The Effects of Modality and Involvement on
Comprehension and Attitudes of Autoplay
Advertisements
Alexa Sanford Student ID: 10915826
Master’s Thesis
Graduate School of Communication Corporate Communication Supervisor: Suzanne de Bakker
Abstract
The present study examines how comprehension and involvement impact attitude formation towards video advertisements on social media platforms where the audio is automatically muted (autoplay). 127 participants via three conditions watched the same video advertisement either with audio, without audio or subtitled without audio. Results confirmed that comprehension of the muted video advertisement was far inferior to that of the sound video and subtitled muted video respectively. Furthermore, comprehension was shown to be a predictor of attitudes toward the advertisement, with stronger comprehension predicting more positive attitudes, however the effect was weak. Results of the experiment suggest that properly executed visuals are important for generating positive attitudes towards advertisements and alternatives to audio, such as subtitles, are important for strengthening ad comprehension in the absence of sound. Discussion and limitations ensue.
The Effects of Modality and Involvement on Comprehension and Attitudes of Autoplay Advertisements
In the US alone, over 65% of adults have social media accounts, and for young adults (19-29 years old) this number jumps to 90% (Perrin, 2015). With a multitude of regular internet users and numbers continually on the rise, social media have become an ideal platform for companies in all industries to draw attention to their products and services. As video watching is becoming one of the most popular activities on social media, companies such as Facebook, which counts 1.65 billion users worldwide (Frier, 2016), have experienced a recent boom in video advertising.
When and how online advertisements are viewed varies depending on the platform and user. With the autoplay function on platforms such as Twitter, Youtube, Instagram and
Facebook, to the benefit of companies, video advertisements automatically play as users scroll through their feed. However, so as not to disrupt the user, the video is also muted. If an
interesting element of the video catches the user’s eye, the user then has the option to activate the sound. However, users at work, using public transportation, or even those at social gatherings may indeed not activate the sound as not to disturb those around them. Today as much as 85% of Facebook videos are being viewed without sound (Patel, 2016), which is why some advertisers have added captions and subtitles to facilitate users’ viewing experience. This begs the question, how well are viewers understanding online video ads, both muted and subtitled? How does their comprehension then affect their attitudes towards the brand or product? It is therefore important for businesses and advertisers to understand the advantages and disadvantages of different
modalities (e.g. audio or visual) and how social media users process and store information as a result of these modalities.
Dijkstra et al. (2005, p.378), when discussing modality, said that “each sensory mode may potentially affect processing by directly evoking cognitive and affective reactions or indirectly influencing the processing of and reactions to other sensory modes.” The importance of comprehension in advertising therefore lies in its ability to effectively communicate a message to consumers. Product attributes (e.g. utility, aesthetics, functions etc) relevant to potential consumers are expressed in ads in order to lure in customers. An advertiser’s job is to then ensure that those attributes are understood so positive attitudes are generated and so consumers can make a decision about the product with the greatest amount of information possible.
Wyer and Shrum (2015) explain that from exposure an individual will take the verbal and visual parts (e.g. words and images) of a message in order to first understand the literal meaning before context and other factors are taken into consideration. This first stage of literal processing is followed by a second stage of “deliberation” wherein emotional reactions, or attitudes, are generated in accordance with the context and relevance for the viewer. Kim and Lennon (2008) found that when customers considered themselves more knowledgeable about a product, their attitudes towards the product were affected in a positive way. Therefore, it is important to understand how comprehension about a product and its uses, as communicated in an
advertisement, is able to affect attitudes. It is, after all, these attitudes that will potentially drive viewers to research or buy the product in question.
Many studies have examined differences in comprehension while comparing different modalities with results on the superiority of one modality varying significantly (Crigler & Neuman, 1994; Sundar, 2011; Edell & Keller, 1989; Pezdek & Stevens, 1984) and even less
have studied the comprehension of advertisements of different modalities (Sundar, 2011). However, the main focus of the majority of these studies was how the presence of visuals added or detracted from the comprehension of an audio message (e.g. radio versus TV) rather than how removing audio or substituting text for audio had impacted it. There therefore exists a gap in knowledge about advertisement comprehension and specifically comprehension of audiovisual ads where the audio is removed or replaced. Left unstudied, advertisers and researchers may never understand to what degree watching online autoplay ads differs from regular online or TV commercial viewing and how this new type of online advertising affects persuasion. Moreover, given the recent creation and adoption of autoplay by social media networks, no studies to date have explored the efficiency of this function and its potential effects. It is for this reason that very little is known about how audio and visual information from ads aids or impedes cognitive processing and how comprehension relates to the appreciation of a brand or product.
Furthermore, the removal of audio for videos on social media sites has stimulated a growth in subtitled and captioned video content. It is therefore important to understand the effects of subtitled video advertisements, where research is few and far between in the academic world. Overall, delving into the effects of autoplay video ads can bring greater insight into the inner workings of persuasion through modality and the effectiveness of alternatives to audio such as subtitles.
The present study will discuss how the modality of online video advertisements affects the comprehension of and attitudes towards these ads. The key to comprehension and persuasion is not found solely in the words or images of a given message, but in respect to the personal and situational relevance of the viewer to the content of the advertisement, otherwise known as involvement. The Elaboration Likelihood Model (Petty et al., 1983) illustrates that the
processing of stimuli may take one of two routes depending on how involved a viewer is with a certain product or brand. In the first route the more relevant cues are in the message, the more motivated and involved the viewer becomes and the more attention and deliberate elaboration is put into understanding the message and making a decision. In the second route, cues are
perceived as aesthetic and more superficial, which subsequently causes the viewer to put less energy into the analysis and comprehension of the message. Studies such as Macias (2003) found that comprehension of website content was strengthened by participants who were more
motivated and involved in the content of the website. Involvement could therefore be a key to generating stronger comprehension when information through audio is lost.
In this study I will first discuss how information from various modalities is processed cognitively and results in the storage and recall of information, or comprehension. Secondly, I will discuss how comprehension lays the foundation for the creation of attitudes towards a brand or product after the diffusion of a persuasive message. I will also argue that the level of
involvement, characterized by the amount of relevant schema in a message, is able to enhance or weaken the comprehension of the advertisement as well as generate more positive or negative attitudes toward the advertisement. I will then detail the results of the experiment that I conducted on autoplay video ad comprehension and attitudes, followed by a discussion of the results and the future of research in this domain.
Theoretical Framework
Comprehension and modality
Comprehension is defined as “the retrieval and integration of information about something of which one is aware and occurs in concrete and abstract contexts” (Shim &
Grabowski, 2010, p. 45). When information is presented to us we use one or a combination of our 5 senses: olfactory (smell), kinesthetic (touch), visual (sight), auditory (sound) and gustatory (taste) to create sensory data which we interpret or process in a visual or analogous way.
