1
“The Framing of Corporate Promotional Video”
A quantitative content analysis of viral and non -viral corporate advertising from a communication sciences perspective
Author: Joram S. Falkenburg Student ID: 10731237
MASTER THESIS
Master’s Program Communication Science Supervisor: dr. Piet Verhoeven Date of Submission: 26-6-2015
2
Abstract
This explorative quantitative content analysis of corporate promotional video –
deliberated to go viral, studies the occurance of framing factors (i.e. framing types, the framing of emotions, and valence framing) within video content, and its relation on viewing and sharing rates of this video. The overall aim is to obtain insights into what content related elements within corporate promotional video could influence its viewing and / or sharing rates; i.e. make it go viral. The data gathering method allows the study of a full dataset of 80 video’s, of which 40 are considered viral, and 40 non-viral video’s; hence also allowing comparative study. Viewing and sharing rates of video are determined, whereafter results reveal that all eight framing types (situations, attributions, actions, choices, issues, news, responsibilities and morality) occurr within corporate video. Hereby the framing of situations, attributions and news occur most frequently, and the framing of issues least. Correlation is found between numerous framing types and respective viewing / sharing rates. Highly significant and moderately strong predictive relationship is found between the framing of morality and higher viewing rates; including viral viewing rates. Highly significant, yet weak predictive relationship is found between framing of
issues and higher (viral) sharing rates. Of the nine discrete emotions (Joy/Happiness,
Intrigue/Interest, Awe/Excitement, Surprise / Shock, Anxiety / Fear, Anger, Sadness, Disgust and Pity), Joy/Happiness, Intrigue/Interest, Awes/Excitement and Surprise/Shock occur most
frequently overall. Without exception, viral video sees more emotional framing than non-viral video. Anxiety/fear, Anger and Sadness only occur in viral video, while Disgust and Pity in neither. Correlation was found between numerous discrete emotions and viewing / sharing rates. The framing of Surprise/shock sees a significant and moderately strong predictive relation to increasing both viewing and sharing rates. Positive valance occurs more than double as much as
3
negative valence, overall – and in viral as opposed to non-viral video. Correlation with viewing /
sharing rates is found between very positive, positive, and very negative valence. Very negative valance sees weak yet highly significant predictive relation to the increase of sharing rates.
4
Preface
First and foremostly, I would like to thank dr. Piet Verhoeven for the supervision of this research. I am grateful to have been able to study the subject of viral corporate video while under your guidance. Many thanks also to prof. dr. Ed Tan, for the feedback session on
operationalizing emotions within film. Many thanks to Rutger de Graaf for answering my questions regarding the operationalization of corporate communication theories. Lastly, many thanks to Nathaly van Heyningen for helping me code the data, and thereby allowing the establishment of inter-coder reliability.
5 Distrust in institutional legitimacy is a growing problem for corporate and media actors in getting across their messages effectively (Allsop, Bassett & Hoskins, 2007; Hinz, Skiera, Barrot, & Becker, 2011; Porter & Golan, 2006). Rapid evolution of information and communication technologies (ICTs) give publics unprecedented choice around the consumation and utilization of media (Deacon & Stanyer, 2014; Porter & Golan, 2006). This media diversification, however, leads to an increase in advertising – irrelevant to much of its audiences. As a result, publics are developing an increasing resistance towards corporate promotional content (Hann, Hui, Lee, & Png, 2008; Porter & Golan, 2006). Simultaneously, individuals have gained increasing control over the content they get exposed to, so that unwelcomed advertising can be disengaged with instantaneously (Camarero & San José, 2011; Hann, et al., 2008; Porter & Golan, 2006). These fastmoving changes within the media landscape ask for new ways to look at corporate
communication, publicity and promotion (Grunig, 2009; Overton-de Klerk, & Verwey, 2013). A form of communication born out of these new developments is Electronic Word-of-Mouth (EWOM). EWOM describes multi-directional discussion between actors on digital platforms and social networks, and is shown to be one of the most effective communication channels in the market place (Allsop, Bassett, & Hoskins, 2007; Barnes & Jacobsen, 2014; Camarero & San José, 2011; Overton-de Klerk, & Verwey, 2013). With EWOM, discussions specifically concern opinions and experiences with organizations, products and services (Allsop, Bassett, & Hoskins, 2007; Camarero & San José, 2011). Hereby non-commercial messages concerning brands, products or services are exchanged (Alon & Brunel, 2006).Empirical evidence reveals that EWOM is increasingly influential towards brand attitudes and
organizational reputation (Barnes & Jacobson, 2014; Hinz, et al., 2011; Willemsen, 2013; Utz, Kerkhof, & Van den Bos, 2012). EWOM is argued to pertain an entirely new communication
6 paradigm, and demonstrates a blurring between marketing and broadcasting functionality
(Barnes & Jacobsen, 2014; Overton-de Klerk, & Verwey, 2013; Welker, 2002).
As EWOM allows consumers to talk freely without organizational filtering, publics often assume their own EWOM activity to be free of corporate influence (Grunig, 2009; Overton-de Klerk, & Verwey, 2013). Contrastingly, research shows that organizations have utilized online EWOM dynamics successfully for their own gain (Van Noort & Willemsen, 2011). Such
attempts can be refered to as viral marketing (Berger & Milkman, 2012; Hinz, et al., 2011). Viral marketing describes a type of marketing effort, whereby corporate content (often video) is
delibirately created and disseminated on the World Wide Web, to play into individuals’
tendencies to share and thereby broadcast this content within their networks (Hinz, et al., 2011; Van der Lans, Van Bruggen, Eliashberg, & Wierenga, 2010). As such, the non-commercial nature of EWOM is arguably exploited for commercial purposes.
As a medium, video especially is found be highly immersive and thereby influential on viewers’ emotional states (Visch, Tan, & Molenaar, 2010). If done effectively, this corporate promotional video content may be shared and viewed millions of times within a matter of days (Berger, et al., 2014; Hinz, et al, 2011). The term ‘viral’ is essentially an adjective, which describes the abnormally high sharing and viewing rates of content. With viral marketing, these effects are highly desired, as the reach and viewership of viral content easily competes with the most costly and powerful traditional media outlets.
Despite its apparent effectiveness, much ambiguity still surrounds online EWOM and virality (Botha & Reyneke, 2013; Van der Lans, et al., 2010; Watts & Peretti, 2007). Not all content that is deliberated to go viral, actually goes viral (Botha & Reyneke, 2013; Van der Lans, et al., 2010). Marketing literature on virality concerns itself largely with how – and why content
7 appeals to audiences; with dissemination strategies, and other factors that could spurr viral spreading (Berger, et al., 2014; Camarero & San José, 2011). Although gainful insights have been made, a large element of unpredictability surrounds the spreading of viral messages effectively; especially in relation to message content (Botha & Reyneke, 2013). Despite ample study, a knowledge gap remains around how exactly publics are influenced by organisations through viral communication content.
Perhaps a reason for the prolonged existence of this knowledge gap, is that online virality is almost exclusively considered from a marketing perspective. This while even marketing literature describes viral marketing to actually concern a broadcasting phenomenon (Camarero & San José, 2011). From this alternate media-centric viewpoint, viral marketing efforts piggy-back the underlying broadcasting structures and dynamics of EWOM (Welker, 2002). Hereby publics are unawaringly influenced by marketing efforts to broadcast disguised commercial messages on their own digital media platforms (Welker, 2002). Such influence can occur regardless of
individuals being consciously aware of this. Individuals whom share this content for entirely different – non-commercial purposes. These inherently opposing forces that are involved in viral marketing, might lead to various intended, but also unintended effects – yet remain unexposed due to lack of study.
