Download Media uses and their effects and more Thesis Mass Communication in PDF only on Docsity!
Media Effects: Theory and Research Patti M. Valkenburg,^1 Jochen Peter,^1 and Joseph B. Walther^2 (^1) Amsterdam School of Communication Research, University of Amsterdam, the Netherlands email: p.m.valkenburg@uva.nl, j.peter@uva.nl (^2) Wee Kim Wee School of Communication and Information Nanyang Technological University, Singapore email: jwalther@ntu.edu.sg
Corresponding author: Patti M. Valkenburg University of Amsterdam Spui 21, 1012 WX Amsterdam, the Netherlands tel: +
Please cite as: Valkenburg, P. M., Peter, J., & Walther, J. B. (Research. Annual Review of Psychology, 67. doi: 10.1146/annurev-psych-122414-033608 2016 ). Media effects: Theory and
Keywords Media effects theory, selective exposure, media violence, computer-mediated communication (CMC), mass communication, mass media, meta-analysis.
Abstract The aim of this review is to analyze trends and commonalities among prominent theories of media effects. On the basis of exemplary meta-analyses of media effects and bibliometric studies of well-cited theories, we identify and discuss five features of media effects theories as well as their empirical support. Each of these features specifies the conditions under which media may produce effects on certain types of individuals. Our review ends with a discussion of media effects in newer media environments. This includes theories of computer-mediated communication, the development of which appears to share a similar pattern of reformulation from unidirectional, receiver-oriented views, to theories that recognize the transactional nature of communication. We conclude by outlining challenges and promising avenues for future research.
INTRODUCTION
Research on the effects of media originated under the umbrella term mass communication research. The last five reviews on the effects of media that appeared in the Annual Review of Psychology include the word ‘mass’ in their titles (Liebert & Schwartzberg 1977, Roberts & Bachen 1981, Schramm 1962, Tannenbaum & Greenberg 1968, Weiss, 1971). The concept of mass communication arose during the 1920s as a response to new opportunities to reach large audiences via the mass media: newspapers, radio, and film (McQuail 2010). However, mass refers not only to the size of the audience that mass media reach, but also to uniform consumption, uniform impacts, and anonymity, notions that are progressively incompatible with contemporary media use. Since the 1980s, media use has become increasingly individualized and, with the advent of the internet, has also taken a decidedly personalized character. This increase in individualization and personalization of media use has enabled a form of communication that Castells (2007) has called mass self- communication. Mass self-communication shares with mass communication the notions that messages are transmitted to potentially large audiences, and that the reception of media content is self-selected: Media users select media content to serve their own needs, regardless of whether those needs match the intent of the generator of the content (McQuail 2010). However, whereas mass communication research focuses only on media reception processes, mass self-communication focuses on media reception and generation processes, and, thus, on the effects of media generation on the generators themselves (Castells 2007). The current co-existence of mass communication (e.g. via newspapers, radio, and television) with mass self-communication (e.g. via social media) is reflected in the structure and content of this article. The aim of this review is to assess the most important media effects theories that have emerged in the past three decades, and to chart the development
of media effects thinking from its roots in assumptions about unidirectional effects to contemporary recognition of complex reciprocal interactions. To do so, we do not aim to discuss each of the theories of media effects that have emerged successively. Instead, we start with a brief overview of approaches and their summary by way of several exemplary meta-analyses of media effects. We then organize our review around five important features of media effects theories, including their analytic implications and empirical support. Subsequently, we describe the effects of mass self-communication in the newer media environment. We briefly discuss the historical development of theories of computer- mediated communication (CMC), including the state of present-day CMC theories and research. We conclude by outlining challenges and promising avenues for future research. Meta-Analyses of Media Effects Research on the effects of media emerged between the 1920s and 1930s, but it became a prominent focus only at the end of the 1950s, after the introduction of television and the emergence of academic communication departments in Europe and the US (but see: Hovland et al 1953, Katz & Lazarsfeld 1955, Lazarsfeld et al 1948). These developments generated a proliferation of media effects theories and research, albeit initially, as in other social science disciplines, at a fairly basic level. By the 1980s, thousands of empirical studies had been published investigating the cognitive, emotional, attitudinal, and behavioral effects of media on children and adults (Potter 2012, Potter & Riddle 2007). Moreover, since the 1990s, a sizeable number of meta-analyses have synthesized the results of these empirical studies. Table 1 presents a list of 20 examples of meta-analyses on media effects that have appeared in the past two decades. These meta-analyses were selected because together they cover the broad plenitude of media effects that have been investigated since the 1960s, ranging from the effects of exposure to media violence on
