Democratic Disarray: Organizational Messaging Coherence and the Local Echoing Press During the 2020 Iowa Democratic Caucus

Abstract

The Iowa caucuses are the inaugural event of the American presidential nomination process. When the state Democratic Party failed to report the 2020 caucus results in a timely manner, researchers began to assess how party and press officials co-constructed the events. This research presents an in-depth case of the Iowa state and local Democratic Party’s public communication response regarding an event described by the Des Moines Register as “hell” and a “results catastrophe.” Specifically, we were interested in how the Iowa Democratic Party responded to the crisis event and the extent to which the party organization was successful in disseminating favorable messaging about the caucus process to the local press. Drawing on organizational crisis management and echoing press perspectives, this analysis uses network and qualitative analytic approaches to assess message development, dissemination, and ultimately adoption. A local event with national implications presents a critical case in investigating how a political party, due to its institutional role in American elections and unique organizational structure, struggled to respond to the crisis.

Keywords

political communication echoing press situational crisis communication theory semantic network analysis

The Iowa caucuses are the inaugural event of the American presidential nomination process. The traditionally midwinter election event values a retail style of politics based heavily on campaign organizing, mobilization, and party building (Redlawsk et al., 2011; Squire, 2008). Held on February 3, the 2020 Iowa caucuses continued a 44-year tradition where members of the Republican and Democratic parties physically gather at hundreds of precinct locations statewide. In gymnasiums, libraries, school cafeterias, and other venues, caucusgoers publicly announce their support for a presidential candidate by gathering (caucusing) in a set location in each indoor space (Klinefeldt et al., 2020). The event is a culmination of a year’s long process where prospective candidates visit with local party officials, debate opposing candidates in the party, engage local media, and hold campaign events of varying sizes. The caucus electorate tends to differ from the primary or general election electorate: more politically attentive, activist, and partisan (Redlawsk et al., 2011). In addition to choosing candidate delegates for county, district, and state conventions, the caucuses build up the internal organization of the local and state political parties ahead of the general election.

The caucuses hold practical significance as part of the presidential nominations process and serve as an important site for research efforts in political science, organizing, and communication, prompting two of the authors to attend the 2020 caucuses for field research. Previous research analyzed Iowa’s betterment of the nominations process (Redlawsk et al., 2011), candidate television advertising (Gronbeck, 1989), online organizing (Hull, 2006), campaign rallies and trust in democratic institutions (Warner et al., 2019), caucusgoers’ perceptions of the candidates (Winfrey, 2017), and attitudes toward presidential communication (Scacco & Coe, 2017). In the midst of observing a well-run caucus site, the authors also witnessed the unexpected implosion of the statewide caucus reporting process.

When the state Democratic Party failed to report the 2020 caucus results in a timely manner, researchers began to assess how party and press officials co-constructed the events. Specifically, we were interested in how the Iowa Democratic Party responded to the crisis event and the extent to which the party organization succeeded in distributing favorable and coherent messaging about the caucus process to the local press. This research presents an in-depth case of the Iowa state and local Democratic Party’s public communication response regarding an event described by the Des Moines Register as “hell” and a “results catastrophe” (Pfannenstiel, 2020). Drawing on organizational crisis management and echoing press perspectives, we chronicle the days around the caucuses in how party messages developed and how local news outlets covered the events. This mixed-methods case study provides insight into a local political event with national implications: The Democratic Party—due to its role as an institution and unique organizational structure—struggled to respond to a crisis.

The 2020 Iowa Caucus Results Delay

The use of digital and social media technologies has only recently been added to the traditionally interpersonal, grassroots of the 2020 Iowa Democratic caucuses. For the 2020 caucuses, the state party had contracted with a then relatively unknown technology company, Shadow, to develop a new smartphone application called IowaReporterApp to digitally report results from caucus sites and streamline the process (Romm et al., 2020). IowaReporterApp was not the first reporting app since an app designed by another company had been used for the 2016 caucus. Yet, even as the party valued the efficiency that digital technologies could offer, party officials were sensitive to concerns about the security of caucus results. Foreign actors had hacked the Democratic National Committee and Clinton presidential campaign in 2016 (Lipton et al., 2016), and party officials wished to prevent any further technological issues that could affect the veracity of the election results. In response, the national Democratic Party sent a team of staffers to Iowa to monitor and prevent the spread of digital disinformation and misinformation during the 2020 caucus (Marks, 2020). In addition, the Department of Homeland Security’s cybersecurity division comonitored election efforts with federal and state party officials.

Technology use often divided state and federal party officials, revealing different organizational goals for conducting the voting event. The after-caucus audit sponsored by the state party notes that the Democratic National Committee vetoed or altered state technology-related plans at several points prior to the caucus (Klinefeldt et al., 2020). First, the Iowa Democratic Party developed plans for six virtual or tele-caucuses designed to promote accessibility and enhance voter turnout (Parks & Detrow, 2019). The Democratic National Committee vetoed these plans due to security concerns. Second, and most important in understanding the ultimate delay in results, the national party committee made a “late demand” for additional technological features from Shadow, Inc. that would allow the app to report results directly to federal party officials for monitoring and cross-checking (Klinefeldt et al., 2020). Although late feature changes are common in software development, such volatility when combined with a limited time frame can severely deteriorate technical and user experience testing as well as training.

On February 3, a combination of user and technical errors created a delay in reporting the results on caucus night. A combination of inadequate volunteer training, individual use errors, and technical issues led to a cascading set of events that stymied the receipt of results from caucus sites to the state Democratic Party. Shadow, Inc., acknowledged and regretted the delay on the afternoon following the caucuses in a series of tweets while framing its role as one of support relative to the party infrastructure.

We will apply the lessons learned in the future, and have already corrected the underlying technology issue. We take these issues very seriously, and are committed to improving and evolving to support the Democratic Party’s goal of modernizing its election processes. (Romm et al., 2020)

This delay left party officials, candidates, voters, and the press searching for a prevailing narrative about the results following the caucus. The results of the Iowa Caucuses provide an initial filter of presidential candidates and sparks momentum for the following months of U.S. presidential campaigning. With a crowded candidate field and a lack of immediately coherent results, the 2020 Iowa caucuses felt more chaotic than ever before.

