Self-Monitoring, Opinion Leadership and Opinion Seeking: a Sociomotivational Approach
- First Online:
- Cite this article as:
- Rose, P. & Kim, J. Curr Psychol (2011) 30: 203. doi:10.1007/s12144-011-9114-1
- 462 Views
In complex markets characterized by abundant choice, many people assume the roles of opinion leaders and opinion seekers. Understanding people who gravitate toward these roles is a priority for consumer psychologists, because the effectiveness of large-scale persuasion often depends on word-of-mouth or peer-to-peer communication. In this study we tested a model, inspired by prior research, that included self-monitoring, status motivation and belonging motivation as predictors of both opinion leadership and opinion seeking. Self-monitoring was a significant predictor of opinion leadership and status motivation mediated this relationship. Self-monitoring was not a significant predictor of opinion seeking, but belonging motivation was. The study highlights motivations associated with self-monitoring and also suggests that the sociomotivational bases of opinion leadership and opinion seeking differ.
KeywordsConsumer psychologySelf-monitoringOpinion leadershipOpinion seeking
Many of today’s consumers face a bewildering number of choices. In 2010, American consumers who chose to receive television programming via satellite could choose one of four packages on the Dish Network’s web site. The most popular of these packages promised consumers no less than 200 television channels to choose from. In the same year, a search on “tortillas” in Amazon.com’s grocery department yielded more than 500 results which included not only dozens of kinds of tortillas but also more tortilla-related products than most people can imagine. The abundance of choice that is available to many consumers is not always satisfying and in some cases may be distressing (Schwartz 2004). It spurs consumers to seek clarifying information that can reduce uncertainty and aid decision-making (Rogers 2003). With so many choices available to them, many consumers—called opinion seekers—consistently invite others’ influence as they construct their own opinions about products and services. Many consumers also act as opinion leaders by behaving in ways that have the potential to influence the opinions of others.
Influencing others and being influenced by others is common among humans, who are often described as extremely social animals. It has been demonstrated that two arbitrarily selected Americans who may not know each other are typically indirectly connected through about five social connections (Travers and Milgram 1969; see also Dodds et al. 2003). Influence flows through these connections (for example, a person is more likely to be obese if the friends of that person’s friends are obese; Christakis and Fowler 2009), which is one of the reasons why word-of-mouth processes are of significant interest to persuasion professionals (Dichter 1966; Godes and Mayzlin 2004). Word-of-mouth marketing, after all, can be substantially more effective than traditional marketing (Trusov et al. 2009).
But as influence spreads through social networks, it can be observed that not all people are equally influential and some people are easier to influence than others. Opinion leaders influence others’ behavior by giving advice and direction (Flynn et al. 1996). Compared to other consumers, opinion leaders tend to be centrally located in their social networks, more cosmopolitan and socially active, more conscious about their appearance, more involved with mass media and, of course, more likely to share information with others (Katz 1957; Baumgarten 1975; Venkatraman 1989; Chan and Misra 1990; Weimann 1991; Rogers 2003). Understanding who opinion leaders are and what motivates them may improve the means through which policy makers, marketers, activists, health professionals and others change people’s behavior.1
To affect the greatest change, however, it is also important to understand the characteristics of people who seek to be influenced. Opinion seekers tend to be highly attentive to social comparison information and low in need for uniqueness (that is, relatively conforming; Bertrandias and Goldsmith 2006). They also tend to be less educated and lower in self-esteem than others (Pornpitakpan 2010). Some opinion seekers are also opinion leaders, although the correlation between opinion leadership and opinion seeking is weak (Flynn et al. 1996). Given that behavior change often requires a degree of readiness to change, it may be that understanding the characteristics of those who want to be influenced is every bit as important as understanding the characteristics of those who provide influence.
Bertrandias and Goldsmith (2006) have noted that an understanding of why some consumers want to influence or be influenced “should be found at a psychological level” (p. 280). Little research has focused on the motives of opinion leaders and opinion seekers, but understanding what energizes these consumers’ behavior (cf. Bargh et al. 2010) should assist professionals tasked with creating large-scale behavior change. One framework which emphasizes two motives of possible relevance to opinion leadership and opinion seeking is socioanalytic theory (Hogan 1983, 1996; Hogan and Holland 2003), a theory which recognizes how much human activity occurs within groups that are hierarchically structured. According to socioanalytic theory, two overarching motivations in work and life in general are getting along (achieving belonging) and getting ahead (achieving status). Digman (1997) obtained evidence that suggests how central individual differences in these motives may be in social living when he discovered two higher-order factors associated with the Big Five personality dimensions. The two factors were tentatively labeled α and β, but they essentially represented individual differences in tendencies to get along and get ahead.