These explanations are grounded in five specific features of media effects theories. Four of these features have been identified earlier by Valkenburg and Peter (2013a), albeit in less detail. This review both complements and extends Valkenburg and Peter’s analysis by adding more evidence, and seeking parallels between the mass communication and mass self-communication literature. FIVE FEATURES OF MEDIA EFFECTS THEORIES The focus of this review is on micro-level media effects theories. Several bibliographic analyses have tried to document the state of the art of both micro- and macro-level media effects theories in the scholarly journals (Bryant & Miron 2004, Potter 2012, Potter & Riddle 2007). Table 2 lists the micro-level media effects theories that have been identified as particularly well-cited in these bibliographic studies. Valkenburg and Peter (2013a) have recently attempted to organize existing micro-level media effects theories in terms of their basic assumptions. They observed that these theories differ substantially in how they conceptualize the media effects process. Some theories, particularly the earlier ones, focus primarily on unidirectional linear relationships between media use and certain outcomes (e.g. Cultivation Theory; Gerbner et al 1980). Other, more comprehensive theories (e.g. Bandura 2009, Slater 2007) pay more attention to the interaction between media factors (media use, media processing) and non-media factors (e.g. disposition, social context). When looking over the existing media effects theories, they can be organized along the following five global features that address the relationships between both media factors and non-media factors and specify the boundary conditions of media effects. Feature 1: Selectivity of Media Use A first feature of media effects theories that specifies the boundary conditions of media effects involves the selectivity paradigm. The two propositions of this paradigm are that (a) people only attend to a limited number of messages out of the constellation of messages
that can potentially attract their attention, and (b) only those messages they select have the potential to influence them (Klapper 1960, Knobloch-Westerwick 2015, Rubin 2009). More than 60 years ago, researchers discovered that people do not randomly attend to media, but rather focus on certain messages as a result of specific social or psychological needs or beliefs (Katz & Lazarsfeld 1955). For example, in their classic study of the 1940 U.S. presidential election, Lazarsfeld et al (1948) suggested that people often seek out political content that reinforces their beliefs while they avoid content that was meant to change their opinions. This insight led the researchers to conclude that the power of media to change attitudes or behavior is limited (Klapper 1960; Lazarsfeld et al 1948). The selectivity paradigm, so coined in the 1940s, has been further elaborated into two theoretical perspectives: Uses-and-Gratifications (Katz et al 1973, Rosengren 1974, Rubin
- and Selective Exposure Theory (Knobloch-Westerwick 2015, Zillmann & Bryant 1985). Both the Uses-and-Gratifications and Selective Exposure Theory postulate that individuals select media in response to their needs or desires, and that a variety of psychological and social factors guide and filter this selection. Both theories also propose that media use is a precursor to consequences (named ‘obtained gratifications’ in Uses- Gratifications-Theory and ‘media effects’ in Selective Exposure Theory). An important difference between the two theories is that Uses-and-Gratifications Theory conceptualizes media users as rational and aware of their selection motives, whereas Selective Exposure Theory argues that media users are often not aware, or at least not fully aware, of their selection motives. This difference in conceptualization of the media user has methodological consequences. For example, in line with the assumption that users can articulate their motives for using media, research based on the Uses-and-Gratifications Theory mainly uses self-reports to gauge media use behavior. In contrast, research based on Selective Exposure
