NEED TO EVALUATE MODERATES CULTIVATION EFFECTS 1
Please cite as “Coenen, L., & Van den Bulck, J. (2016). Cultivating the opinionated: The Need to Evaluate moderates the relationship between crime drama viewing and scary world evaluations. Human Communication Research. doi: 10.1111/hcre.12080”
Cultivating the opinionated: The Need to Evaluate moderates the relationship between crime drama viewing and scary world evaluations
Lennert Coenen and Jan Van den Bulck
Leuven School for Mass Communication Research, Faculty of Social Sciences, KU Leuven, Leuven, Belgium
THIS DOCUMENT IS A POSTPRINT VERSION OF THE PAPER PUBLISHED IN HUMAN COMMUNICATION RESEARCH. FULL ARTICLE AVAILABLE FROM: http://onlinelibrary.wiley.com/doi/10.1111/hcre.12080/abstract
Author note
E-mail addresses: ;
Correspondence concerning this article should be addressed to Lennert Coenen, Leuven School for Mass Communication Research, KU Leuven, Parkstraat 45 BOX 3603, 3000 Leuven, Belgium. E-mail:
Abstract
The literature on cultivation processes assumes that second-order judgments (e.g., attitudes) are repeatedly updated during viewing (i.e., online) and can be reported when asked. In this article, we propose this reasoning only holds for people high in their Need to Evaluate (NTE). Low NTE individuals do not update their opinions online and have to construct their attitudes memory-based, limiting the likelihood of finding second-order relationships. Results from a cross-sectional survey among 226 adolescents indicate that crime drama viewing and scary world beliefs are only correlated among high NTE individuals. NTE does not moderate cultivation through non-fiction. The theoretical implications of memory-based attitudes for second-order cultivation are discussed, and suggestions are made for future research on attitudinal media effects.
Keywords: cultivation theory, cognitive processes, Need to Evaluate, first-order, second-order, memory-based, online, attitudes, scary world
Cultivating the opinionated: the Need to Evaluate moderates the relationship between crime drama viewing and scary world evaluations
Cultivation theory is one of the most quoted theories in communication research (Bryant & Miron, 2004), but our understanding of the psychological processes explaining its effects is incomplete. While ample research supports the notion that first-order influences of television viewing on respondents’ factual beliefs (i.e., prevalence estimates of crime or violence) occur through memory-based heuristic processing, the mechanisms behind second-order cultivation effects on evaluative beliefs (i.e., opinions and attitudes; Oskamp & Schultz, 2005) are less well understood (Shrum & Lee, 2012). It is commonly assumed that second-order judgments take shape online, during viewing. Over time, heavy viewers should therefore develop content-consistent opinions that are accessible when a judgment is needed (Shrum, Burroughs, & Rindfleisch, 2004; Shrum & Lee, 2012; Shrum, Lee, Burroughs, & Rindfleisch, 2011). This part of the cultivation typology may be problematic, however, as it is at odds with attitude theory: survey respondents are often unable to retrieve previously formed evaluations and therefore have to rely on memory-based processing when prompted for their opinion (e.g., Schwarz, 2007).[1]
The present paper will discuss the relevance of memory-based attitude construction for second-order cultivation processes. In particular, it will focus on the Need to Evaluate (NTE), a personality trait reflecting a chronic motive to form and maintain strong evaluations (Jarvis & Petty, 1996; Petty & Jarvis, 1996). Whereas high NTE individuals fit with the assumptions made by the online model of cultivation - they spontaneously evaluate information online and hold strong opinions about a variety of subjects – low NTE individuals often only construct an opinion when required (i.e. memory-based; Tormala & Petty, 2001). Because such attitude constructions are the outcome of temporarily salient information in memory, numerous incidental factors can influence the eventual judgment. This would seem to limit the probability of finding second-order cultivation relationships, and it suggests that correlations between television viewing and attitude reports may only exist among high NTE individuals
Literature Review
Cultivation theory
Cultivation theory proposes that, over time, the stories television tells about the world should shape viewers’ perceptions of reality. Because television programs overuse violence as an element of storytelling, such perceptions should be heavily biased. Early studies indeed found television viewing to correlate with perceptions of, and beliefs in, a dangerous world (e.g. Gerbner & Gross, 1976). Nonetheless, the correlations were small and no explanations were offered for the psychological mechanisms at work. This left open the possibility that so-called cultivation effects were no more than methodological artifacts (e.g., Hirsch, 1980), and scholars realized the psychological “black box” (Shanahan & Morgan, 1999, p. 172) needed to be addressed.
