the cognitive thinking (Edwards et al., 2002). In this case, an ad that is perceived to be intrusive interferes with the goals of the consumer. Since the goal of a YouTube visitor is to watch a video, having to watch a pre-roll ad first interferes with this goal. Several authors identified goal impediment or intrusiveness as the most important antecedent of ad avoidance (Cho & Cheon, 2004; E. Kim et al., 2011; Speck & Elliott, 1997). The more a viewer’s freedom is taken away, i.e. the more intrusive an ad is perceived to be, the more resistance this viewer is likely to show (Brehm, 1966; C. Li & Meeds, 2007; Silvia, 2006; Song, 2014).
H2b: Intrusiveness is positively related to active avoidance
H2c: Intrusiveness is positively related to passive avoidance
Because an increase in perceived intrusiveness is possibly related to an increase in ad avoidance (e.g. Cho & Cheon, 2004), it may also be related to a decrease in ad approach strategies, which, in this paper is defined ad engagement (intentions to engage with the pre-roll) (Tang & Zhang, 2013). This leads to the following hypotheses:
H2d: Intrusiveness is negatively related to ad engagement
Since choice is hypothesized to have an effect on intrusiveness (H2a) and intrusiveness is expected to be related to ad- avoidance and engagement (H2b/c/d), the effect of choice on
9
active and passive ad avoidance (H1a & H1b) and ad engagement (H1c) may be mediated by intrusiveness. This leads to the following hypothesis:
H2e: Intrusiveness mediates the effect of choice on active avoidance
H2f: Intrusiveness mediates the effect of choice on passive avoidance
H2g: Intrusiveness mediates the effect of choice on ad engagement
In the next paragraph, the relationships between choice, avoidance and control will be discussed.
Control
According to Inesi, Botti, Dubois, Rucker and Galinsky (2011), power and choice are both sources for personal control. Power refers more to control over indivduals while choice concerns the ability to select a preferred course of action. This can be related to selecting an ad out of mulitple options. Control is traditionally divided into three dimensions: behavioral, cognitive and decisional control (Averill, 1973). Behavioral control refers to the ability to influence the outcome of an event. Cognitive control consists of the predictability and
reinterpretation of a situation. Decisional control refers to choice in the selection of outcomes. Arguably, offering viewers a choice in the pre-roll ad they want to watch could increase both behavioral control (because they have the ability to click on the ad themselves), and
decisional control (because they can select the outcome of what ad will be watched).
H3a: Participants in the choice condition perceive to be more in control than participants in the no choice condition
Perceived control has a positive impact on the physical and psychological well-being of people (Hui & Bateson, 1991). YouTube is aware of the consumers’ desire to feel in control;
10
the introduction of a skip feature in pre-roll advertising to empower viewers is proof of this. Introduction of the skip feature led to a more positive attitude towards pre-roll advertising by consumers (S. C. Kim & Yoon, 2015). Industry creatives (i.e. the people responsible for creating the ads) even listed skippabilty as the number one quality of a good advertising experience (Stanley, 2017). The option to skip an ad can be seen a tool to exert behavioral control. Research among students has shown that increasing decisional control, by offering choices, increases a feeling of responsibility which leads to less employment of resistance strategies (Freedman & Steinbruner, 1964). In this research project, ad avoidance as a dependent variable is the resistance strategy that is studied. This leads to the following hypotheses:
H3b: Control is negatively related to active avoidance
H3c: Control is negatively related to passive avoidance
Because an increase in perceived control is hypothesized to be related to a decrease in ad avoidance (e.g. Freedman & Steinbruner, 1964), it may also be related to an increase in ad approach strategies, which, in this paper is defined ad engagement (intentions to engage with the pre-roll) (Tang & Zhang, 2013). This leads to the following hypothesis:
H3d: Control is positively related to ad engagement
Because choice is hypothesized to have an effect on control (H3a) and control is expected to be related to ad- avoidance and engagement (H3b/c/d), the effect of choice on active and passive ad avoidance (H1a & H1b) and ad engagement (H1c) may be mediated by control. This leads to the following hypotheses:
H3e: Control mediates the effect of choice on active avoidance
11
H3g: Control mediates the effect of choice on ad engagement
In the next section, the relationships between choice, avoidance, ad engagement and personal relevance will be discussed.
Personal Relevance
The third factor that might explain a possible effect of choice on ad avoidance and -engagement, is personal relevance. Perceived personal relevance refers to a person’s belief that an object associates with this person’s lifestyle, values or self-image (Celsi, Chow, Olson, & Walker, 1992; Kang, Liu, & Kim, 2013). Skaters, for instance, might be more inclined to choose the brand Vans because they feel their products associate with being a skater. Good advertising has to be personally relevant to consumers (Stanley, 2017). Personally relevant ads are considered less intrusive than non-relevant ads because they offer more relevant information to the viewers (Edwards et al., 2002). Ads with more relevant information are considered more useful. This leads to more positive attitudes towards online advertising (Martí Parreño, Sanz‐ Blas, Ruiz‐ Mafé, & Aldás‐ Manzano, 2013). Empowering viewers by letting them choose the ad they want to watch has been found to increase perceived personal relevance because they can choose the ad they feel might contain interesting, useful or relevant information themselves (Katz, 2010). This leads to the following hypothesis:
H4a: Participants in the choice condition perceive ads to be more personally relevant than participants in the no choice condition
When it comes to the relevance of advertising, personal ad relevance was found to be more important to consumers than contextual relevance (Katz, 2010). In the case of YouTube, the pre-roll does not have to fit perfectly into the rest of the YouTube environment, or relate perfectly to the video that is streamed but it does need to be personally relevant. Several researchers have shown that avoidance of ads can be reduced by making ads more personally
12
relevant (Hussain, 2014; E. Kim et al., 2011). Pre-rolls that are perceived as more interesting, useful or relevant are therefore expected to be avoided less than boring, useless or irrelevant ones. It makes sense that, when a pre-roll is personally relevant, people will be less inclined to avoid this pre-roll since it contains information that is perceived valuable. This leads to the following hypotheses:
H4b: Personal relevance is negatively related to active avoidance
H4c: Personal relevance is negatively related to passive avoidance
Also, when services or products connect with personal values, beliefs and lifestyles more, it is expected people will not only avoid ads for these products less, but that they will also intend to engage with these ads more (Celsi et al., 1992; Kang et al., 2013). Because an increase in perceived personal relevance may be related to a decrease in ad avoidance (e.g. E. Kim et al., 2011), it may also be related to an increase in ad approach strategies, which, in this paper is defined ad engagement (intentions to engage with the pre-roll) (Tang & Zhang, 2013). This leads to the following hypothesis:
H4d: Personal relevance is positively related to ad engagement
Because choice is hypothesized to have an effect on personal relevance (H4a) and control is expected to be related to avoidance (H3b & H3c) and ad engagement (H4d), the effect of choice on active and passive ad avoidance (H1a & H1b) and ad engagement (H1c) may be mediated by control. This leads to the following hypotheses:
H4e: Personal relevance mediates the effect of choice on active avoidance
H4f: Personal relevance mediates the effect of choice on passive avoidance
13
All hypotheses are combined into a hypothetical model (Figure 1).