whereas men are more likely to select sports, horror, and action-adventure than females are (for more evidence see Knobloch-Westerwick 2015, Oliver et al 2006, Oliver & Krakowiak 2009). The evidence of the effects of proximal dispositions on selective exposure is more complex. Since the work of Lazarsfeld et al (1948) and Klapper (1960), the selectivity paradigm has predominantly been inspired by Festinger’s (1957) cognitive dissonance theory, which argues that people typically avoid discomforting cognitive dissonance caused by information that is incompatible with their existing dispositions (e.g. beliefs, attitudes). To avoid or solve this state of dissonance, they may actively seek information that reinforces their dispositions and they avoid potentially contradictory information that would exacerbate dissonance. However, although there is ample evidence for the mechanism that individuals seek congenial information (Hart et al 2009), cognitive dissonance reduction is not as consistent a cause of selective exposure as it was previously assumed to be (Donsbach 2009, Hart et al 2009, Smith et al 2007). First, it seems to hold more consistently for political than for health messages (Hart et al 2009, Knobloch- Westerwick 2015). Second, subsequent evidence showed that under specific conditions, people are willing or even eager to attend to uncongenial information, for example, when the perceived utility of information is great, when they are uncommitted to an attitude, or when the reliability of the offered information turns out to be poor (Hart et al 2009). In the realm of media entertainment, counterintuitive findings also challenged the consistency assumption. For example, when it comes to fearful and tragic entertainment, people often expose themselves to content that is inconsistent with their moods and existing values, and that may even elicit uncomfortable reactions, such as fear and sadness. Several more recent theories have proposed plausible explanations for people’s occasional attitude- inconsistent selective exposure to information and entertainment, for example Information-
Utility Theory (Atkin 1973); Mood Management Theory (Zillmann & Bryant 1985); and Eudaimonia Theory (Oliver 2008). Developmental factors. As for development, research has shown that individuals typically prefer media content that is only moderately discrepant from their age-related comprehension schemata and experiences (e.g. Valkenburg & Cantor 2001). If they encounter media content that is too discrepant, they will allocate less attention to it or avoid it. This moderate-discrepancy hypothesis explains, for example, why: (a) toddlers are mostly attracted to media with a slow pace, familiar contexts, and simple characters; (b) preschoolers typically choose a faster pace, more adventurous contexts, and more sophisticated characters; and (c) adolescents are the most avid users of social media and seek entertainment that presents humor based on taboos and irreverent or risky behavior (Valkenburg & Peter 2013a). Although developmentally-related media preferences are most evident in childhood, they also extend to adulthood. For example, in comparison to younger adults, middle and older adults more strongly prefer nonarousing, meaningful, and uplifting media content, whereas younger adults more strongly prefer arousing, violent, and frightening media (Mares et al 2008, Mares & Sun 2010, Mares & Woodard 2006). Social context factors. Most media effects theories recognize the importance of social context at the micro, meso, and macro level in encouraging or discouraging media use (Klapper 1960, Prior 2005, Slater 2007). Social influences can occur deliberately and overtly, when institutions, schools, or parents restrict or regulate media use (Nathanson 2001, Webster 2009). On the macro level, structural aspects of the media system (e.g. channel availability) can affect media choices (Webster 2009), whereas on the micro level, adults can forbid children to watch violent content and encourage them to use educational media (Nathanson 2001).
Modality. Since the early days of mass communication research, it has been common to study the differential effectiveness of modalities for information processing and learning. Marshall McLuhan (1964) is best known for his theory of the differential impact of modalities. By means of his aphorism, “the medium is the message,” he argued that media affect individuals and society not by the content delivered, but primarily by their modalities. Inspired by McLuhan’s theories, a myriad of media comparison studies have tested whether information delivered via auditory or textual modalities encouraged learning, reading skills, or imagination more (or less) than information delivered through audiovisual media (e.g. Beentjes & van der Voort 1988, Greenfield et al 1986). These media comparison studies largely lost their appeal in the new millennium, probably because they often failed to produce convincing results especially when it comes to learning (Clark 2012). Many content and structural properties related to the presentation of information (e.g. difficulty, repetition, prompting) turned out to be more important for learning and information processing than modality (Clark 2012). In the new millennium, due to advances in technology, research interest in the differential effects of media modalities have shifted to, for example, a comparison of the effects of interfaces that differ in their degree of interactivity on engagement, information processing, and learning (Sundar et al 2015). Media comparison studies also started to focus on the differential effects of reading on paper versus screens (via tablets or e-readers) for learning and information processing (e.g. Mangen et al 2013, Small et al 2009). This rapidly growing literature has to date yielded small and inconsistent differences in favor of reading on paper (cf. Mangen et al 2013, Rockinson-Szapkiw et al 2013). Content properties. The contribution of media content to guide selective exposure or to predict media effects has received relatively little attention on both the theoretical and the empirical level. For example, in an edited book about selective exposure (Hartmann 2009),
not a single chapter focused on specific media content that may trigger or enhance the likelihood of selective exposure. Likewise, a comprehensive edited volume on media effects contained no integrating theory on how media content may enhance or constrain media effects (Bryant & Oliver 2009). Although related fields (e.g. cinematography, advertising) have paid more attention to content properties that may attract attention or enhance effects (e.g. Boerman et al 2011), media effects researchers typically assess the effectiveness of media content/messages from the psychological reactions they elicit (O'Keefe 2003, Slater et al 2015). For example, in experiments investigating the differential effects of fear-provoking messages, the extent of fearfulness is typically evaluated via pretests or manipulation checks in which subjects’ reactions are observed or surveyed (O'Keefe 2003). Such an effect-based approach, however, offers little understanding of the specific content/message properties that have evoked these states in media users. The complexity faced in formulating a comprehensive theory of content properties that guide selective exposure is particularly challenging because the attractiveness and effectiveness of content is strongly contingent upon individual users or, at best, subtypes of users. After all, what keeps one’s attention to media content is the result of a complex and intertwined set of dispositional, developmental, and social-context factors. For example, the nature of characters, narratives, contexts, and humor that attract the attention of early adolescents may be unappealing or even distasteful to other age groups. Still, the literature reveals some notions about media content that may guide selective exposure. For example, it has often been found that people pay more attention to negative media content than to positive content, especially when it comes to news (Zillmann et al 2004). These results are consistent with theories that argue that people are “hardwired” for attention to danger- conveying stimuli (Shoemaker 1996). People attach more weight to negative information because such information probably contrasts with their baseline positive reactions to social
environment, such as a bright flash of light or a sudden noise. It is accompanied by an attentional process that has been called stimulus-driven or transient attention (e.g. Corbetta & Shulman 2002). This type of attention contrasts with goal-directed or sustained attention, which is not driven by stimulus properties, but directed by the goals and experiences of the media user him- or herself. Stimulus-driven automatic attention is already present in infants and is less contingent on audience factors than sustained attention is (Bradley 2009, Valkenburg & Vroone 2004). However, although stimulus-driven automatic attention can instigate selective exposure, it is unlikely a sufficient condition for sustained selective exposure. First, after repeated exposure to a novel or otherwise salient stimulus, people’s attention toward it becomes weaker, even if the stimulus is strong (Bradley 2009). Second, selective exposure is primarily guided by the goals and experiences of media users, and, hence, it is more sensitive than stimulus-driven attention to dispositional, developmental, and social context differences in the media users. Feature 3: Media Effects are Indirect A third feature of many media effects theories that may specify the boundary conditions of media effects is that most media effects are indirect rather than direct (e.g. McLeod et al 2009, Petty & Cacioppo 1986). An indirect effect is one in which the influence of an independent variable (e.g. media use) on other variables (e.g. outcomes of media use) works via its influence on one or more intervening (mediating) variables. The conceptualization of indirect media effects is important for two reasons. First, intervening variables provide important explanations for how and why media effects occur, and, therefore, they can be helpful when designing prevention and intervention programs. Second, ignoring indirect effects can lead to a biased estimation of effects sizes in empirical research and, thus, meta-analyses (Holbert & Stephenson 2003). After all, it is the
combination of direct and indirect effects that makes up the total effect of an independent variable on a dependent variable. Thus “if an indirect effect does not receive proper attention, the relationship between two variables of concern may not be fully considered” (Raykov & Marcoulides 2012, p. 7). Media effects theories have identified three types of indirect effects. In the first type, which we discussed in the section about selectivity (feature 1), media use itself acts as an intervening variable between pre-media use variables (development, dispositions, and social-context factors) and outcome variables. In the second type of indirect effects, the cognitive, emotional, and physiological processes that occur during and shortly after exposure act as mediators. It has often been posited and shown that the way in which individuals