This lack of results created a spirit of uncertainty that flowed through the election night remarks of the top five Democratic candidates for President of the United States: former Vice President Joe Biden, former Mayor of South Bend Pete Buttigieg, and Senators Amy Klobuchar, Bernie Sanders, and Elizabeth Warren. Beyond similar tactics in addressing the uncertain reality of the moment, the candidates diverged sharply in how to interpret Iowa, as well as the state party’s role in organizing the caucus. Buttigieg and Sanders both framed a victory while leaving the organizational efforts unacknowledged. Warren noted the race as “too close to call” and Klobuchar said her campaign was “punching above our weight.” All four candidates used Iowa to construct momentum for the next nominating contest in New Hampshire. Biden was the only candidate to not only frame a close contest but also to acknowledge the efforts of the state Democratic organization. “Well, the Iowa Democratic Party is working to get this result, get them straight, and I want to make sure they’re very careful with their deliberations.” A lack of results meant all five candidates claimed momentum going into the New Hampshire primary.

As the first opportunity to set the tone and momentum for the presidential nominating process, and a salient example of local politics, the Iowa caucuses appeal to local and national news audiences. When the state Democratic Party failed to report results in a timely manner, a crisis developed that threatened the reputation of the state party, but also the legitimacy of the process itself. How the party and the local press co-constructed these events provides case for understanding how organizational structure can influence crisis response and the importance of local news in election sensemaking.

Party and Press Co-construction of Caucus Reality

Party Organization and Crisis Response

A crisis creates a need for information as people seek to understand why an event occurred and who is responsible (Coombs, 2007; Weiner, 2006). The evolving and often fast-paced nature of crises necessitates a swift response in “one voice” from the entities affected (Coombs, 2017). In a digital media environment, the initial organizational response should be as quick as possible, offering accurate information, even if it is incomplete (usually within an hour; Coombs, 2007, 2017; Sohn & Lariscy, 2014). The initial information—and how that information is framed—are crucial to public and press evaluations of organizational reputation (Fombrun & van Riel, 2004).

Situational crisis communication theory is one of the primary approaches used to respond to crises. It recommends organizations match their response to the crisis type and likely attributions of responsibility to protect reputation and preserve stakeholder relationships (Coombs, 2007, 2017). An appropriate response either works to distance an organization from an event (denial); minimize responsibility by framing lack of control (diminishing); and/or apologize and compensate (rebuild). In some instances, an organization also may choose nonresponse. Choosing an appropriate response depends on understanding the type of crisis. An appropriate response matches the type of crisis (e.g., using denial when an organization is responsible will increase reputational damage) and is effectively implemented (Coombs, 2017). In addition to clarifying responsibility, protecting organizational reputation also requires addressing other issues such as managing information flow to stakeholders (Ma & Zhan, 2016).

The organizational structure of the Democratic Party complicates the necessity of quickly explaining why a crisis occurred and who is responsible in a unified way. First, the Democratic Party is both an institution and organization. Institutions are the more enduring “constellations of established practices guided by enduring, formalized rational beliefs that transcend particular organizations and situations” (Lammers & Barbour, 2006, p. 357). In contrast, organizations can be viewed as groups of people who work together to serve particular functional objectives. As taken-for-granted ideas and ideals, institutions create expectations that people expect to be fulfilled—in the case of political parties—that expectation includes election management and bolstering the legitimating function of elections in a democracy. As Jarvis (2005) notes, government textbooks treat political parties as largely institutional forces made up of political elites engaged in action, the unification of disparate members, and ideological clarification of party positions.

Second, the distributed and decentralized organizational structure of the Democratic Party also complicates a timely, unified response. Officially, the structure of American political parties is a hierarchy of federal, state, and local officials dedicated to winning office and controlling government (Schlesinger, 1984). In practice, political parties are structured based on different political functions and geographic locations. The federal party, located in Washington, D.C., is primarily responsible for national messaging, resource management, and federal elections (Hershey, 2017). In contrast, each state central committee is located in the capital (or another major city) and is responsible for statewide candidate recruitment, messaging, resource management, and partnering with local committees to engage in voter mobilization. Finally, each local committee is situated in its respective cities or towns close to their communities from which they recruit and coordinate the volunteers vital to voter registration, canvassing, and mobilization activities. In Iowa, the state party committee (the Iowa Democratic Party) is based in Des Moines, with 99 county party committees each managing their “gathering of neighbors” and strangers for the caucuses. Moreover, unlike the Republican party, the Democratic Party couples this distributed structure with a coalitional and pluralistic organizational culture (Freeman, 1986). These distributed responsibilities and local “center of gravity” for key functions of the democratic process (Key, 1948, pp. 289-290) make the party structure more akin to multiple organizations interconnected by a single institutional identity than a singular hierarchical organization.

Although these structural differences may be salient to people working within the federal, state, and local party organizations, public understandings of the political party often draw from the hierarchical identification of the political party as institutions. Herein lies the third challenge: Stakeholders, including candidates, news media, and general public, often view—or treat—a political party as if it were a singular organization with a unified message. Candidates, for instance, talk of political parties as hierarchical entities disconnected from voters (Jarvis, 2005). Journalists treat political parties as if they “own” particular issues—a set of expectations that carry over to coverage of a party’s presidential candidates (Hayes, 2008). This idea of a singular Democratic Party likely compounds expectations that “the party” should offer a coherent message and voice. Left unmet, such expectation gaps undermine organizational credibility and reputation (Sohn & Lariscy, 2014).

The Echoing (Local) Press and the Construction of Reality

News media serve a critical function in the co-construction of reality alongside the political actors covered, including in how journalists use language in reporting an event (Ogden & Richards, 1923; Sarkar, 2016). How news media frames a crisis influences the public’s view of an organization and the institution it represents. The actors that frame reality quickest and most effectively will more than likely be the ones to establish the prevailing reality of the situation for the public (Edelman, 1985). Initial information offered to and by the media—with their corresponding explanations and attributions of responsibility—tends to be sticky, even if the initial information and corresponding frames are inaccurate or unfair. An inability to provide a fitting and quick response will lead news outlets to fill the information vacuum created by a crisis. This alternative narrative may rely on information problematic to the organization’s interests. Whether and how an organization responds to a crisis influences how the event is likely to be reported and interpreted by news audiences.

