Do superdiffusers argue differently? An analysis of argumentation style as a function of diffusion ability.
In addition to the practical need to examine the methods of influentials, due to a dearth of research in the area there is value in advancing argumentation research by examining how different persuasive situations and individual differences combine to affect how it is that influentials convince others (Hample, 2005, p. 183). The recent development of an extensively validated and highly reliable measure of three dimensions that assess the extent to which one has the characteristics of an opinion leader facilitates the exploration of what makes persuasive attempts by opinion leaders different from the persuasion attempts of those who are not opinion leaders (Boster, Kotowski, & Andrews, 2006).
Diffusion scholars have demonstrated that the diffusion of a new practice or product is often dependent on particular, unique people in a network (Bullet, et al., 2001; Kelley et al., 1997; see Rogers, 2005, p. 308-330; Valente & Pumpuang, 2007 for reviews). They are variously known as influentials, opinion leaders, champions, change agents, and peer leaders, and they are able to persuade many people to change their behavior much more quickly than others. Previous research has identified these people in a number of ways including sociometric methods, asking those who are knowledgeable about a network, and by the construction of self-report surveys (Valente & Pumpuang). The self-report survey promises to be a highly reliable and easily implemented method of identifying them, although previous measures have failed to measure all of the important facets of the construct (Boster et al., 2006).
To improve upon the self-report method, Boster et al. (2006) presented evidence consistent with a three-dimensional conceptualization of opinion leaders. They argue that someone who can persuade many people to change their behavior, here termed a superdiffuser, must be high on each of the three dimensions described subsequently.
First, such a person must be a connector. Connectors are people who enjoy meeting new people, maintaining contacts with many others, and who often provide a link between disparate groups. These characteristics give them access to many people in many different groups, greatly increasing their reach and influence.
Boster et al. (2006) argue that the second characteristic required is to be a persuader. Persuaders are highly influential because they are skilled at argument. They enjoy arguing and seek out opportunities to influence others. They are able effectively to frame arguments for particular audiences. Boster et al. presented evidence consistent with their concept of persuaders as strongly correlated with, but conceptually distinct from, trait argumentativeness.
The third required characteristic is to be a maven (Boster et al., 2006). Mavens are those who have in-depth knowledge of a particular content area, such as healthy lifestyles, politics, or particular kinds of consumer goods. Mavens are recognized by others as experts in a given area. Others come to them for advice in the content area in which they specialize. They also enjoy helping others learn about their area of expertise. They are particularly effective diffusers in their area of specialization because when they recommend a new idea or practice, their expertise helps them construct compelling arguments that are likely to persuade others to adopt it.
Boster et al. (2006) provided evidence consistent with the hypothesis that these three dimensions are able to capture aspects of opinion leadership that previous measures have failed to include. The construction of a highly valid scale to identify superdiffusers allows detailed examination of how the persuasive strategies that superdiffusers use are different from those who are less influential. Previous research from the compliance gaining literature concerning argument production allows the generation of hypotheses concerning the types of persuasive strategies that superdiffusers might use. This analysis will review the research on the relationship between individual differences and compliance gaining strategies prior to examining the ways in which these dimensions of individual difference and variation in persuasive context combine to affect the generation of arguments.
Research on compliance gaining message production found that both personality traits and cognitive traits are related to different kinds of compliance gaining message production. Many studies have tried to use personality traits to predict the different argumentation strategies people report they would use or the strategies they use in an interaction (Boster & Stiff, 1984; Canary, Cody, & Marston, 1986; Lusting & King, 1980; see Wilson, 2002, pp. 99-102 for a review). Argumentativeness is one of the strongest predictors of the ways people will try to gain the compliance of others (Boster, Levine, & Kazoleas, 1993; Levine & Boster, 1996). Boster et al. (1993) found that argumentativeness is associated positively with strategy diversity in actual compliance gaining interactions. Past research examining superdiffusers has consistendy found that the extent to which one is a connector, persuader, and maven are all associated positively with argumentativeness, especially the persuader scale (Boster et al., 2006). Therefore, these measures are expected to associate positively with the diversity of arguments that people produce in a compliance-gaining context.
Like personality traits, a number of cognitive traits have been used to predict the kinds of compliance gaining messages people produce (Hample, 2003; O'Keefe, 1988; see Hample, 2005, pp. 176-240 for a review). One important trait is inventional capacity, which Hample defines as the extent to which people vary in the number of possible compliance gaining messages produced. When asked to generate as many ways as possible to change a target's behavior, Hample (2003) found that some people generate many messages whereas others generate very few.
Given the conceptualization of a superdiffuser as someone highly connected, very persuasive, and highly expert in a content domain, this study predicts that the inventional capacity of superdiffusers is higher than that of non-superdiffusers. By virtue of their breadth of association, this analysis suggests that connectors have been exposed to a greater number and variety of arguments in a given domain, and predicts that persuaders have developed a wider repertoire of persuasive messages and tactics so that they can adapt their arguments to a variety of audiences. Mavens are expected to have a wider repertoire of arguments for their specialty based on their wide knowledge base in that domain.
This study will also make predictions concerning how superdiffusers might respond to different argumentation contexts. Several lines of research have examined the impact of the complexity of the persuasive task on the diversity of arguments that are produced (Hample & Dallinger, 2002; O'Keefe, 1988). One feature of the complexity of the persuasion task is the extent to which the persuasion target resists the influence attempt. Pertinent to this issue Hample and Dallinger's secondary analysis of nine compliance gaining studies found that anticipated resistance was one of the strongest predictors of the number of strategies that participants claim they would use to change a target's behavior. Thus, this study predicts that superdiffusers will respond to more complex compliance situations with greater argument diversity.
When explicating the concept of message design logics, O'Keefe (1988) noted that when dealing with tasks in which persons only need to pursue one simple goal, individual differences in message construction ability are unrelated to the kinds of messages produced. Conversely, when goals become more difficult to achieve, people with more complex message production ability distinguish themselves by producing more, and more complex, messages. Thus, this study predicts that when faced with a relatively simple compliance gaining situation, superdiffusers will not produce messages that differ in quality or quantity from non-superdiffusers, but when the compliance gaining goal becomes more complex superdiffusers will produce more complex messages than non-superdiffusers using a wider diversity of argument strategies.
Given the preceding reasoning, if persons are presented with compliance gaining tasks that vary in complexity, those identified as superdiffusers and those identified as non-superdiffusers by the Boster et al. (2006) measures will be similar in the arguments that they generate to persuade a target person when the task is a relatively simple one. In contrast, however, it is expected that superdiffusers will produce more arguments, more varied arguments, and arguments more sensitive to context than non-superdiffusers when the task is relatively complex. An experiment was designed employing two topics, smoking and obesity, to examine these predictions.