process media forms the route to media effects. For example, research based on the Elaboration Likelihood Model (Petty & Cacioppo 1986) has found that attitude change is more enduring when a message leads to a high level of attention and elaboration (i.e. the central route). Anderson and Bushman’s (2002) General Aggression Model predicts indirect effects of exposure to media violence on aggression through three response states: cognition, emotion, and arousal. Finally, experiments based on Zillmann’s (1996) Excitation-Transfer Model have demonstrated that residual arousal that results from media- induced sexual excitement can intensify positive (e.g. altruistic feelings) and negative feelings and behavior (e.g. anger, aggressive behavior). The third type of indirect effects that has been identified conceptualizes post- exposure variables that may themselves be dependent variables (e.g. attitudes and beliefs), as mediators of other post-exposure variables. Especially in political and health communication, it has repeatedly been found that effects of media use on political and health behavior are mediated by certain beliefs and attitudes (Holbert & Stephenson 2003). For example, recent work in political communication increasingly conceptualizes the
Gratifications Theory (Rubin 2009), Reinforcing Spiral Model (Slater 2007), the Conditional Model of Political Communication Effects (McLeod et al 2009), the Elaboration Likelihood Model (Petty & Cacioppo 1986) and the Differential Susceptibility to Media Effects Model (Valkenburg & Peter 2013a). For example, in the Elaboration Likelihood Model, need for cognition, the tendency to enjoy effortful information processing, is seen as a moderator of media effects on attitudes. Some theories have proposed that the same factors that can predispose media selection can also modify the direction or strength the effects of media use (e.g. Bandura 2009, McLeod et al 2009). Valkenburg and Peter (2013a) argued that dispositional, developmental, and social context factors have a double role in the media effects process: They not only predict media use, but in interaction with media properties they influence the way in which media content is processed. In other words, properties of media affect how media content is processed (i.e. property-driven processing), but the effects of this property-driven processing are contingent upon specific dispositions, developmental level, and social context factors of the media user. As discussed, individuals have the tendency, at least to a certain extent, to seek out congenial media content (Hart et al 2009, Klapper 1960), that is, content that does not deviate too much from their dispositions, developmental level, and the norms that prevail in the social groups to which they belong. It is conceivable that these same factors can also moderate the way in which media content is processed. Qualitative critical audience research has often emphasized that audiences differ in their interpretations of media content (e.g. Hall 1980) and that these interpretations partly depend on gender, class, race, and age (e.g. Kim 2004). However, in social science-based media effects theories such interactive influences on media processing have, to our knowledge, received less attention. There has been ample research on selective exposure and selective recall, but less on selective
reception processes (Hart et al 2009). Moreover, the scarce research that is available has mainly focused on cognitive processing of media content and less on emotional processing, despite the growing evidence that emotional processes, such as identification with characters and involvement in the narrative, are important routes to persuasion (e.g. Slater & Rouner 2002). As for dispositions, research indicates that trait aggressiveness moderates media violence effects on cognitive (e.g. misinterpretation of ambiguous non-violent acts) and emotional processing (e.g. a decreased empathy with characters; Bushman 1995, Krcmar 2009). A high need for cognition has been shown to moderate message effects on cognitive processing (Cacioppo et al 1996, Shrum 2009). Trait empathy and need for affect can enhance emotional processing when watching sad or frightening films (Krcmar 2009, Oliver & Krakowiak 2009). Finally, bodily needs such as hunger may significantly alter the way in which individuals perceive food products presented on a screen. Such products may seem bigger when subjects are hungry than when they are not (McClelland & Atkinson 1948). The moderatin g role of dispositional variables can be explained by the disposition- content congruency hypothesis (Valkenburg & Peter 2013a), which argues that dispositionally congruent media content may be processed faster and more efficiently than incongruent content because it can be assimilated more readily to the media user’s existing cognitive schemata. Because congruent content requires less cognitive effort, it leaves more resources available for the processing of less salient content (Alba & Hutchinson 1987). Dispositionally congruent content can also affect emotional processing through processing fluency. Congruent content enhances the media users’ experience of familiarity or at least their illusion of familiarity. This (illusion of) familiarity may in turn enhance positive affect