How reporters respond to crises affects how news consumers understand their community (Leupold et al., 2016; Richards, 2013) at national and local levels. Clearly signaling a perceived reality is important for influencing journalists and the public (Edelman, 1985). The opportunity for influence can occur when crafting a message about an object or event, and also when consuming and processing a message about an object or event. For example, if Iowa county party chairs want to reassure the public that their caucus sites ran smoothly and transparently, they must intentionally craft a message that frames caucus operations in a desirable and understandable manner. The ability to set the frame of a crisis, and thus shape the reality of that crisis (Coombs, 2017; see also Edelman, 1985) is a powerful tool worth investigating in the context of the Iowa caucuses and the local Iowa news media.

To explore further the local news media’s choices to adopt the communicative reality of political elites, we look to the notion of the echoing press. An echoing press perspective examines how news entities adopt the semantic structure—including words and phrasing—of political actors covered (Domke, 2004; Wiemer & Scacco, 2018). Direct quotation of political elites as sources is a prominent means by which echoing is accomplished and studied by researchers (e.g., Coe & Bradshaw, 2014). Lack of cohesion in a source’s messaging or absence of visible messaging decreases the ability for the press to effectively echo that source’s messaging which in turn will compromise an organization’s ability to manage how people perceive the reality of the situation.

Perspectives on the echoing press argue that the media will most likely take on the dominant discourse of relevant political elites when the current political moment is disrupted suddenly, such as a moment of crisis (Coe et al., 2004; Domke, 2004). The media’s role is generally seen as the fourth estate—traditionally making the media a watchdog of the government; however, times of crisis have historically led the media to drop more adversarial functions and act instead as a megaphone for relevant political actors (Bennett et al., 2007; Graber, 2001; Zaller, 1994). Although this line of research predominantly examines how agenda-setting news outlets cover national crises, one might expect that local press would echo the narrative frame of political elites in a localized crisis when that local crisis has national implications.

Analyzing political elites’ messaging compared with the local Iowa press’s messaging during the 2020 Iowa caucuses offers a fruitful case study in which to consider whether and how press echoing occurs in local crises with national implications. The local news in Iowa were tasked with a unique challenge of reporting on the chaotic events as they unfolded, both to inform their immediate local communities, as well as to satisfy the national attention placed on this consequential event. Although the media’s general role in influencing public opinion on political issues and events is openly acknowledged (e.g., Graber & Dunaway, 2015), local news has faced challenges that pose opportunities for theorizing around the echoing press.

Identified by the Federal Communication Commission as a troubled area of the information ecosystem, local news outlets have experienced declining resources and audiences in recent years (Pew Research Center, 2016; Waldman, 2011). In particular, local newspapers bear the brunt of decline (Martin & McCrain, 2019). Said declines deteriorate journalistic coverage, such as increasing the reliance on a single source (e.g., O’Neill & O’Connor, 2008) and cutting reporting beats like local politics and government (Hayes & Lawless, 2018). The interacting lack of resources and declining community interest means local outlets increasingly prioritize national stories and national sources at the expense of local journalism (Thorson et al., 2020). Although the extent to which such challenges directly influenced Iowa caucus coverage are beyond the scope of this study, a strained local news environment could nonetheless limit the ability of political organizations to adequately distribute messages to journalists.

Crisis moments can illuminate the possibilities and limits of local news coverage. Given the strained state of local journalism and the media’s typical responses to crises, we might expect that local Iowa news would heavily rely on, and echo, the messaging from the state party. Yet, that said, the crisis-ridden reality of the caucuses had the potential to distort the normal practices of press echoing and the means by which journalists could source information. Party messaging in the crisis may not have provided journalists a coherent message. Journalists, in turn, may have lacked access to party officials at higher levels of the organization even as they had access to official social media accounts. As the postcaucus audit produced by the state party noted, journalists “indirectly interfered with the results reporting process by reaching out to county chairs directly for information and calling the boiler room hotline, tying up the already busy phone bank” (Klinefeldt et al., 2020, p. 25).

Given the mixed expectations of how local Iowa news media and the state party chose to communicate about the caucuses, we propose the following research questions.

Research Question 1: To what degree was the Iowa Democratic Party’s digital messaging echoed in the local Iowa newspapers’ digital coverage of the 2020 Iowa caucuses?

Research Question 2: How did the Iowa Democratic Party’s digital messaging construct crisis attributions of responsibility and organizational legitimacy?

Research Question 3: How did local Iowa newspapers cover the 2020 Iowa caucus results delay?

Method

Procedure

To investigate caucus messaging from the state party in relation to the local Iowa press, we employed semantic network analysis and qualitative thematic analysis approaches. This mixed-method approached allows us to assess quantitative patterns in message use that can be contextually explained by qualitative analysis of digital messages offered by stakeholders (Creswell, 2014). We collected data from Twitter using a web scraping package for python 3.7.1 called twint (Twint, 2018). Twint can pull tweets from specific accounts during specific timeframes while bypassing the traditional limitations of the Twitter Application Programming Interface (Twint, 2018). We collected all relevant tweets from party officials and local news outlets sent on caucus day and the 2 days following the caucuses (February 3-5, 2020), a time frame that allowed for exploration of immediate and developing messages associated with the result delays.

The “voices” of party officials and local news outlets on Twitter were compared with assess message cohesion and echoing. For the party voice, all Iowa county party officials listed on the state party website with a Twitter account represented local sources. The state party communications chair, Mandy McClure, also was included to fully represent the state and local party. For the press voice, we selected the top 10 newspapers by circulation in Iowa with a Twitter presence (Agility, 2020).

The tweets were compiled into several corpora. Party tweets were grouped into a text corpus that included all tweets for all 3 days of the time frame (n = 115), as well as three separate corpora for each individual day in the time frame (February 3, n = 64; February 4, n = 32; February 5, n = 19). This same process was repeated for the combined tweets from all sampled Iowa newspapers during the time frame (All 3 days, n = 273; February 3, n = 129; February 4, n = 102; February 5, n = 42).