In this study we examined the possibility that the motives which play such a central role in socioanalytic theory may account for individual differences in opinion leadership and opinion seeking. One of our specific hypotheses was that opinion seeking should be associated with belonging motivation. Conforming behaviors (and the behaviors that facilitate conformity, such as seeking information from others) are often used to gain others’ acceptance or avoid their rejection (cf. Deutsch and Gerard 1955). Even mimicking others’ gestures can facilitate harmonious interactions and increase liking (Chartrand and Bargh 1999). People who have an especially strong motivation to belong, therefore, may seek and use information from others to ensure that their own behavior is socially acceptable (cf. Williams et al. 2000).
A second hypothesis was that opinion leadership should be associated with status motivation. Lampel and Bhalla (2007) have argued that online opinion sharing (for example, posting product reviews) is motivated by status seeking; when opinion providers post their messages, those messages affect the level of influence and respect they achieve within virtual groups. Similarly, Roberts et al. (2006) found that computer program developers who voluntarily spend inordinate amounts of their time and effort contributing to open source software exhibit stronger status motivation.
In our analysis of trait and motivational predictors of opinion leadership and seeking, we focused on opinion leadership and seeking that is domain-general (cf. King and Summers 1970) rather than specific to a particular product or service domain (cf. Flynn et al. 1996; Richmond 1980). Although domain-specific tendencies are important in some marketing contexts (Flynn et al. 1996; Piirto 1992), one advantage of examining domain-general opinion leadership and opinion seeking is that doing so can help identify characteristics of consumers who are likely to be influential or influenceable (or both) in a variety of contexts (cf. King and Summers 1970). We followed the specificity-matching principle (Ajzen and Fishbein 1977; Epstein 1977) and matched general opinion leadership and seeking tendencies with general motivations (rather than specific goals) and a broad personality trait.
The personality trait we included in our analysis was self-monitoring (Snyder 1974, 1987), a trait that may envelop both belonging and status motivation. Early research on self-monitoring focused on how readily high self-monitors adapt their behavior to social situations (see Fuglestad and Snyder 2009). More recently, however, Gangestad and Snyder (2000) called for greater understanding of the motives associated with self-monitoring and drew particular attention to status motivation. Other recent work has shown that self-monitoring may also be associated with belonging motivation (Rose and DeJesus 2007; see also Day and Schleicher 2006). Given that belonging and status motivation may predict opinion leadership and opinion seeking, self-monitoring may represent a single trait that is positively associated with both providing and seeking influence.
Most of the evidence that self-monitoring is associated with a desire to attain dominant, respected positions within social hierarchies is indirect. Compared to low self-monitors, high self-monitors are more likely to emerge as leaders within organizations (Day and Schleicher 2006). High self-monitors are also more extraverted (Gangestad and Snyder 2000) and extraversion is characterized by dominance- and recognition-striving (Costa and McCrae 1988). In addition, when selecting dating partners, high self-monitors, more than low self-monitors, strongly value the physical attractiveness and social status of potential partners (Snyder et al. 1985; Jones 1993; see also Sigall and Landy 1973). High self-monitors are also more persuaded by a product’s image than the product’s content and quality (DeBono 2006) and are more sensitive to social cues when a product is a luxury product such as golf clubs rather than a necessity product like toothpaste (Brinberg and Plimpton 1986). High self-monitors mimic others’ gestures more when the other person is superior, compared to inferior, to them, whereas low self-monitors’ mimicry is not influenced by an interaction partner’s status (Cheng and Chartrand 2003, Study 2).
Although less attention has been paid to the link between self-monitoring and belonging motivation, Rose and DeJesus (2007) demonstrated in two studies that these two constructs are positively correlated. Hill (1987, 2009) showed that self-monitoring positively correlates with individual differences in four different affiliation goals that should relate to belonging motivation: seeking emotional support, seeking attention, seeking social comparison information and seeking positive social stimulation. In addition, Harris and Rosenthal (1986) found that high self-monitors are more susceptible to interpersonal expectancy effects than low self-monitors, presumably because high self-monitors are more concerned with behaving appropriately in social situations. Evidence that high self-monitors, compared to low self-monitors, use more second- and third-person pronouns (relative to first-person pronouns) during conversation (Ickes et al. 1986) is also consistent with the hypothesis that there is a relationship between self-monitoring and belonging motivation.
With the aim of identifying a single trait predictor of both opinion leadership and opinion seeking, our model also includes self-monitoring as an exogenous predictor of opinion leadership and opinion seeking. As implied previously, self-monitoring may be indirectly linked to opinion leadership and opinion seeking through status and belonging motivation. (The indirect path through status motivation is depicted in the upper part of Fig. 1; the indirect path through belonging motivation appears in the lower part.) If self-monitoring is associated with both opinion leadership and opinion seeking, it would suggest that large-group persuasion might be efficiently executed by focusing communications on people who are relatively high in self-monitoring.
Our hypotheses concerning mediation (that is, the indirect effects through status and belonging motivation) have important implications, especially because motivation has well-documented effects on cognition and attitude change (Kunda 1990; Wood 2000). Understanding what motivates a particular group of people allows targeted messages to be made more appealing and effective to that group.
Participants and Procedure
Two-hundred fifteen undergraduates (71 male, 131 female, 13 unspecified) at a public, southeastern university in the United States completed an anonymous online survey in exchange for extra credit. Most participants were European American/White (159). The remainder were African American (25), Middle Eastern American (6), Hispanic American (5), Asian American (4) or Native American (2). Fourteen participants did not describe their ethnicity.
The survey contained two quality-check items (for example, “If you are carefully reading this, please respond with the ‘neither’ answer in the middle of the scale”). Our data-quality rule, developed before the data were analyzed, led to the exclusion of seven participants’ data. In addition, several additional participants did not answer some questions, so n’s vary somewhat across analyses.
We used the updated 18-item, true-false version of the Self-Monitoring Scale (Snyder and Gangestad 1986; Gangestad and Snyder 2000) and found it to be acceptably internally consistent (α = .69). A sample item is “In different situations and with different people, I often act like very different persons.” The self-monitoring total score and all other total scores described below were created through summation (after appropriate transformation of any reverse-scored items). Thus, for all variables, higher scores reflect higher standing.
Belonging motivation was assessed with the 10-item Need to Belong Scale (Leary et al. 2003), which has been successfully used in a variety of published studies (for example, De Cremer and Leonardelli 2003; Pickett et al. 2004) and was internally consistent in this study (α = .80). Participants responded to statements such as “I try hard not to do things that will make other people avoid or reject me” on a 1-to-7 Likert-type scale.
Drawing on theorizing which posits that social status consists of (a) holding a dominant, influential position in a hierarchy as well as (b) having the admiration of others (Ridgeway and Walker 1995; Ridgeway 2001), we developed a composite measure of status motivation consisting of two subscales. The motivation-to-dominate subscale (α = .70) consisted of 10 dominance/influence items from the Interpersonal Adjective Scales (such as “assertive” and “dominant”; Trapnell and Wiggins 1990) preceded by the stem “I am the sort of person who strives to be…” and followed by a 7-point Likert-type scale. The motivation-to-be-admired subscale was the fame/social recognition component of Kasser and Ryan’s (1996) Aspiration Index. This scale (α = .89) consists of five items such as “to be admired by many people” (rated on a 7-point scale of importance as a long-term life goal). Total scores on each of the two subscales (which correlated positively, r  = .22, p = .002) were standardized and then summed to form a composite measure of status motivation. Although scores from the two subscales did not correlate very strongly, this is to be expected when two measures assess different sub-motives, or components (cf. Costa and McCrae 1995, pp. 22–23), of a broader motivation (see Spector 1992, p. 39). By creating a composite status motivation scale consisting of both dominance and admiration subscales, we strengthened the measure’s construct validity by ensuring that it assessed motivation toward an end-state characterized by both dominance/influence and admiration/social recognition (cf. Ridgeway and Walker 1995; Ridgeway 2001).
The opinion leadership scale we used (Flynn et al. 1996) consisted of six items (for example, “I often influence people’s opinions about what to buy”) to which participants responded on a seven-point Likert-type scale. The items referred to purchases in general and the scale was internally consistent (α = .79).
Flynn et al. (1996) also wrote and tested six opinion-seeking items which we used in this study (α = .81). These items also referred to purchases in general (for example, “I like to get others’ opinions before I purchase something”) and participants responded using a seven-point Likert-type scale.
Descriptive statistics and correlations among key variables
2. Status motivation
3. Belonging motivation
4. Opinion leadership
5. Opinion seeking
The model depicted in Fig. 1 requires two single-predictor regression analyses and two dual-predictor regression analyses. As already implied by the correlations, self-monitoring was a significant predictor of status motivation, β = .45, p < .001, but our hypothesis that self-monitoring would significantly predict belonging motivation was not supported, β = .11, p = .14. When considered in conjunction with prior research on the social motivations linked to self-monitoring (such as Rose and DeJesus 2007; and Gangestad and Snyder 2000), these two regressions suggest that self-monitoring may be more strongly linked to status motivation than to belonging motivation.
In the next regression analysis, status motivation (the primary predictor) and belonging motivation (a covariate) were entered as simultaneous predictors of opinion leadership. Status motivation was a significant predictor of opinion leadership, β = .37, p < .001; and although this was not predicted, belonging motivation was as well, β = .14, p = .048. Because we did not expect both predictors to be significant, we used a z-test for correlated correlation coefficients (Meng et al. 1992) to confirm that the part/semi-partial correlation between status motivation and opinion leadership (part r = .37) was significantly stronger than the part/semi-partial correlation between belonging motivation and opinion leadership (part r = .14), z = 2.33, p = .01. These results imply that opinion leadership may emerge out of desires to achieve social status and social acceptance, but the desire for social status may play a stronger role.
In the fourth regression analysis, we found, as hypothesized, that belonging motivation was a significant predictor of opinion seeking, β = .17, p = .03, but status motivation (a covariate in this analysis) was not β = .03, p = .68. Thus, while opinion leadership may emerge most strongly out of status motivation, opinion seeking may emerge most strongly out of belonging motivation. These results support our earlier suggestion that the sociomotivational origins of opinion seeking and opinion leadership differ.
Our hypothesized model proposes that status motivation mediates the association between self-monitoring and opinion leadership (see the upper portion of Fig. 1). (Although the figure might be interpreted to imply that belonging motivation also mediates the association between self-monitoring and opinion leadership, we did not consider such an indirect effect theoretically plausible. The path from belonging motivation to opinion leadership serves only to show that the effect of status motivation on opinion leadership was examined independently of any possible effect of belonging motivation.) To provide initial evidence that status motivation may mediate the association between self-monitoring and opinion leadership (cf. Baron and Kenny 1986), we simultaneously regressed opinion leadership on self-monitoring, status motivation and, to control for this variable, belonging motivation. Status motivation (the mediator) was significant, β = .31, p = <.001, belonging motivation (a covariate) was only marginally significant, β = .12, p = .095, and, as would be expected if the mediation is full rather than partial, self-monitoring was not significant, β = .12, p = .15. These results suggest that status motivation accounts for much of the shared variance between self-monitoring and opinion leadership.
To test the significance of the indirect effect of self-monitoring on opinion leadership through status motivation, we used Preacher and Hayes (2008) bootstrapping technique for mediation models. Unlike earlier mediation methods (for example, Sobel 1982; Baron and Kenny 1986), the Preacher-Hayes technique does not rely on the questionable assumption that the total and indirect effects are normally distributed. In addition, the technique allows multiple mediators to be evaluated and produces point estimates and bias-corrected and accelerated (BCA) confidence intervals (see Efron 1987) for each indirect effect. A confidence interval that does not include zero suggests that the indirect effect is significant and mediation may be present.
The indirect effect of self-monitoring on opinion leadership through status motivation was significant, with a point estimate of .26 and a 95% BCA confidence interval of .08 to .45. (In contrast, the indirect effect through belonging motivation was not significant, with a point estimate of .02 and a 95% BCA confidence interval of −.01 to .10.) Combined with the results of the earlier regression analyses, these indirect effect tests suggest that status motivation, not belonging motivation, significantly mediates the self-monitoring-opinion-leadership relationship.
Had a positive correlation between self-monitoring and opinion seeking emerged, the next step in our analyses would have been to examine whether belonging motivation mediates this relationship. However, as noted, self-monitoring and opinion-seeking were not significantly related (see Table 1), nor was self-monitoring significantly related to belonging motivation. Belonging motivation (the proposed mediator) was a significant predictor (β = .18, p = .03) of opinion-seeking in a simultaneous regression in which self-monitoring (β = .10, p = .25) and status motivation (β = −.02, p = .82) were included, but a significant relationship between the mediator and the dependent variable is not sufficient to suggest that mediation is present (Baron and Kenny 1986; Preacher and Hayes 2008). Not surprisingly, the indirect effect of self-monitoring on opinion seeking through belonging motivation was not significant (point estimate = .02, 95% BCA confidence interval = −.02 to .10), nor was the indirect effect through status motivation (point estimate = −.02, 95% BCA confidence interval = −.17 to .13). Thus, although belonging motivation significantly predicted opinion seeking, self-monitoring had neither a direct nor indirect effect on opinion seeking.
When mediated relationships are hypothesized, it is prudent to examine plausible alternative pathways to those proposed. In this study, it seemed appropriate to consider whether the relationships between the motives we measured (status, belonging) and the opinion tendencies we measured (leading, seeking) might be mediated by self-monitoring. Most personality traits are at least partially heritable (including self-monitoring; Gangestad and Simpson 1993), manifest early in life and influence a broad range of specific behaviors, so we do not consider it especially plausible that self-monitoring mediates any associations between the motives we measured and the opinion tendencies we measured. Nevertheless, it is debatable whether motives arise from traits or traits arise from motives. To summarize briefly the evidence against self-monitoring as a mediator, analyses already reported showed that when opinion tendencies (leading, seeking) were simultaneously regressed on motives (status, belonging) and self-monitoring, the motives we expected to be significant were significant and self-monitoring was not. Moreover, there was no evidence in indirect effect tests of self-monitoring as a mediator.
The primary aims of this study were to examine whether self-monitoring predicts opinion leadership and seeking and whether these relationships are mediated by social motives. We hoped to identify a single trait that is linked to both opinion leading and opinion seeking and perhaps most important, we hoped to better understand the sociomotivational antecedents of opinion leading and seeking. Our results suggested that self-monitoring is linked to opinion leading (as well as status motivation), but not opinion seeking (or belonging motivation). Also, status motivation mediated the relationship between self-monitoring and opinion leading, which suggests that consumers who are high self-monitors may share their opinions with others because of their motivation to achieve high social status within the marketplace.
Although belonging motivation was hypothesized as a mediator of the self-monitoring-opinion-seeking relationship, no relationship between self-monitoring and opinion seeking (or between self-monitoring and belonging motivation) emerged. A priority in future research should be identifying trait predictors of opinion seeking. Given our finding that belonging motivation correlates with opinion seeking, a focus on traits that may be associated with a strong need to belong (for example, agreeableness and some of its components) may be fruitful. An additional priority in future research should be to confirm that belonging motivation is linked not only to opinion seeking (a result we predicted and found) but also opinion leadership (a result we found but did not predict). We consider this hypothesis tentative, but high belonging motivation may be a characteristic that some opinion leaders and some opinion seekers have in common.
Our study confirms the hypothesis (see Gangestad and Snyder 2000) that high self-monitors are motivated to attain high social status. The pattern of relationships that emerged suggests that getting ahead may be more important to high self-monitors than getting along. Prior research has shown small-to-moderate relationships between self-monitoring and belonging motivation (Hill 1987; Rose and DeJesus 2007), but in our study the relationship was not significant (even though it was in the hypothesized direction). The true relationship between self-monitoring and belonging motivation may be small.
Our results concerning the motivational correlates of self-monitoring may suggest that a slight shift in our understanding of this trait is in order. Traditional self-monitoring research (Snyder 1974, 1987) has focused on the tendency of high self-monitors to modify their behavior to meet situational demands. High self-monitors were described in early work as individuals who are susceptible to others’ influence in a variety of situations. With this perspective in mind, we hypothesized, but then did not find, that high self-monitors are more likely than low self-monitors to seek information from others to meet their motivation to belong. What we found, however, was that self-monitoring is more clearly linked to status motivation and opinion leadership. It may be less accurate to think of high self-monitors as modifying their behavior in follower-like roles than to think of high self-monitors as modifying their behavior in order to rise within social hierarchies.
An important limitation of our correlational study was that we could not identify causal relationships. Even though we statistically evaluated the pathways we predicted as well as alternative pathways, without experimental evidence, the support we obtained for some of our hypotheses should be interpreted with caution. Furthermore, for theoretical and practical reasons, we focused only on one trait (self-monitoring) and two motives (status seeking and belonging motivation) to predict opinion leadership and opinion seeking. We hope that future research will identify additional traits and motives underlying opinion leadership and opinion seeking and thus suggest more effective means of leveraging peer-to-peer communication to create large-scale behavior change.
Market mavenism (Clark and Goldsmith 2005; Feick and Price 1987), which is characterized by a high level of involvement with marketplace information and a strong interest in novel products and services, is similar to opinion leadership. Hoyer and Stokburger-Sauer (2007), however, have documented some ways in which market mavens and opinion leaders differ.