Figure 1: Hypothesized model METHOD
Sample
Participants were collected by distributing the experiment via e-mail and social media. The experiment was designed to measure the effect of choice on ad avoidance for YouTube pre-roll advertising. The online experiment used a single factor (choice/no choice) between subjects design and was created using Qualtrics. Because the experiment was in Dutch, only people with sufficient Dutch language skills to participate in the experiment were contacted. The final sample consisted of a convenience sample of 298 people (221 male, 77 female). 60.1% of the participants were 35 years old or younger (N = 179), 50.7% had a higher
14
vocational education or university degree (N = 151) and 46.3% had a gross yearly income of €35.000 or higher (N = 138).
Procedure
Participants were told they were participating in research about processes related to the remembering of information. Next, they were informed they were going to be presented with (1) an ad and (2) a short clip. Participants in the choice condition could choose between 2 ads: a Nikon pre-roll ad and a T-mobile pre-roll ad, participants in the no choice condition were randomly assigned to either of the two ads. After watching or skipping the ad participants had to watch a 23-second clip featuring Dutch politician Geert Wilders. Next they had to answer four multiple choice questions about the clip. After these questions, participants answered open and multiple recall questions about the ad which were used to measure passive avoidance. Subsequently, participants answered several questions that were designed to measure perceived intrusiveness, control, personal relevance and demographic variables. Finally, participants were debriefed on the goal of the experiment and asked to leave their e-mail address to have a chance at winning a €30 lottery ticket.
Choice
Choice was manipulated by letting participants choose which ad they wanted to watch by clicking one of two pictures with the image of a brand (T-mobile or Nikon). A short descriptive text was also included in the picture. Both ads were thirty seconds long. In the choice condition, participants had seven seconds to choose between the Nikon or T-mobile ad. If no choice was made within seven seconds, participants were randomly assigned to either the T-mobile or the Nikon ad. In the no choice condition participants were immediately, randomly, assigned to one of the two ads. In both conditions, participants had the option to skip the ad at any time using the skip button. To make the ad run, participants had to click
15
play to start the video themselves. On YouTube, pre-rolls play automatically. In Qualtrics, videos can be set to start playing automatically but this autoplay function is unfortunately blocked on most smartphones for security reasons (“Why does Video Autoplay on Mobile Devices not Work?,” 2015). The manipulation of choice is based on previous research with ad selection models where participants had 10-15 seconds to choose an ad, and they could
choose between two or three ads (Katz, 2010). The selection time and number of options have been reduced to seven seconds and strictly two options in this research because people got annoyed with (1) waiting for the ad to start and (2) having to choose an ad according to Stross (2011). A timeframe of seven seconds should be more than enough to choose between two ads when viewers are motivated to do so (Bryce & Bratzke, 2017).
Cover story
After watching or skipping the ad participants had to watch a 23-second clip featuring Dutch politician Geert Wilders as a cover story. The cover story was designed in such a way
participants would focus on the clip as they normally would when choosing to watch a clip on YouTube. The goal was to get participants to treat the pre-roll ad as they would in an
everyday situation.
Active avoidance
Active avoidance of the advertisement was measured with a metric in Qualtrics which counted after how many seconds a participant submitted the page with the ad. Participants could submit the page by clicking the ‘SKIP AD’-button. Active avoidance was coded ‘1’ when the ‘SKIP AD’-button was clicked within thirty seconds (the full runtime of the ad). Otherwise active avoidance was coded ‘0’. A score of 1 on the variable active avoidance means the participant skipped the ad, a score of 0 means the participant did not skip the ad. Overall, 23.5% of the participants (N = 70) actively avoided the ad by skipping it.
16 Passive avoidance
Passive avoidance was measured via a free recall question and five multiple choice questions about the pre-roll ad based on work by Katz (2010). Passive avoidance corresponds with not paying attention to the pre-roll (cognitive avoidance). Participants who paid less attention to the pre-roll can be expected to score lower on both recall questions than participants who paid more attention to the pre-roll. For the open recall measure participants were asked to think back on the ad and list as many details as they could in a text box. Afterwards, every correctly remembered detail was coded with a score of 1 resulting in a score for free recall (N = 170; M = 6.68; SD = 2.33). For example: seven correctly remembered unique details about the ad would result in a score of 7 on the scale for active approach. The codebook used can be found in the Appendix. Eleven participants were coded missing because the ad would not run or because they listed details about the clip with Geert Wilders instead of the ad. Directly after the free recall question about the ad, participants were asked to answer five multiple choice questions about the ad. Because the correct answer is present in the answer options, multiple choice questions can be answered with less cognitive effort than an open recall question (Katz, 2010). Every single correct answer on the multiple choice questions was coded with a score of 1 resulting in a scale ranging from 0 (avoidance) to 5 (approach) for passive
approach/avoidance (N = 50; M = 3.26; SD = 1.14). The multiple choice questions and
corresponding answer options for the T-mobile and Nikon ads can be found in the Appendix. Ad engagement
After the questions measuring passive avoidance, ad engagement was measured using a single question. Specifically, participants were asked the following question about the ad: ‘in an everyday situation, how likely is it you would engage with this ad (i.e. click on the ad to for instance: share, bookmark, like, visit the website, look for more information, look to buy the
17
product)’. The item was measured on a seven point scale (M = 1.99; SD = 1.28) ranging from 1 (very unlikely) to 7 (very likely).