Data Analysis

The first part of the analysis employs semantic network approaches to assess the degree of shared associations of communicative phrasing between different party officials and local news media. This approach identifies reoccurring connections of words, such as transparent+process, and gives greater weight to more commonly repeated phrases compared with less commonly repeated phrases. In evaluating the degree of messaging overlap between party officials and local press outlets, we attempt to capture the degree of message echoing.

The text of the tweets was preprocessed using AutoMap, a text mining software (Carley, 2001; Carley et al., 2010). This process involved removing common metadata and high frequency, low meaning words for this context (e.g., stop words such as “the,” “and,” “a”). Last, we created a custom thesaurus to reconcile various words or phrases that hold the same meaning for the purposes of this study. For example, “Bernie Sanders,” “Senator Sanders,” and “Sen. Sanders” were all recoded as “berniesanders” to accurately represent connections between other words and reference to Bernie Sanders. We completed these techniques iteratively to ensure that the preprocessing of the text was compatible with the natural structure of the messaging (For a full review of preprocessing techniques in text mining, see Lambert, 2017).

A word co-occurrence list was produced for each of the text corpora to calculate all of the words that appeared together in the same two sentence window. The co-occurrence list gives a comprehensive list of all words that appear together in the text as well as how often those word associations occur in each corpus. To standardize the weights of the repeated associations, the frequency of word associations in each corpus was divided by half the number of total associations that existed in the corpus, then multiplied by 1,000. This method is common in network analysis as network metrics are often relative to the network within which they exist (Wasserman & Faust, 1994). This standardizing technique also allows for comparisons across semantic networks. Co-occurrence lists for each entity were imported into a network analysis software, NodeXL (Smith et al., 2012). Then, semantic networks were constructed for each entity’s corpora and exported as network matrices into UCINet, a software tool that offers more statistical capabilities for network-level analyses (Borgatti et al., 2002).

We used a network analysis technique called semantic saturation (Wiemer & Scacco, 2018) that employs Quadratic Assignment Procedure (QAP) to evaluate the communication echoing between the state party and the local Iowa press. QAP is a statistical technique that correlates two matrices of the same boundaries together and compares the statistical uniqueness of those matrices against 5,000 random permutations of a matrix with the same boundaries (Krackhardt, 1987). The unit of analysis in semantic saturation is the connection between individual words used in tweets, not the whole tweets themselves. Semantic saturation also provides a correlation coefficient for the similarity of two matrices. It therefore allows us to compare the direction and magnitude of the correlations between the entities of interest.

Network analytic results were supplemented by a qualitative thematic assessment of the messaging to provide an in-depth description of how the party’s public outreach and potential adoption by news sources manifested online. Building on the responsibility and legitimacy focusing necessary to crisis communication (Coombs, 2007, 2017), these areas served as sensitizing concepts to orient the analysis of party communication and how the local press responded (Patton, 2015). Tweets (and additional public communication materials where noted) were analyzed using emergent category designation, a qualitative approach where latent theory is used to abductively assess message construction (Erlandson et al., 1993; see also Tavory & Timmermans, 2014 on abduction). We turn to the results of the network analyses first, followed by a qualitative evaluation of the actual messages themselves.

Findings

The Press-Party Semantic Relationship

To investigate Research Question 1—which asked about the degree of message echoing between party sources and the local Iowa press in their online messaging about the caucuses—we used semantic saturation to assess the similarities and differences in language structure. The overall QAP results can be seen in Figure 1.

Figure 1.

Message correlations between party officials and local newspapers by day.

In discussions on Twitter that occurred within the 3-day time frame during and after the Iowa caucuses, there is a statistically significant overlap of messaging used by both party officials and the local press (r = .12, SD = .001, p < .001). The magnitude, however, indicates that a small number of similar words, combinations of those words, and frequencies of those combinations of words were echoed when discussing the caucuses on Twitter.

To investigate the observed relationship in message overlap further, we disaggregated the data by day to observe the evolving pattern of communication between party officials and the local press. Given the party’s concern over caucus integrity and existing research on crisis communication and message echoing, we might expect that message echoing may be strongest closest to the onset of the crisis (caucus day) with the relationship decaying as time and alternative narrative messaging appears. Caucus day, February 3, 2020, showed the biggest relationship (but still weak) between the messaging of party sources and the press (r = .14, SD = 0.002, p < .001) compared with the following 2 days. On February 4, 2020, the relationship remained significant but continued to weaken (r = .04, SD = 0.003, p < .001). By February 5, 2020, the echoing effect had completely disappeared.

These results, including the evident but initially weak observed echoing, begin to illuminate the messaging vacuum between party officials and the press, a precarious position for an organization attempting to shape a crisis narrative. Despite the statistically significant relationships on caucus day and the following day, meaningful semantic saturation (as indicated by the strength of the QAP correlations) between state and local party officials and the top Iowa print news outlets never occurred. Favorable conditions were present for the party organization to successfully relay messages to the press. The caucuses were a newsworthy event due to the national consequences of the presidential election and the local angle of Iowa’s prominence as the inaugural state in the nominating contest. Moreover, local political elites would serve as critical sources for reporting (Entman, 2003; Schudson, 2011). However, when relevant actors fail to provide a coherent narrative for journalists, or when organizational structure potentially influences incoherent crisis messaging, press sources will search for and frame events without the interpretation of organizational personnel.

The dissolution of communicative echoing between party officials and the local press warrants a deeper examination into the evolution of caucus messaging—especially given the importance placed on this particular event. In the following section, we present qualitative accounts from county party chairs and the press’s online reporting of the events, to help contextualize and explain the QAP results.

Party Messaging

Our second research question asked how state and local Democratic Party officials responded to the crisis. To answer this question, we qualitatively evaluated the tweets sent by individual Iowa county party chairs during the same 3-day period. We focused specifically on how party officials attributed responsibility and attempted to construct organizational legitimacy, to identify critical components of the crisis response (Coombs, 2007).