Sample. Participants were 164 students (50 male, 114 female) from a large Midwestern university. They were recruited from undergraduate communication courses and they were given course credit in exchange for their participation. Their mean age was 19.42, SD = 1.47.
Boster et al. (2006) found that student and non-students samples responded to the connector, maven, and persuader items in the same manner. The same factor structure fit the data, standard deviations were similar, and approximately the same percentage of the sample was high on all three measures, although the means differed slightly. This study explores the relationships among task difficulty, superdiffuser status, and several dependent variables. Lacking a reason to believe that this relationship is in any way affected by demographic characteristics, such as the age or the education level of the sample, a more diverse sample was not sought.
Procedures. After the participants had signed up for the experiment, they were given a link to log on to an online data collection website where they viewed one of the scenarios concerning persuading a young woman to adopt a healthier lifestyle and responded to a set of items pertinent to them. The first page asked for their sex and age. Measurement order was counterbalanced so that one-half of the participants filled out the connector, persuader, and maven (CPM) scales first (Boster et al., 2006) and then were exposed to one of the compliance scenarios and the associated questions about how they would respond to it. The other participants were given the scenario and associated questions before filling out the CPM scales.
Design. This experiment employed a 2 (complexity of persuasive task: simple v. complex) X 2 (topic: avoid smoking v. lose weight) X 2 (question order: CPM scales before or after exposure to the scenarios) independent groups design. Participants were assigned randomly to conditions. The complexity of the persuasive task was induced by making the proposed target more or less resistant to persuasive attempts and indicating that she either had been receptive or not to the goal of the influence attempt in the past (see Appendix A for all four scenarios). The topic was either to convince the target to avoid smoking cigarettes (not starting in the simple task condition, quitting in the complex task condition) or to lose weight (some weight in the simple task condition, losing 90 pounds in the complex task condition). Two separate topics were chosen to increase the generalizability of the results. These topics both cover serious health risks for college age students (Lowrey et al., 2000; Rigotti, Lee, & Wechsler, 2000).
Instrumentation. The CPM scales were taken from Boster et al. (2006). Responses to the items were made on 8-point Likert response scales anchored by the phrases "disagree strongly" and "agree strongly." The maven items focused on health so as to identify health mavens, i.e., those knowledgeable about healthy lifestyles. Responses to each of the three measures were averaged to form a connector index, a persuader index, and a health maven index for each participant (the items for the CPM scales and their factor loadings can be found in Table 1).
For the smoking topic the distribution of the connector index was skewed modesty in a negative direction (-.60) with [alpha] = .92, M = 5.47, and SD = 1.58. For the weight loss topic the distribution of scores the connector scores exhibited a modest negative skew (-.57) with [alpha] = .90, M = 5.38, SD = 1.39.
For the smoking topic the distribution of persuader scores was skewed negatively (-1.16) and leptokurtic with [alpha] = .94, M = 5.77, and SD = 1.34. For the weight loss topic the distribution of persuader scores was slightly negatively skewed (-.51) with [alpha] = .93, M = 5.40, and SD = 1.21.
For the smoking topic the distribution of the maven index exhibited a modest negative skew (-.69) with [alpha] = .93, M= 5.14, and SD = 1.65. For the weight loss topic the distribution of the maven index approximated closely the normal distribution (skew = -.04) with [alpha] = .91, M = 4.92, and SD = 1.49.
Thus, the scales demonstrated ample reliability for both topics. Moreover, confirmatory factor analysis replicated the three-factor structure found previously (Boster et al., 2006) with a low RMSE of .08 indicating a good fit of the model to the data.
After the scenario the participants were first asked if they would try to persuade the target. If the answer was yes, they were then asked what they would say to the target to convince her. Of the 164 participants who participated, 142 (86.6%) said that they would try to persuade the target.
Three trained undergraduate raters and two of the authors coded the various open-ended responses. The undergraduates were trained together by reviewing the instructions for coding each variable with one of the authors and resolving any disagreements about how to define each variable. Each rater coded all of the responses and worked independently. In cases where a rater's responses lowered the inter-rater reliability coefficient, that rater's counts were dropped for that variable. For the dichotomous items inter-rater reliability was assessed with Cohen's Kappa (K), and disagreements were resolved by discussion. For continuous items inter-rater reliability was estimated with Ebel's Coefficient (EC; Ebel, 1951) and responses were averaged (mean).
Responses to the question of what they would say to the target were coded for the number of arguments (smoking scenario M = 2.20, SD = 1.55, EC = .94; weight loss scenario M = 1.34, SD = .89, EC = .73; see Table 2 for correlations among the dependent variables for each topic). The instructions for the raters for this variable stated, "Here we want you to examine how many separate arguments the participant listed to persuade the target to pursue a healthy lifestyle." They also coded the number of distinct themes in the arguments (smoking scenario M = 1.30, SD = .74, EC = .82; weight loss scenario M = 1.32, SD = .71, EC = .88). For the number of themes the instructions stated, "Here we want you to count how many overall themes they covered. A theme is a broad type of argument used to change someone's mind and may contain several arguments." The coders also coded the number of references to elements of the scenario (avoid smoking scenario, M = .76, SD = .68, E.C. = .91; weight loss scenario, M = .55, SD = .66, EC = .73). For this variable, the raters were told to count, "The number of parts of the scenario they mentioned."
There were several types of arguments that participants said they would use to persuade the target. Each topic (avoid smoking or lose weight) produced different types of arguments. The most common argument for the smoking scenario was that smoking causes health problems (47% used this argument, Percentage Agreement = 88%, [kappa] = .67). The most common argument for the weight loss scenario was an offer to exercise with the target (36% used this argument, Percentage Agreement = 93%, [kappa] = .84).
The responses of the participants who said they would not try to persuade the target were coded as zero for both the number of arguments and the number of distinct themes. Because those who produced arguments but did not reference the scenario were qualitatively different than those who simply did not try at all, only the responses of participants who said that they would try to persuade the target were analyzed for the number of elements of the scenario that were addressed.
Superdiffusers were defined as those whose scores were above the 75th percentile on each of the three CPM scales. To be considered a superdiffuser among the participants who responded to the smoker scenario, a participant had to score more than 6.80 on the connector scale, 6.75 on the persuader scale, and 6.38 on the maven scale. To be considered a superdiffuser among the participants who responded to the weight loss scenario, a participant had to score above 6.20 on the connector scale, 6.25 on the persuader scale, and 6.16 on the maven scale. Among the participants responding to the smoking scenario, there were two people in the simple task condition and two people in the complex task condition who met this criterion. There was only one person identified as a superdiffuser in the simple weight loss condition and five people identified as superdiffusers in the complex weight loss condition. Because the random assignment process resulted in only one superdiffuser being assigned to the simple weight loss topic condition, some formal statistical analyses pertinent to this topic were precluded.
Number of arguments. There was no effect of question order for any dependent variable; thus, it was dropped from all subsequent analyses. For the number of arguments produced in response to the smoking scenario, the main effect for the complexity of the scenario was substantial, F(1, 78) = 4.05, p = .05, r = .22, d = .44, such that more arguments were generated by Ss in the complex scenario (Weighted: M = 2.24, SD = 1.89, Unweighted: M = 3.63, SD = 3.50) than in the simple scenario (Weighted: M = 2.15, SD = 1.15, Unweighted: M = 2.08, SD = 3.50). (1) The main effect for superdiffuser status did not reach the conventional .05 level of statistical significance, but was of similar magnitude, F(1, 78) = 3.57, p = .06, r = .20, d = .40, such that more arguments were generated by superdiffusers (Weighted = Unweighted: M = 3.58, SD = 2.04) than by non-superdiffusers (Weighted = Unweighted: M = 2.12, SD = 1.50). There was no evidence that the number of arguments was correlated more than would be expected by sampling error with the connector and persuader measures (connector r = .10, t(80) = .90, ns, persuader r = .18, t(80) = 1.64, ns). There was evidence of linear association between the number of arguments generated and the health maven measure (r = .36, t(80) = 3.45, p = .001), such that those higher on the health maven index generated more arguments.
The main effects were qualified by an important interaction effect, F(1, 78) = 4.37, p = .04. Examining the means and standard deviations in Table 3 shows that the two participants in the complex/superdiffuser condition produced two and one-half times more arguments than the average participant in any other condition, none of the means in the other three conditions differing substantially from each other. To illustrate the matter in a different way, number of arguments (A) was regressed onto the dichotomous variable, superdiffuser status (S), for those Ss in the simple scenario condition and those in the complex scenario condition separately. These analyses indicated that although there was no effect of superdiffuser status on argument generation in the simple scenario condition (regression equation: A = 2.15 .15 S, r = -.03, d = -.06), there was a substantial effect in the complex scenario condition (regression equation: A = 2.09 + 3.07 S, r = .36, d = .76). To estimate the magnitude of this interaction effect a contrast analysis was employed, and the sampling distribution statistics were transformed to an effect size estimate. The outcome of this analysis indicated that the effect was substantial (t(71) = 2.85, p = .006, r = .31, d = .76), particularly when corrected for the radical unequal split in the superdiffuser measure (r' = .72; see Hunter & Schmidt, 2004, p. 280).
For the participants who responded to the weight loss scenario, superdiffusers produced more arguments (m= 2.17, SD = .81) than non-superdiffusers (m= 1.32, SD = .87), t (79) = 2.31, p = .02, r = .25, d = .52, r' = .44, by a factor of 1.64. The main effect of scenario complexity was neither statistically significant nor substantial. For this scenario, there was no evidence of linear association between the number of arguments and the connector and persuader measures (connector r = .10, t(79) = .89, ns, persuader r = .01, t(79) = .09, ns). There was again evidence of positive association with the health maven measure (r = .22, t(79) = 2.00, p = .05).
Number of themes. A similar pattern was found in the count of the number of different themes the participants produced in their attempt to persuade the target in response to the smoking topic. The main effect for the complexity of the task was again statistically significant, F(1, 78) = 3.57, p = .06, r = .20, d = .40, and such that slightly more diverse themes were generated by participants in the simple scenario (Weighted: M = 1.38, SD = .51, Unweighted: M = 1.20, SD = 1.61) than in the complex scenario (Weighted: M = 1.23, SD = .91, Unweighted = 1.99, SD = 1.61). The main effect for superdiffuser status did not reach the conventional .05 level of statistical significance, but was similar in magnitude, F(1, 78) = 3.25, p = .08, r = .20, d = .40, and was such that more themes were generated by superdiffusers (Weighted = Unweighted: M = 1.92, SD = 1.07) than by non-superdiffusers (Weighted = Unweighted: M = 1.27, SD = .71). There was no evidence that the number of themes was correlated with the connector index (connector r = .09, t(80) = .81, ns), but there was evidence of positive association with the persuader (r = .28, t(80) = 2.61, p = .01) and health maven measures (r = .39, t(80) = 3.79, p = .0004).
Again, the main effects were qualified by an important interaction effect, F(1, 78) = 8.57, p = .004. Examining the means and standard deviations in Table 4 shows that the two participants in the complex/superdiffuser condition produced twice as many different themes as the average participant did in any of the other three cells in the design, the other three means not differing substantially from one another. Put differently, although there was no difference between superdiffusers and non-superdiffusers in the simple scenario condition (regression equation: T = 1.40 - .40 S, r = -.17, d = -.34, where T denotes number of themes), superdiffusers generated substantially more themes than non-superdiffusers in the complex scenario condition (regression equation: T = 1.15 + 1.69 S, r = .41, d = .89). To estimate the magnitude of this interaction effect a contrast analysis was employed, and the sampling distribution statistics were transformed into an effect size estimate. The outcome of this analysis indicated that the effect was substantial (t(71) = 3.45, p = .001, r = .33, d = .70, r' = .75). This outcome indicates that not only do superdiffusers produce more arguments; they also cover more topics when they are faced with a complex persuasion task.
Superdiffusers who responded to the weight loss scenario did not produce substantially more themes than non-superdiffusers. The main effect for the complexity of the scenario was substantial, (t(68) = 2.05, p = .04, r = .24, d = .50). The participants in the simple condition produced fewer distinct themes in their responses (M = 1.35, SD .41) than the participants in the complex condition (M = 1.59, SD = .55). For this scenario, there was no evidence that the number of themes was correlated with the individual CPM scales to an extent greater than what would be expected by sampling error (connector r = .08, t(79) = .71 ns, persuader r = .10, t(79) = .89, ns, health maven r = .20, t(79) = 1.81, p = .08).
Number of references to the scenario. Participants referenced more elements of the scenario when they were responding to the complex smoking topic. The main effect for complexity was statistically significant, F(1, 69) = 16.82, p < .001, r = .20, d = .40, such that slightly more references to the scenario were made by participants in the complex scenario (Weighted: M = .78, SD = .80, Unweighted: M = 1.66, SD = 1.40) than in the simple scenario (Weighted: M = .75, SD = .58, Unweighted: M = .4, SD = 1.40). There was also a notable main effect for superdiffuser status F(1, 69) = 3.91, p = .05, r = .20, d = .40, such that more references were made to the scenario by superdiffusers (Weighted = Unweighted: M = 1.33, SD = 1.54) than by non-superdiffusers (Weighted = Unweighted: M = .73, SD = .61). There was no evidence of association between the number of references and any of the CPM scales (connector r = .06, t(71) = .51, ns, persuader r = .10, t(71) = .85, ns, health maven r = .14, t(71) = 1.19, ns).
Moreover, the interaction was ample, F(1, 69) = 20.55, p < .001 (see Table 5 for the means and standard deviations). Examination of the cell means shows that superdiffusers in the complex task condition addressed more than two and one-half times more of the specifics of the scenario than any other group, the other three means not differing substantially from one another. Although there was a negative effect in which the two superdiffusers in the simple condition did not reference the scenario at all (regression equation: R = .79 - .79 S, r = -.30, d = -.63, where R denotes the number of references) superdiffusers generated more arguments than non-superdiffusers in the complex scenario condition (regression equation: R = .66 + 2.01 S, r = .61, d = 1.54). To estimate the magnitude of this interaction effect, a contrast analysis was again employed, and the sampling distribution statistics were transformed to an effect size estimate. The outcome of this analysis indicated that the effect was substantial, t(71) = 4.49, p = .00003, r = .47, d = 1.07, r'= .87).
Superdiffusers who responded to the weight loss scenario did not make substantially more references to the scenario than non-superdiffusers. Furthermore, the main effect for scenario complexity was neither substantial nor statistically significant. For this scenario, there was no evidence of association between the number of references and the CPM scales (connector r = 0, t(69) = 0, ns, persuader r = .02, t(69) =. 17, ns, health maven r = .09, t(69) = .75, ns).
The data are consistent with the hypothesis that superdiffusers argue differently than people who are not superdiffusers. The data are also consistent with the hypothesis that superdiffusers are more likely to distinguish themselves from those who are not superdiffusers when the persuasion task is complex than when it is simpler.
Superdiffuser status. Unfortunately it was not possible to examine how superdiffuser status interacted with task complexity for the weight loss scenario, although the finding that superdiffusers produced more arguments than non-superdiffusers is encouraging. The results for the smoking scenario were consistent with the hypothesis that superdiffusers would argue differently when the persuasion task was complex. This pattern emerged for both the number of arguments the participants made and the number of distinct themes they covered. This outcome is consistent with the hypothesis that when the task is complex, superdiffusers tap their wider repertoire of arguments to persuade the target. Also, insofar as the tendency to refer to the scenario is an indication of audience adaptation, superdiffusers made a greater attempt to adapt their message to the specifics of the target when the task was complex. By addressing more elements of the scenario, the superdiffusers were demonstrating superior listener adaptation. These results are consistent with the hypothesis that superdiffusers' effectiveness stems from their ability to adapt to a complex target by producing a wider repertoire of arguments that address the target's unique situation.
Limitations. The complexity of the task requires a more precise induction in future research. Examination of the differences between the simple and complex scenarios reveals a number of possible ways that the complexity of the task may have been induced. Future research would profit by attempting to systematize the ways that a persuasive task's complexity can be induced and which elements combine non-additively with the superdiffuser scales to produce different argumentation strategies. This study lacked an induction check for the degree to which the participants perceived the task to be complex. Although future work might include one in order to help determine which elements of a persuasive situation cause superdiffusers to change their argumentation strategy, such a measure could be reactive and investigators would be served well by examining this possibility.
This study was also limited in its focus on health topics. Although the maven scale items focus on knowledge of healthy lifestyles, the items can be adapted to any number of topics. Future work would profit by using different versions of the maven scale and asking the participants to argue for positions in other areas than health issues. This study was also limited in that it employed scenarios instead of face-to-face
interaction. Participants may react differently when they are forced not only to produce arguments, but also to address counterarguments in an interactive situation. Unlike face-to-face interactions, self-report measures allow more time for deliberation and rely on the participants' cognitive skill in imagining the scenario (Miller, Boster, Roloff, & Seibold, 1987). Furthermore, the superdiffuser measures were designed to predict individual ability to diffuse information in interpersonal interactions so that a better examination of the argumentation abilities of superdiffusers would require an interpersonal setting. Thus, Study 2 was conducted to determine if superdiffusers distinguish themselves from non-superdiffusers in a face-to-face context. By creating a situation that allows for, and in fact demands, elaboration one is better able to ascertain how superdiffusers' thinking is translated into argument. The elaboration of this process might have the additional benefit of improving the training of non-superdiffusers.
Sample. The second study sampled 35 female students who were at least 18 years old and enrolled in undergraduate Communication classes at a large Midwestern university. These students received compensation for their participation in the form of course credit. Participants were on average 19.66 (SD = 1.63) years old, in the 2.31 (SD = 1.11) year of college, with 23% reporting that they had held at least one leadership position since beginning college.
Procedures. Participants volunteering to participate in a study described as an investigation of communication and social behavior arrived individually at a laboratory for a scheduled session lasting approximately 45 minutes per participant. Participants were greeted by an experimenter (E) and seated at a desk. The E informed the student that participation involved the completion of two main tasks. The first task involved the completion of a questionnaire containing several measures of social opinion and social behavior, which took approximately 15 minutes to complete. The second task was a structured audio taped interview of the participant focusing on healthy lifestyles. This task took approximately 30 minutes to complete.
After informing the participant of the tasks, the questionnaire was administered and a brief set of instructions regarding how to complete the items on the questionnaire was reviewed before allowing the participants to begin. After the questionnaire was completed and returned to the E, the tape recorder was started and the participants were interviewed on their position towards dieting according to the protocol in Appendix B. After completing both tasks participants were dismissed.
Instrumentation. Several demographic items were administered, as well as the connector, persuader, and health maven measures (Boster et al., 2006), responses to the latter three measures being assessed on 100 point response scales. Connector, persuader, and healthy lifestyle maven indices were formed by calculating the mean response across all items on each factor. A confirmatory factor analysis showed that the items again replicated the three-factor structure (factor loadings for each item are shown in Table 1). Though the RMSE of .14 for the CFA is a bit higher than would be desirable, it is probably attributable to the small sample size rather than poor fit given the smaller fit indices found in the studies with larger samples in Boster et al. The distribution of connector scores approximated closely the normal distribution with a mean of 46.09 (SD = 21.84). The reliability of this index was estimated as [alpha] = .91. The distribution of persuader scores also approximated the normal distribution, M = 63.21, SD = 13.82, [alpha] = .91. The distribution of the maven index approximated the normal distribution, M = 50.33, SD = 22.10, [alpha] = 91.
Interview. Argumentative complexity was measured by the participant's interview responses. Participants were interviewed by the E for their position on the issue of dieting. Dieting was chosen because a pretest sample indicated that dieting was a substantial health concern of the female population that was sampled. Consequently, participants would be likely to have relatively crystallized attitudes towards dieting. These extant attitudes were important because interview questions asked the participants for detailed reports of reasons for holding their position. Furthermore, given that the interview asked participants to report counterarguments against their position, it was also important that an issue was chosen in which arguments and evidence both for and against the issue exist in abundance.
The structure of the interview was an adaptation of Kuhn's (1991, pp. 299-300) protocol which elicited responses indicative of several components of argumentative complexity. Kuhn (pp. 14-20) identified several components of argumentative complexity through a content analysis of 160 interviews taken from people across four age groups (teens, 20s, 40s, 60s), which asked participants to argue for their position on social problems such as crime, failure in school, and unemployment. In addition to identifying several concepts related to the participants' cognitive complexity, the analyses revealed six components particular to argumentative complexity: causal theories, evidence, alternative causal theories, alternative evidence, counterargument, and rebuttal.
Causal theories are chains of premises that proceed logically to the respondent's position on an issue. The quality of a causal theory, according to Kuhn (1991, pp. 28-36) is based on the length of the causal theory judged by the number of premises and the amount of integration between parallel causal theories. For example, it is possible that a participant could have multiple chains of premises leading to the same conclusion. The extent to which the multiple chains share premises is equivalent to the amount of integration. The more premises in the chain and the greater the amount of integration between chains were found by Kuhn (1991, p. 43) and Campo (1999) to be indicative of greater argumentative complexity. Alternative causal theories can be described in a similar fashion except that the participants' alternative causal theories argue for a position on an issue opposing their own (Kuhn, p. 97).
The third component of argumentative complexity is the evidence respondents provided to justify their position on an issue. Kuhn (1991) categorized evidence based on strength into three forms: genuine evidence (p. 45), pseudoevidence (p. 65), and nonevidence (p. 81). Genuine evidence, being the strongest of the three forms, is pertinent and can stand apart from the participant's causal theory. For example, a participant's statistical claim of a correlation between a cause and effect would be classified as genuine evidence. Pseudoevidence is a scenario or script illustrating how an issue might occur. A participant's anecdote used as evidence for why an outcome occurred would be classified as pseudoevidence. Finally, the weakest form of evidence according to Kuhn (pp. 81-82) is nonevidence. A participant provides nonevidence when it is implied that evidence is not necessary for the position taken, or, if in a circular manner, the position itself is claimed as evidence. The greater the amount of genuine evidence and pseudoevidence to nonevidence given by the participant indicates greater argumentative complexity according to Kuhn. Alternative evidence is defined by the same categories with the exception that the evidence is provided as justification for a position on an issue opposing one's own.
The final two components identified by Kuhn (1991) are counterarguments (p. 117) and rebuttals (p. 145). Counterargument is the participants' ability to generate conditions that falsify the reasons for why they hold their position on an issue. The strength of a counterargument is the extent to which it falsifies the participants' own reasons for their position on an issue. Rebuttals on the other hand are the participants' attempt to rebuff counterarguments and alternative causal theories opposing their position. To the extent that the rebuttal argues against opposing positions by integrating the participant's causal theory, evidence, and counterargument the rebuttal is said to be strong.
Kuhn (1991, pp. 264-268) argued that these six components are indicators of a single factor, argumentative complexity. Therefore, it is hypothesized that as the degree to which these six components were present in the participant's interview responses increased, so did the argumentative complexity of the response. Argumentative complexity is thought to be related to persuasiveness because research has found that people presenting more complex arguments are not only more persuasive, but also perceived by others as more competent persuaders (Campo, 1999; Kuhn, 1991, p. 273; Shestowsky, Wegner, & Fabrigar, 1998). Moreover, as argued previously, mavens' expert knowledge provides information that is able to facilitate the development of more complex arguments, and the more extensive and diverse contacts that connectors have provide exposure to more and diverse arguments in a domain of interest. Therefore, it was expected that superdiffusers' arguments would be more complex than those of participants who were not superdiffusers.
Interview Rating. The audio tapes from all of the interviews were transcribed by one of the authors. Following transcription, three interview raters who were blind to the study's hypotheses were trained. Training occurred in several stages. First, over several sessions raters read and discussed the parts of Kuhn's (1991) text relevant to defining the six components of argumentative complexity. When definitional agreement among the E and the raters was achieved, focus turned to developing the raters' ability to identify the components of argumentative complexity in the interview transcripts with a high degree of inter-rater reliability. Interviews were rated employing the coding sheet presented in Appendix C. Referring to the definition of each component provided by Kuhn, raters were instructed to assess each interview for the components of argumentative complexity using eight-point Likert response scales.
Several interview transcripts not included in the study's sample were used for training. Initially, training interview transcripts were rated collectively and raters, along with the E, discussed any rating discrepancies until agreement was reached. Then, working independently, the remaining training transcripts were rated. When discrepancies were found between raters, a discussion was held in which raters offered their perspective until agreement was reached. Substantial inter-rater reliability for each of the components in the training interview transcripts was achieved (ECs > .85) and raters began coding the Study 2 interviews.
The inter-rater reliabilities of the six argumentative complexity components were examined before forming the argumentative complexity index. Each rater was treated as one item on a three-item measure. Thus, for each of the six components of causal theory, evidence, alternative causal theory, alternative evidence, counterargument, and rebuttal there were three items measuring the extent to which the construct was present in the participants' interviews. Six indices were created by taking the mean of the rater ratings on each of the six components.
Scores on the causal theory index were distributed approximately normal, ranged from 2.00 to 6.67, and had a mean of 3.99 (SD = 1.14). The inter-rater reliability of the index was estimated as EC = .89. The evidence index ranged from 2.33 to 6.33, scores approximating closely the normal distribution. The mean of this index was 4.17 with a standard deviation of 1.19. The inter-rater reliability, was estimated as EC = .84. The distribution of the alternative causal theory index ranged from 1.00 to 6.00 and approximated normality as well, M = 3.90, SD = 1.16, EC = .80. The distribution of scores approached normality on the alternative evidence index. Scores on this index ranged from 1.00 to 4.67, had a mean of 2.90 (SD = .87) and EC = .71. The counterargument index was approximately normally distributed and ranged from 1.00 to 5.00. The mean response on the index was 2.46 (SD = 1.01) and EC = .83. The distribution of the rebuttal index approached normality, scores ranging from 1.50 to 7.50 (M = 4.07, SD = 1.48, EC = .80).
Confirmatory factor analysis was employed to test the hypothesis that these six indicators of argumentative complex fit a unidimensional model (Hunter & Hamilton, 1992). Initial examination of the correlation matrix indicated that both evidence and alternative evidence did not exhibit internal consistency with the other components. Therefore, the evidence and alternative evidence components were discarded from the model.
Subsequent to removing these two components, CFA was reemployed to test the unidimensionality of the causal theory, alternative causal theory, counterargument, and rebuttal components. The analysis showed that all factor loadings were above .48 and the RMSE of .11 was well within sampling error of zero. Given the fit of the model, the mean of these four components was computed for each participant to form an argumentative complexity index. Scores on this index, which was approximately normally distributed, ranged from 1.50 to 5.04. The mean of the index equaled 3.60 (SD = .89), [alpha] = .73.
Superdiffuser Status. Superdiffusers were again identified as those whose scores were above the 75th percentile on each of the CPM scales. To be considered a superdiffuser a participant had to score above 66.65 on the connector scale, 72.34 on the persuader scale, and 68.88 on the maven scale. There were two superdiffusers in the sample based on this criterion. The argument complexity scores of the superdiffusers (M = 4.67, SD = .53) were higher than the scores of the non-superdiffusers (M = 3.60, SD = .81), t (32) = 1.82, p = .08, r = .31, d = 1.58. Because of the small number of superdiffusers, this effect is artifactually attenuated substantially due to range restriction in the dichotomous superdiffuser status variable (Hunter & Schmidt, 2004), and correcting it for the value that would be expected given an equal split yields an exceptionally large correlation, r' = .91. This result is consistent with the hypothesis that superdiffusers produce more complex arguments. The correlation matrix of argumentative complexity with the three dimensions of opinion leadership is presented in Table 6.
Additional Analyses. Determining if leaders and non-leaders are distinguishable using superdiffuser status as a predictor has the potential to be informative for future development of the model. Therefore, it would be interesting to examine if self-reported leadership can be predicted from scores on the connector, persuader, and maven measures. Because of the small proportion of subjects self reporting as leaders (.23) and the small number of superdiffusers (2) in this study, the sample is underpowered, and the confidence one could have in any conclusions regarding the relationship between self-identified leadership and superdiffuser status is minimal. Keeping this limitation in mind, the mean score of self-identified leaders was within sampling error of those who did not self-identify as leaders for each of the CPM measures (see Table 7).
The data are consistent with the hypothesis that superdiffusers produce more complex arguments than non-superdiffusers. The substantial correlations between each of the CPM scales and argument complexity scores are also consistent with this hypothesis.
The Study 2 findings must be interpreted with caution. The small sample size limits the confidence one can have in the stability of these findings. Also, the generalizability of the findings needs to be probed by employing different topics.
Both of these studies demonstrated that superdiffusers employ different argumentation strategies than non-superdiffusers. Study 1 found that superdiffusers produce more arguments, cover more topics, and adapt to the target more than non-superdiffusers. This relationship was moderated by task complexity such that superdiffusers only produced more complex messages when they were addressing a complex persuasion task. Study 2 found that superdiffusers produce more complex arguments in a face-to-face interview setting.
These studies were exploratory and thus only examined two healthy lifestyle contexts: smoking and weight management. Other contexts must be examined to determine which features interact with the individual differences associated with the ability to diffuse healthy behavior changes. The set of skills that superdiffusers possess may facilitate diffusing some health behavior changes more than others. Measuring the perceptions that participants have of the context would be a helpful next step in determining in a more systematic manner the elements of the scenarios that interact with the connector, persuader, and maven scales to affect argumentation strategies. Such data would also expand the very limited research on trait by context interactions in persuasion.
These studies contribute to the study of opinion leadership by exploring the ways that attributes associated with successfully persuading others to adopt positive health behaviors interact with the kind of persuasive task. These interactions can predict the persuasive methods that superdiffusers will use if and when they try to persuade others to adopt healthier lifestyles. Moreover, this research contributes to expanding argumentation research to explore how individual differences and situations can affect argument production. Future research in this area will enrich both fields.
Simple Avoiding Smoking Scenario
Kelly is a commuter student. You consider her an acquaintance, but you have known her for a long time and think she's a good person. She is on campus during the day, but when she's off-campus she hangs out with her high school friends; most of these friends got jobs instead of going to college. Many of Kelly's friends are cigarette smokers; however, Kelly is a non-smoker. She dislikes the taste of cigarettes and the smell from the smoke. Furthermore, Kelly understands that smoking can eventually lead to diseases such as lung cancer. Although her friends tease her about it, she has resisted the urge to start smoking, although sometimes she wavers.
Complex Avoiding Smoking Scenario
Kelly is a commuter student. You consider her an acquaintance, but you have known her for a long time and think she's a good person. She is on campus during the day, but when she's off-campus she hangs out with her high school friends; most of these friends got jobs instead of going to college. Many of Kelly's friends are cigarette smokers, as is Kelly. Kelly likes to smoke because it's relaxing and it's part of socializing with friends. She believes that it takes a long time for smoking to affect a person's health, and she feels there is plenty of time before she needs to quit. In the past, she has made attempts to quit smoking and has failed.
Simple Weight Loss Scenario
Pat is an acquaintance, but you have known her for a long time and think she's a good person. She has been told by doctors that she needs to lose weight, or else potentially suffer health consequences like diabetes. According to her doctor, Pat needs to lose some weight. Pat has never really thought much about her weight, though; it doesn't bother her that she is overweight. She has never tried dieting, and has never attempted to exercise regularly. She doesn't dislike the taste of healthy food, though, and usually has fun on the rare occasions when she does exercise. She realizes that she would probably lose a lot of weight if she ate more healthy foods or exercised, but doesn't really have the motivation. Recently she has been more open to people who have tried to talk to her about her weight.
Complex Weight Loss Scenario
Pat is an acquaintance, but you have known her for a long time and think she's a good person. Doctors have told her that if she does not lose weight, she might suffer future health consequences like diabetes. According to her doctor, Pat is about 90 pounds overweight. She has struggled with her weight her whole life and has tried many different diets. Sometimes they help her lose weight, but she gains it back immediately. She really doesn't like eating healthy food, and constantly craves candy and desserts when on diets. She has tried exercising, but doesn't enjoy it very much and thus has a hard time motivating herself to go to the gym everyday. She is at a point where she views her weight as determined by her genes, and doesn't feel like it's worth trying to be healthy anymore. Recently she has been resistant to people who have tried to talk to her about her weight.
Argumentative Complexity Interview Protocol
1. Do you believe dieting, that is eating a restrictive diet, is a healthy or unhealthy practice?
2. How strongly do you hold your belief?
3. For what reasons do you hold that position?
4. How do you know this supports your position?
5. Just to be sure I understand, can you explain exactly how your arguments support your position?
1. If you were trying to convince someone else that your view is right what evidence [verbal emphasis] would you give to try to show this?
7. Can you be very specific, and tell me some particular facts that you could mention to try to convince the person?
1. Is there anything further you could say to help show that what you have said is correct?
2. Can you remember when you began to hold this view?
10. Can you remember what it was that led you to believe what you do?
Alternative Causal Theory
1. Suppose that someone disagreed with your view. What might they [verbal emphasis] say to try to convince you that you were wrong?
2. What evidence might this person give to try to convince you that you were wrong?
Alternative Causal Theory
3. Just to be sure I understand, can you explain exactly how they would think this would show that you were wrong?
14. In order to support their view, what arguments might this person give?
15. Is there any fact or evidence which, if it were true, would show your view to be wrong?
16. Could someone prove that you were wrong?
16a. (Probe if yes) How?
Alternative Causal Theory
17. Imagine a person like we have been talking about whose view is very different from yours--what might they say is their view?
18. How would you respond? What would you say to try to convince them?
19. Just to be sure I understand, can you explain exactly how this would show the person was wrong?
20. Would you be able to prove this person wrong?
21. (Probe if yes) How?
22. What could you say to show that your own view is the correct one?
Boster, F.J., Kotowski, M. R., & Andrews, K. R. (2006, November). Identifying influentials: Development of the connector, persuader, and social maven scales. Paper presented at the annual meeting of the National Communication Association, San Antonio, TX.
Boster, F.J., Levine, T., & Kazoleas, D. (1993). The impact of argumentativeness and verbal aggressiveness on strategic diversity and persistence in compliance-gaining behavior. Communication Quarterly, 4l, 405-414.
Boster, F.J., & Stiff, J. B. (1984). Compliance-gaining message selection behavior. Human Communication Research, 10, 539-556.
Buller, D. B., Woodall, W. G., Rogers, E. M., Burris-Woodall, P., Zimmerman, D., Slater, M., Pepper, J, Bartlett, K., Hines, J., Unger, E., Han, B., & LeBlanc, M. M. (2001). Formative research activities to provide web-based nutrition information to adults in the Upper Rio Grande Valley. Family and Community Health, 24, 1-12.
Campo, M. L. (1999). Arguing for change: Arguments by activists, non-activists, and in the media regarding domestic partner benefits. Unpublished Doctoral Dissertation. Michigan State University.
Canary, D.J., Cody, M.J., & Marston, P. (1986). Goal types, compliance-gaining, and locus of control. Journal of Language and Social Psychology, 5, 249-269.
Cicero. (1970). On oratory and orators (J. S. Watson, Trans.). Carbondale, IL: Southern Illinois University Press. (Original work published 55 BCE).
Dearing, J. W., Rogers, E. M., Meyer, G., Casey, M. K., Rao, N., Campo, S., & Henderson, G. M. (1996). Social marketing and diffusion-based strategies for communicating health with unique populations: HIV prevention in San Francisco. Journal of Health Communication, 1, 343-363.
DeFleur, M. (1987). The growth and decline of research on the diffusion of news, 1945-1985. Communication Research, 14, 109-130.
Doumit, G., Gattellari, M., Grimshaw, J., & O'Brien, M. A. (2007). Local opinion leaders: Effects on professional practice and health care outcomes. Cochrane Database of Systematic Reviews, 1, 1-28.
Ebel, R. L. (1951). Estimation of the reliability of ratings. Psychometrika, 16, 407-424.
Greeberg, B. S. (1964). Diffusion of news about the Kennedy assassination. Public Opinion Quarterly, 28, 225-232.
Greenhalgh, T., Robert, G., MacFarlane, F., Bate, P., & Kyriakidou, O. (2004). Diffusion of innovations in service organizations: Systematic review and recommendations. The Milbank Quarterly, 82, 581-629.
Hample, D. (2003). Inventional capacity. In F. H. van Eemeren, J. A. Blair, C. A. Willard, & A. F. Snoeck-Henkemans (Eds.), Proceedings of the fifth conference of the International Society for the Study of Argumentation (pp. 437-440). Amsterdam: SicSat.
Hample, D. (2005). Arguing: Exchanging reasons face to face. Mahwah, NJ: Lawrence Erlbaum Associates.
Hample, D., & Dallinger, J. M. (2002). The effects of situation on the use or suppression of possible compliance gaining appeals. In M. Allen, R. Preiss, B. Gayle, & N. Burrell (Eds.), Interpersonal communication: Advances through meta-analysis (pp. 187-209). Mahwah, NJ: Lawrence Erlbaum Associates.
Hunter, J. E., & Gerbing, D. W. (1982). Unidimensional measurement, second order factor analysis, and causal models. Research in Organizational Behavior, 4, 267-320.
Hunter, J. E., & Hamilton, M. A. (1992). CFA--A Program in Basic to do Confirmatory Factor Analysis. East Lansing, MI: Michigan State University.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting for error and bias in research findings (2nd ed.). Thousand Oaks, CA: Sage Publications.
Kelly, J. A., Murphy, D. A., Sikkema, K.J., McAuliffe, T. L., Roffman, R. A., Solomon, L.J., Winett, R. A., & Kalichman, S. C. (1997). Randomized, controlled, community-level HIV-prevention intervention for sexual-risk behaviour among homosexual men in US cities. Lancet, 350, 1500-1505.
Kuhn, D. (1991). The skills of argument. Cambridge, UK: Cambridge University Press.
LaRose, R., & Hoag, A. (1997). Organizational adoptions of the internet and the clustering of innovations. Telematics and Informatics, 13, 49-61.
Levine, T. R., & Boster, F.J. (1996). The impact of self and others' argumentativeness on talk about controversial issues. Communication Quarterly, 44, 345-358.
Lowry, R., Galuska, D. A., Fulton, J. E., Wechsler, H., Kann, L., & Collins, J. L. (2000). Physical activity, food choice, and weight management goals and practices among U.S. college students. American Journal of Preventative Medicine, 18, 18-27.
Lusting, M., & King, S. (1980). The effects of communication apprehension and situation on communication strategy choices. Human Communication Research, 7, 74-82.
Miller, G. R., Boster, F.J., Roloff, M. E., & Seibold, D. R. (1987). MBRS rekindled: Some thoughts on compliance gaining in interpersonal settings. In M. E. Roloff & G. R. Miller (Eds.), Interpersonal processes: New directions in communication research (89-116). Newbury Park, CA: Sage Publications.
O'Keefe, B. (1988). The logic of message design: Individual differences in reasoning about communication. Communication Monographs, 55, 80-103.
Rigotti, N. A., Lee, J. E., & Wechsler, H. (2000). US college students' use of tobacco products: Results of a national survey. The Journal of the American Medical Association, 284, 699-705
Rogers, E. (2003). The diffusion of innovations (5th ed.). New York: The Free Press.
Shestowsky, D., Wegener, D. T., & Fabrigar, L. R. (1998). Need for cognition and interpersonal influence: Individual differences in impact on dyadic decisions. Journal of Personality and Social Psychology, 74, 1317-1328.
Valente, T. W., & Pumpuang, P. (2007). Identifying opinion leaders to promote behavior change. Health Education & Behavior, 34, 881-896.
Wilson, S. R. (2002). Seeking and resisting compliance: Why people say what they do when trying to influence others. Thousand Oaks, CA: Sage Publications Inc.
(1) As an insightful reviewer pointed out, given the small mean difference this effect does not appear to be statistically significant. The analysis of variance algorithm performed for this analysis was an unweighted means algorithm (the Type III sum of squares option in SPSS). When a weighted means analysis is performed this result is no longer statistically significant. Regardless of the algorithm employed, however, the interaction effect was found to be statistically significant.
Christopher J. Carpenter is a doctoral student at Michigan State University. Michael R. Kotowski is an Assistant Professor in the School of Communication Studies, University of Tennessee. Franklin J. Boster is a Professor of Communication at Michigan State University, where Kyle R. Andrews, Kim Scrota, and Allison S. Shaw are doctoral students. Contact the first author by email at email@example.com.
TABLE 1. CPM ITEMS AND FACTOR LOADING'S ON THE PRIMARY FACTOR BY STUDY Factor Study 1 Study 2 Connector I'm often the link between friends in .83 .69 different groups I often find myself introducing .86 .88 people to each other I try to bring people I know together .67 .77 when I think they would find each other interesting I frequently find that I am the .95 .92 connection between people who would not otherwise know one another The people I know often know each .82 .83 other because of me Persuader I am good at thinking of multiple .79 .72 ways to explain my position on an issue I'm able to argue well for a position .82 .62 I believe in When in a discussion, I'm able to .85 .81 make others see my side of the issue I am able to adapt my method of .86 .89 argument in order to persuade someone When my approach to an argument is .89 .79 not working, it is easy for me to quickly come up with something new that does work I can effortlessly offer multiple .82 .70 perspectives on an issue that all support my position More often than not, I am able to .81 .73 convince others of my position during an argument I am skilled at using my read of .77 .72 others to successfully persuade them Health Maven My friends think of me as a good .76 .83 source of information when it comes to healthy lifestyle issues When I know something about a healthy .79 .72 lifestyle topic, I feel it is important to share that information with others I like to be aware of the most up-to- .87 .89 date healthy lifestyle information so I can help others b sharing when it is relevant If someone asked me about a healthy .74 .74 lifestyle issue that I was unsure of, I would know how to help them find the answer Being knowledgeable enough about .85 .86 healthy lifestyles so that I could teach someone else is important to me People often seek me out for answers .87 .72 when they have questions about a healthy lifestyle issue TABLE 2. CORRELATION MATRIX OF STUDY 1 DEPENDENT VARIABLES Arguments Themes Elements Number of Arguments .87 .45 Number of Themes .72 .53 Number of Scenario .34 .22 Elements Note: Correlations in the upper triangle are from the smoking scenario and the lower triangle are from the weight scenario. TABLE 3. MEAN (SD) NUMBER OF ARGUMENTS BY CONDITION Task Superdiffuser Status Complexity Non-S.D. S.D. Complex 2.09 (1.81) 5.17 (.71) Simple 2.15 (1.15) 2.00 (1.41) Note. SD = Superdiffuser TABLE 4. MEAN (SD) NUMBER OF DISTINCT THEMES BY CONDITION Task Superdiffuser Status Complexity Non-S.D. S.D. Complex 1.15 (.85) 2.83 (.24) Simple 1.40 (.52) 1.00 (0) TABLE 5. MEAN (SD) NUMBER OF REFERENCES TO THE SCENARIO BY CONDITION Task Superdiffuser Status Complexity Non-S.D. S.D. Complex .66 (.66) 2.67 (0) Simple .79 (.57) 0 (0) TABLE 6. STUDY 2 CORRELATION MATRIX Connector Persuader Maven Connector Persuader .37 Maven -.02 .11 Argue .45 .23 .38 Notes. Maven = Healthy Lifestyle Maven, Argue = Argumentative Complexity. TABLE 7. MEAN (SD) CPM SCORES AS A FUNCTION OF SELF-IDENTIFIED LEADERSHIP CPM Scales Leadership Status Connector Persuader Maven Leader (n = 8) 52.95 (25.76) 70.36 (12.50) 55.81 (21.97) Non-Leader (n = 27) 44.06 (20.65) 61.09 (13.68) 48.70 (22.29) t(33) = .91, ns t(33) = 1.79, ns t(33) = .86, ns Notes. Maven = Healthy Lifestyle Maven.
|Printer friendly Cite/link Email Feedback|
|Author:||Carpenter, Christopher J.; Kotowski, Michael R.; Boster, Franklin J.; Andrews, Kyle R.; Serota, Kim;|
|Publication:||Argumentation and Advocacy|
|Date:||Jan 1, 2009|
|Previous Article:||The interacting arguments of risk communication in response to terrorist hoaxes.|
|Next Article:||Deliberative Democracy for the Future: The Case of Nuclear Waste Management in Canada.|