Perceived control
Next, perceived control was measured using four statements based on work by Ajzen (2002): ‘I had the situation completely under control’, ‘I found it very easy to decide what was happening in the situation’, ‘what happened in the situation was completely up to me’ and ‘I was not capable of deciding what was happening at all’. The answer scale ranged from 1 (strongly disagree) to 7 (strongly agree). The four items were rescaled so higher scores would reflect greater levels of perceived control and combined into one variable (α = 0.80, M = 4.83, SD = 1.24).
Intrusiveness
After perceived control, intrusiveness was measured. Participants were asked to indicate on a 7-point scale (1 = strongly disagree; 7 = strongly agree) to what degree they found the ad: distracting, disturbing, forced, interfering, intrusive, invasive and obtrusive (Edwards et al., 2002; H. Li, Edwards, & Lee, 2002). The seven items were combined into one variable (α = 0.92, M = 4.16, SD = 1.40).
Personal relevance
Personal relevance was measured in a similar way as intrusiveness. Participants were asked to indicate on a 7-point scale (1 = strongly disagree; 7 = strongly agree) to what degree they found the ad: interesting, useful and relevant. The three items for personal relevance were adapted from Hussain (2014). The three items were used to form a single scale (α = 0.78, M = 3.59, SD = 1.30).
18 RESULTS
Randomization check
Because demographic variables such as age and income may influence ad avoidance, a randomization check was conducted with choice as the independent variable and several demographic variables (age, income, gender and education) as the dependent variable (Speck & Elliott, 1997). There was no significant difference in average age t(296) = 1.33, p = .186, 95% CI = [-1.10, 0.53], income t(296) = 1.39, p = .165, 95% CI [-1.11, 0.63] and education level t(296) = 0.97, p = .332, 95% CI [-0.19, 0.57] between the respondents in the choice and the no choice condition. The final randomization check (a Fisher’s exact test) showed that significantly (p = .012) more males were randomly assigned to the choice condition (57.5%) than to the no choice condition (42.5%). This means differences in the effects of choice on ad avoidance are not because of the age, income and education of the participants while this could not be concluded for gender based on the randomization check. Gender was therefore included in the mediation analyses as a covariate.
Choice
To test the hypothesis that participants in the no choice condition actively avoid ads more than participants in the choice condition (H1a), a Fisher’s exact test was performed using SPSS. A marginally significant difference between participants in the choice and no choice was found for active avoidance (p = .076). Relatively many participants in the no choice condition chose to actively avoid the ad (28.3%) while relatively few participants in the choice condition skipped the ad (19.4%). Although hypothesis 1a is therefore confirmed the relationship between choice and active avoidance was weak (tau = .011).
Next, two independent samples t-tests were conducted to assess if participants in the no choice condition, who did not skip (actively avoid) the ad, passively avoided ads more than
19
participants in the choice condition (H1b). Results showed that participants in the no choice condition, who did not skip the ad, did not score significantly lower on the either the open recall question t(218) = -1.43, p = .15, 95% CI = [-1.35, 0.22] or the multiple choice recall question t(226) = -0.80, p = .43, 95% CI = [-0.36; 0.15] than participants in the choice condition. Because both t-tests produced insignificant results, hypothesis 1b is rejected. To test if participants in the choice condition intended to engage with ads more than participants in the no choice condition (H1c), another independent samples t-test was performed. Results showed that participants in the choice condition indeed scored significantly higher t(296) = -2.58, p = .011, 95% CI = [-0.67, -0.09] on the scale for ad engagement (M = 2.16; SD = 1.38) than participants in the no choice condition (M = 1.78; SD = 1.13), confirming hypothesis 1c.
Intrusiveness
To test the hypothesis that participants in the no choice condition perceive ads as more intrusive than participants in the choice condition (H2a), an independent samples t-test was conducted. There was no significant difference found t(296) = 1.44, p = .15, 95% CI = [-0.09, 0.55] between participants in the no choice condition and participants in the choice condition implying hypothesis 2a could not be confirmed.
To test the hypothesis that intrusiveness is positively related to active avoidance (H2b) a binary logistic regression was performed. Results indicated that intrusiveness had a significant relationship with active avoidance Wald(1) = 9.07; p = .003. Respondents scoring one point higher on the scale for intrusiveness were about 1.36 times more likely to skip the ad. This confirms hypothesis 2b.
Two simple logistic regression analyses were performed to test if intrusiveness had a positive relationship with passive avoidance (H2d). Results from the first regression analysis showed
20
that there was a marginally significant, weak, negative relationship between intrusiveness and open recall b* = -0.13, t = -1.94, p = .054, 95% CI [-0.53, 0.00]. An extra point on the scale for intrusiveness correlates with an expected decrease of 0.26 in the score for the open recall question. The second regression analysis indicated that there was a significant, moderately strong, negative relationship between intrusiveness and multiple choice recall b* = 0.21, t = -3.18, p = .002, 95% CI [-0.23, -0.05]. An extra point on the scale for intrusiveness
corresponds with an expected decrease of 0.14 in the aggregated score for the multiple choice recall questions. Lower open- and multiple choice recall scores correspond with higher levels of passive avoidance, implying a positive relationship between control and passive avoidance. This confirms hypothesis 2c.
A simple regression analysis was performed to test if intrusiveness is negatively related to ad engagement (H2d). The results showed that there was a significant, moderately strong,
negative relationship between intrusiveness and ad engagement b* = -0.29, t = -5.20, p < .001, 95% CI [-0.37, -0.16]. An extra point on the scale for intrusiveness corresponds with an expected decrease of 0.27 on the scale for ad engagement. These results confirm hypothesis 2d.
Because no direct effect of choice on intrusiveness was found (H2a = rejected), no mediation analyses were conducted to measure if intrusiveness mediated the effect of choice on active avoidance (H2e), passive avoidance (H2f) or ad engagement (H2g). Therefore, hypothesis 2e, 2f and 2g were rejected.
Control
To test the hypothesis that participants in the choice condition perceive to have more control than participants in the no choice condition (H3a), an independent samples t-test was
21
0.67, p = .50, 95% CI = [-0.19, 0.38] between participants in the choice condition and participants in the choice condition. Therefore, hypothesis 3a could not be confirmed. To test the hypothesis that control is negatively related to active avoidance (H3b) a binary logistic regression was performed. Results indicated that control had no significant
relationship with active avoidance Wald(1) = 0.12; p = .73, rejecting hypothesis 3b.
