International Journal of Advertising
The Review of Marketing Communications
ISSN: 0265-0487 (Print) 1759-3948 (Online) Journal homepage: https://www.tandfonline.com/loi/rina20
User generated content presenting brands on
social media increases young adults’ purchase
Mira Mayrhofer, Jörg Matthes, Sabine Einwiller & Brigitte Naderer
To cite this article: Mira Mayrhofer, Jörg Matthes, Sabine Einwiller & Brigitte Naderer (2020) User
generated content presenting brands on social media increases young adults’ purchase intention,
International Journal of Advertising, 39:1, 166-186, DOI: 10.1080/02650487.2019.1596447
To link to this article: https://doi.org/10.1080/02650487.2019.1596447
© 2019 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Published online: 28 Aug 2019.
Submit your article to this journal
Article views: 3964
View related articles
View Crossmark data
Citing articles: 1 View citing articles
Full Terms & Conditions of access and use can be found at
INTERNATIONAL JOURNAL OF ADVERTISING
2020, VOL. 39, NO. 1, 166–186
User generated content presenting brands on social
media increases young adults’ purchase intention
€rg Matthesa, Sabine Einwillerb and Brigitte Naderera
Mira Mayrhofera, Jo
Advertising and Media Effects Research Group, Department of Communication, University of Vienna,
Vienna, Austria; bCorporate Communication Research Group, Department of Communication,
University of Vienna, Vienna, Austria
On Facebook, companies not only actively spread branded content
themselves, they also encourage users to do so. Hence, persuasive
messages blend into the stream of content, making it increasingly
difficult for users to identify and cope with this covert advertising
content. In an experimental study, we confronted users to disclosed
advertisements; brand; and user-generated posts allowing us to
discern effects on persuasion knowledge, affective reaction and, in
turn, purchase intention. Furthermore, we manipulated viewer’s
attention to the posts. In line with the Persuasion Knowledge
Model, we found that user-generated content did not trigger
persuasion knowledge and a subsequent negative affect. Thus,
user-generated content led to higher purchase intention compared
to disclosed advertisement and brand posts. Surprisingly,
participants’ heightened attention decreased their negative
affective reaction towards the advertisement post compared to the
brand post. We conclude that policy makers should consider
employing advertising disclosures for user-generated content.
Received 28 June 2018
Accepted 13 March 2019
Social media; covert
In 2017, Facebook had 2 billion daily users worldwide (Socialbakers 2017). Given its
massive number of users, Facebook leads a pack of social media sites in marketing
spending, which from 2009 to 2016 increased by a staggering 234%. Today, a whopping 72.5% of companies in the U.S. use Facebook for advertising purposes (Moorman
2016). With a vast majority of social media marketers being convinced that Facebook
‘delivers the best ROI among the social networks’ (Newberry 2018), it is an important
and established marketing channel (Choi 2011).
On Facebook’s newsfeed, users are confronted with branded content in three different ways. First, paid advertisement posts can be placed by companies in the target
group’s news feed, in which case brands appear as the sources of the posts. Such
posts are comparable to traditional advertising insofar as companies invest financial
resources into exposing a broad target base of consumers to their persuasive
CONTACT Brigitte Naderer
Advertising and Media Effects Research Group,
Department of Communication, University of Vienna, W€ahringerstr. 29, 1090 Vienna, Austria.
! 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives
License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in
any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.
INTERNATIONAL JOURNAL OF ADVERTISING
messages. To comply with Facebook’s policy for advertisers, such posts have to be
marked as advertisements by a disclosure. In 2018, five million businesses actively
placed paid advertisement campaigns on Facebook (Newberry 2018). Second, brands
can create a brand page to solicit a brand community, i.e. users who ‘like’ the brand
page. Over 70 million businesses operate their own Facebook page, thus taking advantage of this promotion opportunity (Newberry 2018). In those instances, however, only
users who have already ‘liked’ the brand on Facebook are exposed to the companies’
posted content. Third, users can post content that contains brand references to their
Facebook pages and, appearing as its sole source, thereby exposing their entire
Facebook network to the branded content (Facebook 2017).
The latter option, user-generated brand content, is a highly discussed marketing
tool. Especially business media, such as Forbes (Olenski 2017) or Adweek (Merckel
2017) praise user-generated content as highly advantageous covert marketing tool for
companies, as it blends into the editorial social media content. Regulators might be,
however, rightfully concerned that users are no longer able to identify persuasive content on social media. This is crucial though, as we know based on the Persuasion
Knowledge Model (PKM) by Friestad and Wright (1994) that in the case of traditional
media, only if viewers identify a certain content as commercial, coping mechanisms
are triggered, which may lead to more critical processing of the message. Prior
research in the online realm found similar results for blog posts (e.g. van Reijmersdal
et al. 2016), vlog posts (De Jans, Cauberghe, and Hudders 2018), native advertising in
articles (Campbell and Evans 2018), or Instagram posts (e.g. Evans et al. 2017). The U.S
Federal Trade Commission advises social media users to employ media literacy techniques to identify branded content. Thus, users should reflect on questions like ‘who
created or paid for the ad, and why?’ (FTC 2013). However, answering those questions
might not be so easy anymore in times of social media marketing, when the explicitly
stated source of a posting can be a user and FTC guidelines of disclosing advertised
content are bluntly ignored by businesses (Fletcher 2017). Hence, it is crucial to generate insights on how the processes proclaimed by the PKM manifest for the different
forms of branded posts on social media.
Furthermore, even if branded posts are disclosed as advertisement, studies on
source and disclosure effects suggest that viewers have a hard time identifying persuasive content because they do not pay enough attention to corresponding indicators (Wojdynski and Evans 2016). In fact, the European Commission (2018) stated that
increasing viewers’ visual attention to advertising disclosures can serve as a remedy
against disguised advertising practices. This could be the case, because heightened
attention can enhance persuasion knowledge (Boerman, van Reijmersdal, and Neijens
2015). Two highly related constructs in this regard are visual attention and cognitive
involvement (Pieters and Wedel 2007). That is, when humans concentrate on a content
(i.e. are highly cognitive involved), their visual attention to the different aspects of this
content is also high. In light of this, one would theorize that higher attention helps
users to understand that they are confronted with persuasive messages.
Yet from literature on information processing, we know that humans only have limited cognitive processing capacities (Lang 2000). Studies indicate that social media use
has led to an extensive increase in the amount of information a user is exposed to,
M. MAYRHOFER ET AL.
greatly increasing the cognitive load (Gomez-Rodriguez, Gummadi, and Schoelkopf
2014). Consequently, highly attentive users might not have enough capacity to experience an affective reaction towards a content (Matthes, Schemer, and Wirth 2007;
Janssen et al. 2016). Thus, attention may foster persuasion knowledge, but not necessarily negative affect.
With this study, we are the first to investigate the effects of attention on persuasion
knowledge and affective reaction in the social media environment. In addition, this is
also the first comprehensive experimental study testing the effects of three different
types of branded Facebook posts, i.e. disclosed advertisement posts, brand posts, and usergenerated posts. This allows us to assess viewers’ persuasion knowledge and affective
reaction towards these three different types of posts and subsequent brand outcomes.
The power of user-generated content
Media outlets have often highlighted the value of user-generated branded content.
Forbes has recommended that companies ‘take proactive steps to stimulate the creation of user-generated content’ (Olenski 2017), and the subtitle of an article in the
advertising online journal Adweek has argued that ‘Not only is UGC [user-generated
content] much cheaper to implement, but it is also much more effective’ (Merckel
2017). Citing metrics such as hashtag usage and retweets, the latter publication especially underscored the power of user-generated brand images (Merckel 2017). In the
MIT Sloan Management Review, businesses even get advised to redefine their social
media marketing goals in a way that includes brand engagement, i.e. customers posting branded content, as an ROI indicator (Hoffman and Fodor 2010). Consequently,
numerous studies have explored under which circumstances users are willing to post
branded content. Researchers who have investigated the reasons why users contribute
and create brand-related content have identified personal identity, integration, and
social interaction as major motivations. Many users upload pictures displaying brands
in order to express their connection to a brand’s image and popularity as well as their
inclusion in the social group that uses the brand (Muntinga, Moorman, and Smit
2011). Common among such posts are so-called ‘brand-selfies’, which Sung, Kim, and
Choi (2018, 10) have dubbed an ‘effective means of self-expression’. In a content analysis of posts connected to two apparel brands, Smith, Fischer, and Yongjian (2012)
found that personal identity, integration, and social interaction motivated a third of
user-generated Facebook posts.
More importantly, user-generated brand posts result from marketing efforts such as
real-world tie-ins or contests. In their content analysis of several social media networks,
Ashley and Tuten (2015, 22) found that ‘26/28 brands invited users to share content’.
One industry that often uses covert marketing techniques is the alcohol industry. Realworld tie-ins are the most important content generators on alcohol brand community
pages. Alcohol brands for instance launch branded event series (Nicholls 2012), sponsor sport events (Pinsky et al. 2017), and corresponding hashtags. Another technique
of the alcohol industry, described by Lobstein et al. (2016), are photo- or video-competitions. People are therefore encouraged through these events and competitions to
take pictures and post them on their own social media channels: ‘The personal brand
INTERNATIONAL JOURNAL OF ADVERTISING
experience of the fan leads to the circulation of branded information within the
broader mediation of their everyday life within their circle of friends’ (Brodmerkel and
Carah 2013, 279). Carah and Shaul (2016) analyzed two highly successful Smirnoff campaigns that mobilized user-generated branded content on Instagram to promote the
Smirnoff brand. Regarding how user-generated content connected the brand to the
users, they wrote that users’ ‘use of hashtags places the brand within a wider flow of
images related to their own bodies and identities’ (Carah and Shaul 2016, 74).
However, even with the reasons for users to create and share brand-related content
established, it remains unclear whether users exposed to such content will react similarly to how they react to brand-related content posted by commercial sources. In the
area of effect studies, research on the increasingly recognized value of user-generated
brand content is scarce. Hence, we sought to assess how the source of a branded
message affects viewers’ evaluations of the brand.
Several scholars have investigated the effects of having users as the source of persuasive messages. Central to their studies was, however, the likeability, credibility, and
perceived quality of user-generated advertisements (Ertimur and Gilly 2012; Steyn
et al. 2011; Thompson and Malaviya 2013). For instance, Ertimur and Gilly (2012, 126)
found that when viewers were aware that a user-generated post was commercial content, they did not cope with the persuasive intent of the post. Instead, they took the
role of an ad critic and, for example, gave suggestions about the lighting or editing of
the presented pictures.
Other authors have examined the role of users as the source of advertisements,
when testing advertisement disclosures (e.g. De Jans, Cauberghe, and Hudders 2018;
Hwang and Jeong 2016; Kim and Song 2017). Their results do not tackle, however
how different sources affect the activation of persuasion knowledge, because they
depend on the effectiveness of the disclosure manipulation itself – that is, whether
the disclosure made participants realize that the source of the content was a company
and not the user (Kim and Song 2017) and whether that realization triggered negative
affect (Matthes and Naderer 2016). For example, Kim and Song (2017) compared user
generated brand posts with and without disclosure. This however cannot clearly establish the effect of a user as the source of branded content in comparison to brand messages by the company itself.
At the same time, although other scholars have focused on the effects of reviews
or product-related stories shared by users, such forms of advertising are hardly comparable to user-generated brand posts on Facebook, which do not always have a persuasive or informative intent (Smith, Fischer, and Yongjian 2012). This brings us to the
possibly biggest advantage for companies in employing user-generated brand posts
as marketing effort and an area strongly lacking research: the alleged incapability of
social media users to identify it as persuasive content and in turn to cope with it.
Coping with brand messages
A model analyzing the different factors influencing consumers’ understanding of persuasion tactics and their knowledge of how to cope with persuasive attempts is the
PKM (Friestad and Wright 1994, 1995). Persuasion knowledge develops gradually, due
M. MAYRHOFER ET AL.
to consumers’ regular exposure to persuasive content. Persuasion knowledge is based
on knowledge about the topic, understanding the persuasive process, and knowing
the agents respectively communicators. In this study, we specifically focus on the latter, i.e. agent knowledge. This aspect consists of the consumer being aware that the
source of a message has a commercial background and, hence, a persuasive intent.
This knowledge should lead users to identify a message as a persuasive attempt and
in turn trigger coping mechanisms. The agent ‘represents whomever a target identifies
as being responsible for designing and constructing a persuasion attempt’ (Friestad
and Wright 1994, 2). In the case of Facebook posts, the source of the posting is always
stated above the content, i.e. the picture or statement. In other words, the PKM suggests that knowing that a brand or company is behind a communication effort will
trigger coping mechanisms aimed at potentially blocking a persuasive intent (Friestad
and Wright 1995).
On Facebook, however, there are two types of posts that present a brand as the
source. Brands can pay to appear in users’ newsfeeds that fit their target group. In
this case, a short disclosure alerts the user to the fact that they are confronted with
paid content. If a user or one of his or her ‘friends’ however, follows a brand page on
Facebook, they agree to be confronted with branded posts and as such no ad-disclosure is presented. Embedded in a stream of miscellaneous content, both types have, at
first glance, only one main difference: The disclosure of the advertising message.
Consequently, it is questionable if users correctly identify persuasive content presented
on social media at all or only in the case of paid posts that contain an ad-disclosure
(Evans et al. 2017).
Past research indicates that the appearance of a disclosure can alert users’ attention
to the fact that they are confronted with a persuasive message, thus trigger users’ persuasion knowledge. This has been indicated for product placement disclosures on TV
(e.g. Boerman, van Reijmersdal, and Neijens 2012, 2015; Boerman and van Reijmersdal
2016; Matthes and Naderer 2016; Tessitore and Geuens 2013), disclosures in advergames (Evans and Hoy 2016; van Reijmersdal et al. 2015), and social media outlets
such as blogs (van Reijmersdal et al. 2016) or Instagram (Evans et al. 2017). Results
hitherto remain inconclusive, however, mainly due to the differences in disclosure
manipulations (see Evans et al. 2017). In other words, whether the disclosure factually
made participants understand that the underlying source of the content is a company
and not a user is of key importance. Some studies, however, have put the effectiveness of certain ad disclosures in covert marketing into doubt (e.g. Wojdynski and
Evans 2016). Following the logic of the PKM, we therefore hypothesize:
H1: Persuasion knowledge is activated a) the least when participants are exposed to usergenerated posts and b) the most when exposed to disclosed advertising posts. Hence, we
expect brand posts to evoke more persuasion knowledge than user-generated posts and
less than disclosed advertising posts.
Activated persuasion knowledge may in turn encourage users to critically contest
the content itself, the source of the content, and the persuasive tactics of the content
(Fransen, Smit, and Verlegh 2015). In a qualitative study, Zuwerink and Cameron
(2003) assessed the variety of coping mechanisms that can be triggered by persuasion
knowledge. Cognitive resistance, such as attitude bolstering, which is the reassurance
INTERNATIONAL JOURNAL OF ADVERTISING
of one’s belief, and affective resistance, such as negative affect against the persuasive
attempt, were among the most frequent resistance strategies. In the context of sponsored content, extant research (van Reijmersdal et al. 2016) suggests that affective
resistance co-occurs with cognitive resistance for two reasons. First, research showed
that content which is perceived as advertising often directly triggers negative attitudes, annoyance, and avoidance behaviour (e.g. Mittal 1994; Moriarty and Everett
1994). Second, Friestad and Wright (1994) pointed out that identifying a communication as persuasive intent is a change of meaning. Thus, the persuasive attempt is considered an intruder into the communication context. In other words, if a post is
identified as a persuasive intent while someone is browsing through entertaining and
informative content on their Facebook wall, it might intrude in the Facebook browsing
experience and thereby trigger resistance. The meaning of the persuasive message is
scrutinized, which in turn leads to negative effects on brand evaluations (Evans et al.
€derlund 2015). This was
2017; Hwang and Jeong 2016; Liljander, Gummerus, and So
supported by research on advertisement intrusiveness which has linked the intrusiveness to viewers being annoyed by advertisements (e.g. Ha 1996; Truong and Simmons
2010). Consequently, we pose the hypothesis:
H2 : The higher a participant’s level of persuasion knowledge regarding a Facebook post
the stronger is his/her negative affective reaction against the Facebook post.
Last but not least, the goal of a marketing communication is to influence marketing
outcomes, such as, for example, purchase intent. Extant research has demonstrated in
diverse channels how increased persuasion knowledge diminished marketing outcomes (Boerman, van Reijmersdal, and Neijens 2012; Matthes, Schemer, and Wirth
2007). Interestingly, Wei, Fischer, and Main (2008) have found the effect to especially
hold for unfamiliar brands. We predict that this appears on account of the explained
triggered coping mechanism of negative affect (Zuwerink and Cameron 2003).
Therefore, we suggest that:
H3 : The stronger participants’ negative affect against a Facebook post, the stronger the
decrease in their purchase intention of the brand presented within the post.
The role of attention in the persuasive process
In extant research, different aspects of attention have been studied (Park and McClung
1986; Wojdynski and Evans 2016). Two concepts that researchers have connected are
visual attention and cognitive involvement (Pieters and Wedel 2007). For instance,
involvement scales measure aspects as ‘I concentrated on the story’ (Matthes,
Schemer, and Wirth 2007, 499) and cognitive involvement manipulations demand participants to ‘pay specific attention’ to the depicted content (Park and McClung 1986,
546). This is highly connected with participants looking more attentively at all aspects
of a content (i.e. visual attention, Pieters and Wedel 2007).
Why is the degree of attention important? Especially in the social media environment, attention to posts may be crucial for the persuasion process, as commercial and
non-commercial content are strongly interconnected. In fact, several studies on the
topic of source and disclosure effects have demonstrated that viewers are unaware
M. MAYRHOFER ET AL.
that a content has a commercial source, even if the source itself is disclosed or an
advertising disclosure is included in the content (e.g. Boerman, van Reijmersdal, and
Neijens 2012, 2015; Nelson and Park 2015; Kim, Pasadeos, and Barban 2001; Wojdynski
and Evans 2016). Eye-tracking studies have established that one issue is a lack in visual
attention of the viewers. If users don’t pay attention to the posts, they may not process the information provided in, for instance, disclosures. As a consequence, the
intended effect of disclosures on persuasion knowledge can hardly be achieved (see
Wogalter and Laughery 1996 for results on warning signs and labels in general, see
Boerman, van Reimersdal, and Neijens 2012 for a study on product placement disclosures). For instance, Wojdynski and Evans (2016) found that in the case of a news
story, viewers did not visually engage with an advertising disclosure positioned above
the story. This position of the disclosure, however, matters for Facebook posts, as the
source of a message is positioned above the content. Hence, when viewers are confronted with advertising or brand posts, a lack of attention to the posts makes it
unlikely that sources and disclosures are processed. This, in turn, may impede the generation of persuasion knowledge. Therefore we hypothesize:
H4: In the disclosed advertising and brand post conditions, opposed to the usergenerated content condition, users in the high-attention condition will have a higher
persuasion knowledge, compared to users in the normal-attention condition.
Regarding negative affect, the role of attention, however, is less clear. As already
mentioned, cognitive resources of viewers are typically limited (Nairn and Fine 2008).
That is, an information overload could occur when the amount of input to a system
exceeds its processing capacity (Janssen et al. 2016). Yet negative affect towards an
advertising attempt requires a substantial amount of cognitive resources. Following
this assumption, users that process posts with high attention have only limited resources left to experience an affective reaction towards messages (Nairn and Fine 2008). In
fact, this mechanism was found for traditional television programming (e.g. Park and
McClung 1986), as well as for integrated content, such as product placements (e.g.
Janssen et al. 2016; Matthes, Schemer, and Wirth 2007). For instance, Matthes,
Schemer, and Wirth (2007) observed that levels of involvement exerted an effect on
negative affective outcomes: Highly involved viewers with low levels of persuasion
knowledge were least critical towards the embedded brands. Along the same lines,
Janssen et al.’s (2016) results indicate that depletion can impact the critical processing
of information. The authors did find that depleted participants were less critical
towards the embedded brands. Again, these findings can be explained on the theoretical basis of the limited capacity model which suggests that humans have only a limited capacity of critically processing the abundance of information provided (Lang
2000). Especially in the social media environment, heightened attention to posts might
lead to a cognitive overload for users, as they are confronted with a variety of contents in those posts. This cognitive overload could impede their experience of a negative affect towards content. In other words, in the case of processing social media
content, attention to the posts as a whole may require substantial cognitive resources.
These resources, however, are necessary to raise negative affect in response to advertising or brand posts. That is, even though viewers may realize the persuasive attempt,
they may lack the necessary cognitive resources to generate negative affect. However,
INTERNATIONAL JOURNAL OF ADVERTISING
to our knowledge, there is no research yet on the effects of attention on negative
affect in regard to social media content. We therefore refrain from formulating an
hypothesis, and ask:
RQ1: How will attention influence participant’s affective reaction regarding the (a)
disclosed advertising post, (b) brand post and (c) user-generated post?
To conclude, there is no research yet comparing the effects of user-generated
posts, brand post, and disclosed advertising posts as sources of branded Facebook
posts. We expect that users as sources of branded Facebook posts will trigger less persuasion knowledge regarding the post compared to brand posts as well as disclosed
advertising posts, therefore impeding negative affect. Hence, user-generated brand
posts should more positively affect purchase intention than disclosed advertising posts
or brand posts. Furthermore, we test the role that attention could play in the persuasion process. For the full effect model see Figure 1.
The experiment had a 3 (source: disclosed advertisement versus brand post versus
user-generated post) ! 2 (attention: high versus normal) between-subject design.
Analyses were conducted with a total of N ¼ 293 college students above the legal
drinking age (Mage ¼ 22.57; SD ¼ 3.58; 78.2% female) in the university laboratory of the
Department of Communication at the University of Vienna in May 2017. We followed
the APA ethical guidelines; hence all participants signed an informed consent form
before entering our university laboratory, where the study was conducted. Then, we
randomly assigned them to an attention condition consisting of a textual instruction.
We operationalized attention via two aspects, cognitive involvement adapted from
Park and McClung (1986) and visual attention. In the high-attention condition, we
asked the participants to concentrate on all aspects of the posts, to memorize them
and to look carefully at all aspects of them. In the normal-attention condition, we
encouraged their usual browsing behaviour (see Appendix). In a second step, we randomly assigned participants to one of the source conditions. Participants went
through the 20 Facebook posts in randomized order, fifteen identical nonrelated posts
and five source manipulated alcohol posts. After the study, they were debriefed about
the topic of the study and the risks of excessive alcohol consumption.
All participants saw fifteen identical, nonrelated posts aimed at embedding the five
manipulated posts in a realistic setting. Furthermore, participants saw five posts featuring the ‘Ciroc’ brand. Each post presented a different photo related to the ‘Ciroc’
brand. The ‘Ciroc’ posts were identical, except for the source being a disclosed advertising by the brand ‘Ciroc’, a brand post by the same brand, or a user-generated brand
post featuring the brand ‘Ciroc’. In other words, the remaining five posts were identical except for the source of the posts and, respectively, the included ad disclosure
M. MAYRHOFER ET AL.
Figure 1. Model of all hypothesized paths and research questions.
(Evans et al. 2017). Hence, each participants either saw all five different ‘Ciroc’ posts
by a disclosed advertising, a brand post or a user-generated post, the source being
treated as between-subject factor (Stimulus material is available upon request).
We chose an alcohol brand as the target brand because we deemed it important
that the user-generated content seems realistic to the participants. Regarding the post
engagement rate – which involves response rates to fan posts, counted likes, comments and shares – the alcohol industry is the clear forerunner (Socialbakers 2012).
Brand pages post content up to twice a day (Carah 2014), strongly encouraging users’
engagement with real-world tie-ins; interactive games; sponsored online events and
invitations to drink’ (Nicholls 2012, 487). Hence we assumed that users are used to
this kind of content posted by brand as well as user sources.
We pre-tested the stimulus pictures to assure that their content did not appear more
realistic for either a post placed by a company or a user-generated post. This was of
high importance, in order to exclude effects from the content itself. Student participants (N ¼ 73) saw twenty Facebook posts, among them the manipulated posts. They
were randomly assigned to a condition with the user being the post source (n ¼ 25) or
the brand being the source (n ¼ 48). We showed participants only the content of the
manipulated posts, i.e. a statement and a picture but without any source. We did this
for all manipulated posts one by one. Subsequently, we asked participants, if they felt
that a user or brand would be equally probable to post such content (1 ¼ not at all,
5 ¼ very much). Participants evaluated the probability for picture one (M ¼ 3.74,
SD ¼ 1.11), picture two (M ¼ 3.53, SD ¼ 1.27), picture three (M ¼ 3.44, SD ¼ 1.21), picture
four (M ¼ 3.22, SD ¼ 1.29), and picture five (M ¼ 3.74, SD ¼ 1.33). All content was
deemed to be as probable to be posted by a user as by a brand. Additionally, participants were asked, if they could imagine seeing such a post on Facebook (1 ¼ not at
all, 5 ¼ very much). All pictures were deemed expectable as a Facebook post: picture
INTERNATIONAL JOURNAL OF ADVERTISING
one (M ¼ 4.33, SD ¼ 0.77), picture two (M ¼ 3.89, SD ¼ 1.04), picture three (M ¼ 3.78,
SD ¼ 1.24), picture four (M ¼ 4.0, SD ¼ 1.08), and picture five (M ¼ 4.38, SD ¼ 0.72). As
such, the content of the post was deemed appropriate for the study.
We assessed participants’ attention using two statements on a 5-point scale adapted
from Park and McClung’ cognitive involvement manipulation (1986) (‘I tried to memorize the posts’, ‘I concentrated on the posts’) and three statements focusing on participants’ visual attention (‘I looked at all the aspects of the post’, ‘I looked at the posts
attentively’, ‘I took time to look at the posts’). Following Carpenter (2018), we conducted an explorative factor analysis with oblique rotation with the five items, which
indicates that both elements are part of the same concept, which we employed as a
measure of attention. The analyses yielded one factor, explaining 71.71% of the variance. Hence, the statements formed a reliable index (a ¼ 0.90; M ¼ 3.87, SD ¼ 0.95).
Furthermore, we measured source recall as an open question. Comparable to van
Reijmersdal et al. (2016), we measured the participants’ persuasion knowledge using
three items (‘These posts were advertising’, ‘These posts were posted without commercial interest’ (recoded), ‘These posts were posted to advertise a product’) on a 5-point
scale (a ¼ 0.90; M ¼ 4.18, SD ¼ 1.15). Negative affective reaction (van Reijmersdal et al.
2016) to the post was also measured using three items on a 5-point scale (‘I was irritated/annoyed/enraged by those posts’). The formed index was reliable (a ¼ 0.78;
M ¼ 2.30, SD ¼ 1.04). In addition, we assessed participants’ purchase intention for the
embedded brand, Ciroc, using three items (‘I would buy Ciroc Vodka’, ‘I would by
other alcoholic products of the Ciroc brand’, ‘I am interested in where to buy Ciroc
Vodka’) on a 5-point scale (a ¼ 0.78; M ¼ 1.92, SD ¼ 0.93). As a control variable, we
included participants’ alcohol consumption, stated on a single 5-point scale item
(1 ¼ never, 2 ¼ less than once a month, 3 ¼ two to four times a month, 4 ¼ two to three
times a week, 5 ¼ four times a week or more; M ¼ 2.94, SD ¼ 0.85). It was to be expected
that participants’ current level of alcohol consumption might predict their interest in
buying alcohol (e.g. Alhabash et al. 2015). Further, we included age (mean-centred) as
a control variable, as extant studies have shown that age can affect viewers’ perception of alcohol related content (Mayrhofer and Naderer 2019).
Randomization and manipulation checks
A randomization check for gender (v2 ¼ 5.75, df ¼ 5, N ¼ 293, p ¼ .33) was successful.
Furthermore, the manipulation check of attention in the main study was successful.
Hence, in the normal attention condition, users indicated a significantly lower level of
attention (M ¼ 3.50, SD ¼ 0.91) compared to the high attention condition (M ¼ 4.27,
SD ¼ 0.83) (t(291) ¼ #7.55, p
Purchase answer to see full