On February 4, the day after the caucuses, individual county party chairs largely communicated their gratitude for the hardworking volunteers and their confusion about what was happening at the state-party level with reporting the results. Boone County (@boone_iowa) tweeted

Not sure what is happening at the state-level right now, but our local caucuses went very well. Most quite smoothly—thanks to all of our dedicated volunteers! Special thanks to precinct chairs, secretaries, and others who put in so much time to get trained. (#IowaCaucuses)

Similar messages of thanks were tweeted from Winnebago County to the “volunteers who worked hard to see that things ran smoothly” and from Clay County leadership that wanted to ensure volunteer “efforts are not unnoticed.” By centering their message on the volunteers and party officials necessary for local efforts, the county party chairs attempted to reassure followers of the routine, normal, and therefore trustworthy, operation of their county’s caucuses.

Such messaging by local county party chairs was paired with an acknowledgement of the uncertainty surrounding the state-level procedures, who used diminishing and denying responses to distance themselves from any responsibility for the reports delay. Clay County acknowledged the delay with “While we wait for the results” and Franklin County messaged that a lack of results was “frustrating.” Boone and Dubuque counties both made a separation between the county-level operations of the caucuses and the state-level reporting of the caucus results. Dubuque County (@dbqdems) tweeted

All Dubuque County Precinct Chairs either reported their results via the app, or phone, or were on the phone until the call was dropped. We do not know what happened but do know our county caucuses ran well, and transparently.

In highlighting the process, Dubuque County attempts to pinpoint where failure occurred. Responsibility is shifted away from county efforts (“our county caucuses ran well”) to technology and state infrastructure (“until the call was dropped”).

In recognizing local interdependency with the state party while maintaining distance, the county party committees were careful to note the legitimacy of the process itself and the areas where local officials felt the state party succeeded. Dubuque County clarifies in a separate tweet on February 4 that “The IDP provided good training. We have raw data. We will continue to work to report out our results.” Polk County highlighted that its own local processes emanating from the caucuses continued apace. “Fearless credentials folks march on with data entry for county convention committee members.” Poweshiek County preferred to link to an article explaining why the public could “trust the Iowa caucus results once they’re released.” The messages prioritize the legitimacy of the caucuses, the results, and ultimately the state party itself while also bolstering the local party’s reputation and distancing themselves from responsibility (i.e., well-trained, credentialed volunteers). Describing successful party actions and processes helps bolster Iowa’s reputation, place, and institutional legitimacy, in this stage of American elections.

In contrast to local party officials, the state party issued little official communication online related to the crisis on caucus day or the 2 days following. The Iowa Democrats Twitter account did not post a tweet during this time frame that in any way acknowledged the status of the situation. There also were no publicly available press conferences or press releases on the official website during these critical few days following the caucuses. Troy Price, the then-chair of the state party, held a meeting with several Democratic candidate campaigns’ staff in the early morning of February 4, but reportedly did not provide satisfying answers to the campaigns and or take any questions from the press (Merica, 2020). The after-caucus audit identifies that the state party released three separate press releases on February 3 and 4 that set expectations for releasing results as well as noted the secure nature of the caucus process and app (Klinefeldt et al., 2020). Price also held a press conference on February 4. These outreach materials are not located on the party website as of this writing. Such communication decisions illustrate that the state party identified the presidential campaigns as a key stakeholder to be placated (due to the time and money invested in Iowa) and that public outreach prioritized traditional media channels.

In assessing the press releases and press conference issued by the state party offline, messaging focused on the legitimacy of the process and the party’s role in securing it. The February 4 press release notes the accuracy of the results and process five times while highlighting the pending investigation of the reporting delays. “It became clear that there were inconsistencies with the reports. The underlying cause of these inconsistencies was not immediately clear, and required investigation, which took time” (CNN, 2020). From a crisis communication standpoint, the initial press release does not identify any source of responsibility. It is not until the afternoon of February 4 that then-chair Price in a brief press conference admits that the delays were “unacceptable” and he “apologized deeply” by accepting responsibility as chair (C-SPAN, 2020). Price’s message then defaulted to the “accuracy and integrity” as well as “paper trail” of the results while answering press inquiries about whether his leadership could be trusted and if Iowa’s lead off status was threatened.

The apparent lack of internal message cohesion among local and state party members helps explain and reinforce findings from the network analysis. Due to a lack of party consensus regarding who was responsible for publicly messaging about the events and who were the key stakeholders, a unified message and frame did not materialize for the public, the presidential campaigns, or the press during the crisis. Local party officials diminished or denied responsibility for the delays while accepting responsibility for “smooth” community caucus operations. The state party did not accept responsibility in the first set of statements to legacy media and did not use digital media venues to rapidly disseminate messages and correct misinformation. The leading candidates were left publicly searching for an explanation in their caucus night remarks. So too were local news outlets in their reporting. With little information regarding the results and absent a unified message from the multiple voices of local and state sources, the local press were left to frame a narrative about the caucuses.

The Press Response

Our third research question asked about the extent to which local Iowa newspapers adopted party messaging. Traditionally, both political elites and the news media are dependent on one another for information distribution and each have their own agenda and incentives for what that information might be (Schudson, 2011). Political leaders have a desire for journalists to amplify their message and the news media often rely on political elites as sources, even though both entities largely maintain independence from one another.

When the state party left a narrative vacuum surrounding the caucus events, our qualitative analysis of local press coverage highlighted a competing set of narratives for such a newsworthy event. The first of these narratives was a perceived behavioral split in how local and state party officials responded to the caucus delays. Messaging inconsistencies within the party organization were covered by press outlets as evidence of conflict and procedural inconsistencies. On February 4, the Gazette (@gazettedotcom)—following the messaging of local party leaders—tweeted that “Site leaders say Iowa caucuses went fine. Getting results to the state party didn’t.” The Sioux City Journal (@scj) reported on concerns expressed about the app by local party officials. “The head of the Woodbury County Democratic Party said he encountered problems when he tested a new app to [report] presidential caucus night results to the state party.” The Des Moines Register (@dmregister) tweeted an interview regarding “Chair of the Scott County Democrats details #IAcaucus experiences, issues.” After the state party chair gave an afternoon press conference on February 4, the Register began to prominently feature state-level messaging on Twitter; however, the initial information vacuum was filled by newspaper reports based on explanations offered by county party chairs.

Absent actual results to game the presidential horserace, local news reports described confusion and inconsistencies—frames that chip away at the (institutional) Democratic party’s reputational goal of transparent, secure, and therefore trustworthy electoral processes. Normally, despite the news media’s perceived role as an unbiased intermediary between elites and the public, certain circumstances can arise where journalists will echo elites’ narrative frames (Domke, 2004; Entman, 2003). Some stories and online messaging adopted the state party’s crisis messaging that described the results as “inconsistent.” On February 4th, the Iowa City Press-Citizen (@presscitizen)—in a partial echo of party messaging that the results were secure—nonetheless noted discrepancies. “The Iowa Democratic Party said early Tuesday the ‘underlying data’ collected on the smartphone app used to report caucus results ‘was sound’ but the system was only reporting out partial data.” The Quad-City Times (@qctimes), Waterloo-Cedar Falls Courier (@wcfcourier), and Sioux City Journal (@scj) all adopted the state party language of “inconsistencies” to describe the results. Moreover, on February 5, the Des Moines Register (@dmregister) tweeted about the inability for the state party to accurately report out the results claimed to be secure and accurate. “After its struggles to release proper results from Monday’s #IAcaucus, the party released some mistaken results Wednesday afternoon.” Although these instances illustrate press officials using materials and messaging provided by state party sources, the narrative of “confusion” counteracted the party message emphasizing the accuracy and therefore trustworthiness of caucus returns.

When met with a lack of proactive and coherent Democratic Party messaging from state and local officials, the press was left to describe the nature of the caucuses as chaotic; imperiling Iowa’s coveted lead-off status. The Gazette (@gazettedotcom) tweeted stories labeling the caucus as a “fiasco,” “meltdown,” and “crisis.” For instance, on February 4, the newspaper noted “Online conspiracy theories flourish after Iowa caucuses fiasco.” The Sioux City Journal (@scj) tweeted “When everyone here and across America finally learn the final votes, those results will be overshadowed by frustration, anger and conspiracies about how they were tabulated and released. The Iowa story will be largely negative.” The Waterloo-Cedar Falls Courier (@wcfcourier) and the Telegraph Herald (@telegraphherald) added stories about the consequences for future caucuses. The Herald summarized on February 4 that “Iowa’s coveted voting status in doubt after delay on results.”

From an echoing press perspective, the ability for the press to leverage its symbiotic relationship with political elites hinges on the presence of some semblance of a narrative frame to echo in the first place (Entman, 2003). Without a coherent, proactive state party message, control of the narrative deteriorates. Local newspapers adopted party messaging of results “inconsistencies” and “delays,” but such adoption also reinforced the reporting delay and perceived organizational failures. The Democratic Party could then be easily labeled as “incompetent,” as the Sioux City Journal tweeted in a letter to the editor on February 5. Press treatment of the contradictions between local and state party leaders as well as the description of the caucuses themselves validates the communication divergences observed in the network analysis between party officials and the local press.

Discussion and Conclusions

The 2020 Iowa caucuses illustrate how a decentralized and distributed organization complicates an institution’s ability to adequately distribute and manage a unified public message that speaks to appropriate responsibility attributions and institutional legitimacy. A messaging vacuum allowed local (and national) press to construct a narrative that focused on conflictual signals within the party organization, results inconsistencies, and ultimately chaos given public expectations of a unified institutional message. Such party recriminations continued well into the general election contest. However, 9 months after the caucuses took place, the state party released its postcaucus audit that shifted responsibility to the federal Democratic Party for programming issues with the app and the inability for the state party to train volunteers with the new technology.

In the last days or weeks leading up to the caucuses, the [Democratic National Committee] DNC demanded that it have access to the [Iowa Democratic Party] IDP’s incoming reporting data in real time on caucus night in order to perform its own caucus results calculations and confirm the IDP’s results before the IDP could report them to the public. In doing so, the DNC took control over the reporting of results. When the DNC’s database conversion tool failed to work correctly, it caused the DNC to wrongly stop the IDP from reporting its results, and the IDP’s entire planned reporting process was thrown into disarray. The DNC’s interjection was the catalyst for the resulting chaos in the boiler room and in the IDP’s attempts to manually collect and confirm caucus results by hand. If the DNC had not interjected itself into the results reporting process based on its erroneous data conversion, caucus night could conceivably have proceeded according to the IDP’s initial plan. (Klinefeldt et al., 2020, p. 23)

This audit contradicts the apology and acceptance of responsibility by then-chair Troy Price on the afternoon of February 4. Moreover the 26-page report dedicates a mere 35 words to recommend the state party “proactively [manage] media expectations regarding the timing of caucus results reporting” (Klinefeldt et al., 2020, p. 25). Ultimately, this research goes beyond the failure to set media expectations to how (and what) the party organization communicated in the days following the crisis.

In our first step toward untangling the communication environment surrounding the caucuses, we found that the party’s online messaging about the caucus debacle and the local press’s online reporting on the caucuses had little overlap. Even when the messages from each entity were disaggregated and examined by day, the semantic relationship between party officials and the local newspapers was quite small and completely deteriorated only 2 days after the caucuses. The semantic saturation analyses allowed us to take a large amount of textual data over several days and assess the degree of message echoing in the communication environment. This quantitative inductive method offered macro-level insights about message dissemination and adoption happening at the same time about the same event.

To explore further the dynamics of crisis messaging and echoing, as well as why message echoing failed to occur, we then qualitatively assessed the messaging from party officials and the top Iowa newspapers. For party officials, messaging was differences occurred along organizational lines and failed to meaningful cohere around a credible, prompt, and unified message necessary for effective crisis response. The lack of narrative frame provided by the state party in the initial hours of the crisis left local county officials without a template or frame for how to communicate about the unfolding events. In response, county party chairs tweeted expressions of gratitude toward their volunteers who ran the individual precincts and caucus sites, diminished or denied responsibility for the delays, and constructed the legitimacy of their local caucus processes. Given the configuration of the Democratic Party as heavily reliant on volunteer labor, as well as the decentralized and distributed structure at the state level (particularly with regard to caucus and elections management), this sentiment from county party chairs is appropriate from an organizing standpoint. As individuals representing an institutional force in American politics, county party officials need to be able to continue to recruit volunteers and voters from their local communities going forward. In the absence of a clear narrative, local officials’ personal evaluations of what happened and who was responsible began to fill the information vacuum—to the detriment of the institution’s reputation. When these communications began to contradict state party messaging, including some acceptance of organizational responsibility, the conflictual responses compounded the crisis situation and undermined the reputation of the Democratic Party as institution and the Iowa caucuses.