Ultimately, a mere marketing perpective might be too narrow to fully understand the complexities surrounding the influence of publics through corporate viral content. The
communications field has a long tradition in the study of corporate influence on publics, and as viral marketing actually pertains a broadcasting phenomenon it might result in more satisfying insights. Could a communication sciences approach show itself helpful in establishing more certainty in this novel communication paradigm?
8 Studying the influence of viral corporate video content from a communication sciences perspective would require the use of established theory for analyzing influence exertion through communication content. Fitting this requirement perfectly, is framing theory.
Framing theory can be approached from various angles; including defining the types of
framing used, the framing of emotions and valance framing (Gross & D’Ambrosio, 2004;
Hallahan, 1999; Kellaris, Kardes & Dinovo, 1995). As emotions and valence are shown to be a determinant of viral spreading of content, the latter two theoretical approached could also prove very insightful (Berger & Milkman, 2012; Botha & Reyneke, 2013). Before skipping ahead, however, the question first arises as to whether certain framing factors could be identified within viral corporate video content in the first place. Answering such a question could be a first aim towards obtaining empirically based clarity on the matter. Once this would be established, the influence of viral corporate video on its audiences would need to be measured, and therefore the sharing and viewership rates of such video would need to be determined. Only when these various framing factors within corporate viral video are positively identified, and the influence of these messages on their audiences have also been measured, can deeper evaluative aims be realized. These might include the exploration of relationship between framing, valance framing and the framing of emotions in video content, and their sharing and viewership rates.
Similarities, differences, or reoccurring patterns in these factors; in relation to their sharing and viewership outcomes, could imply possible correlation – and thereby measure influence – between these specific content elements and subsequent viral results. Greater understanding of how influence is exerted upon publics through viral communicative efforts could hereby be achieved. The forces behind certain content going viral, while other does not, may be explained
9 with higher certainty through this long over-due theoretical perspective shift. In order to meet these aims, the following main research questions would need to be studied and answered: RQ1: How are corporate videos – deliberated to go viral – framed in terms of types, emotions and valance, and how many views and shares did these videos receive?
RQ2: How do framing types, the framing of emotions and valance framing relate to the viewing and sharing rates of these videos?
Theoretical background
Mediatization
Studying the influence of corporate viral video content on its audiences from a communication sciences (framing) theory perspective, might be helped and optimized by consulting marketing-based literature and findings. Such synergy between these two distinct academic fields, is something that forebodes the rapidly emerging concept of mediatization. Mediatization addresses the increasing pervasion of media into every day life, and the influence of this on both public and organizational actors (Deacon & Stanyer, 2014; Pallas, & Fredriksson, 2013). From a social-contructivist tradition it describes the evolution of information and
communication technologies, and how this drives the changing communicative construction of culture and society (Deacon & Stanyer, 2014; Hepp, 2013). From an institutionalized
perspective, mediatization describes non-media actors in society adapting to media’s rules, aims, production logics, and constraints (Deacon & Stanyer, 2014; Hepp, 2013). The effects of
mediatization are also felt within academia; the fields of marketing and communications
approach the study of media from entirely different angles, yet mediatization is forcing these two fields to increasingly fuse together. The result is that both marketing and communication
10 sciences are increasingly merging to deal with overlapping subject matter (Overton-de Klerk, & Verwey, 2013). The study of viral marketing content and EWOM from a communication sciences perspective might turn out to examplify this.
Regardless, mediatization is leading to the convergence of traditional and digital media, which continually forces both corporate and non-corporate actors to adapt (Deacon & Stanyer, 2014; Hepp, 2013; Pallas, & Fredriksson, 2013). Media technologies are physically merging and morphing, which requires continous redefinition of media functionality and understanding (Deacon & Stanyer, 2014; Hepp, 2013; Overton-de Klerk, & Verwey, 2013). As the ability for individuals to communicate at global scales is as pervasive as ever, organizations must be aware of the possibility of becoming the subjectmatter within a global discussion in a relative instant. Electronic Word of Mouth (EWOM)
‘Contemporary’ forms of Word-of-Mouth (WOM) has been occurring since long before the advent of electricity and modern civilization, and may be defined as person‐to‐person, oral communication between a receiver and a sender (Abrantes, et al., 2013; Lee & Youn, 2009). Electronic Word-of-Mouth (EWOM) can be understood as quite similarly, although hereby communication takes place on recently developed digital platforms and social networks (Allsop, Bassett, & Hoskins, 2007; Barnes & Jacobsen, 2014; Camarero & San José, 2011; Overton-de Klerk, & Verwey, 2013). Concretely, EWOM is the online exchange of non-commercial messages concerning brands, products or services; whereby opinion, experience or other information are shared (Allsop, Bassett, & Hoskins, 2007; Alon & Brunel, 2006; Camarero & San José, 2011). Product reviews, discussions over email, and select social media activity are all examples of EWOM. Empirical evidence reveals that EWOM is increasingly influential towards brand attitudes and organizational reputation (Barnes & Jacobson, 2014; Hinz, et al., 2011;
11 Willemsen, 2013; Utz, Kerkhof, & Van den Bos, 2012). This effectiveness being mainly due to the believability of peer recommendation over messages received directly from corporate actors (Allsop, Bassett, & Hoskins, 2007; Camarero & San José, 2011).
EWOM dynamics
Despite its straightforward definition, (E)WOM is a complex phenomenon that cannot be understood in isolation, but rather in relation to other business processes and dynamics (Allsop, et al., 2007). For instance, if an organization behaves unethically, or produces faulty products, negative (E)WOM will be likely the result. Conversely, ethical behavior and reliable products are generally thought to result in neutral or positive (E)WOM (Allsop, et al., 2007). Hence, (E)WOM is the result of both organizational behavior and communication. This also suggests certain interrelatedness between (E)WOM and institutional legitimacy. EWOM ought to be approached wholistically, and such an approach could even be favor organizational reputation (Allsop, et al., 2007). On the other hand, utilizing it as an isolated publicity or marketing tool might not result in the desired result, since EWOM can take place far beyond the control, influence, or awareness of the organization (Allsop, et al., 2007). Marketers tend to wrongfully utilize the online arena, by approaching it as one-way communications broadcasting platform, while failing to recognize the uncontrollability of EWOM dynamics (Grunig, 2009; Overton-de Klerk, & Verwey, 2013). Such unintened insensitivity surrounding corporate utilization of EWOM structures and dynamics could explain the knowledgegap that still exists around the issue.
EWOM as a broadcasting phenomenon
EWOM is a relatively recent phenomenon, and its popular growth runs parallel to previously mentioned evolutions in ICTs (Allsop, et al., 2007; Hinz, et al., 2011). Accordingly, EWOM is the direct result of mediatization processes. Although it can be be utilized for viral
12 marketing purposes; EWOM is essentially a broadcasting phenomenon (Camarero & San José, 2011). With viral marketing, the creator of the message tailors its content in such a way it plays into certain predictable content sharing tendencies and the behaviors that make up EWOM. The desired effect of repeated message sharing is herein the broadcasting process. Ultimately, the receivers of the seeded message re-send the messsage within their social network. If the content is considered to be worthy of sharing this process will be repeated; hereby acting as itirative broadcasting channels towards their constitutients.
Virality
The term EWOM is sometimes used interchangebly with term ‘viral marketing’ (Camarero & San José, 2011). In reality, however, the two are entirely seperate phenomenon. Viral marketing describes the mutual sharing and spreading of marketing-relevant information by individuals, initially sent out deliberately by private actors to stimulate and capitalize on non-commercial EWOM activity(Hinz, et al, 2010; Van der Lans et al., 2010). Viral marketing is a business endevour, which attempts to infiltrate non-commercial communication between publics on online platforms. Hereby high sharing and viewing rates are the desired result. Viral
marketing occurs online, although WOM discussions as a result of this online content may continue offline (Helm, 2010; Welker, 2002). Noteworthy, is that the adjective ‘viral’ also implies certain result. In fact, what distincts ‘regular’ EWOM, from viral EWOM, is the magnitude of content viewing and sharing rates.
There is no official threshhold that signifies a video officially having “gone viral,” but viral viewing rates can run into the millions within a matter of days. These figures are reached through repetitive individual sharing at exponential-like rates (Berger & Milkman, 2012; Camarero & San José, 2011; Welker, 2002). Like virusses, viral marketing strategies take
13 advantage of rapid multiplication to explode the message to thousands or even millions
(Camarero & San José, 2011). Virality also implies a certain kind of secrecy, and insidiousness (Welker, 2002). A virus can be caught before the host realizes sees any symptoms. Similarly, a viral marketing message is often disguised to not be immediately recognized as a typical advertisement (Welker, 2002). The dynamics of resistance and distrust in advertisement is
expected to play a crucial role in the effectiveness of viral spreading. Much like a virus, there is a piggy-back effect taking place where the viral message spreads by using the resources of its hosts (Welker, 2002). Viral content is initially released in select – often strategic – points within online networks; a marketing strategy also known as seeding (Hinz, et al., 2011). The content is designed in such a way, that viewers are likely to view and then spread on this message within their network – often without knowing or caring to be facilitating a marketing or publicity campaign (Hinz, et al., 2011; Welker, 2002).
For organizations, the added value of using viral marketing, as opposed to traditional promotion and publicity methods starts with cost-efficiency (Hinz, et al., 2011). This economic benefit can often be achieved since overall dispersion of the message is done - and left upto - the consumer (Hinz, et al., 2011). To enjoy viral results, individual appeal towards the message, the seeding strategy, and ultimately the content are all influential factors (Camarero & San José, 2013). Despite this multitude of factors, the highest determinant for content spreading virally is found to be content-related (Berger, et al., 2012). The influence of content on its audiences is therefore a pertinent matter within the viral communications paradigm. When studying the influence of attempted viral content on publics, viewing; and especially sharing rates could be a good measure. Functioning as a stepping stone within the main research questions, the start of such study would require a sub-research question much to the like of:
14
Sub-RQ #1: How many views and shares were received by corporate videos – deliberated to go viral?
Framing
Entman (1993) and Coombs (2007) are known for their work on framing theory. Framing essentially involves salience and selection (Coombs, 2007; Entman, 1993). To frame is to select certain aspects of perceived reality, and make them more salient within a communication text (Entman, 1993). In doing so, certain issues, objects and actors within communication content receives more or less emphasis than others (Hallahan, 1999; Valkenburg & Semetko, 2000). “Frames are used to define problems – determine what causal agents is doing what with what
cost and benefit, usually measured in terms of common cultural values; diagnose causes – identify the forces creating the problem; make moral judgements – evaluate causal agents and their effects; and suggest remedies – offer and justify treatments for the problems and predict their likely effects.” (Entman, 1993, p. 52). Framing analysis can thereby reveal the precise ways
in which influence and power is exerted onto human conciousnous in the transfer of the communication at hand (Entman, 1993).
Entman quotes Edelman (1993) in describing how frames exert power and alter perception: “The character, causes, and consequences of any phenomenon becomes radically
different as changes are made in what is prominently displayed, what is repressed, and
especially in how observations are classified. Far from being stable, the social world is therefore a chameleon, or, to suggest a better metaphor, a kaleidoscope of potential realities, any of which can be readily evoked by altering the ways in which observations are framed and categorized.”
(Edelman, 1993, p. 232). Framing happens on two levels: the framing of communication and the framing of thoughts. Frames used in texts can shape frames in the thoughts of the receiver of
15 these messages (Coombs, 2007). Hence, framing analysis might serve as a theoretical bridge across the current knowledge gap surrounding influence of viral corporate video content on publics.
Framing Types
Hallahan (1999) identified 7 general types of framing, which could be applied to
analyzing and typifying various communication content: The framing of situations, attributions,
choices, actions, issues, responsibilities and news. Framing situations pertains relationships
between individuals in everyday life situation. This type of framing is at times labeled as relational framing, which suggests that situational framing revolves heavily around human interaction, relationships, and negotiation (Hallahan, 1999). As the transmission of viral content serves ultimately a social purpose, correlation between how human interactions are framed within such content might exist (Berger & Milkman, 2012). The framing of attributes is done when certain emphasis is put on specific characteristics of objects and individuals; hereby potentially leading to biased projection and reception of information and people involved. Attribute framing can shed favorable or unflattering light upon its subject at hand, and therefore be beneficial or harmful to message sponsors in persuasive communications; including viral video (Hallahan, 1999). The framing of choices is often done when concerning uncertain situations, and puts emphasis on the losses and gains based on choices made within the
situational context. The framing of choices and actions is often done in persuasive messages; a category fitting viral promotional videos. As actors within the video are presented alternative actions towards reaching certain goals, certain actions are framed more positively or negatively than others. Thereby the perceived probability with which an actor meets his or her goal, can be skewed. Since the framing of choices and actions are commonplace in advertisements, these are expected to occur less within video deliberated to go viral, as too many shared characteristics of
16 advertising would be implied. The framing of issues emphasizes social problems and disputes, and the responsibility frame argues certain actors to be a causal factor in relation to an outcome. Various findings suggest that both positive and negative emotions can be driving factors behind sharing content online (Berger & Milkman, 2012; Bothka & Reyneke, 2013). Disputes, conflict and social issues might perhaps give rise to intense negative emotions in most people.
Nevertheless, copious viral marketing and psychology literature shows that negative emotions and negatively framed messages can have a very strong effects on its audiences (Berger & Milkman, 2012; Bizer & Petty, 2005). It is therefore of interest to study the framing of issues within corporate viral video, and how it potentially correlates with viewing and sharing rates. Attributing responsibility puts actors in a certain light, which similarly to the framing of attributes can be utilized within promotional content towards portraying the message sponsor favorably. Finally, the framing of news does not just limit itself to daily news updates, but also happens when communication texts contain familiar or culturally resonating themes, allowing information to be seen through currently actual and familiar contexts (Hallahan, 1999). In their study on framing within press and television news, Semetko & Valkenburg (2000) typified
morality framing in addition to many overlapping findings with regards to Hallahan (1999).
Messages containing moral statements could be a good example of reoccurring, culturally resonating themes. Moreover, communication literature often acknowledges the influence of morality within communication dynamics (Schultz, Castello, & Morsing, 2013). As such, morality framing is of interest in a study of framing within corporate viral video content.
Each type of framing could be present within viral video content. Exploring and
identifying the exact distribution in which this occurs will be a key area of interest, whereby the following sub-research question would need answering:
17
Sub-RQ #2: How do framing types occur within corporate video deliberated to go viral?
Sub-RQ #3: How do framing types relate to viewing and / or sharing rates?
Emotions in Viral Content
Emotionality within content is a greatly influential towards its viral spreading (Berger & Milkman, 2012; Botha & Reyneke, 2013). The immersive effects of film as a medium is known to enhance response to emotional stimuli, and therefore might play an influential role in the spreading of viral corporate video also (Visch, Tan, & Molenaar, 2010). Within marketing research, the study of emotion sees emotion as either seperate discrete events, or to occur along a continuum (Barsade & Gibson, 2007; Botha & Reyneke, 2013). From a discrete point of view, emotions are distinct from one another, and might spur individuals to act (Barsade & Gibson, 2007; Botha & Reyneke, 2013). Hence discrete emotion within corporate video content could influence their viral spreading. Discrete emotions include joy, anger, fear, sadness, disgust, and surprise (Barsade & Gibson, 2007; Botha & Reyneke, 2013). Findings by Berger & Milkman (2012) very much support this theory, which reveals that arousing content, is the most significant content- related factor in messages going viral. Arousing emotions essentially involve discrete emotions such as joy, interest, awe, excitement, and surprise. However anxiety, shock, and especially anger have been found very arousal and influential in viral sharing and viewing of video content (Berger & Milkman, 2012). From the continuum-centric approach, emotion can be scaled as positive or negative; also known as valence (Botha & Reyneke, 2013). From this perspective, the intensity with which the emotion is felt can be also considered, yet its incorporation within this research would reach beyond its scope (Botha & Reyneke, 2013). Accordingly, discrete emotionality and valence could deserve analysis from a framing perspective.
18 Framing emotions
Messages being framed a certain way can result in-, and alter a wide variety of emotional reactions within audiences (Gross & D’Ambrosio, 2004). Frames mould interpretation and perception of messages, and thereby induce or alter emotions according to viewer evaluation of their content (Gross & D’Ambrosio, 2004). Analysing emotions within video content can be done from a qualitative perspective, whereby the effect of emotion within such content is
measured by personal feedback (Visch, Tan, & Molenaar, 2010). One the other hand, measuring effect of emotions within content does not always have to be the main focus of study. Due to the complexities surrounding viral viewing and sharing results of corporate video, establishing significant relation in the first place could be more in order. Determining and measuring relation between the framing of emotions, and viral viewing and sharing results brings up the following sub-research questions:
Sub-RQ# 4: How are emotions framed within corporate video – deliberated to go viral?
Sub-RQ#5: How does the framing of emotions relate to viewing and / or sharing rates?
Valence Framing
Valence describes the labelling of the same critical information in either a positive or a negative light (Levin, Schneider & Gaeth, 1998). Valance framing of content is shown to have strong effect on audiences (Bizer & Petty, 2005; Levin, Schneider & Gaeth, 1998). Valance is shown to influence sharing and spreading of content (Berger & Milkman, 2012). Positively valanced messages are most likely to spread, but also negatively valanced messages are almost equally likely to spread rapidly (Berger & Milkman, 2012). In fact, many studies reveal that negatively framed messages see a stronger effect on their audiences than positively framed messsages (Bizer & Petty, 2005; Levin, Schneider & Gaeth, 1998). When applied to EWOM, negative messages do not always clearly produce immediate economic value for the message
19 seeder, nor does it necessarily reflect favorably on individuals that pass on the content. Most crucially, this suggests that EWOM may be less about motivation and more about the
transmitter’s internal states (Berger & Milkman, 2012).
Sub-RQ #6: How are corporate video’s – deliberated to go viral framed in terms of valance?
Sub-RQ #7: How does valance framing relate to the viewing and / or sharing rates of video?
Methods
Research DesignThis explorative study concerns a quantitative content analysis of corporate promotional videos – deliberated to go viral. Deliberation to go viral is hereby simply assumed, as it simply pertains video which was released on the World Wide Web. ‘Going viral’ implies high viewing rates; an implication which for the sake of this research is a presupposition – and not an area of question. Deliberation to go viral is hereby an umbrella-phrase which allows for the inclusion of both viral and non-viral within the study scope.
Content analysis is a staple tool within communication sciences, and used frequently within the study of both written and visual communication materials (Riffe, Lacy & Fico, 2014). The study of content framing has been done since at least the late 1970’s (MacQuail, 2010, pp. 380); a subject area becoming increasingly popular with regards to film, yet seemingly
understudied with regards to viral video. This specific research revolves around the study of framing factors (i.e. framing types, the framing of emotions, and valance framing) within corporate video – deliberated to go viral; and their relation to viewing and sharing rates of this video. The viewing rates of video are assumed to be an indicator of video popularity; a measure of virality, and perhaps an indicator of influence. The sharing rates of video could be another
20 indicator of the influence of content on its audiences, since sharing a video requires behavioral action as opposed to ‘passive’ viewing.
Datasets and Research Model
In attempt to obtain more evaluative results, these potential relations are studied comparatively with regards to both viral, and non-viral video. Within this comparison, ‘non-viral’ video was intended to act as a potential control. Accordingly, the data-collection method played a crucial role within this study. The total pertaining dataset of this study contained (N=80) corporate promotional video’s released online. More specifically, this dataset was made up of ‘viral’ (n=40) and ‘non-viral’ (n=40) video. Each framing factor potentially representing an independent variable towards viral viewing and sharing results. The dependent variables are hereby the viewing and sharing rates each video received. Where possible, the comparison between results from the viral and non-viral datasets will be made. For clarity sake, figure 1 displays a schematic representation of this research model:
21 Data collection
A total of (N=80) corporate video’s – posted online were obtained and subsequently analyzed. The viral and non-viral datasets were obtained in two distinct manners. Data collection was initiated by entering the search query “viral advertising” into the YouTube search engine. The query resulted in over 85.700 search results; most of which were irrelevant. Ultimately, the overall population size of corporate viral promotional video (i.e. relevant data) is rather limited. Although quite some viral video exists, most of it does not concern promotional video of corporate brands. The gathering of data was therefore contingent upon obtaining (n=40) promotional video which actually saw viral results, and pertained corporate brands. Video with the highest amounts of views were obtained first, where after the dataset was increased in size by including video that saw less high - yet still viral - viewing rates. To give an indication, the highest view counts were in excess of 79 million views. Less high – yet still viral viewing rates were still well into the 20-50 million. The cut-off point of what could be considered viral results was anything in excess of 1.000.000 views. Although no official cut-off point exists regarding the virality of video, these ranges were chosen based on the viewing results of the viral videos returned in the search query. The number of shares was not directly available and not a decisive factor in the collection of data.
The other half of video’s (n=40) were purposefully chosen for not having received viral-like results; meaning they saw less than 1.000.000 views. Moreover, these videos were not obtained with the same data-collection method. Instead, the already collected dataset of (n=40) viral corporate promotional video’s was considered, and every corporate channel on which the
viral videos were posted was searched for promotional video’s that saw considerably less views.
The purpose was to obtain contrasting data; yet still from the same channel and corporate brand. This was done with the aim to eliminate potential moderating or mediating effects from seeding
22 and other non-content related factors on viewing and sharing rates. Another reason for this was to obtain clearly comparable data. An example of this would be a viral video by Volvo Trucks, which received in excess of 79 million views. Its non-viral counterpart was also from Volvo Trucks, and found on the same YouTube channel. This latter video received a mere 80.000.
A total of (n=33) corporate brands are represented in the dataset, which includes
Volkswagen, Coca-Cola and Dove, but also lesser known (yet legitimate) brands such as Dollar Shave Club, Turner Network Television (TNT) and the motion picture Carrie. Since the preferred
focus lies with highly established corporate organizations, multiple viral video’s from certain individual brands were included [i.e. Dove (n=7) TomTom (n=4) and Volkswagen (n=7)]. See Figure 2 for the full list of brands, including the number of video’s represented. Occasionally, this list looks to show only (n=1) video for a brand name represented (e.g. Carrie the movie). However, the actual corresponding (non-viral) brand; in this case TechnoMarine (n=1), is created by the same organization and seeded on the same channel. The analysis required the collection of video’s, which were systematically analyzed with help of a codebook (see Appendix A). This codebook was transferred into Qualtrics for more efficient data entry. The completed dataset was then exported and analyzed using SPSS Statistics 22 software.
23 Figure 2: Complete list of corporate brands and number of videos (n) represented
Operationalization
Firstly, basic stats were coded for every video; starting with the corporate brand, and whether this was stated in the title, video, or not. Information on the hosting platform (i.e.
YouTube in each and every case), posting date, video length, number of views and shares, and
2 1 2 1 2 2 1 1 4 3 1 3 7 4 3 1 1 3 1 2 2 5 1 1 1 1 4 1 1 5 7 4 2 80 0 10 20 30 40 50 60 70 80 90 Always Amtrak Blendtec BlendTec & Doritos Bodyform Cardstore Carrie the movie Cathay Pacific Coca Cola Coca-Cola Zero Degage Ministries Dollar Shave Club Dove Durex Evian Ford Ford Mustang KLM Love has no labels Nike Old Spice Pepsi Pugedon Qualcomm Recycling Samsung T-Mobile TechnoMarine Telenet TNT TomTom Volkswagen Volvo Trucks WestJet Total
24 whether the video was seeded on a broadcasting channel different or similar to the corporate actor represented in the video were all collected.
The number of shares was obtained through Sharescount (http://www.sharescount.com), whereby the number of shares for each video were a sum of the number of Facebook shares, Tweets and LinkedIn shares. At times the same video was posted on multiple YouTube channels, whereby only the video with the highest viewing rates was sampled. The number of views and shares of every video is essentially a factor of the length of time it has been online: as time allows the number of views and shares to accumulate. The timespan each video was online widely varied, with the most recent video posted on March, 2015, and the oldest going back to January, 2009. Each video’s viewing and sharing rates had to be standardized in order to be measured comparatively. Therefore the ‘daily viewing rates’ (DVR) and ‘daily sharing rates’ (DSR) were established by dividing the number of views and shares with the number of days online. This figure was then rounded to the nearest full number. The DVR and DSR could subsequently be correlated to data regarding types of framing, the framing of valance and framing of emotions.
Framing Types
The operationalization of framing types, was based on theory by Hallahan (1999) and Semetko & Valkenburg (2000). Hereby the following eight types of framing were included: the framing of situations, attributions, choices, actions, issues, responsibilities, news and morality. Operationalizing framing theory has at times been found difficult, but Hallahan (1999) served as the foundation for this operationalization. This was done by asking three dichotomous (yes or no) sub-questions (i.e. indicators) for every framing type; with the exception of the framing of
25 was analyzed by asking whether the story showed relationships between actors; actors within
everyday life situations; and / or every day life situations that are interupted by an unusual event.
For the full codebook – including exact definitions, please refer to the Appendix A. As morality was thought to play a potential role with regards to viral spreading of video, morality framing as explained by Semetko & Valkenburg (2000) was also added and operationalized in manner similar to the other framing types.
Framing Emotions
The exact discrete emotions were based on the emotions measured within viral content by Berger & Milkman (2012) and Botha & Reyneke (2013). These were: Joy/Happiness,
Intrigue/Interest, Awe/Excitement, Surprise / Shock, Anxiety / Fear, Anger, Sadness, Disgust and Pity. Defining these emotions in terms of framing was done based on a study by Verhoeven
(2008) which - amongst other elements - analyzed and measured framing and discrete emotions in medical films. The emotion of joy/happiness, for instance, was determined when the story showed one or more actors laughing and / or smiling. This smiling could have been accompanied by crying, as long as smiling or laughter occurred simultaneously. Joy/Happiness was also chosen when actors verbalized to be happy or feel joy, or when the narrative made this explicit. Similar method for analyzing the framing of emotion in content is also seen in Gross &
D’Ambrosio (2004). For the operationalization of all emotions, please see the Appendix A; Definitions.
Framing Valance
Besides measuring discrete; individual emotions, emotions were also measured on a continuum by measuring valance. Valance Framing was operationalized as inspired by Bizer & Petty (2005). Valance framing can be operationalized in many different ways, yet in such cases
26 often concern other manners of framing; including choices, actions, risks and attributions (Levin, Schneider & Gaeth, 1998). Since framing types believed relevant for viral video are already covered in other areas of the analysis, valence was analyzed in terms the degree of positity, negativity or neutrality. This was done through a 7-point likert-scale from ‘very positive’ to ‘very
negative.’ The frequency that positive or negative emotions were portrayed by the actors
determined the degree of valence. For instance, if only one actor portrayed joy by laughing out loud once, then the grade of ‘slightly positive’ was assigned. However if five actors portrayed joy, the video was deemed ‘very positive.’ This valance operationalization is very similar to, and inspired by Bizer & Petty (2005); whom published articles on valance on numerous occasions. Coding procedure
Each video was coded seperately, and in its entirity. Framing factors were coded dichotomously as either present within the video, or not. Video was pauzed when found necessary, but no predetermined intervals were made with this regard. For framing types and
framing of discrete emotions, no extra measurement was taken regarding the frequency at which
such framing factors were present, or not. With valance framing, however, the degree of
positivity, negativity (or neutrality) was taken into consideration. To examplify: the valence of a video was found ‘very positive’ if emotions such as joy/happiness and/or awe/excitement were shown three or more times. However if such emotion was shown only once, the video was found ‘slightly positive.’
Coder-reliability
In order to meet and overshoot the 10% quota for establishing inter-coder reliability through Krippendorff’s Alpha, 13 videos were coded by two separate coders. These coding results were analyzed with SPSS, using the procedure as described by De Swert (2012), and their referenced SPSS Macro. The coincidence interval was 95% and each run was bootstrapped 9999
27 times. Relevant to note is that the dependent variables: views and sharing rates were not suitable to test for ICR, as they are inherently subject to change. For these variables careful observation ensured that no unusual anomalies occurred. Overall, the KALPHA values were somewhat low – all the while percentage-based agreement rates were very high. Please refer to Appendix B for the full list of variables; their respective ICR KALPHA values and agreement rates. To
exemplify: a 92.3% agreement was the case, while the KALPHA value was (α=.60). Another example is an 84% agreement (11 out of 13 cases were in agreement), in both cases of Q22 and Q29. The first had (α=.62), while for the latter (α=.37). In one case α=negative, with again an 84% agreement rate. As became clear, these highly volatile and relatively low values are only slightly influenced by the overall (very good) agreement rate. Besides the small dataset, the real culprit is the rarity with which the dichotomous categories that make up the data occur; thereby arguably not being able to properly represent actual reliability (De Swert, 2012). With
dichotomous data like this, it is actually normal to have low KALPHA values (De Swert, 2012). To more accurately display the reliability, a percentage-based agreement/disagreement rate was included for every KALPHA value (see Appendix B). Considering the issues surrounding the ICR values due to the dichotomous data, the ICR for this codebook can overall be found satisfying; especially when considering the percentage-based agreement rates.
Results
We set out to answer the following two main research questions: 1) How are corporate videos – deliberated to go viral – framed in terms of types, emotions and valance, and how many views and shares did these video’s receive? and 2) How do framing types, the framing of emotions and valance framing relate to the viewing and sharing rates of these videos? In order
28 to facilitate the analysis and enhance explanatory clarity, these two main research questions were divided into sub-research questions, and answered below.
Corporate Video – Views and Shares
Sub-RQ #1: How many views and shares were received by corporate videos – deliberated to go viral?
A total of 80 corporate videos – seeded online were obtained and analyzed. Table 1 below displays every viral video coded; including the respective brands, number of views, number of shares, daily viewing rates (DVRs) and daily sharing rates (DSRs). Table 2 displays the same for non-viral videos.
Table 1: Overview of viral video – ranked by total number of views
Video Title Brand Views DVR* Shares DSR*
Evian Roller Babies international version
Evian 79,559,399 36,919 1009,353 468
Volvo Trucks - The Epic Split feat. Van Damme (Live Test 6)
Volvo Trucks 79,092,766 141,743 1024,213 1,836 Dove Real Beauty Sketches |
You’re more beautiful than you think (3mins)
Dove 65,652,768 85,153 32,824 43
Telekinetic Coffee Shop Surprise Carrie the movie 62,597,715 105,383 868,440 1,462
Always #LikeAGirl Always 57,563,339 172,345 407,243 1,219
A DRAMATIC SURPRISE ON A QUIET SQUARE
Telenet TNT 52,813,920 46,369 1608,374 1,412 Diversity & Inclusion – Love Has
No Labels
Love has no labels 52,282,263 622,408 179,765 2,140 Old Spice | The Man Your Man
Could Smell Like
Old Spice 50,777,006 26,228 498,435 257 Jeff Gordon: Test Drive | Pepsi
Max
Pepsi 44,416,363 55,313 924,984 1,152
WestJet Christmas Miracle: real-time giving
WestJet 41,023,426 76,967 715,725 1,343
29 Durex - #Connect - Official Durex 37,196,703 495,956 20,556 274 World's Toughest Job Cardstore 24,467,451 60,265 670,345 1,651 Piano stairs - TheFunTheory.com -
Rolighetsteorin.se
Volkswagen 21,551,732 10,477 324,269 158 Homeless Veteran Timelapse
Transformation
Degage Ministries 21,501,493 38,056 110,120 195 Dove Patches | Beauty is a state of
mind (4mins)
Dove 21,029,969 51,168 28,689 70
KLM Lost & Found service KLM 19,883,671 81,158 139,616 570 DollarShaveClub.com - Our
Blades Are F***ing Great
Dollar Shave Club 19,173,702 16,318 - -
Will It Blend? – iPad Blendtec 17,644,688 9,411 129,008 14
The T-Mobile Welcome Back T-Mobile 14,880,598 8,916 241,260 145 Speed Dating Prank | 2015 Ford
Mustang
Ford Mustang 12,392,663 113,694 41,320 380 Unlock the 007 in you. You have
70 seconds!
Coca-Cola 11,156,820 950 332,152 350
Hearing Hands - Touching Ad By Samsung
Samsung 10,319,161 125,843 23,616 278
Durex #TurnOffToTurnOn – OFFICIAL
Durex 9,225,345 21,062 36,668 84
Durex Fundawear -- Touch over the Internet [OFFICIAL]
Durex 8,025,205 10,436 61,249 80
Harlem Shake | Jeff Gordon Edition | Pepsi
Pepsi 7,170,497 8,629 24,393 29
Coca-Cola Happiness Machine Coca Cola 6,871,797 3,506 93,078 47 Dove Choose Beautiful | Women
all over the world make a choice
Dove 6,465,383 134,695 26,326 548
Alexandria Morgan runs strapless TomTom Runner Cardio
6,229,275 20,225 7,550 25 Darth Vader recording for
TomTom GPS - behind the scenes
TomTom 5,727,267 3,101 132,991 72
Bodyform Responds :: The Truth Bodyform 5,685,095 5,972 49,193 52 The world's deepest bin -
Thefuntheory.com - Rolighetsteorin.se
Volkswagen 4,600,924 2,237 41,669 20
Speed up your life - take the Volkswagen slide!
30 Coke Zero Presents: Zero Clue Coca-Cola Zero 3,432,446 757 3,286 4 Yoda recording for TomTom GPS
- behind the scenes
TomTom 3,355,704 1,882 65,968 37
Fast Lane - The Shopping Carts Volkswagen 1,194,718 660 18,734 10
TomTom (almost) makes a viral TomTom 1,167,533 1,289 2,578 3
Fast Lane - The Elevator Volkswagen 1,135,718 579 4,177 2
*Note: DSR (Daily Sharing Rate) and DVR (Daily Viewing Rate) values Table 2: Overview of non-viral video – ranked by total number of views
Video Title Brand Views DVR* Shares DSR*
Will it Blend? - Doritos Jacked BlendTec & Doritos
977,573 982 1,087 1
Coke Hug Machine Coca-Cola 884,512 796 4,614 4
Cathay Pacific Flashmob @ HKIA 2013
Cathay Pacific 801,215 1,506 12,998 24
T-Mobile Dance Rehearsal T-Mobile 712,368 307 153 0
Volvo Trucks - The extensive test program behind the new Volvo FE
Volvo Trucks 690,538 1,218 921 2
Underwater Nightclub NYC / TechnoMarine
TechnoMarine 646,058 598 8,256 8
Always #LikeAGirl – Stronger Together
Always 508,234 6,123 6,117 74
Peter Crouch does the T-Mobile Dance
T-Mobile 498,705 217 606 0
Jeff Gordon & Kasey Kahne: Go-Kart Challenge
Pepsi 491,784 1,490 3,671 11
KLM Surprise KLM 425,055 261 5,800 4
Uncle Drew: Where In The World Did The Big Man Go?
Pepsi 297,247 638 65 0
Baby | Pepsi | Commercial Pepsi 264,330 359 882 1
Dad Casting - World's Toughest Job Cardstore 233,243 650 2,758 8 Tranq Dart - TV Commercial | Dollar
Shave Club
Dollar Shave Club
219,523 1,109 828 4
Pay Up - TV Commercial | Dollar Shave Club
Dollar Shave Club
31
Evian Bottle Service Evian 163,724 620 206 1
NikeFuel London Nike 163,425 323 270 1
Recycling – Smile :30 Recycling 155,768 1,876 117 1
Coca-Cola Happy Beep Coca-Cola 143,724 368 1,431 4
Nike Sport Research: The Art of Science
Nike 114,229 438 771 3
Old Spice | Stairs Old Spice 77,047 342 35 0
A mother's body Dove 76,950 200 1,002 3
Volvo Trucks - Subscribe to our YouTube channel today!
Volvo Trucks 75,746 195 282 1
Volvo Trucks - Subscribe to our YouTube channel today!
Volvo Trucks 75,709 195 282 1
Coca-Cola Zero – Taste it to believe it Coca-Cola Zero 72,303 479 229 2
Durex - 50 Games to Play Durex 69,926 624 192 2
Why Blendtec? Blenders are our Passion!
Blendtec 63,562 169 66 0
evian Twitter followers ask Maria Sharapova...with Jonathan Ross
Evian 59,350 17 14 0
Dove Inner Critic | Are you your own worst critic?
Dove 55,799 109 227 0
Dog Sled Surprise Qualcomm 48,081 58 268 0
Mother's Day Dream Car Ford 37,430 98 718 2
A typical day at the airport. KLM 33,094 121 268 1
I wanna be a Rockstar Coca-Cola Zero 22,540 17 28 0
TomTom Traffic Manifesto TomTom 17,448 11 75 0
WestJet and Free The Children We Day experience
WestJet 14,578 38 78 0
Das Auto. Magazine - e-Golf – electrified
Volkswagen 9,599 21 32 0
32
Volkswagen Smiles Ad Volkswagen 6,484 6 129 0
This Vending Machine Gives Food To Stray Dogs
Pugedon 6,097 31 22 0
Times Have Changed with Bodyform Bodyform 2,407 9 23 0
*Note: DSR (Daily Sharing Rate) and DVR (Daily Viewing Rate) values
The videos were anywhere between 8 and 542 seconds long. (M = 125.96, SD = 89.61). Of the video’s which saw viral results, the highest number of views were in excess of 79,000,000 (M =23,533,401.12, SD =23,134,394.95) and 1,000,000 shares (M = 268,436.03, SD =
374,672.75). The other half of video’s did not see similar viral-like results, whereby the lowest number of views were 2407 (M = 235,276.52, SD = 275,268.01), and shares 23 (M = 1,394.25,
SD = 2,686.64).
The daily viewing rates (DVR) and daily sharing rates (DSR) are also included for every video. These rates relativize the timespan each video has been online, so that drawing
comparison between these rates becomes more practical and clear. Viral video saw DVR (M= 65,776.85, SD=124,540.09) and DSR (M= 434.72, SD =598.03) values. Non-viral video saw DVR values (M=589.45, SD=1013.84) and DSR values (M=4.10, SD=12.13). In order to give additional perspective to the matter, table 3 and table 4 show the top-10 video’s ranked by highest DSR and DVR values. When investigating these tables, it becomes apparent that
numerous times (n=7) same video’s appears in both tables. The appearances, however, are not in the same order. The video by Ad Council (USA) is ranked at first place in both tables, but for instance Volvo Trucks is ranked slightly lower on DSR’s than in DVR’s.
The three video’s that are present within the DVR-ranked table, but not in the DSR-ranked one represent the brands of Samsung, Ford Mustang and Durex. The video’s contained within the DSR-ranked table, but not the DVR-ranked one are Pepsi, Cardstore and Telenet
33
TNT. This could allude to the possibility that certain video’s might be watched more often (i.e.
multiple times) after being shared, but also that certain video might be easier to find than other video, or even pertains a stronger brand than the other. Investigating such possibilities goes beyond the scope of this research, however.
Table 3: Top 10 most viral video’s – ranked by DVR
Video Title Corporate Brand Daily Viewing Rate Views
Diversity & Inclusion – Love Has No Labels Ad Council (USA) 622,408 52,282,263
Durex - #Connect – Official Durex 495,956 37,196,703
Always #LikeAGirl Always 172,345 57,563,339
Volvo Trucks - The Epic Split feat. Van
Damme (Live Test 6) Volvo Trucks 141,743 79,092,766 Dove Choose Beautiful | Women all over the
world make a choice Dove 134,695 6,465,383
Hearing Hands - Touching Ad By Samsung Samsung 125,843 10,319,161
Speed Dating Prank | 2015 Ford Mustang Ford Mustang 113,694 12,392,663
Telekinetic Coffee Shop Surprise Carrie the movie 105,383 62,597,715 Dove Real Beauty Sketches | You’re more
beautiful than you think (3mins) Dove 85,153 65,652,768
KLM Lost & Found service KLM 81,158 19,883,671
Table 4: Top 10 ‘most viral’ video’s – ranked by DSR
Video Title Corporate Brand Daily Share Rate Shares
Diversity & Inclusion – Love Has No Labels Ad Council (USA) 2,140 179,765 Volvo Trucks - The Epic Split feat. Van
Damme (Live Test 6) Volvo Trucks 1,836 1,024,213 World's Toughest Job - #worldstoughestjob -
Official Video Cardstore 1,651 670,345
34
A DRAMATIC SURPRISE ON A QUIET
SQUARE Telenet TNT 1,412 1,608,374
WestJet Christmas Miracle: real-time giving WestJet 1,343 715,725
Always #LikeAGirl Always 1,219 407,243
Jeff Gordon: Test Drive | Pepsi Max Pepsi 1,152 924,984
KLM Lost & Found service KLM 570 139,616 Dove Choose Beautiful | Women all over the
world make a choice Dove 548 26,326
Framing Type Occurrences
Sub-RQ #2: How do framing types occur within corporate video deliberated to go viral?
The eight framing types (situations, attributions, choices, actions, issues, responsibilities,
news and morality) each had three indicators within the codebook, with the exception of the
framing of attributions; which saw two. Table 5 below displays all operationalized indicators of framing types, including the frequency by which these occurred. For instance, the framing of
situations was measured through the coding of the following three indicating questions: ‘Does the story show relationships between actors?’ (n=67), ‘Does the story show actors within everyday life situations?’ (n=58), and ‘Does the story show everyday life situations that are interrupted by unusual events or unexpected happenings?’ (n=35). Hereby N=80. These results
reveal that 67 out of 80 video’s showed relationships between actors; i.e. interaction between two or more people in a shared context. Hereby actors are shown within everyday life situations 58 out of 80 times. The story shows such everyday life situations being interrupted by unusual events unexpected happenings 35 out of 80 possible occurrences. Everyday life situations being the type of situations which are highly likely occurring at least once, somewhere in Earth at any
35 and every moment in time. An example being a woman (i.e. actor) ordering a coffee at a café. An example of an unexpected or unusual interruption, would be when this woman would see another woman use telekinetic powers to throw a man up onto the ceiling. The situation here exemplified was an actual situation framed within one of the viral videos, from the motion picture Carrie. The telekinetic powers which the woman in the video seemingly possessed, is also measured as the framing of attributions; through the indicating question ‘Does the story put certain emphasis
on specific characteristics of individuals?’ (n=45). For further definitions also see Appendix A
– Codebook.
The results reveal that the framing of situations occurred the most within corporate video – deliberated to go viral, since the two most frequently measured framing indicators were both indicators of the framing of situations; in 83.8% and 72.5% of all video’s respectively. The framing of news; i.e. the video story containing familiar contexts was also popular (n=54, 67.5%). The framing of attributions; which means certain emphasis was put of specific
characteristics of objects and individuals, took up fourth (n=48, 60%) and fifth (n=45, 56.3%) respectively. The video’s showed very few actors involved in conflict (n=4, 5%) and disputes (n=2, 2.5%); thereby the framing of issues saw two out of three framing type indicators in the bottom of the list; thereby combined the least occurring framing type. Ultimately, the bottom place was shared with one of three indicators for the framing of morality (n=2, 2.5%) Table 5: Framing Types Frequencies – Ranked by highest occurrence – Full Dataset
Framing Type Sub- Question Occurrence*
(n)
n of total* (%)
Situations Does the story show relationships between actors? 67 83.8
36
News Does the story contain familiar contexts? 54 67.5
Attributions Does the story put certain emphasis on specific characteristics of objects?
48 60
Attributions Does the story put certain emphasis on specific characteristics of individuals?
45 56.3
News Does the story contain culturally resonating themes? 37 46.3
Situations Does the story show everyday life situations that are interrupted by an unusual event or unexpected happenings?
35 43.8
Choices Does the story put emphasis on gains due to choices made by actors within the situational context?
33 41.3
Responsibility Does the story portray certain actors to be a causal factor in relation to an outcome?
30 37.5
Choices Does the story show a scenario concerning uncertain situations? 27 33.8
Actions Are certain actions portrayed by the story to have a more positive outcome than others?
27 33.8
Morality Does the story offer special social prescriptions about how to behave?
25 31.3
Morality Does the story contain any moral message? 21 26.3
Actions Does the story show actors that are presented alternative actions towards reaching certain goals?
17 21.3
Choices Does the story put emphasis on losses due to choices made by actors within the situational context?
14 17.5
Issues Does the story show actors involved in societal problems? 11 13.8
Responsibility Does the story portray certain actors to be a causal factor of any social problems, conflicts or...
10 12.5
Responsibility Does the story suggest a societal problem, conflict or dispute requires urgent action?
8 10
Actions Are certain actions portrayed by the story to have a more negative outcome than others?
7 8.8
News Does the story contain currently actual contexts? 7 8.8
Issues Does the story show actors involved in disputes with one another? 4 5
Issues Does the story show actors involved in conflict? 2 2.5
Morality Does the story make references to God, and / or other religious tenets?
2 2.5
*Note: N= 80
37 When comparing how framing types occur within viral vs non-viral video’s (as also displayed in table 6 below), we see only two cases where no difference occurs; namely when video’s show actors within everyday life situations (n=29) (framing of situations), and actors shown in conflict (n=0) (framing of issues). The framing of choices (i.e. when a scenario is shown which concerns uncertain situations), was measured to occurr in 55% of all viral video, while only in 12.5% of all non-viral video. Viral video sees 22.5% more culturally resonating themes (the framing of news), and 27.5% more everyday life situations which are interrupted by unusual events or unexpected happenings (framing of situations) than non-viral video.
Ultimately, it becomes clear that there are noticeable differences between viral and non-viral videos, when it comes to the frequency that a framing type occurs.
Table 6: Framing Types Frequencies – Non-viral vs Viral (Ranked by % difference) Framing Type Frame indicator Viral
(n) Viral n of total (%) Non-viral (n) Non-viral n of total (%) Differe nce n (% ) Choices A scenario concerning
uncertain situations
22 55 5 12.5 42.5
Situations Everyday life situations that are interrupted by an unusual event or unexpected happening
23 57.5 12 30 27.5
News Culturally resonating themes 23 57.5 14 35 22.5
Morality Story containing any moral message
15 37.5 6 15 22.5
Morality Special social prescriptions about how to behave
17 42.5 8 20 22.5
Attributions Certain emphasis on specific characteristics of individuals
26 65 19 47.5 17.5
Choices Emphasis on losses due to choices made by actors within the situational context
11 27.5 3 7.5 20
Responsibility Certain actors to be a causal factor in relation to an outcome
19 47.5 11 27.5 20
Responsibility Certain actors to be a causal factor of any social problems, conflicts or disputes
7 17.5 3 7.5 20
Actions Certain actions portrayed to have a more positive outcome than others
38 Choices Emphasis on gains due to
choices made by actors within the situational context?
18 45 15 37.5 7.5
Issues Actors involved in societal problems
7 17.5 4 10 7.5
Attributions Certain emphasis on specific characteristics of objects
22 55 26 65 10
News Story containing familiar contexts
29 72.5 25 62.5 10
Situations Relationships between actors 34 85 33 82.5 2.5
Actions Actors that are presented alternative actions towards reaching certain goals
9 22.5 8 20 2.5
Actions Certain actions portrayed to have a more negative outcome than others
6 15 1 2.5 2.5
News Currently actual contexts 4 10 3 7.5 2.5
Issues Actors involved in disputes with one another?
3 7.5 1 2.5 5
Responsibility Suggestion of societal problem, conflict or dispute requires urgent action
5 12.5 3 7.5 5
Morality References to God, and / or other religious tenets
2 5 0 0 5
Situations Actors within everyday life situations
29 72.5 29 72.5 0
Issues Actors involved in conflict? 2 5 0 0 0
Sub-RQ #3: How do framing types relate to viewing and / or sharing rates?
Correlation Framing Types - Viewing and Sharing Rates (Full Dataset)
First, bivariate correlation analyses were done between each framing type indicator variable - and the daily viewing rates (DVR) and daily sharing rates (DSR) variables. For the full lists of every framing type and its corresponding correlations with DVR and DSR, refer to Appendix C. Below in table 7 and table 8 you will find only the significant correlations between framing types and DVR / DSR rates respectively. Table 7 reveals that the framing of morality sees the strongest and most significant correlations with DVR. Each of its three indicators are found in the top three (r=0.51, p<.001), (r=0.41, p<.001), and (r=0.30, p=.006). These
39
situations also see moderately strong correlation, within slightly less significance. The framing
of attributions, news and situations occurred with rather high frequency (see table 5 above) compared to the framing of morality, so these correlation results are perhaps even more telling. Table 7: Significant bivariate correlations between framing types and DVR – Full Dataset
Framing Type Indicating Question r p
Morality
Does the story make references to God, and / or other
religious tenets? 0.51* <.001
Morality Does the story contain any moral message? 0.41* <.001 Morality
Does the story offer special social prescriptions about how to
behave? 0.30* .006
Attributions
Does the story put certain emphasis on specific
characteristics of individuals? 0.26* .022
News Does the story contain culturally resonating themes? 0.24* .032 Situations Does the story show actors within everyday life situations? -0.24* .033 *Note: (p<.05)(2-tail)
The bivariate correlations between framing types and DSR rates in Table 8 show similar moderately strong correlations, whereby the framing of issues seems to have the strongest correlation with sharing rates. When the video shows actors involved in conflict or disputes sharing behavior might be probable. Hereby r=0.42, p=<.001, and r=0.24, p=.037 respectively. Video’s which contain references to God, religious tenets (r=0.33, p=.003), moral messages (r=0.25, p=.026), and stories within video containing culturally resonating themes (r=0.29,
p=.011) all seem moderately strong predictors of sharing rates. Contrasting to viewing rates,
stories showing scenarios of uncertain situations (framing of choices) do see significant correlation (r=0.22, p=.048) with sharing rates.
40
Framing Type Indicating Question r p
Issues Does the story show actors involved in conflict? 0.42* <.001 Morality
Does the story make references to God, and / or other
religious tenets? 0.33* .003
News Does the story contain culturally resonating themes? 0.29* .011 Responsibility
Does the story portray certain actors to be a causal factor
of any social problems, conflicts or... 0.28* .013 Morality Does the story contain any moral message? 0.25* .026 Issues
Does the story show actors involved in disputes with one
another? 0.24* .037
Responsibility
Does the story suggest a societal problem, conflict or
dispute requires urgent action? 0.23* .041
Attributions
Does the story put certain emphasis on specific
characteristics of individuals? 0.22* .047
Choices
Does the story show a scenario concerning uncertain
situations? 0.22* .048
*Note: (p<.05) (2-tail)
Predictive Relation Framing Types – Viewing and Sharing Rates (Full Dataset)
Next, multiple regression analyses between the significantly correlating framing types, and DVR/DSR rates was done. Hereby potential predictive relationship between framing types and DVR/DSR can be revealed. For DVR (M = 33183.15, SD = 93451.88), R² = 0.372; which suggests that as a group, all framing types account for 37.2% of the variance in DVR. The overall regression model was significant, F (6, 73) = 7.22 (p<.001), R² = 0.372. This implies a moderately strong predictive relation of these framing types and DVR. Refer to table 9 below, for individual beta values for each individual framing type on DVR. These results reveal that only two of the morality framing indicators are significant predictors. Only when a story makes references to God, and / or other religious tenets (β= 0.27, p=.043), or when the story contains a moral message (β =0.39, p<.001) does it look to predict viewing rates (DVR).