Processing in a visual way denotes creating images associated with the stimuli, while processing in an analogous way entails the use of language. The smell of apple pie might evoke the image (visual processing) of our grandmother, while a picture of the Louvre could stimulate a lexical response (analogous processing) such as the largest museum in the world. Therefore, our senses detect information which is then processed cognitively through our thoughts and stored in our memory. The world around us provides information and we use our senses combined with our concept of the world to analyze and store this information. While this process is systematic in its way of functioning, it is also largely unconscious. Comprehension is therefore characterized by our capacity to recall information that appealed to our senses and that we then interpret and process cognitively (Kök, 2014).
While our senses process the stimuli at hand, the modality is what determines how our senses process the information, which subsequently affects our comprehension. Modality refers to the medium used to communicate information to a receiver (Mohr & Nevin, 1990). At the most basic level modality, as used in the media, refers to information that is transmitted either visually, audibly or through a combination of the two (multimodal). Information is coded in the form of a sound or image, which activates our senses to process and store the information cognitively, either visually or semantically. Sounds and images as modalities are then capable of relaying information that is processed in different ways in the form of pictures, sound, spoken language or textual language. However, another level exists where modality is broken down into sub-modalities such as text and image, which are different ways of transmitting information
through a visual modality. “Each medium is not just an information source, but is also a complex information-conveying channel” ( Mohr & Nevin, 1990). The brain therefore encodes
information differently according to the modality through which information is transmitted. Being confronted by many stimuli from multiple modalities every day, our brain is able to dissect and register incoming information through our senses, which are linked to many cognitive processes in charge of comprehension and recall. The brain uses certain codes to store and retrieve information when needed (Paivio, 2010), which differ according to the modality and how that modality is processed in the brain. On a cognitive level the exact regions that are activated during comprehension will vary depending on the sensory modality of the information source. Information conveyed through an auditory modality is processed in the left frontal cortex while visual stimuli are processed in the right frontal cortex (Childers & Jiang, 2008). It is in this way that modality and comprehension are linked since the frontal regions of the brain have an important function in processing modal information as well as storing it in the memory for later recall (Buckner et al., 1999).
At the commencement of comprehension, information is received in its original modality: visual, auditory or both. Next, as that information is being processed or when it is needed for retrieval, the modality is able to change according to the needs of the situation and the learning strategies of the individual (Wyer et al., 2008). A photo may need to be described audibly or an image may be drawn of a story being told. Likewise, hearing a sentence read aloud (audio) can generate a visual image, stimulating cerebral processes for visual as well as audio processing (Wyer et al., 2008). One study, comparing written word versus spoken word tasks concluded that text comprehension produced stimulation in the left and right inferior frontal gyrus while spoken word comprehension only activated the left inferior frontal gyrus located in the frontal cortex.
This suggests that semantic content from a visual modality can stimulate information processing in more than one region of the brain simultaneously.
Overall, the cognitive workings of the brain lead me to assume that as the brain processes different modalities in different ways, certain modalities will have a different effect on
comprehension. Moreover, as video advertisements on social media platforms can be viewed with and without audio, it is important to understand how well or poorly these ads are being understood by their target audience in order to see what impact comprehension has on brands and products. When it comes to multimodal comprehension, some research suggests that “recognizing the meaning of information in one mode, may facilitate the interpretation of meaning in another (Edell & Keller, 1989, p. 150). In the next section I will elaborate on what past studies have concluded about comprehension through different modalities.
Comprehension of visual, audio and audiovisual stimuli
In the past several decades many studies have looked at the effects of modality on comprehension of stimuli in order to establish what effects these modalities have on memory, recall and overall comprehension. Visual modality content, as far as media is concerned, refers to images, both moving and still, such as video footage, printed text and photography. Studies have shown that images, when associated with semantic content (e.g. text), can reinforce memory and recall (Wyer et al., 2008, Garnham, 1981). Radio, music and audible language are examples of auditory modalities characterized by their reliance on audition. Visual and auditory modalities presented together can result in audiovisual stimuli such as television and even subtitled film. Multimedia studies have compared the recall and comprehension of both visual and auditory modalities separately and together in televised news (Crigler & Neuman, 1994; Pezdek &
Stevens, 1984), in combination with subtitled films (Hayati & Mohmedi, 2011; Hinkin, Harris, & Miranda, 2014) and in online news (Sundar, 2011; Gunter, 1985). However, conclusions on the most effective modality differ greatly depending on the content, quality and sequence of the stimuli transmitted.
For instance, extensively studying cognitive experiments on mental lexicon processing, Paivio’s (2010) theory of dual-coding (DCT) postulates that stimuli presented in two modalities (visual and audio) activate both the imagery and verbal systems, and when presented together in a relevant way these modalities should lead to enhanced information processing and therefore better recall (Perez et al., 2013). This leads to the first of three major conclusions about comprehension and modality, which has been corroborated by many studies (Dijkstra et al., 2005; Edell & Keller, 1989; Drew & Grimes, 1987; Young & Belleza, 1982; Basel & Gips, 2014): the combination of multiple modalities, specifically audio and visual, are more effective than visual alone due to the amount of information available to the audience.
In contrast, studies on child and adult comprehension of televised news have established that the audio of the segment was the overall most important factor leading to good
comprehension and that the comprehension of the audiovisual condition did not improve but was solely deemed more salient (Pezdek, 1984; Crigler et al. 1994). This can be explained by the content and creation of news segments, where audio information, more than visual information, is often made to be the dominant information carrying modality and images are often added solely for entertainment purposes (Drew & Grimes, 1987). The second conclusion therefore states that media content where the audio and visual stimuli are not complementary may result in one modality that is more essential to comprehension than the other. Along the same lines, when modal information, or information transmitted through one or more modalities, is not congruent
this leads to confusion on the part of the viewer and weakens comprehension (Edell & Keller, 1989). For example, if a video segment includes narration about a war while showing beautiful images of a lush countryside, this might contradict the narration and distract the viewer from the narration about war. Research has determined that memory can be strengthened when a certain degree of redundancy exists between modal information (e.g. visual and verbal) (Kim & Lennon, 2008). Sundar (2000) explains that this is because “repetition of content, in different forms, contributes to cognitive rehearsal, thereby enhancing its likelihood of storage in memory” (p. 482)
Comprehension of modal content has also been shown to improve or decrease depending on the type and sequence of images conveyed. In videos where narrative images were used, participants in the visual-only condition comprehended just as much as those in the audio-only condition. The third conclusion explains that recall and learning of visual stimuli are
strengthened when images are presented in a story format with a logical sequence of events that participants could follow (Crigler et al., 1994).
In summation, the three aforementioned conclusions demonstrate that the combination of modalities (e.g. audiovisual), the content of the stimuli and its presentation can therefore have an impact on comprehension, leading to various outcomes depending on the execution. Given the persuasive nature and shorter length of advertisements, does modal information processing and comprehension differ for advertisements?
TV advertising, combining both audio and video to present information, may provide more information to viewers, but some postulate that content from more than one modality requires more cognitive effort as there are more stimuli to process (Edell & Keller, 1989). The
question remains whether viewers obtain more information from audiovisual advertisements, such as videos, than unimodal advertisements where only visual or only audible modalities diffuse information?
Some research suggests that perceiving the same information in different ways can strengthen the memorability of the stimuli and that one modality can facilitate the
comprehension of another (Young & Bellezza, 1982) thus increasing comprehension. One experiment (Sundar, 2011) concluded that audio advertisements generated more item recognition than text and picture advertisements separately and that audiovisual advertisements stimulated the most item recognition compared to all other modalities. Similarly, while directly comparing radio and TV advertisements, it was established that viewers of TV advertisements had higher recall for brands and brand objectives than those who listened to the radio advertisement (Edell & Keller, 1989). The presence of multiple modalities therefore did not penalize comprehension, but strengthened it just as Paivio’s (2010) theory suggests.
These results are in line with the first major conclusion above, which states that multiple modalities reinforce comprehension due to information coming from multiple sources to multiple senses. The combination of content that addresses both the eyes and the ears therefore guarantees a greater chance of message comprehension and recall. Further corroborating these findings, Dijkstra et al. (2005)) established that TV advertisements were more effective communicators compared to text and pictorial internet advertisements. This confirms previous studies (Bulkin & Groh, 2006); Paivio, 2010) which concluded that a combination of both modalities presented together are complementary and can provide more accurate information to receivers through reinforced modal processing. An example of reinforced modal processing is when a narrator of a
TV show discusses a topic, which is also portrayed visually through images. The brain can therefore use the visual as well as audible stimuli to process and understand what is happening.
However, these results differ from studies on selective attention that have suggested that multiple modalities can interfere with one another (Edell & Keller, 1989), especially when audio and visual stimuli are not congruent (Bither & Wright, 1979). As the second conclusion on complementarity and congruency maintains, this is most likely due to the fact that receivers who perceive conflicting modal information (e.g. incongruent image and narration) will result in less effective information processing and will struggle to arrive at a coherent conclusion. Overall, I can postulate that audiovisual ads where modalities are not conflicting should lead to stronger comprehension than unimodal ads where information is conveyed through one modality. Therefore,
H1a Online audiovisual video advertisements generate stronger comprehension than muted video advertisements.
Comprehension through subtitling
Audiovisual modalities are not restricted to audio and images, but may also contain the presence of sub-modalities such as text, which can replace audio information transmission with visual text. Subtitling or captioning refers to transcribed language in text format that appears at the bottom of the screen during an audiovisual program and is either the transcribed or translated script of the audio portion of a program (Gernsbacher, 2015). Subtitling can be an important function for auto-play advertisements on social media sites where the sound is not activated, as it could indeed compensate for the missing audio.
Certain studies on the comprehension of subtitled films point to flaws in comprehension due to location of text and time allotted to read the text (Dias-Cintas & Remael, 2007). However, others maintain that subtitles strengthen comprehension (Hinkin et al., 2014; Hayati & Mohmeti, 2009; Brasel & Gips, 2014), aid in making inferences (Gernsbacher, 2015) and aid recall of specific information (Bird & Williams, 2002; Gernsbacher, 2015). Mayer’s (2001) multimedia theory was born from Paivio’s DCT (2010) and states that comprehension through information processing is facilitated by the presence of visuals (e.g. images) and text rather than solely by visuals (Perez et al., 2013). Therefore, I can make the assumption that multimodal media containing subtitles lead to better comprehension than unimodal media and that subtitling alone is more effective for comprehension than visuals alone.
This assumption was confirmed in a study examining film comprehension, which concluded that participants subjected to audiovisual redundancy (English audio with English subtitles) had the strongest comprehension and performed better than the audio-only and subtitle-only (no audio) conditions (Hinkin et al., 2014). In this way, subtitles can be used to ensure or reinforce comprehension even when audible dialogue is present since individuals are not likely to understand the entirety of a program or ad. This conclusion further supports the first major conclusion where information through multiple modalities reinforces comprehension.
Most studies on subtitling largely focus on foreign language acquisition and
comprehension. While my research does not focus on language acquisition, such studies provide further evidence of the benefits of subtitling for comprehension. One such study consisted of Persian mother tongue participants who understood a film with English audio and English subtitles better than English audio-only or even English audio with subtitles in their mother tongue (Hayati & Mohmeti, 2009). These studies suggest that even the presence of 3 modalities
did not overwhelm viewers and that visual text, specifically, is able to enhance comprehension. Furthermore, congruent modal redundancy, as stated in conclusion two, is important for
reinforcing comprehension as illustrated in the above study where English audio and English subtitles were found more effective for comprehension than English audio and mother tongue subtitles (Hayati & Mohmeti, 2009).
Previous studies have concluded that viewers read subtitles rather automatically, regardless of the relevance or salience of the featured content (D’Ydewalle et al., 1987), the same being true for advertisements (Brasel & Gips, 2014). “Indeed, both reading and visual processing are efficient and partly automatized activities requiring little conscious effort” (Brasel & Gips, 2014, p. 323). As our eyes are naturally drawn to subtitles (even in our mother tongue) their presence may also direct attention to an ad that may be otherwise ignored. Moreover, as individuals process subtitles through “verbal working memory” rather than visually via images, they are able to comprehend audio, visual and text simultaneously without overload (Brasel & Gips, 2014, p. 325). Brasel and Gips (2014) demonstrated that adding subtitles to video lead individuals to focus on the ad as well as read the subtitles, leading to stronger comprehension and enhanced memory of the brand. Such a combination has been proven to strengthen recall and recognition for over the counter drug commercials where risks were presented both through audio and subtitles in comparison to audio without subtitles (Wogalter et al., 2014).
Similarly, an extensive study by Brasel and Gips (2014) examined recall and brand attitude of three different subtitled versus non-subtitled advertisements. Their results confirmed that participants having viewed any of the three subtitled ads produced between 13% and 52% higher brand recall than the non-subtitled ads along with higher brand attitudes. Interestingly, brand recall was higher for ads where subtitles only appeared for pertinent information in the
middle of the ad and not for the entire duration of the ad. This could be due to a pop-out effect where the sudden appearance of subtitles increases attention and thus recall. Moreover, subtitled ad viewers also exhibited higher recall of verbal ad information except in the case of highly complex ads. These results suggest that subtitles, as viewers instinctively read them, can play an important role in boosting comprehension of video ads and that complex ads with subtitles can prove to be overwhelming for viewers. This being said, if online advertisements are not being viewed with audio, subtitles remain an important way to transmit information that would otherwise be lost. Taking into account the above research, DCT (Paivio, 2010) and Mayer’s (2001) theories on multimedia modalities I hypothesize that
H1b Online audiovisual video advertisements will generate stronger comprehension than muted video advertisements and muted subtitled advertisements.
However,
H1c Muted subtitled advertisements will generate stronger comprehension than muted advertisements.
In summary, modality plays an important role in ad comprehension as it has the capacity to stimulate information processing and strengthen or weaken memory depending on its
execution. As stated in the first major conclusion, multiple modalities are able to reinforce comprehension by providing more information to the viewer, much like puzzles pieces aid in completing a puzzle. However, unless the information provided by each modality is
complementary and coherent, comprehension will be weakened. Furthermore, when one
the missing modality. In the next section I will discuss how ad comprehension acts as a stepping stone to attitude formation.
Comprehension and attitude formation
The overarching goal of advertisement exposure is to stimulate attitude formation that further results in a form of consumer behavior by the viewer. Attitude formation is the result of message persuasion whereby salient and non-salient information are deemed relevant and then evaluated and stored by the individual (MacInnis & Jaworski, 1989). Understanding how exposure affects the attitudes of viewers and listeners is therefore essential to ad creation and diffusion.
McGuire (1968) maintained that in order for a message to persuade an individual, persuasion is required to pass through certain stages, beginning with attention to the ad and continuing on to comprehension of the message. According to him, greater focus on the message leads to greater understanding, which is exemplified through recall. Comprehension, therefore, becomes the necessary link between exposure and attitude formation.
While comprehension may seem a logical step to persuasiveness, all studies do not concur on the relationship between comprehension and attitude formation (Cacioppo & Petty 1983; Petty & Cacioppo, 1979; Ratneshwar & Chaiken, 1991). This can be due to many factors such as deliberate, short and easy to understand stimuli, unreliable comprehension/memory scales and the fact that during lab experiments participants are aware of their task and therefore make an effort to concentrate on and remember the stimuli they are exposed to-- reducing external validity. Unlike the present study, previously cited studies all presented complete stimuli, such that participants had all the necessary modalities to understand the message in
question, thus providing for stronger comprehension. With the removal of an essential modality, I assume that comprehension will be affected to such a degree that it will affect attitudes. Other studies have indeed found a significant comprehension-attitude relationship (Stewart & Koslow, 1989; Stewart & Furse, 1985; MacInnis & Jaworski, 1989; Wyer & Shrum, 2015; Eagly & Warren, 1976; Chattopadhyay & Alba, 1988).
The path from exposure to attitude formation was exhibited by Greenwald and Leavitt’s (1984) and MacInnis and Jaworski’s (1989) frameworks on ad exposure where comprehension (syntactic analysis and cognitive processing respectively) must be achieved before attitudinal effects are manifested. Comprehension as a mediator was also confirmed by Eagly and Warren (1974) who established that higher intelligence led to stronger comprehension of an argument and therefore a change in opinion about the subject when two arguments were presented
together. Participants that better understood the arguments were therefore more likely to form an opinion. Further studies have shown that when participants were either distracted from a
message by aesthetic details or the message was too complex, participants’ comprehension suffered as did positive attitude formation (Stewart & Koslow, 1989; Carpenter & Boster, 2013; Ratneshwar & Chaiken, 1991). I can therefore postulate that comprehension of a video
advertisement will be a precursor to attitude formation. If individuals are not able to receive and adequately understand all parts of a persuasive message due to a missing channel of information (e.g. no audio) then weaker comprehension should inhibit attitude formation.
H2 Stronger ad comprehension will generate more positive attitudes toward the ad while poor ad comprehension will generate less positive attitudes.
Therefore,
While modality plays an important role in information processing and comprehension, other factors such as involvement can affect how viewers comprehend advertisements and what information is stored and recalled. The role of involvement can help explain the divergent results on modality and comprehension discussed above. Furthermore, simply because a message is understood, does not guarantee that an individual will form a positive or negative opinion about it. What motivates an individual to form an opinion is involvement, which possesses a carry-over effect, affecting both comprehension during exposure and attitudes towards a message after initial comprehension. The next sections will discuss what involvement entails and how it impacts comprehension and attitude formation respectively.
The Role of Involvement in Comprehension: Elaboration Likelihood Model
The Elaboration Likelihood Model (ELM; Petty et al., 1983) explains that there are two different routes of processing messages that lead to persuasion. The first route, the peripheral route, denotes a path to persuasion characterized by less attention to the stimulus and a low level of information processing based on heuristics of the environment or context. The second, the central route, cues deeper information processing as individuals are motivated to seek
information and examine it in order to arrive at a meaningful conclusion. What distinguishes the peripheral route from the central route is the level of elaboration. Elaboration or elaborative processing refers to previously stored information that is associated with new information (Huhmann & Stenerson, 2008). It is therefore elaborative processing that denotes involvement.
The factor that determines the level of information processing during exposure is known as felt involvement. Felt involvement entails the presence of personal or situational relevance or
motivation where the content of the advertisement aids in the attainment of personal goals, interests or values (Celsi & Olson, 1988; Behe et al., 2015). Moreover, “the personal relevance of a product is represented by the perceived linkage between an individual's needs, goals, and values (self-knowledge) and their product knowledge (attributes and benefits)” (Celsi & Olson, 1988, p. 211). Level of involvement is therefore contingent upon situational factors such as a 50% off coupon, which may be relevant to someone on a budget, or intrinsic factors that are more stable such as a wine connoisseur’s affinity for purchasing wine.
In the context of ELM, when involvement is stimulated through personal relevance the central route cues deeper information processing as individuals are motivated to seek information and pay attention to persuasive stimuli. When attention on an ad increases, more effort is put into cognitive processing and memory storage of the ad, which in return has an impact on
comprehension by reinforcing recall (MacInnis & Jaworski, 1989). The peripheral route, on the other hand, relies on contextual cues such as the visual appeal of a product. Being on a more surface level due to less relevance, this route stimulates less attention and memory storage of stimuli and results in more superficial information processing (Han, 1992). Brasel and Gips (2014) came to the same conclusion with advertisements, where ads with little personal relevance to viewers were shown to cause low default processing, whereby the brain made a reduced effort to process certain stimuli.
A number of studies have examined the influence of involvement and established a significant effect in regards to information processing (Park et al., 2007; Matthes et al., 2013; Han, 1992; Behe et al., 2015). Celsi and Olson (1988), while looking at the role of involvement through situational and intrinsic relevance to tennis products, found that participants that felt more involved in tennis devoted more attention to the ads and as involvement increased so did
their focus on product-relevant information. Furthermore, they established that more involved participants also put more effort into comprehending the ads, measured by the number of free responses generated. Similarly, in a study about interactive website experience, participants that were more involved exhibited better comprehension and memory of the site (Macias, 2013). The moderating role of involvement was also confirmed by Han (1992) where higher involvement resulted in more focus on the ad as well as better recall of the ad. These studies attest not only to the capacity of involvement to gain and keep attention on stimuli, but its role in stimulating higher information processing which resulted in better recall and in more free responses.
In conclusion, the level of elaboration or involvement of an individual with a product or brand can affect how that individual processes a persuasive message about that product. When individuals deem content more important, they put more energy into focusing on and
comprehending stimuli and therefore generate deeper meaning about such stimuli, which further acts as a memory reinforcer. I propose that the level of involvement of an individual with a product or brand in an online video advertisement will alter their degree of comprehension and recall of that advertisement.
H4 Involvement will act as a moderator between ad exposure and ad comprehension such that more involved participants will generate stronger comprehension of the ad and lesser involved participants will generate weaker comprehension.
Involvement and attitude formation
As stated above, the personal and situational relevance of a product or brand of an individual can strengthen comprehension through enhanced attention and memory, however, the role of involvement does not stop there. When relevance motivates an individual to pay attention
and comprehend an advertisement, that relevance is then taken into account when evaluating the message and making a decision. Once the central route is activated, the levels of higher
processing that are activated continue through to affect attitude formation.
MacInnis and Jaworski’s (1989) extensive framework on attitude formation of
advertisements details how motivation (or involvement) moderates the relationships between exposure to an ad, information processing and attitude formation. When an individual perceives that the content is relevant to his or her life it causes his or her attention to intensify and
systematic processing (elaboration) to deepen to allow for the evaluation of information in comparison to existing beliefs and argument quality. The final step in this process is attitude formation. Attitudes generated from central route processing are more enduring and more likely to cause positive consumer-related behavior (Greenwald & Leavitt, 1984). It is for this reason that studies have shown that viewers gave ads a positive assessment when they found them personally relevant (MacInnis & Jaworski, 1989; Lalwani et al., 2009).
The peripheral route on the other hand, being characterized by low motivation to process and less attention, stimulates a response to superficial elements such as aesthetics of ad
execution, music and humor, which in turn can lead to a momentary attitude change or no change at all. Higher cognitive processing about the message or argument is therefore not activated. While a temporary attitude change may arise, in the long run low involvement has been shown to inhibit further knowledge seeking about a product and lowers purchase intention (Cho, 1999). These effects were exemplified by Cho (1999) who found that low-involved participants were more likely to click on online banners due to size and animation, whereas participants with positive attitudes toward the site were more likely to click on banners containing relevant information. Overall, banners with relevant information were the most clicked on. Therefore, by
first processing and comprehending the content of the message, higher involved individuals, in contrast to lesser involved individuals, will generate attitudes towards the message based on content, which will ultimately alter their behavior. Taking into consideration the role of involvement as proposed by ELM I hypothesize that
H5 Involvement will moderate the relationship between ad comprehension and attitude formation such that more involved participants will generate more positive attitudes about the ad and lesser involved participants will generate less positive attitudes.
To conclude, it is the same involvement in a message that is able to strengthen or weaken comprehension that is also able to alter attitudes towards that message. When individuals find a message relevant to their own needs and experiences, careful deliberation and processing are activated in order to form an opinion of the message. A lack of relevancy, on the other hand, leads to a fixation on superficial details that in the long run does not lead to message or product interest.
In the scope of this study, I argue that involvement is not only the key to better
comprehension, but to attitude formation towards online video advertisements. Today’s online platforms such as Facebook and Twitter have changed online ad viewing by activating auto-play and deactivating the sound. When individuals are faced with an online ad that is muted, and therefore missing a modality that conveys information as well as its attention-grabbing quality, the content of relevance in the ad is what will draw attention to the ad, strengthen comprehension of the ad, which will further stimulate higher cognitive processing where evaluation of the content is able to generate a product/brand attitude. However, if involvement during exposure to a muted ad is low, the effects could be detrimental for the comprehension of the ad, inhibiting
the formation of brand and product attitudes. It is for this reason that understanding how comprehension is affected by modality as well as involvement in advertisements is vital information for brands and advertisers as they design and diffuse ads.
Method
Design and Procedure
The experiment was a between-subjects design with three levels for the independent variable. The experiment was administered online through Qualtrics survey software. Participants took part in one of three conditions, all diffusing the same video ad. The first condition watched the regular audiovisual ad (N = 54) without any adjustments. The second condition watched the video ad without sound (N = 41). The third condition watched the video ad with subtitles on the bottom of the screen and without sound (N = 32). Participants could choose to take the experimental survey on a computer, tablet or mobile phone as social media platforms are often accessed from those devices. The survey contained demographic questions, four sets of 8 questions about involvement, the video ad diffused via Youtube, 6 multiple choice questions assessing comprehension and a set of 6 questions about attitudes towards the ad. The end of the survey included a multiple choice manipulation check to confirm that participants watched the correctly manipulated video. The last question asked whether participants had already seen the video in order to filter participants that have an advantage in understanding the content of the ad.
Stimulus material
The American video advertisement, Countertop, was used as the stimulus and independent variable and was retrieved from a website diffusing video ads from startup
companies. A video was chosen from a start-up website as it reduces the likelihood that participants have already seen the ad or know the brand and therefore have a preconceived attitude toward the product and a prior understanding of the content of the ad. The video advertises for an innovative health and cooking apparatus that connects to mobile phones, cooking appliances and fitness trackers. The product helps individuals eat healthily according to their needs, offers recipes and shows individuals how to prepare meals with appliances in their kitchens.
It is important to keep in mind that the ad differs from typical ads one would be exposed to while watching Youtube. As this study focuses on sites using autoplay, platforms such as Facebook do not require users to watch ads before accessing other content, as is common on Youtube. Therefore it is important that the ad possessed enough appeal so as to entice users to watch it of their own accord. The specific ad was also chosen for its length (1 minute 37 seconds)—comparable to the length of videos on social media platforms, but longer than video ads on Youtube. Lastly, it was chosen for its product, which could be used by both adult men and women of almost any age.
Participants
127 participants voluntarily took part in the online experimental survey, of which 37 males and 90 females between the ages of 18 and 80 with an average age of 33. They were from 23 countries (4 continents). 49.6% of participants were from the U.S. (N = 63), 13.4% from the Netherlands (N = 17), 7.1% from Canada (N = 9) and 6.3% from Germany (N = 8). Participants were recruited via Facebook, Twitter and e-mail and were therefore not representative of the
entire population. All participants were randomly assigned to one of three conditions, with 32 people in the subtitled condition, 54 people in the sound condition and 41 in the muted condition.
Dependent measures
Comprehension
Comprehension was assessed through 6 cued-recall multiple choice questions that were designed according to the product description and its uses as shown in the video advertisement. This form of measurement is similar to scales used by Kok (2013), Macias (2003) and Eagly and Warren (1974). An example of a question asked is “What is the name of the product in the advertisement?” 5 of the 6 questions proposed 5 possible answers including “I don’t know.” The 6th was a true or false question (“Countertop learns as it goes.”). The total score right out of 6 was attributed to each participant where no correct answers equals a score of 0 and all correct answers equals a score of 6. The overall mean comprehension score was 4.96 (SD = 1.53) with scores ranging between 0 and 6 (N = 127).
Attitude toward the advertisement
Attitude toward the advertisement was measured with one scale by De Pelsmacker, Geuens and Anckaert (2002). The scale included 6 statements (e.g. “I got a positive impression” and “I found it really credible”) that participants responded to on a 7 point likert scale (strongly disagree to strongly agree). The fifth item (“I found the advertisement exaggerated”) was reverse recoded. A principal component factor analysis with a varimax rotation showed these 6 items form two components, with all items but one (“I found the advertisement exaggerated”) loading on one component. This item was therefore deleted and the remaining 5 items all loaded on one
component with an eigenvalue above 1 (eigenvalue = 3.47) and the point of inflexion after this component is clearly visible on the scree plot. All items positively correlate to this component. The variable “I found the advertisement interesting” had the strongest correlation (factor loading is .87). The Cronbach’s alpha of .92 indicates that the scale is reliable. Furthermore, it appears the scale adequately measures attitudes towards the advertisement with low scores indicating less favorable attitudes.
Involvement
Involvement was measured through two separate scales from the Personal Involvement Inventory (Zaichkowsky, 1985), which originally contained 20 items. Ofir (2004), shortened that scale to 8 items, which is the version used in this survey. The first scale assessed to what degree living a healthy lifestyle was relevant to participants. Participants responded to 8 statements on a 7 point likert scale with 7 points between the two adjectives (e.g. “For me, living a healthy lifestyle is not important to me --- important to me” and “worthless--- valuable”). The second scale assessed to what degree cooking was relevant to participants. This second scale used the same 8 items on a 7 point likert scale (e.g. “For me cooking is uninteresting --- interesting). Healthy lifestyle and cooking were chosen as those are the main benefits and functions of the product that can apply to individuals’ lives.
A principal component factor analysis determined that the 8 items for the “ involvement in living a healthy lifestyle” scale do indeed load positively on one factor with an eigenvalue above 1 (eigenvalue = 5.7) with the point of inflexion being after this component on the scree plot. The item “Living a healthy lifestyle is important to me” had the highest factor loading (.92).
With a Cronbach’s alpha of .93 the scale is therefore reliable, indicating that lower scores equal lower involvement or relevance for living a healthy lifestyle.
A principal component factor analysis was also run for the second involvement scale (“Involvement in cooking”) and indicated that the 8 items load on two components. 2 items (“cooking is essential” and “cooking is needed”) load on a second component and were therefore deleted. A second principal component factor analysis determined that the remaining 6 items load positively on one component. Furthermore the eigenvalue is above 1 (eigenvalue = 4.28) with the inflexion on the scree plot after this component. The item that loads the highest
(“cooking is important to me”) has a factor loading of .92. With a Chronbach’s Alpha of .92 the scale is deemed reliable with lower scores indicating lower involvement in cooking.
Manipulation check
In order to analyze whether participants correctly experienced the manipulation a Chi square analysis was performed with conditions as the independent variable and the multiple choice manipulation check question “Which video did you watch”. Results indicate that participants were correctly exposed to the manipulation x² (4, 122) = 207.10, p = .000.
Randomization check
In order to assess if participants were randomly assigned to the three experimental conditions, a series of univariate ANOVAs was carried out with the experimental conditions as the independent variable and age, healthy lifestyle involvement, cooking involvement and
attitude as the separate dependent variables. Results indicate that there was no main effect for the conditions on attitude toward the advertisement F(2, 127) = 0.83, p = .438, health involvement
F(2, 127) = 0.72, p = .489 and cooking involvement F(2, 127) = 1.36, p = .261. However the
univariate Anova with age as the dependent variable indicated that there was a significant difference between conditions F(2, 126) = 4.12, p = .019. Therefore, age will be controlled for.
To further check for successful randomization a chi square test with the experimental conditions and gender was performed. No significant differences in gender exist between conditions x²(2, N = 127) = 0.265, p = .876, therefore gender will not be controlled for.
Analysis and interpretation of data
In order to test hypotheses one through five I used IBM’s SPSS statistics 20 software. Frequencies and descriptives were used to determine the sex, average age, nationality and number of participants per condition. Hypotheses 1a, b and c were tested with a univariate ANOVA to determine if comprehension between conditions differed. To assess whether higher comprehension scores corresponded with more positive attitudes (H2) and vice versa I employed a linear regression analysis. SPSS’ PROCESS tool was used to determine if comprehension mediates the relationship between video exposure and attitude formation (H3). The first moderation analysis (H4) with involvement as a moderator between exposure and
comprehension was analyzed with SPSS’s PROCESS tool, as was the second moderation analysis with involvement as a moderator between comprehension and attitude formation.
Results
Hypotheses 1a, b and c
To measure whether audiovisual advertisements with sound generate stronger comprehension than muted and subtitled muted advertisements and whether subtitled muted
advertisements generate stronger comprehension than muted advertisements, a univariate ANOVA was performed. Experimental conditions were used as the independent variable, comprehension score (average score of the number of correct answers) was used as the
dependent variable and age as a control variable. Participants that watched the muted video had the weakest comprehension scores (M = 3.61, SD = 1.46). Those that watched the video with sound had a near perfect comprehension score (M = 5.72, SD = 0.66), which was slightly higher than those that watched the muted subtitled video (M = 5.48, SD = 1.52). Results indicate that there is indeed a significant difference between groups and the effect size is rather large F(3, 126) = 37.17, p = .000, η2 = 0.38. However, age did not have a significant effect F(1, 126) = 0.04, p = .847. It is important to take into account that Levene’s test of homogeneity of variances was violated (p = .001). A Bonferroni post-hoc indicated that there was a significant difference between the muted video condition and the sound video condition (Mdifference = -2.11, p = .000) and the muted video condition and the subtitled video condition (Mdifference = -1.80, p = .000). No significant difference was confirmed between the sound and subtitled video conditions (p = .743). I can hereby conclude that watching audiovisual video advertisements leads to
significantly stronger comprehension than watching muted advertisements and watching muted subtitled video advertisements generates significantly stronger comprehension than watching muted video advertisements, but not as strong as the audiovisual advertisement. Therefore, H1a and H1c were confirmed, but not H1b as no significant difference existed between
comprehension for audiovisual and muted subtitled video advertisements.
A linear regression analysis with comprehension score as the independent variable and attitude toward the advertisement (average score of the scale) confirmed that the stronger the comprehension score the more favorable attitudes towards the video advertisement are, F(1, 126) = 3.92, p = .05, however these results only barely reach significance. Comprehension score can therefore be used as a predictor of attitudes toward the ad, but the strength of the prediction is very weak. Only 3% of the variance in attitude towards the ad can be predicted based on comprehension b* = 0.17, t = 1.98, p = .05, 95% CI [0.00, 0.27]. Per unit increase in
comprehension score, which runs between 0 and 6, attitude towards the advertisement (between 1 and 7) increases by 0.03 units. H2 is therefore confirmed as higher comprehension leads to a slight increase in attitude toward the advertisement.
Hypothesis 3: mediation
In order to measure whether comprehension fully mediates video exposure and attitude toward the video advertisement Andrew Hayes’ PROCESS SPSS macro (Model 4; 1000 bootstraps; 95% BcaCI) was used with video exposure (experimental conditions) as the independent variable, attitude towards the advertisement as the dependent variable,
comprehension score (0 to 6) as the mediator and age as a control variable. As the independent variable consists of three levels, INDICATOR was used to create a dummy variable (muted= D1, sound= D2, subtitled= 0). No significant indirect effect was found between exposure to a video ad with sound (D2) and attitude through comprehension (b = 0.03, SE = 0.05, CI [-0.02, -0.18]) nor between exposure to a muted video ad and attitude through comprehension (b = -0.26, SE = 0.18, CI [-0.76, 0.02]). However, the direct effect between muted video exposure and
video exposure (see table 1). Direct effects for video exposure and ad attitude as well as
comprehension and ad attitude were also not significant (see table 2 below). When controlled for, age did not produce any significant effect (see tables 1 and 2). Thus, H3 is not confirmed since comprehension is not a necessary step to form attitudes toward the ad from video ad exposure.
Table 1
Direct effect of ad exposure on comprehension while controlling for age
Note: N = 122, * p ˂ .05
Table 2
Direct effects of ad exposure and ad comprehension on ad attitude while controlling for age
B SE t R² F Model summary 0.03 (4, 121) = 0.99 Constant 5.54 18.92 0.29 Comprehension 0.14 0.09 1.52 Muted video -0.05 0.33 -0.15 Sound video -0.05 0.27 -0.17 Age: control -0.001 0.01 -0.07 B SE t R² F Model summary 0.39* (3, 122) = 26.15 Constant 9.19 19.23 0.48 Muted video -1.88 0.29 -6.44* Sound video 0.24 0.28 0.88 Age: control -0.001 0.01 -0.19
Note: N = 121, * p ˂ .05
Hypothesis 4: moderation
H4 postulated that higher levels of involvement will alter the relationship between ad exposure and ad comprehension so that ad comprehension will be stronger-- this relationship being strongest for the audiovisual ad, followed by the subtitled muted ad and finally by the muted ad. A moderation analysis through PROCESS SPSS macro (model 1, 1000 bootstraps, 95% BcaCI) was carried out with video exposure as the independent variable, using
INDICATOR to create the dummy variable (subtitled muted ad, D1= muted, D2= audiovisual). Ad comprehension was the dependent variable, healthy lifestyle involvement was the moderator and age was controlled for. While the model summary was significant (R² = 0.40, F(6, 119) = 13.21, p = .000), there were no significant effects to show that healthy lifestyle moderates ad comprehension from ad exposure (see Table 3).
To determine whether cooking involvement is a moderator, a moderation analysis was carried out through PROCESS SPSS macro (model 1, 1000 bootstraps, 95% BcaCI) with INDICATOR creating a dummy variable for the subtitled muted ad (as mentioned above). Again, the model summary was significant (R² = 0.39, F(6, 119) = 12.83, p = .000), however no significant interaction effect between cooking involvement and ad exposure on ad
comprehension (see Table 4) was able to prove that being involved in cooking can increase ad comprehension after exposure. Therefore, H4, H4a and H4b were not confirmed, concluding that whether an individual is involved in cooking or living a healthy lifestyle, does not alter their comprehension of the video advertisement.
Table 3
Healthy involvement as a moderator between ad exposure and ad comprehension while controlling for age
B SE t Constant 11.45 19.99 0.57 Health involvement 0.06 0.32 0.20 Muted video 0.45 2.58 0.18 Sound video 1.33 2.39 0.56 Interaction: muted x involvement -0.37 0.41 -0.91 Interaction: Sound x involvement -0.18 0.34 -0.46 Age: control -0.003 0.01 -0.32 N = 126, * p ˂ .05 Table 4
Cooking involvement as a moderator between ad exposure and ad comprehension while controlling for age
B SE t Constant 9.13 19.48 0.47 Cooking involvement -0.08 0.25 -0.30 Muted video -2.26 1.56 -1.45 Sound video 0.20 1.61 0.12 Interaction: 0.07 0.29 0.24
(muted x involvement) Interaction: 0.01 0.30 0.97 (Sound x involvement) Age: control -0.002 0.001 -0.17 N = 126, * p ˂ .05 Hypothesis 5: moderation
Andrew Hayes’ PROCESS SPSS macro (model 1, 1000 bootstraps, 95% BcaCI) was used to determine if healthy lifestyle involvement moderates the relationship between
comprehension (IV) and ad attitude (DV) so that stronger comprehension leads to more positive attitudes and vice versa while controlling for age. The analysis indicates that there is no
significant interaction effect between ad comprehension and healthy lifestyle involvement on ad attitudes (see table 5), therefore whether a person is involved in leading a healthy lifestyle or not does not affect their attitudes toward the video ad.
To see whether involvement in cooking moderates the relationship between
comprehension and ad attitudes, the PROCESS SPSS macro tool (model 1, 1000 bootstraps, 95% BcaCI) was used once again with comprehension as the independent variable, attitudes toward the ad as the dependent variable, cooking involvement as the moderator and age as a covariate. Similar to healthy lifestyle involvement, there was no significant interaction between
comprehension and cooking involvement on ad attitudes (see table 6). I can therefore conclude that whether a person considers him/herself involved in cooking does not change the relationship between ad comprehension and attitudes toward the ad. For the reasons stated above, H5 was not confirmed.
Table 5
Healthy lifestyle involvement as a moderator between ad comprehension and ad attitude while controlling for age
N = 126, *p ˂ .05
Table 6
Cooking involvement as a moderator between ad comprehension and ad attitudes while controlling for age
N = 126, *p ˂ .05 B SE t R² F Model summary 0.05 (4, 121) = 0.99 Constant -0.54 18.89 -0.29 Health involvement 0.23 0.62 0.38 Comprehension 0.16 0.74 0.22 Interaction -0.002 0.12 -0.02 Age: control -0.001 0.01 0.17 B SE t R² F Model summary 0.05 (4, 121) = 1.72 Constant 6.31 18.32 0.34 Cooking involvement -0.001 0.31 -0.002 Comprehension -0.04 0.35 -0.13 Interaction 0.03 0.06 0.53 Age: control -0.001 0.01 -0.12
Discussion and Conclusion
Firstly, with the recent emergence of autoplay which mutes social media videos, this study aimed to understand how the modality of online video advertisements, or the lack thereof, affects viewers’ comprehension. Indeed, the modality of online video ads does significantly affect comprehension. Participants that watched the audiovisual advertisement comprehended on average 47% more than those who watched the muted video advertisement. And those who watched the muted subtitled advertisement had on average a 37% stronger comprehension score than the muted advertisement. And while there was no significant difference between
comprehension scores for the muted subtitled ad and the audiovisual ad, these results indicate just how much comprehension can vary according to the modality which diffuses the content.
As both the audiovisual and subtitled videos generated significantly stronger comprehension than the muted video, this confirms the first major conclusion that more modalities provide more information to the viewer, which in return reinforces comprehension (Dijkstra et al., 2005; Edell & Keller, 1989; Drew & Grimes, 1987; Brasel & Gips, 2014). These high comprehension scores are also in line with conclusion two, as the audio and subtitles were both congruent and redundant, thus contributing to stronger comprehension (Kim & Lennon, 2008; Hayati & Mohmeti, 2009). Furthermore, these results are proof that subtitled content in advertisements can be used as a good alternative to audio in order to put emphasis on the specific content of the ad. This in turn aids audiences in better understanding products and their features, as portrayed in a muted autoplay video. These results add to the little research done on
advertisement comprehension and modality.
Secondly, I sought out to determine how comprehension of an advertisement affects viewers’ attitudes towards that advertisement. While the effect was weak, it does appear that an
increase in comprehension can have a slight effect on attitudes, with stronger comprehension scores generating more positive attitudes. Comprehension can therefore contribute in a small way to more positive or negative attitudes and should not be overlooked. However, as the ad differs greatly from Youtube ads, this conclusion may not hold true for ads diffused on the video platform giant. This being said, as no mediation could be proven, comprehension is not an essential stepping stone to attitude creation. More concretely, a product or ad may be deemed appealing by an individual without the individual knowing all the specific functions shown in the ad. As even attitudes for participants in the muted condition were positive, these results can be explained by Hastie and Park’s (1986) idea of on-line memory, whereby an overall affective judgment is formed and stored despite the fact that specific details are not recalled. Therefore, during exposure the individual developed a general positive feeling toward the ad, however, the specific details that led to that feeling are mostly unretrievable or forgotten. Furthermore, this study shows the power of visuals to contribute to positive attitude formation. It is therefore important that companies take factors other than specific content details into consideration when striving to generate positive attitudes towards a brand or product through a video ad.
I also aimed at understanding if involvement is able to alter the relationship between ad exposure and ad comprehension, arguing that a relevant product or brand can cause participants to focus and process content more deeply, leading to stronger comprehension. Nevertheless, involvement did not appear to have an impact on participants’ comprehension. Therefore, according to my results, the factor that has the biggest impact on attention and comprehension is modality. Considering the scales adequately measured involvement, it could stand to reason that involvement has a direct effect on comprehension, comparable to results found by Celsi and Olson (1988). However, it is also possible that as participants were aware that they were
watching the video advertisement for a study, they paid more attention to the content than they otherwise would have when surfing Facebook, despite their level of involvement.
An alternative conclusion could be that muting a video ad is so detrimental to
understanding specific features of the ad that involvement was not able to provide any help to participants when watching and comprehending the ad. For advertisements without subtitles, visuals then become essential to comprehension and narrative video (as seen in the third
conclusion) and could provide for stronger comprehension in the absence of audio and subtitles (Crigler & Neuman, 1994).
Contrary to my argumentation, involvement did not alter attitudes towards the ad based on ad comprehension either. As average involvement scores in each condition for healthy lifestyle and cooking were moderately high to very high and not one participant generated an even slightly negative or neutral attitude toward the ad, one could infer that the appeal of the ad (or product) was portrayed well enough through visuals, that the degree of comprehension and involvement then became superfluous. Participants may have then determined they liked the ad due to features presented visually, without needing to know the finer details about the product as presented in the audio or subtitles. Alternatively, Petty et al.’s (1983) idea of elaboration and involvement through free responses could be the missing link between comprehension and attitude formation. The number of elaborative thoughts that a person generates about a subject may be a stronger indicator of involvement. These results once again emphasize the importance of visual appeal, but raise the question if involvement has any impact at all on attitude formation.
In conclusion, despite being able to confirm all the proposed hypotheses, this study attests to the importance of advertisement design and execution for video advertisements
plays a big role in comprehension and that muted autoplay videos can cause a significant decrease in recall of product attributes. While stronger comprehension can slightly increase attitudes towards advertisements, comprehension is not the end all be all for companies and advertisers. Subtitles and appealing visuals can serve as good alternatives to ensure stronger comprehension and more positive attitudes respectively. Furthermore, this study raises important questions about free recall, advertisement execution and involvement, which could be addressed in further research. Overall, this research is a first step in understanding how autoplay video advertisements affect the viewing experience as well as attitudes of viewers.
Limitations and Further Research
Limitations to this study include the lack of external validity due to a convenience sample from Facebook and e-mail. This being said, the sample includes a range of ages and nationalities, which could benefit international companies and advertisers that target consumers around the world. A further limitation to external validity includes that the video was viewed on the Qualtrics website rather than a social media platform. Future studies could integrate videos into a social media platform where the full viewing experience could be taken into account (e.g. distractions and the possibility to scroll).
As the present study provided one video and required all participants to watch it in its entirety, future research could also delve into the amount of time users spend watching videos, which videos they choose to watch and whether they activate the sound or subtitles while watching. With the plethora of stimuli on social media, it is possible that viewers only watch videos partially and are quickly distracted by posts and photos around them, which has been shown to affect information processing (Han, 1992). Moreover, some content may be more
desirable than others to ensure that viewers watch video advertisements longer, and with more focus.
Furthermore, the present study focuses on attitudes, but future studies could include purchase intention and brand image as there exists a large gap in research about the effects of muted autoplay on ads and even more so on purchase intention. It is to be noted that while most participants generated a positive attitude toward the video, they were not asked to make a decision on whether or not they would purchase it. Perhaps, as Zaichowsky (1994) and Behe et al. (2015) suggest, involvement has a bigger role when it comes to making a decision about purchasing a product rather than simply forming an attitude, as purchasing requires careful examination of a product and the buyer’s needs.
Future research looking into involvement and comprehension could take factors such as free responses into account as well as more in depth comprehension measures with open questions or free recall to ensure both variables are adequately measured.
Lastly, the randomization function from Qualtrics survey software was not entirely consistent, producing a somewhat uneven number of participants per condition, which should be avoided in future studies.
In conclusion, as video ads become more prevalent on social media platforms and as billions of dollars are continuously being invested, it is important for companies and advertisers to understand what viewers are taking away from these autoplay ads and how the degree of comprehension of viewers affects the persuasiveness of the ad. The following experiment is the first of its kind to expose how viewers comprehend autoplay video ads and how their
comprehension relates to their attitudes towards them. This study is a great addition to current advertising and modality research and will help pave the way for future research on autoplay
video advertising as well as help companies and advertisers in creating and adapting video content for social media audiences.