Consistent with media responses to crises, the local Iowa press sought relevant political elites to report on the narrative of the events as they unfolded. Although these circumstances may have normally led to large amounts of echoing in the press from political elites, the relevant leadership did not effectively communicate an immediate crisis response reinforcing the information vacuum created by the crisis. Consequently, journalists looked within the perceived hierarchy of the institutional Democratic Party for local sources to frame the initial state of the presidential race. Given that the institutional hierarchy of political parties is often seen as the organizational structure of the party by political elites, including candidates and news outlets (Jarvis, 2005), journalists attempted to source information given the constraints of a rapidly evolving situations. Yet the dynamics of the distributed and functionally divided organizational structure in practice created a state-local disconnect in goals and stakeholders—despite their interdependent public reputation. For audiences who perceived the party as a unified hierarchy, Local sources interviewed by the press offered seemingly contradictory takes about smoothly run county operations and state-level backlogs in reporting results. Most notably, a caucus precinct chair in Story County, Iowa was being interviewed on air by CNN’s Wolf Blitzer and missed the opportunity to report results to the state party headquarters (Bowden, 2020). The semiautonomy of local committees and their perceived success openly conflicted with the state organization’s relative silence and delays in managing the results. This conflict, and the chaos it communicated, easily constituted newsworthy content for journalists, even as it damaged the party’s reputation.

Theoretically, this study contributes to echoing press theory and situational crisis communication theory in several key ways. The consideration of situations where the press’s echoing is expected to be high or low is a component that researchers interested in echoing press processes should continue to explore. Using the 2020 Iowa caucuses as a case study, typical expectations of high echoing press processes did not come to fruition. Local news would be expected to heavily echo Democratic party leaders under normal circumstances, and even more so in a moment of political crisis. Yet, considerations of organizational composition and norms for both the Democratic party and local Iowa news illuminate the near impossibility of substantial coherent press echoing during the 2020 Iowa caucuses. The structure of relevant organizations matters for how reality is constructed around a crisis event involving those organizations. The organizational composition of all actor groups involved in a crisis event plays a crucial role in how the public subsequently learns about that event.

Practically, the findings from this research further demonstrate the unique role of local journalism during moments of local and national crisis, as well as the potential pitfalls of certain organizational compositions. According to our results, local journalists would benefit from a firm understanding of the role they play in moments of crisis when reporting on events as they unfold. Particularly when crisis events have national implications, local journalists may be at a disadvantage when relying on their traditional norms of echoing their sources. This is especially true amid declining resources and motivations for local journalism (Pew Research Center, 2016; Waldman, 2011). Along with this, organizations as large and consequential as the state and federal Democratic parties should consider how the structure of their organization has implications for the strategies at its disposal to combat an ongoing crisis.

Even as this study offers insights into the need to consider organizational structure and institutional perceptions when understanding press echoing during crisis response, our study is not without limitations. First, the semantic saturation analyses are correlational, meaning we cannot make causal claims about the relationship between party officials’ and local Iowa newspapers’ digital messaging. To further validate the semantic saturation correlations and present a more holistic picture of each entity’s messaging strategies (or lack thereof), we supplemented our quantitative analyses with qualitative accounts of the messages themselves. We note that the lack of substantial digital messaging from the state party during the period of analysis also complicates the ability to observe such effects. Moreover, a lack of material necessitated the researchers to engage in post hoc data collection of press releases and press conference transcripts not featured online to fill in details of how the state party responded. Furthermore, as a critical case analysis of a single crisis situation, these results would benefit from further study in other crisis situations and contexts.

In assessing the communicative relationships between party officials and local news outlets, this research builds on organizational crisis management (Coombs, 2007, 2017) and echoing press approaches (Domke, 2004) to focus on how relevant groups of actors construct the discursive environment in a rapidly developing crisis situation. By examining a political party organization, that is viewed simultaneously as a democratic institution (Lammers & Barbour, 2006), and how it operated in relation to local news entities during a crisis, this research contributes to literatures in organizational and political communication. Additionally, identifying some of the organizational and crisis dynamics that may make message echoing more or less likely to occur builds on previous work that interrogates the issue and political conditions where national media echo political elites (Coe & Bradshaw, 2014; Domke, 2004). The dynamics that led to a lack of echoing in the local press are a consideration future scholarship should continue to assess.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Eric C. Wiemer (MA, University of Illinois at Urbana-Champaign) is a doctoral candidate in the Brian Lamb School of Communication at Purdue University. His research is concerned with political communication, media effects, and the relationship between political elites, the news media, and the public.

Joshua M. Scacco (PhD, University of Texas at Austin) is an associate professor in the Department of Communication at the University of South Florida. He also serves as a Faculty Research Associate with the Center for Media Engagement. Scacco is an expert on political communication and news media, having published more than 50 academic articles, book chapters, and public research papers.

Brenda Berkelaar (PhD, Purdue University) is an organizational communication scholar who has authored more than 50 research articles and chapters on work and technology.

References

Agility. (2020, August). Top 10 Iowa daily newspapers by circulation. https://www.agilitypr.com/resources/top-media-outlets/top-10-iowa-daily-newspapers-by-circulation/

Bennett

W. L.

Lawrence

R. G.

Livingston

(2007). When the press fails: Political power and the news media from Iraq to Katrina. University of Chicago Press. https://doi.org/10.7208/chicago/9780226042862.001.0001

Borgatti

S. P.

Everett

M. G.

Freeman

L. C.

(2002). Ucinet 6 for Windows: Software for social network analysis. Analytic Technologies.

Bowden

(2020, February 3). Iowa precinct chair hung up on when calling state party while live on CNN. The Hill. https://thehill.com/homenews/campaign/481311-dnc-officials-hang-up-on-iowa-precinct-chair-while-live-on-cnn

Carley

(2001). AutoMap (version 3.0.10.41) [Computer software]. CASOS, Carnegie Mellon University. http://www.casos.cs.cmu.edu/projects/automap/index.php

Carley

K. M.

Columbus

Bigrigg

Diesner

Kunkel

(2010). AutoMap user’s guide 2010 (CMU-ISR-10). Carnegie Mellon University. http://www.casos.cs.cmu.edu/publications/papers/CMU-ISR-10-121.pdf

CNN. (2020, February 4). Iowa Democratic Party statement on caucus reporting. https://www.cnn.com/2020/02/04/politics/iowa-democratic-party-statement/index.html

Coe

Bradshaw

S. C.

(2014). Toward a fuller understanding of the echoing press: Presidential addresses and The New York Times, 1933-2013. Communication Theory, 24(3), 272-290. https://doi.org/10.1111/comt.12037

Coe

Domke

Graham

E. S.

John

S. L.

Pickard

V. W

. (2004). No shades of gray: The binary discourse of George W. Bush and an echoing press. Journal of Communication, 54(2), 234-252. https://doi.org/10.1111/j.1460-2466.2004.tb02626.x

10.

Coombs

W. T.

(2007). Protecting organization reputations during a crisis: The development and application of situational crisis communication theory. Corporate Reputation Review, 10(3), 163-176. https://doi.org/10.1057/palgrave.crr.1550049

11.

Coombs

W. T.

(2017). Revising situational crisis communication theory. In Austin

Jin

(Eds.), Social media and crisis communication (pp. 21-37). Routledge. https://doi.org/10.4324/9781315749068-3

12.

Creswell

J. W.

(2014). A concise introduction to mixed methods research. Sage.

13.

C-SPAN. (2020, February 4). Iowa Democratic Party news conference on caucus results. https://www.c-span.org/video/?468988-1/iowa-democratic-party-holds-news-conference-caucus-results

14.

Domke

D. S.

(2004). God willing? Political fundamentalism in the White House, the “War on Terror,” and the echoing press. Pluto Press.

15.

Edelman

(1985). The symbolic uses of politics. University of Illinois.

16.

Entman

R. M.

(2003). Cascading activation: Contesting the White House’s frame after 9/11. Political Communication, 20(4), 415-432. https://doi.org/10.1080/10584600390244176

17.

Erlandson

D. A.

Harris

E. L.

Skipper

B. L.

Allen

S. D.

(1993). Doing naturalistic inquiry: A guide to methods. Sage.

18.

Fombrun

C. J.

van Riel

C. B. M.

(2004). Fame and fortune: How successful companies build winning reputations. Prentice Hall.

19.

Freeman

(1986). The political culture of the Democratic and Republican parties. Political Science Quarterly, 101(3), 327-356. https://doi.org/10.2307/2151619

20.

Graber

D. A.

(2001). Processing politics: Learning from television in the internet age. University of Chicago Press. https://doi.org/10.7208/chicago/9780226924762.001.0001

21.

Graber

D. A.

Dunaway

(2015). Mass media and American politics (9th ed.). Sage/CQ Press.

22.

Gronbeck

B. E.

(1989). Mythic portraiture in the 1988 Iowa presidential caucus bio-ads. American Behavioral Scientist, 32(4), 351-364. https://doi.org/10.1177/0002764289032004003

23.

Hayes

(2008). Party reputations, journalistic expectations: How issue ownership influences election news. Political Communication, 25(4), 377-400. https://doi.org/10.1080/10584600802426981

24.

Hayes

Lawless

J. L.

(2018). The decline of local news and its effects: New evidence from longitudinal data. Journal of Politics, 80(1), 332-336. https://www.dannyhayes.org/uploads/6/9/8/5/69858539/decline.pdf

25.

Hershey

M. R.

(2017). Party politics in America. (17th ed.). Routledge. https://doi.org/10.4324/9781315544427

26.

Hull

C. C.

(2006). Online organization: Dean, Kerry, and internet politicking in the 2004 Iowa caucus. In Williams

A. P.

Tedesco

J. C.

(Eds.), The internet election: Perspectives on the web in Campaign 2004 (pp. 57-66). Rowman & Littlefield.

27.

Iowa Democrats. (2020). Iowa democratic county parties. https://iowademocrats.org

28.

Jarvis

S. E.

(2005). The talk of the party: Political labels, symbolic capital, and American life. Rowman & Littlefield.

29.

Key

V. O.

(1948). Politics, parties, and pressure groups (2nd ed.). Thomas Y. Crowell Co.

30.

Klinefeldt

Yoshimura

Campbell

(2020). Internal review for the Iowa democratic party: 2020 Iowa caucuses. https://assets.documentcloud.org/documents/20423516/2020-11-10-internal-review-report-for-the-iowa-democratic-party-2020-iowa-caucuses.pdf

31.

Krackhardt

(1987). QAP partialling as a test of spuriousness. Social Networks, 9(2), 171-186. https://doi.org/10.1016/0378-8733(87)90012-8

32.

Lambert

N. J.

(2017). A text mining tutorial. In. Pilny

Poole

M. S.

(Eds.), Group processes: Computational and data driven approaches (pp. 93-118). Springer. https://doi.org/10.1007/978-3-319-48941-4_5

33.

Lammers

J. C.

Barbour

J. B.

(2006). An institutional theory of organizational communication. Communication Theory, 16(3), 356-377. https://doi.org/10.1111/j.1468-2885.2006.00274.x

34.

Leupold

Klinger

Jarren

(2016). Imagining the city: How local journalism depicts social cohesion. Journalism Studies, 19(7), 960-982. https://doi.org/10.1080/1461670X.2016.1245111

35.

Lipton

Sanger

D. E.

Shane

. (2016, December 13). The perfect weapon: How Russian cyberpower invaded the U.S. The New York Times. https://www.nytimes.com/2016/12/13/us/politics/russia-hack-election-dnc.html

36.

Zhan

(2016). Effects of attributed responsibility and response strategies on organizational reputation: A meta-analysis of situational crisis communication theory research. