In many respects, groups had similar opinions, such as being pessimistic about the state of the country and negative feelings towards government, but their narrative constructions differ when examining characters, policy solutions, and, to a certain extent, causal mechanisms, the building blocks of what constitutes a policy narrative (Shanahan et al. 2018a). Consistent with our expectations, a unique narrative emerged from each focus group.Footnote 9 We reveal our findings with a network analysis and more detailed node-level explorations of characters, plot, and causal mechanisms.
Pre-information narrative network comparison
We begin by comparing the narrative networks from the focus group discussion prior to the introduction of specific campaign finance information. Since we want to discover if different narratives emerge from the groups (RQ1) we deploy social network analysis (SNA) to map and compare the characteristics of the narrative generated by each CC focus group. The analyzed networks were generated using the affiliation command in UCINET 6 for Windows which transforms a two-mode dataset (rows representing coded text and columns representing narrative elements) into a 1-mode adjacency matrix where the value of each cell reflects the number of links between each narrative element and all other narrative elements using the sums of cross-products. Each node represents a narrative element (heroes, villains, victims, policy solutions, or causal mechanisms). Larger nodes indicate greater degree centrality of a node (narrative element) in the network. For instance, the hierarchical-individualist pre-information narrative network map identifies a private individual as a villain with 20 links. The node is larger than the node for money which has 2 links. Links between the nodes represent instances where an element co-occurred with another element (in one coded paragraph). The darker lines indicate a greater frequency of co-occurrence.
The density measure for each focus group’s narrative network was computed. Density is a measure that conveys the general level of connectivity between the nodes in the network and is calculated by dividing the total number of dyadic ties present in the network by the total number of all possible ties in the network (Yang et al. 2017, p. 58) (Table 3).
Table 3 Density measures for focus group’s narrative network Prior to information dissemination by the moderator, we see egalitarian-communitarians have the densest network (19.1%), followed by the hierarchical-communitarians (15.9%). Hierarchical-communitarians and egalitarian-individualists demonstrate much lower narrative network density measures of 9.5% and 4.0% respectively. To discover whether the differences are statistically significant, we utilized the compare densities function (paired networks in UCINET 6) with 10,000 bootstrapped samples and found the hierarchical-communitarian network was significantly denser than the egalitarian-individualist network by 11.9% (Sig. 0.030 p < .05). Hierarchical-communitarian network density was significantly less dense then the egalitarian-communitarian one, by 9.7% (Sig. 0.045 p < .05). The egalitarian-individualist network was significantly less dense than both the hierarchical-communitarian and egalitarian-communitarian narrative network densities, the latter by 15.2% (Sig. 0.011 p < .05) (Table 4).
Table 4 Statistical significance of density between group comparisons (pre-information) Density differences are important pieces of information for understanding how CC and the NPF combine to generate insights into narratives for at least two reasons. First, density may represent a cognitive foundation amenable for future information incorporation (more on this below). It is the contention of both the NPF’s model of individual cognition (see Shanahan et al. 2018b, c, pp. 179–183) and CC that existing beliefs and identity commitments impact the assimilation of new information, so understanding the embeddedness of relationships between existing narrative elements could be vital to anticipating the likely power of a particular narrative to persuade a given group. Second, the density of a narrative network may indicate a more cohesive story, with greater ties between narrative elements indicating connections between concepts, or at least the existence of more elements the audience can identify with. This may represent an approach to exploring narrativity, which is the notion that more complete stories are more persuasive (e.g., Crow and Berggren 2014).
If the density of the narrative networks varies between cultural groups, what might this mean for the study of narratives? We can begin by generating network maps for each of the groups. The network diagrams (Fig. 2) show that differences in density translate to “fuller” graphs for hierarchical-individualists and egalitarian-communitarians, as more nodes are connected than in networks assembled by hierarchical-communitarians or egalitarian-individualists.
In Fig. 2, the size of a node is directly related to the number of ties it has to other nodes (degree centrality). Darker lines indicate more instances of connection existing in the same paragraph. Larger nodes and darker lines help identify a distinction between network core and periphery. The core/periphery measurement identifies nodes which constitute a “community” with dense connections to other nodes and which have sparser connections with nodes not in their community (Rombach et al. 2014). The core was identified using the MINRES algorithm and the continuous function with 1000 iterations. The core is important for using SNA to analyze policy narratives because it brings to the forefront those narrative elements that are most frequently employed together while simultaneously illustrating the elements’ connectedness to the entire network. UCINET 6 determines the core measures for each node in the network by comparing its structure to an idealized core/periphery block where several nodes with direct and dense ties exist without ANY connections to other nodes in the network. If the network under investigation matches this model exactly, values for core nodes will equal 1.0, and nodes that demonstrate more “coreness” return values closer to 1.0 (Everett and Borgatti 2005). If we find that the four narrative cores contain different elements and links, we can consider that differences in CC may contribute to the structuring of narrative networks (Figs. 3, 4, 5).
Drawing on these diagrams, and the density, centrality, and core/periphery measures applied to the four pre-information narrative networks, we can easily identify the similarities and differences between the story structures.
Causal mechanisms
Previous studies have found many narratives use intentional causal mechanisms, as villains are highlighted as an entity to combat (Shanahan et al. 2014). Below are examples of how these mechanisms arose in the focus group narratives.
- 1.
Intentional policymakers might be accused of making policies to increase their personal wealth.
- 2.
Inadvertent the American Recovery and Reinvestment Act of 2009 might be explained as having raised inflation.
- 3.
Mechanical a bad policy might be explained as resulting from an unthinking bureaucracy.
- 4.
Accidental fluctuations in the price of commodities due to the weather.
One interesting aspect revealed by the narrative network maps is that each group expressed that campaign finance was a problem caused by intentional action to accomplish desired goals. That this emerged from all groups may indicate an understanding that the system of campaign finance is a product of human construction, maintenance, and manipulation. Unlike harm caused by natural disasters (accidental cause), machines (mechanical) or carelessness (inadvertence), injury caused by the campaign finance system is seen by all as intentional. This lack of ambiguity renders “the problem” of campaign finance amenable to a policy solution (Stone 2012), which is supported by extant campaign finance literature (e.g., Jorgensen et al. 2018). It is also important to note the HI, EI, and EC narratives connect the network nodes Intentional Cause and Villain: Government in their network’s core. Government does not occupy this key role in the core of hierarchical-communitarians. Perhaps HC’s are simply more favorably disposed toward the federal government (by virtue of their high grid and high group scores) and thus unwilling to craft a story with government as a core villain.
The final element in the pre-information narrative network core shared by EC and EI groups is private individuals as victims.
EI: “Yeah, and then they’re taking it down on us regular folk, and bumpin’ up our property taxes, you know?”
EC: “But it’s just going to destroy a lot of middle-class families when you say, “Hey, you know, it’s illegal to have, you know, unions.” And we called our politicians and you’d think that they would listen to you. And I know we have republicans and democratic teachers that were calling and you know, and fireman and whatnot. And they absolutely said, “We don’t care. This is what the big money people want us to do. So, we’re just going to ignore your vote.”
With the use of “regular folk” and “middle-class families” both groups associate the intentional exercise of power by malevolent government (in the case of EC’s), businesses, and private individuals with harm done to average Americans. This construction, also manifesting along the grid-dimension, suggests egalitarians of both varieties generate more cohesive campaign finance stories prior to possessing policy-relevant information. This may be due to factors related to the universe of characters in these stories, and to these characters we now turn.
Characters
Coded character categories include heroes (government, private individuals, and advocacy groups), villains (government, private individuals, advocacy groups, business, and money) and victims (government, private individuals, rights, and business). Given the CC quadrants defined above, we would expect that focus groups situated on the individualist side, i.e., hierarchical-individualist and egalitarian-individualist, would cast heroes as private individuals, villains as government and advocacy groups, and victims as private individuals. Similarly, we would expect communitarian focus groups, i.e., hierarchical-communitarian and egalitarian-communitarian, would cast heroes as government and advocacy groups, villains, as private individuals, business, and money, and victims as government. However, the pre-information narratives generated by the four different groups failed to demonstrate any statistically significant differences between ANY heroes or victims (see Table 5). When we narrow our attention to the narrative core, we don’t find ANY heroes and only the aforementioned private individual as victim in EI and EC cores. These particular stories are driven by villains. Since the focus groups dealt with campaign finance law, and villains like government and corporations were key to building a story about the issue, it is not surprising that the HC group didn’t generate a more cohesive network; intentional causes require villains and the likely villains in campaign finance reform are manifestations of entities hierarchical-communitarians tend to admire.
Table 5 Kruskal-Wallis and Mann–Whitney U Test Pre-Information Heroes
Heroes are those characters who are cast as those who will fix the problem. The complexity of identifying heroes as individuals or government lay in that an individual must work within the system to affect change. For example, one member of focus group 1 (HI) recognized citizens’ capacity to stand up and make a difference (individualism) but situated this individuality within the community of the tea party: “We’re grassroots, we’re not a big corporation, we’re just people wanting to make a change. And that seems like that would influence the whole corruption big money.” As the network maps show, while each cultural narrative identified heroes that could be situated on opposite ends of the Group continuum, no hero was central enough to the story of campaign finance to exist in the narrative core.
Victims
The victim is the character who suffers at the hands of the villain. The primary victims for campaign finance were the government and the private individual. For the members of the HI group, corporations play a role like that of individuals, valuing freedom and choice for them as well. “They made it sound like corporations, the people, who spend a lot of money in elections is a bad thing. But do we ever stop and think, why do we tax corporations? Corporations shouldn’t have to pay money…what you end up doing is you stifle growth.” As one HI member said, “…in the society where we’re basing our life on freedom and you start telling people what they should and shouldn’t do with their money, which can be a lot farther reaching than any of us realize.” The hierarchical-communitarian and egalitarian-communitarians victims tended to center on money’s threat to the public interest. “…whoever puts in the most money…they’re going to end up on Saturday Night Live…and that is going to influence the public, and then really in the long run their guy is going to get in, and I feel powerless when I hear that kind of thing.” As with heroes, no statistically significant differences in frequency of victims arose between the four groups.
Villains
Villains are the characters responsible for the problem and it is here the differences between the CC groups begin to acquire more resolution. Participants in the focus groups developed narratives that spent considerable time expounding on villains. Focusing on the government as villain across the entire network (not simply examining the core) by using the Kruskal–Wallis H test, we reject the null hypothesis of no difference between the groups (p = .036 < 0.05). Digging deeper, we see that only two groups’ use of government as villain were statistically significant. Applying the Mann–Whitney U test to the four pairs, we find that statistically significant differences are only found between hierarchical-communitarians and egalitarian-individualists (p = .009 < 0.01). The test also found that power as a villain (i.e., power corrupts) demonstrated a significant difference between groups (8.581, p = .035 < 0.05) at the p = 0.05 level, but the stricter p value of 0.01 required by the Bonferroni correction when comparing the four groups individually to one another, failed to identify a significant difference between them.Footnote 10 Concerning villains in the pre-information stage, Grid is the operative scale here, as the HI group demonstrated more instances (n = 29) than the EI group (n = 10) painting government as villain. The HI focused on bad-natured individuals in government, and not the system as a whole. “…the politician in office has such a tremendous advantage because they can squeeze to get the money. They know how to say, ‘yeah, we’ll talk about that after your donation.’” In contrast, when HC narratives did identify government as villain, it was because individuals in government were not acting in the public interest. “Well, right now they are not living up to the role because they are not solving issues before them and it’s very, very divisive. It’s very, very political. And so they are not performing the role that they have or should be doing.” These distinctions provide support for CC’s theory that HI’s, “Marlboro men” will be extremely critical of external forms of authority infringing on their own domains of organized social relations.
However, when we consult the narrative network map, we see that despite the lack of statistically significant differences demonstrated by the Kruskal–Wallis H and Mann–Whitney U tests, three of the four groups (HI, EI, and EC) identified the government as a villain in their network core. The coreness measure was 0.611 in HC narratives, 0.540 for EC, and 0.516 for EI. Recalling that the core measure is derived from comparing the matrix under investigation with an idealized core model (of 1.0), these values indicate that the coreness of these measures is not slight. The government as villain achieved a coreness score of only 0.287 in the HC narrative, which shows that the HC group tended not to blame government, supporting CC’s theoretical stance that HC’s will generally value structure and authority.
These findings indicate that all groups identify the problems arising in campaign finance as generated by intentional human action, and all core narratives, excluding hierarchical-communitarians, lay at least some of the blame at the feet of the government. We can see that a villain that is accessible via existing cultural cognitive understandings allows the formation of more cohesive stories absent additional information. Hierarchical-individualism opposes the imposition of federal government on their local forms of hierarchy, egalitarian-individualism resents what they see as government waste and incompetence demonstrated in the statements like, “I don’t understand why Congress people are receiving a paycheck, when they haven’t obviously passed our previous [sic], budget?” Egalitarian-communitarians, on the other hand, have a more complex relationship that seems to emphasize the intersection between politicians and business interests, as seen in this statement, “But you see it happen every single time a president comes in and then when he leaves office the company that gave him the most money, he becomes a consultant. Probably doesn’t show up but just, you know, to give a speech and he makes a million dollars a year for, you know, the next 20 years or whatever as a consultant, you know, to that organization that gave him all that money.”
To summarize the findings related to the relationship between characters and plot we could describe the initial formulation for HI’s narrative as “government as corrupt”, for EI’s “government as incompetent”, and for EC’s as “government as co-conspirator”. HC’s have no initial core, since they apparently unwilling demonize government, though they too recognize the problem as within human ability to control.
Policy actions or moral of the story
We operationalized the NPF’s moral of the story as policy actions. For example, a pro-government action solution was voiced in the EI narrative when it was stated that “The government that is there for the people, they should step in and say, ‘Hey, this person who can’t afford three hundred a month, should not be given thirteen thousand”; a con-government action solution would have expressed opposition to such a statement. A pro-collective action other than government policy solution called for people to work together in the sphere of civil society to address the issue of campaign finance. A statement made in the EC focus group is indicative this type of solution, “I mean, this makes sense to me because, you know, you do fundraisers and you might get 100 people somewhere and ya’ll donate $200.You do that all over the country and you can fund them like that.” While four policy actions were coded, none demonstrated statistically significant differences using the Kruskal–Wallis H test: statements indicating support for government action (.989 > p. 05) against government action (.131 > p. 05) as well as statements advocating/opposing non-governmental or collective action by citizens (.109 > p .05/.183 > p. 05). This result is, in a certain respect, not surprising since none of the focus groups identified any hero in the narrative core to undertake corrective action. It is important to note this changes when comparing the pre-information narratives to the post-information narratives.
Post-information narrative structure within focus group comparisons
Moving now to the post-information analysis, it is important to remind the reader that the comparisons that will be made for the remainder of the article contrast pre-information narrative structure with post-information structure within individual CC focus groups. The moderator encouraged groups to range fairly freely in their discussion, and thus information provided to individuals prior to presentation of policy relevant information on campaign finance and the Grants and Franklins project (a government-based campaign finance reform policy) was not identical. However, handouts provided to each participant were identical, and the moderator provided three similar hypothetical arguments, one in favor and two opposed to the Grants and Franklins project (see Appendix 2). Comparing the post-information network of any group with its pre-information network is a way to assess how the introduction of new information impacted the development of each CC narrative. We begin by discussing changes in overall network measures.
We see the network density for the EC group was significantly reduced by providing more information, while the other groups showed no significant change (see Table 6). This is surprising since the density of the EC narrative network was highest prior to receiving the policy relevant information. Why this group should display such a significant drop is an interesting question, and one that will be explored in more detail below (Figs. 6, 7, 8, 9).
Table 6 Statistical significance of density within group comparisons (pre/post information) Characters
Heroes
The post-information narratives feature heroes in the core of HI and EI groups. In both cases, the heroes were, as the group dimension of CC would suggest, private individuals.
FG1 (HI): “The thing that, the thing we most need to understand is our founding fathers had a reason why they set up everything the way they set out to do and really quite brilliantly… But they also allowed us to spend whatever money, each person, each corporation wanted to at that time because that’s another way to make…almost like the three different branches of government, how they all work together. So you’re not just having the voting public, whether it’s a ignorant vote or whatever making their vote based upon whose pretty or who says the right thing. You got some people that if you got money, not very often do you get because it’s lucky. You get it because you got some kind of intelligence and that helps guide that somewhat.
FG3 (EI): Okay, but to me, what was good about that is, he was out among the people. Had been for a long time, and reflected a most of the state. I’m not saying everyone, I’m not even saying me, but a huge portion of our state. Without massive corporations backing him. Like one corporation getting him in. And if somehow that could play a part on a national level, where, your guy reflected the majority of the people, and somehow those people contributed, or voted, where who actually got in reflected us, not just one corporation, that’s ideal.
The private individual is employed differently, with the HI narrative equating smart individuals with rich individuals and portraying their contributions as a check on ignorant voters. The private individual, for the EI, is a person who works in the system for the good of the majority. This provides support for using CC to understand narrative differences, as both focus on the individual as an agent of positive change. The HI narrative is animated by the idea that an individual success in the economy should translate into greater influence in selecting national leaders, via campaign donations. EI stories do not focus on the distinction between intelligent and ignorant participants in the political process, but on the elected individual actually representing the will of the majority, while eschewing the unequal influence of concentrated economic power. These changes in narrative could be explained by the difference in grid influencing the nature of the hero employed in high- or low-grid worldviews and the high- or low-group aspect influencing whether a hero even enters the core narrative.
Victims
No victims were found in the narrative core of any group in the post-information portion of the focus group. The EI and EC core didn’t carry the private individual as victim over from their pre-information core. It is unclear what this might mean, but it may be that moving from the pre-information stage where the dialogue was conditioned more so by the influence of the variation in CC worldviews toward a discussion of a specific policy proposal sharpened the focus to groups that were unlikely to be viewed charitably by egalitarians.
Villains
The post-information portion of the focus group lead to a substantial change in the villains found in the narrative cores of each of the four groups. The hierarchical-communitarian narrative core retained government as a villain, but it was less of a core node than the new villain, private individuals. The crux of this villain was a direct response to the Grants and Franklin project’s call for the first $50 of federal taxes paid by each voting-age citizen to be transformed into a “democracy voucher” that individuals could direct to the politician(s) of their choice or elect to have it spent on voting infrastructure if they did not approve of any candidate.
FG1 (HI): Overall I don’t see a clear connection between, you know, campaign finance and how politicians get to Washington to the problems that we have because you know, whether a politician, you know, got there on a string bean budget or Kennedy type budget, if they’re corrupt, they’re corrupt. And they reflect the people in their district. So they’re [sic] the people in their district are freedom loving and gun totting or whether they are in the North and like to drink their tea with their pinkies up. The congressmen are going to reflect the values and the morals of their constituents and…
The HI narrative emphasizes that the major problem is that individuals are likely to take advantage of the system and many of the citizens are generally flawed. Again, CC helps to explain this emergence, since many of the reasons the HI group opposes the initiative relates to the need to check the excesses of democracy.
The rise of a new villain also occurs in the narrative core of the hierarchical-communitarians. This villain emerged from the detail in the Grants and Franklin proposal that allowed politicians to opt-out of the system and raise their money from private donations. This generated a consensus that businesses would oppose the system and that public finance couldn’t overcome corporate money without additional restrictions on corporate behavior.
FG2 (HC): But, these big corporations like Halliburton, that get all the government contracts have got to know that there’s millions and millions of dollars for campaign funding. And, if they just said, “Stop, no more, and if we catch you doing it, here is the penalty.”
This call is also in keeping with what CC would predict from hierarchical-communitarians, the call for public authorities to establish rules and regulations that would advance the public good and bring order to the process.
The Egalitarian-Individualist (FG3) narrative core also featured a new villain post-information: the private individual. This villain was, as in the hierarchical-communitarian core narrative, the ignorant voter.
FG3 (EI): But it just comes back to that same whole ignorance thing, that if people don’t know, and then they’re bombarded with this message, and how many of us really, when we have this coverage? I know I did. I’m sorry, I do the research on the candidates, I do the research on the bills, cause’ I’m around to care. I don’t want it to just be a vote. You know? I don’t want to be ashamed. Cause I know what I voted for. At the end of the day, it comes back and I voted for something I really didn’t stand for or believe in, so I mean, I know that I’m going to know what I vote for. How many of the average American’s that vote will?
This seems to indicate that individualists may have reacted to the plan to use public power to redistribute resources to level the playing field in campaign finance by casting doubt on their fellow citizens to wisely exercise this power. The egalitarian-communitarian narrative core did not add any new villains in the post-information stage.
Policy actions or moral of the story
Neither egalitarian-communitarian nor hierarchical-communitarian core narratives contained policy solutions. However, both hierarchical- and egalitarian-individualist core narratives employed a moral to the story. The solution preferred by the HI narrative was one that was promoting opposition to government action, the Grants and Franklin project specifically. This component relied on a distrust of the government and its power over individuals.
FG1 (HI): “This just seems un-American. It just doesn’t seem like it would be in a free market society. I just don’t see it working.”
The egalitarian-individualist core narrative, instead of opposing government action, proposes that collective action outside of government proper can correct some of the weaknesses of the current campaign finance system and its impact on contemporary politics.
FG3 (EI): And there’s a lot of government responsible for that. So, as for Obama being the problem. It’s years, and accumulation of years and years and years of being let down and the people, all of us, relying on the government, instead of being the government.
The use of network mapping and core and periphery measures provided evidence that each group’s core narrative was impacted by the provision of policy-relevant information and a specific policy proposal.