Hawkins and Pingree (1982) were the first to formulate a solution. In their model, the cultivation process was assumed to consist of two distinct steps: acquiring factual information from television (learning) and integrating that information into beliefs about social reality (construction). Based on this logic, some scholars hypothesized that respondents’ answers to factual first-order cultivation questions (category size, frequency, or probability judgments) would mediate the relationship between television viewing and second-order evaluations (opinions and attitudes) (Gerbner, Gross, Morgan, & Signorielli, 1986; Hawkins, Pingree, & Adler, 1987). Studies have failed to find support for this approach, however, suggesting to some that the two types of judgments are unrelated (Hawkins et al., 1987; Shrum & O’Guinn, 1993).
Answers to first-order cultivation questions are now considered to be the outcome of what Hastie and Park (1986) coined memory-based processing. Because judgments of category size, frequency, or probability are not stored in memory prior to a question, respondents have to retrieve salient information to construct an answer on the spot (Shrum & O’Guinn, 1993). Building on the work by Tversky and Kahneman (1973), Shrum and colleagues have shown that the ease with which relevant exemplars come to mind forms the basis of the judgment: the easier the retrieval process, the higher the estimate. Because information becomes more accessible when it is frequently encountered, heavy viewers of crime-laden television programs give particularly high estimates regarding the prevalence of crime and violence in society (Shrum & O’Guinn, 1993).
Second-order effects are believed to result from different processes. Rather than being formed at the retrieval stage, they are assumed to be made and changed online, during encoding of information into long-term memory (Hastie & Park, 1986; Shrum & Lee, 2012). This perspective has been heavily influenced by dual-process models of persuasion that predict attitude change as a function of issue-relevant thought during exposure to a message (e.g., Elaboration Likelihood Model (ELM), Petty & Cacioppo, 1986). If an individual does not engage in much effortful thought, peripheral processing often only results in changes that are relatively fleeting and susceptible to opposing persuasive attempts (e.g., Haugtvedt & Petty, 1992). When elaborating centrally, on the other hand, the amount and valence of a receiver’s cognitive responses to a message regulate a potentially long-lasting change in opinion (Petty & Cacioppo, 1986). Important for cultivation effects, processing during exposure to narratives is likely to be particularly influential. If people are transported into a story, they are processing information thoroughly – as is the case when elaborating - but their focus is on the narrative itself, not on scrutinizing message propositions. As a result, their cognitive responses are less likely to be affected by prior knowledge structures, and more likely to be consistent with the portrayals on screen (Bilandzic & Busselle, 2008; Green & Brock, 2000).
Building on these theoretical insights from persuasion theory, cultivation studies have found that viewers’ attention, elaboration, and transportation during viewing play a pivotal role in shaping viewers’ evaluative judgments (Bilandzic & Busselle, 2008; Shrum et al., 2004, 2011). Thus, the logic underlying second-order cultivation parallels the evaluative processes described in the literature on online impression formation (Ottati, Edwards & Krumdick, 2005; Shrum & Lee, 2012): People are assumed to develop an “online tally” of evaluations that is constantly updated during viewing (McGraw, Lodge, & Stroh, 1990, p. 43). When prompted by an opinion question, the summary judgment is simply retrieved and reported (McGraw et al., 1990).
Attitudes: temporary constructions or enduring dispositions?
In attitude theory, the question whether and how evaluations are stored and retrieved remains debated. The conceptualization of second-order judgments in cultivation research clearly adheres to traditional formulations of attitudes as enduring dispositions (e.g., Fazio, 2007; Petty & Cacioppo, 1986). Wilson and Hodges (1992) coined this perspective the file-drawer model of attitudes:
“When people are asked how they feel about something, such as legalized abortion, Uncle Harry, or anchovies on a pizza, presumably they consult a mental file containing their evaluation. They look for the file marked abortion, Uncle Harry, or anchovies, and report the evaluation it contains” (p. 38).
Some scholars have objected to this view and define evaluative judgments as episodic constructions (Schwarz, 2007; Schwarz & Bohner, 2001; Wilson & Hodges, 1992). Schwarz (2007), for example, argues that respondents use whatever information is most accessible to construct an answer each time an evaluation is required. This construal model better accounts for the well documented instability of attitude reports because it gives a prominent role to temporarily accessible information in memory and non-evaluative factors during response (e.g., question context: Schwarz, 1999).
In turn, however, constructionist views do not adequately address the issue of attitude consistency (Fazio, 2007). Several authors have therefore underlined the importance of considering both retrieval and ad-hoc construction when examining attitudinal judgments (Kruglanski & Stroebe, 2005). Even adherents of a strong constructivist view acknowledge attitudes are not always computed from scratch (Schwarz & Bohner, 2001). Likewise, dispositional models also allow for memory-based judgments when object-evaluations are not easily accessible from memory (Fazio, 2007). For this reason, the belief sampling model describes an attitude as a general database of “considerations”, consisting of “beliefs, feelings, impressions, general values, and prior judgments about an issue” (Tourangeau et al., 2000, p. 179). As in a file-drawer perspective it argues that respondents may simply retrieve an accessible evaluation when prompted for their opinions, but their judgments can also vary if they focus on different considerations on different occasions (cf. construal models).
So far, cultivation research has implicitly considered memory-based processes to be irrelevant for second-order effects (e.g., Morgan, Shanahan, & Signorielli, 2014). This may be surprising given Shrum’s (2004) remark that assuming “all second-order judgments are made on-line is clearly false” (p. 339), but he argues that the judgments measured in cultivation research occur often enough for the assumption to hold: as crime and violence are so commonplace, almost every viewer should have formed and maintained strong issue-relevant opinions over time. Answering attitude questions therefore only entails retrieval, and no construction (Shrum et al., 2004).
As the information processing literature suggests, however, it is doubtful that television viewers are motivated to form new evaluations or update old ones every time they are exposed – even if we assume that they are always elaborating on the implied message. In traditional formulations of the ELM, central processing was equated with online evaluation (Petty, Priester, & Wegener, 1994) but it has been noted that these are independent processes (Carpenter & Boster, 2013; Choi, 2011). Only if people possess an evaluative processing goal, online evaluation and the formation of strong attitudes is likely (Bizer, Tormala, Rucker, & Petty, 2006; Mackie & Asuncion, 1990; Tormala & Petty, 2001). If they do not have a goal to evaluate, they still need to construct a post-hoc judgment based on the specific information they recall from memory (Mackie & Asuncion, 1990).
The literature on cultivation processes has not completely overlooked the importance of cognitive goals, but it considers them to be a secondary issue because evaluations can occur even if a processing goal is absent (Shrum, 2007; e.g., Bargh, Chaiken, Raymond, & Hymes, 1996). This line of reasoning, however, fails to make a distinction between automatic evaluations and the propositional processes that are typically addressed in persuasion models and cultivation studies (Gawronski & Bodenhausen, 2006). The fact that evaluations can be automatically activated during viewing does not imply that the valence or extremity of evaluative judgments changes (Crano & Prislin, 2006). For the cultivation hypothesis this is crucial (Shrum & O’Guinn, 1993): study participants are not merely hypothesized to answer opinion questions using accessible evaluations, it is also expected that the extremity of their evaluative judgments is directly proportional to television exposure frequency. Whether or not a viewer has an evaluative processing goal therefore remains an important determinant of attitudinal cultivation effects, regardless of the automaticity of initial evaluative reactions (cf. Cunningham & Zelazo, 2007).
The explanatory value of the Need to Evaluate
NTE is an individual’s tendency to engage in evaluative thought (Jarvis & Petty, 1996). As such, it is related to the tendency to think thoroughly (Need for Cognition, NFC: Cacioppo, Petty, & Kao, 1982), but NTE’s emphasis on evaluation marks an important conceptual distinction: the motive to think effortfully is not identical to the motive to think evaluatively. When people are asked to describe their own day, for example, those high in NTE tend to use evaluative (e.g., “I had a delicious breakfast”) instead of more factual descriptions (e.g., “I had breakfast at 10 a.m.”). The same is not true for high NFC individuals (Jarvis & Petty, 1996). In other words, differences in NTE reflect variations in the presence of evaluative processing goals: people high in NTE are likely to evaluate spontaneously during exposure to information (online), while those low in NTE still need to construct a memory-based judgment when required (Carpenter & Boster, 2013; Tormala & Petty, 2001).
Not surprisingly, then, high NTE individuals tend to form and develop more (Bizer et al., 2004), more accessible (Petty & Jarvis, 1996; Tormala & Petty, 2001), more extreme (Britt, Millard, Sundareswaran, & Moore, 2009), and stronger opinions (Britt et al., 2009) about a variety of issues than people who are low in their NTE. Because they possess a file drawer of evaluations or, in Tourangeau et al.’s (2000) terminology, a homogeneous sample of accessible considerations (Petty & Jarvis, 1996), high NTE individuals are also less likely than others to choose “no opinion” response categories (Jarvis & Petty, 1996) or to base their attitudinal response on incidental factors of the survey situation (Toepoel, Vis, Das, & van Soest, 2009).
Research in political psychology suggests that these properties of NTE may be relevant for media effects theorizing. Experiments on framing, for example, only appear to affect participants who rely on memory-based processing – those low in NTE. People high in NTE are less influenced by one-shot experimental exposure because they already possess strong opinions beforehand (Druckman & Leeper, 2012; Druckman & Nelson, 2003). Relatedly, Matthes (2007) has argued that an online model of framing effects is needed because the dominant memory-based models do not suit high NTE individuals well. The inverse is true for cultivation research. As we suggest, second-order cultivation studies should consider the implications of memory-based attitude construction for low NTE individuals.