Next, two simple logistic regression analyses were performed to test if control had a negative relationship with passive avoidance (H3c). Results from the first regression analysis showed that there was a significant, moderate, positive relationship between control and open recall b* = 0.16, t = 2.41, p = .017, 95% CI [0.07, 0.71]. An extra point on the scale for control corresponds with an expected increase of 0.39 in the score for the open recall question. The second regression analysis indicated that there was also a significant, moderate, positive relationship between control and multiple choice recall b* = 0.20, t = 3.00, p = .003, 95% CI [0.05, 0.26]. An extra point on the scale for control corresponds with an expected increase of 0.16 in the aggregated score for the multiple choice recall questions. Higher open- and multiple choice recall scores correspond with lower levels of passive avoidance, implying a negative relationship between control and passive avoidance. This confirms hypothesis 3c.
A simple regression analysis was performed to test if control is positively related to ad engagement (H3d). The results showed that there was a no significant relationship between control and ad engagement b* = 0.00, t = 0.00, p = .998, 95% CI [-0.12, 0.12]. These results reject hypothesis 3d.
Because no direct effect of choice on control was found (H3a = rejected) and no relationship between control and active avoidance was found (H3b = rejected), no mediation analysis was conducted to measure if control mediated the effect of choice on active avoidance (H3e),
22
passive avoidance (H3f) or ad engagement (H3g). Therefore, hypothesis 3e, 3f and 3g were rejected.
Personal relevance
To test the hypothesis that participants in the choice condition perceive ads to be more personally relevant than participants in the no choice condition (H4a), an independent
samples ttest was conducted. The results implied there was a significant difference, t(296) = -3.56, p < .001, 95% CI = [-0.82, -0.24]; participants in the choice condition found the ad on average more personally relevant (M = 3.83; SD = 1.24) than participants in the no choice condition (M = 3.30; SD = 1.32), confirming hypothesis 4a.
To test the hypothesis that personal relevance is negatively related to active avoidance (H4b) a binary logistic regression was performed. Results indicated that personal relevance had a significant relationship with active avoidance Wald(1) = 5.82; p = .02. Respondents scoring one point higher on the scale for personal relevance were about 22.7% less likely to skip the ad. This confirms hypothesis 4b.
Next, two simple logistic regression analyses were executed to test if personal relevance had a negative relationship with passive avoidance (H4c). Results from the first regression analysis showed that there was no significant relationship between personal relevance and open recall b* = -0.08, t = -1.23, p = .22, 95% CI [-0.48, 0.11]. The second regression analysis indicated that there was also no significant relationship between control and multiple choice recall b* = 0.11, t = 1.60, p = .110, 95% CI [-0.02, 0.18]. Because the results of both regression analyses were insignificant hypothesis 4c could not be confirmed.
A simple regression analysis was performed to test if personal relevance is positively related to ad engagement (H4d). The results showed that there was a significant, moderately strong, positive relationship between personal relevance and ad engagement b* = 0.38, t = 7.03, p <
23
.001, 95% CI [0.27, 0.48]. An extra point on the scale for personal relevance corresponds with an expected increase of 0.37 on the scale for ad engagement. These results confirm hypothesis 4d.
Subsequently, a mediation analysis was conducted using PROCESS to investigate if personal relevance mediated the effect of choice on active avoidance (H4e). Because significantly more males were randomly assigned to the choice condition (57.5%) than to the no choice condition (42.5%), gender was included in the analysis as a covariate. Results indicated that gender, surprisingly, had a direct effect on active avoidance b = -0.79, SE = 0.36, p = .027, 95% CI = [-1.50, -0.09] when controlling for choice and personal relevance. There was no significant effect of gender on personal relevance b = -0.09, SE = 0.17, p = .586, 95% CI = [-0.43, 0.24] when controlling for choice. When controlling for gender and personal relevance, choice had no direct effect on active avoidance anymore, b = -0.47, SE = 0.29, p = .105, 95% CI = [-1.03, 0.10]. When controlling for gender and choice, the relationship between personal relevance and active avoidance was still significant b = -0.24, SE = 0.11, p = .027, 95% CI = [-0.46, -0.03]. The indirect effect was tested using a bootstrap estimation approach with 10,000 samples. These results indicated the indirect coefficient was significant, b = -0.13, SE = 0.07, 95% CI = [-0.30, -0.02]. This means that personal relevance fully mediates the effect of choice on active avoidance, confirming hypothesis 4e.
Because no relationship between personal relevance and passive avoidance was found (H4c = rejected), no mediation analysis was conducted to measure if personal relevance mediated the effect of choice on passive avoidance. Therefore, hypothesis 4f was disconfirmed.
Finally, a mediation analysis was conducted using PROCESS to investigate if personal relevance mediated the effect of choice on ad engagement (H4g). Because significantly more males were randomly assigned to the choice condition (57.5%) than to the no choice
24
condition (42.5%), gender was included in the analysis as a covariate. There was no
significant effect of gender on ad engagement b = -0.09, SE = 0.16, p = .580, 95% CI = [-0.40, 0.22] when controlling for choice and personal relevance. When controlling for gender and personal relevance, choice had no direct effect on ad engagement anymore, b = 0.18, SE = 0.14, p = .202, 95% CI = [-0.10, 0.46]. When controlling for gender and choice, the
relationship between personal relevance and ad engagement was still significant b = 0.36, SE = 0.05, p < .001, 95% CI = [0.25, 0.46]. The indirect effect was tested using a bootstrap estimation approach with 10,000 samples. These results indicated the indirect coefficient was significant, b = 0.18, SE = 0.06, 95% CI = [0.08, 0.32]. This means that personal relevance fully mediates the effect of choice on ad engagement, confirming hypothesis 4g.
Additional analysis
Because gender was surprisingly shown to have an effect on active avoidance, two additional independent samples t-tests were conducted to investigate if gender also had an effect on passive avoidance. Results from the first t-test showed that female participants, who did not skip the ad, scored significantly higher t(218) = -2.15, p = .032, 95% CI = [-1.79, -0.08] on the open recall question (M = 6.26; SD = 3.16) than male participants (M = 5.32; SD = 2.79). The second t-test indicated that female participants, who did not skip the ad, did not score significantly higher t(226) = 0.74, p = .462, 95% CI = [-0.18; 0.39] on the multiple choice recall question than male participants, who did not skip the ad. These additional analyses show that males passively avoided ads more than females indicated by lower average scores on the open recall question.
CONCLUSION/DISCUSSION
Previous studies have recognized that choice may have an impact on: resistance strategies such as ad avoidance (e.g. Knowles & Linn, 2004), on ad engagement (e.g. Tang & Zhang,
25
2013), intrusiveness (e.g. C. Li & Meeds, 2005), control (e.g. Averill, 1973) and personal relevance (e.g. Katz, 2010). Intrusiveness, control and relevance, in turn, have been found to be significant predictors of ad avoidance by several different studies (e.g. Cho & Cheon, 2004; Freedman & Steinbruner, 1964; E. Kim et al., 2011) while a relationship of personal relevance with ad engagement could also be inferred from scientific literature (Celsi et al., 1992). However, there was no study conducted to date that examined the mediating roles of intrusiveness, control and relevance in the effect of choice on either ad avoidance, or ad engagement. The present study aimed to fill that gap by performing an experiment in which choice was manipulated and perceived intrusiveness, control and personal relevance, active- and passive ad avoidance and intentions to engage with the ad were measured.
Results from the experiment indicated that choice had a small, marginally significant effect on active pre-roll avoidance. This entails that participants who could choose which ad they wanted to watch were on average less likely to skip the pre-roll than participants who could not choose the pre-roll they wanted to watch. Secondly, no significant effect of choice on passive avoidance was found. Participants who could not choose the pre-roll did not score worse on open- and aided recall measures than participants who could choose between ads, implying that participants who did not skip the pre-roll, and could not choose the pre-roll, were not more likely to avoid the pre-roll in a passive way (e.g. by waiting for it to go away) than participants who could choose the pre-roll. An additional mediation analysis (performed to answer RQ1b) showed the effect of choice on active avoidance was fully explained by personal relevance. This means choice increased personal relevance which was in turn related with a decline in pre-roll skip rates.
Results from the experiment showed that participants in the choice condition self-reported they were, on average, significantly more likely to engage with the YouTube pre-roll than participants in the no choice condition. An additional mediation analysis (performed to
26
answer RQ2b) showed this effect was fully explained by personal relevance. This means choice increased personal relevance which was in turn related with an increase in intentions to engage with the pre-roll.
Overall, this study’s key findings suggest choice had an indirect effect on active ad avoidance (skipping the pre-roll) and ad engagement (intentions to click on the pre-roll). Specifically, choice decreased the likelihood of skipping the pre-roll while increasing the likelihood of engaging with the pre-roll. These effects were achieved through increased personal relevance. It has to be noted that the effect of choice on active ad avoidance was very weak. Because reactance is expected to decrease when there are multiple alternatives (Knowles & Linn, 2004), it is possible choice decreases types of reactance other than avoidance in the context of YouTube pre-roll advertising (such as for instance source derogation). It is plausible people see a brand as more credible when they can choose a brand’s ad themselves. Further research is required to test this hypothesis. That fact that choice more significantly (indirectly) impacts intentions to engage with the pre-roll in suggests that in the context of YouTube pre-roll advertising, choice may be more successful in driving approach than decreasing avoidance, implying that: offering viewers the option to choose is an alpha- rather than an omega strategy (Knowles & Linn, 2004).
While these are important theoretical contributions to the scientific literature, the results of this study arguably have the greatest value for practitioners. The first implication for
practitioners is that offering viewers choices may have a positive effect on personal relevance which can in turn decrease skip rates. Although overall skip rates were extremely
(unrealistically) low in the experiment conducted in this research paper, indicating the
experiment was not a realistic recreation of a regular YouTube environment, this research still gives an indication of the effect of choice on active avoidance because skip rates were
27
unrealistically low in both the choice- and the no choice condition. To try to ensure participants would treat the pre-roll as they normally would, a cover story was designed. Participants were told they were going watch and TV-clip (preceded by an ad) and were participating in research on the processes that lead to the remembrance of information. Furthermore they were told the ad was included in the experiment to create a realistic YouTube environment and they could watch the ad if they wanted to, or skip the ad with the ‘SKIP AD’-button. It may be possible that some of the participants were still unsure if they could skip the ad or were more motivated than they would normally be to watch the pre-roll ad because they were voluntarily participating in an experiment and wanted to check out the whole experiment. Also, it was beyond the means of this research to place the skip button as prominently in the ad as it usually is in a regular YouTube environment. When YouTube first introduced the skip-button they were also surprised by unexpectedly low skip rates. Making the skip button very prominent helped to increase skip rates tremendously (Stross, 2011). The second implication for practitioners is that offering viewers choices may have a positive effect on personal relevance which can in turn increase ad engagement. Ideally, participants are observed in a natural environment to see if they engage with the ad. In this experimental situation it was not possible for them to engage directly with the ad. Therefore, participants were asked about their intentions to engage with the pre-roll if they were in a natural environment. Future research could employ a field study to measure ad engagement more accurately.
Passive avoidance was measured in this research paper through open- and multiple choice recall questions. While recollection of information may be a good indicator of avoidance, ideally participants are physically observed. Passive avoidance can then be measured with eye-tracking (e.g. looking away). Using eye-tracking was beyond the scope of this research
28
project, however, future research could be conducted employing such a method to measure passive avoidance even more accurately.
A hypothesis that was tested was that participants in the no choice condition perceive ads as more intrusive than participants in the choice condition. This hypothesis could not be confirmed. Intrusiveness refers to disruption of a flow or interfering with a viewer’s goal (Edwards et al., 2002; Ha, 1996). Apparently, offering participants choices did not give them a sense of flow or a new goal (i.e. choosing the ad). This is possibly the reason that choice has no effect on perceived intrusiveness. In sum, no evidence was found for the expectancy that the situation where there are choice options is found less intrusive than the situation where there are no choice options.
Intrusiveness was found to be positively related to both active- and passive avoidance. This means participants that perceived the ads to be more intrusive may also have been more likely to avoid the ad. This is consistent with previous literature where intrusiveness was frequently listed as the most important antecedent of advertising avoidance (e.g. E. Kim et al., 2011). This research adds to the existing literature by demonstrating intrusiveness also has a positive relationship with YouTube pre-roll avoidance. Furthermore, intrusiveness was found to have a significant negative relationship with ad engagement implying that participants who found the pre-roll intrusive, were less likely to engage with the pre-roll than participants who found the pre-roll less intrusive. This offers tentative support for the idea that avoidance and
approach are related, at least in some cases (Tang & Zhang, 2013).
The hypothesis that participants in the choice condition perceive to be more in control than participants in the no choice condition was also tested. This hypothesis was rejected on the basis of the results. It is possible participants already felt in control because of the availability of the skip option. According to Inesi et al. (2011) there is a threshold for control above
29
which, additional factors that may normally increase personal control contribute less, or not at all, to a greater sense of personal control. This basically means the need for control may be satisfied by the ability to skip the ad, so the option to choose the ad becomes redundant. Fairly high overall average levels of perceived control for both the participants in the choice and the no choice condition offer some tentative support for this claim. These results provide
evidence for practitioners that the skip button is successful in satisfying viewers’ need for control.
Another interesting result emerged from the analysis of the experimental data. Control was found to be negatively related to passive avoidance while no significant relationship between choice and active avoidance was found. Apparently control does lead to less employment of resistance strategies (Freedman & Steinbruner, 1964) but only when it comes to passive avoidance, which in this case is the resistance strategy. Basically, participants who did not skip the ad and felt more in control, were less likely to employ passive avoidance strategies than participants who did not skip the ad and felt less in control. A possible explanation for this result could be that some participants simply did not know how to skip the ad making them feel less in control and also making them unable to actively avoid the ad. Willingness to avoid the ad may in this case have led to passive avoidance (such as not paying attention). This would suggest YouTube made the right move by making the skip button more prominent (Stross, 2011). It can be speculated that: in a situation where viewers might want to skip a pre-roll but are not able to do so, they will still avoid the ad, only passively. Furthermore, in such a situation, it is probable that participants will also become irritated which can have a negative impact on attitudes towards advertising (S. C. Kim & Yoon, 2015; Martí Parreño et al., 2013). Further research is needed to test the validity of this hypothesis.
Congruent with expectations, personal relevance was found to fully mediate the effects of choice on active avoidance and ad engagement intentions. Nowadays, YouTube can provide
30
consumers with personally relevant ads based on information Google collects when consumers browse their partner websites (“How Google infers interest and demographic categories - AdSense Help,” n.d.). The results of this study emphasize the importance
YouTube does so, since increased personal relevance might lead people to skip pre-rolls less frequently and click pre-rolls more. When it is not possible to provide viewers with personally relevant ads based on these viewers’ browsing history, offering them choices can serve as an outcome to increase perceived personal relevance.
A few limitations of this study have to be noted. Although the manipulation of choice was based on existing ad selection models and scientific literature, a pre-test may have been conducted to ensure the manipulation was successful. A pre-test may have also established if multiple choice questions were equally difficult for the Nikon- and the T-mobile ad. Also the open recall question was not coded by multiple coders; however the codebook is included in the Appendix for reproducibility.
Despite these limitations, this study has produced valuable results. The present research project uncovered personal relevance as a mediator of the effects of choice on active avoidance and ad engagement. It offers evidence for practitioners that the current YouTube pre-roll ad model is not expected to be improved in terms of lowering skip rates or increasing engagement by offering viewers choices, that is, if YouTube can successfully provide
personally relevant pre-rolls.
REFERENCE LIST
Ajzen, I. (2002). Perceived Behavioral Control, Self-Efficacy, Locus of Control, and the Theory of Planned Behavior. Journal of Applied Social Psychology, 32(4), 665–683. https://doi.org/10.1111/j.1559-1816.2002.tb00236.x
31
Arrington, M. (2006, October 9). Google Has Acquired YouTube. Retrieved May 15, 2017, from http://social.techcrunch.com/2006/10/09/google-has-acquired-youtube/
Averill, J. R. (1973). Personal control over aversive stimuli and its relationship to stress. Psychological Bulletin, 80(4), 286–303. https://doi.org/10.1037/h0034845 Booton, J. (2017, April 27). Google earnings: A new approach arrives amid YouTube ad
controversy. Retrieved May 15, 2017, from
http://www.marketwatch.com/story/google-earnings-a-new-approach-arrives-amid-youtube-ad-controversy-2017-04-21
Brehm, J. W. (1966). A theory of psychological reactance (Vol. x). Oxford, England: Academic Press.
Bryce, D., & Bratzke, D. (2017). Are participants’ reports of their own reaction times reliable? Re-examining introspective limitations in active and passive dual-task paradigms. Acta Psychologica, 172, 1–9. https://doi.org/10.1016/j.actpsy.2016.10.007 Celsi, R. L., Chow, S., Olson, J. C., & Walker, B. A. (1992). The construct validity of
intrinsic sources of personal relevance: An intra-individual source of felt involvement. Journal of Business Research, 25(2), 165–185.
https://doi.org/10.1016/0148-2963(92)90015-4
Cho, C.-H., & Cheon, H. J. (2004). Why Do People Avoid Advertising on the Internet? Journal of Advertising, 33(4), 89–97.
https://doi.org/10.1080/00913367.2004.10639175
Edwards, S. M., Li, H., & Lee, J.-H. (2002). Forced Exposure and Psychological Reactance: Antecedents and Consequences of the Perceived Intrusiveness of Pop-up Ads. Journal of Advertising, 31(3), 83–95.
32
Eroglu, S. A., Machleit, K. A., & Davis, L. M. (2003). Empirical testing of a model of online store atmospherics and shopper responses. Psychology and Marketing, 20(2), 139– 150. https://doi.org/10.1002/mar.10064
Fransen, M. L., Verlegh, P. W. J., Kirmani, A., & Smit, E. G. (2015). A typology of consumer strategies for resisting advertising, and a review of mechanisms for countering them. International Journal of Advertising, 34(1), 6–16.
https://doi.org/10.1080/02650487.2014.995284
Freedman, J. L., & Steinbruner, J. D. (1964). Perceived choice and resistance to persuasion. The Journal of Abnormal and Social Psychology, 68(6), 678–681.
https://doi.org/10.1037/h0039864
Ha, L. (1996). Advertising clutter in consumer magazines: dimensions and effects. Journal of Advertising Research, 36(4), 76–85.
Handley, L. (2017, March 21). Facebook and Google predicted to make $106 billion from advertising in 2017. Retrieved May 15, 2017, from
http://www.cnbc.com/2017/03/21/facebook-and-google-ad-youtube-make-advertising-in-2017.html
How Google infers interest and demographic categories - AdSense Help. (n.d.). Retrieved May 19, 2017, from
https://support.google.com/adsense/answer/140378?hl=en&ref_topic=1628432 How to earn money from your videos - YouTube Help. (n.d.). Retrieved May 15, 2017, from
https://support.google.com/youtube/answer/72857?hl=en&ref_topic=6029709 Hui, M. K., & Bateson, J. E. G. (1991). Perceived Control and the Effects of Crowding and
Consumer Choice on the Service Experience. Journal of Consumer Research, 18(2), 174–184.
33
Hussain, D. (2014). Online Video Advertisement Avoidance: Can Interactivity Help? Retrieved from
http://search.proquest.com/openview/731f989157f01b6a8b4dd0aa0f6f7c6e/1?pq-origsite=gscholar&cbl=30135
Inesi, M. E., Botti, S., Dubois, D., Rucker, D. D., & Galinsky, A. D. (2011). Power and Choice: Their Dynamic Interplay in Quenching the Thirst for Personal Control.
Psychological Science, 22(8), 1042–1048. https://doi.org/10.1177/0956797611413936 Janssen, L., Fransen, M. L., Wulff, R., & van Reijmersdal, E. A. (2016). Brand placement
disclosure effects on persuasion: The moderating role of consumer self-control. Journal of Consumer Behaviour, 15(6), 503–515. https://doi.org/10.1002/cb.1587 Johnson, T. (2016, January 5). YouTube Advertising for Retailers - CPC Strategy. Retrieved
January 12, 2017, from http://www.cpcstrategy.com/blog/2016/01/youtube-advertising/
Kang, J., Liu, C., & Kim, S.-H. (2013). Environmentally sustainable textile and apparel consumption: the role of consumer knowledge, perceived consumer effectiveness and perceived personal relevance. International Journal of Consumer Studies, 37(4), 442– 452. https://doi.org/10.1111/ijcs.12013
Katz, H. (2010). The Pool Lane One. Journal of Interactive Advertising, 10(2), 72–77. https://doi.org/10.1080/15252019.2010.10722172
Kim, E., Kim, S., Yeh, Y.-H., & Choi, S. M. (2011). Exploring the antecedents of advertising avoidance on online video sites. Presented at the American Academy of Advertising. Conference. Proceedings (Online), Lubbock: American Academy of Advertising. Retrieved from
http://search.proquest.com/openview/099565daa17c343c42f14f34d30cb996/1?pq-origsite=gscholar&cbl=40231
34
Kim, S. C., & Yoon, D. (2015). 5-second persuasion: the effect of the skip function in pre-roll advertising. Presented at the American Academy of Advertising. Conference.
Proceedings. Retrieved from
http://search.proquest.com/openview/366c14b1beb82019be6877c4a3248cd1/1?pq-origsite=gscholar&cbl=40231
Knowles, E. S., & Linn, J. A. (2004). Resistance and Persuasion. Psychology Press. Knowles, E. S., & Riner, D. D. (2007). Omega approaches to persuasion: overcoming
resistance. In A. R. Pratkanis (Ed.), The science of social influence: advances and future progress.
Li, C., & Meeds, R. (2005). Different forced-exposure levels of internet advertising: an experimental study on pop-up ads and interstitials. Presented at the American Academy of Advertising. Conference. Proceedings. Retrieved from
http://search.proquest.com/openview/9343dfde05df479cc6182376053bcfdf/1?pq-origsite=gscholar&cbl=28878
Li, C., & Meeds, R. (2007). Factors affecting information processing of internet
advertisements: a test on exposure condition, psychological reactance, and advertising frequency. Presented at the American Academy of Advertising. Conference.
Proceedings (Online), Lubbock: American Academy of Advertising. Retrieved from http://search.proquest.com/openview/099565daa17c343c42f14f34d30cb996/1?pq-origsite=gscholar&cbl=40231
Li, H., Edwards, S. M., & Lee, J.-H. (2002). Measuring the Intrusiveness of Advertisements: Scale Development and Validation. Journal of Advertising, 31(2), 37–47.
35
Majhi, S. (n.d.). Ultimate Tips and Tricks for New YouTube Channel Owner. Retrieved May 15, 2017, from https://www.ampercent.com/ultimate-tips-and-tricks-new-youtube-channel-owner/15870/
Martí Parreño, J., Sanz‐ Blas, S., Ruiz‐ Mafé, C., & Aldás‐ Manzano, J. (2013). Key factors of teenagers’ mobile advertising acceptance. Industrial Management & Data Systems, 113(5), 732–749. https://doi.org/10.1108/02635571311324179
Mollen, S., Rimal, R. N., Ruiter, R. A. C., & Kok, G. (2013). Healthy and unhealthy social norms and food selection. Findings from a field-experiment. Appetite, 65, 83–89. https://doi.org/10.1016/j.appet.2013.01.020
Silvia, P. J. (2006). Reactance and the dynamics of disagreement: multiple paths from threatened freedom to resistance to persuasion. European Journal of Social Psychology, 36(5), 673–685. https://doi.org/10.1002/ejsp.309
Song, Y. J. (2014, January 16). Let me skip the ads! Revisiting the psychological reactance theory in an advertising context through examining the role of expectation (Master Thesis). University of Illinois at Urbana-Champaign.
Speck, P. S., & Elliott, M. T. (1997). Predictors of Advertising Avoidance in Print and Broadcast Media. Journal of Advertising, 26(3), 61–76.
https://doi.org/10.1080/00913367.1997.10673529
Stanley, T. L. (2017, April 30). Infographic: What Content Creators Really Think About Branded Video. Retrieved May 9, 2017, from
http://www.adweek.com/tv-video/infographic-what-content-creators-really-think-about-branded-video/
Stross, R. (2011, October 29). On Hulu and YouTube, Commercials by Multiple Choice. The New York Times. Retrieved from http://www.nytimes.com/2011/10/30/business/on-hulu-and-youtube-commercials-by-multiple-choice.html
36
Tang, J., & Zhang, P. (2013). Forced or Inspired: Understanding Consumers’ Cognitive Appraisals and Behavioral Responses towards Online Advertising. SIGHCI 2013 Proceedings. Retrieved from http://aisel.aisnet.org/sighci2013/21
Tang, J., Zhang, P., & Wu, P. (2013). Passive or Active: Understanding Consumers’ Behavioral Responses to Online Advertising. PACIS 2013 Proceedings. Retrieved from http://aisel.aisnet.org/pacis2013/188
Why does Video Autoplay on Mobile Devices not Work? (2015, April 25). Retrieved June 16, 2017, from https://www.aerserv.com/why-does-video-autoplay-on-mobile-devices-not-work/
APPENDIX Open recall codebook
Instructions:
- Every unique detail should be rewarded with a score of 1 point
- Unique details are words or descriptions that are not synonyms for example: ‘camera’ and ‘fototoestel’ are considered synonyms and count for a maximum of 1 point
- Words than are a combination of two unique details count as two points for example: ‘wildlifefotograaf’ consists of ‘wildlife’ and ‘fotograaf’ and has to be rewarded with two points
1 = score with 1 point, 0 = score with 0 points Nikon ad: 1 = Bewolkt/regeachtig/wind/grijze lucht Bioloog Camera/fototoestel/fotocamera/digitale camera/toestel Donkere kleuren Eiland/Noordpool/spot/plek/kustgebied/natuurgebied/kust/IJsland Geduld/geduld belangrijkste wapen/wachten/rust
Foto/Fotograaf/fotografeert/Shot (als in foto)/plaatje/fotograferen/wachten op dat ene shot/perfecte kiekje
IamFreek/I’m Freek/iamfreek.com/webpagina met I am Freek/ Intense stem/typische manier van praten/irritante stem
Lach/liggend/bek open/bekvechten/hij liep, stond/tegen over elkaar in rust/gapende Moeilijk/lastig
Muts/kapuchon Nikon/nukon/Nikkon
37 Meeuwen/zee-vogels
Penguïn
Rood (Vliegtuig) rugzak
Snel/shots (als in montage van de ad)/snelle shots/close up
Spannende muziek/meeslepend gebeuren/avontuur/gepassioneerd/ontdekking Thermopak/zwarte jas
Vliegtuig/Vliegtuigje/propellervliegtuig/klein vliegtuig/watervliegtuig Volgen/Volg zijn verhaal op...
Water/rotsen/rotsachtig/golven/zee/steenachtig/zeewier/zeewater/ijskoud Wildlife/natuur/natuurgebied/wild Zeeleeuwen/zeehonden/dieren/wilde dieren/zeerob/zeekoeien/walrussen/dieren/robben/zoogdieren Zeeolifanten/zeehonden/dieren/wilde dieren/zeerob/zeekoeien/walrussen/dieren/robben/zoogdieren Zoomlens/lens/groot T-mobile ad: 1 =
€35 per maand/35 euro/vast bedrag Auto/politieauto/bus tegen een auto Collega’s
Colliseum/colleseum Dab/dapp/deb
Druk/actief/herrie/onduidelijk/irritant/veel beweging/veel door elkaar heen
DVD gooien/CD gooien/achterstevoren CD in een recorder gooien/CD/DVD in speler gooien Emmer/druppelen in een emmer
Explosie/Ongelukken/ontploffingen/incident/ongeluk/vuur/crashen
Filmpjes/clipjes/bekende internetfilmpjes/fragmenten/korte fragmenten/YouTube fragmenten/beelden/verschillende momenten/films
Festivalfilmpje/Pinkpop/concert/festival/house feest/lowlands fragment/festivaltent Fiets
Flesjes flippen/bottle flip/flesjes gooien/2 waterflesjes/flessen Frituur
Gamen/games/video gamen Gatenkaas
Hand in het haar/haar trekken Hip/snel/populair/grappig/fun Jingle/T-mobile geluidje Kantoor
Koffie Laptop
Logo/Letters/veel letters/witte letters/begin en eind gelijk
Mensen springen op een veld/dansende kinderen op straat/dansend persoon op
grasveld/juichende mensen/vrienden bij elkaar/feestende mensen/blije mensen/dansen in het park/kinderen spelen
Meneer Krabs/alles afgebrand/meneer krab
Muziek (irritante muziek)/zenuwslopende muziek/muziek van Crystal Fighters Ninja kindje/ Aziatische mensen/Jackie chan nadoen
Obama/Barack Obama/Obama die iets laat vallen Oude mensen/oma/man/vrouw
Patrick
38 Reizen/buitenland/vakantie
Robben
Roze (Roze van kleur)/rood/paars/donker/witte letters/zwarte kleuren/magenta/fuchsia T-mobile
Tafel
Telefoonmaatschappij/provider van telefoon/telefonie/telecomaanbieder Toren
Trein/trein tegen tank
Trump/Donald Trump/Trumph Schieten/schot van buiten de 16 Schrikken/moeder schrikt/
scoren/scoort/doelpunt/bal in de linker kruising Selfiestick/go pro Sneeuw Snowboarden/sporten/extreme sporten/skipiste/wintersport/skateboard/sportbeelden/skaters/snowboarder/skiën/ski kleding Social media SpongeBob/strips/Sponsbob/spongebob scenes Spray stadion Streamen/surfen/stremmen
Unlimited data, bellen, alles/unlimited bellen/unlimited data/data/onbeperkt data/onbeperkt data, bellen, alles/bellen/alles/alles in 1/onbeperkt
internetten/voordeel/bundel/telefoonabbonnement, plan/zoveel als je wilt/onbeperkt/unlimited/abbonnement/Mb’s
Stunts/verschillende acties
Vangen/glas bier dat gevangen wordt/bier en band moment/biervanger podium/zanger staat op publiek en vangt biertje/biertje vangen/bierzuipen
Voetbal/voetballer/sporten/sportbeelden/sport VR-brillen/VR/virtual reality/z’n bril
Vrije trap 0 = €25 €30 per maand 2-jarig contract Autorace Basketball Canon Flamengo Geen goedkoper Kodak Helikopter Hollands nieuwe Mooiste Nek Sms’en Niet op gelet Reclame Tennis Vodaphone Vrachtwagen
39 Vrije tijd
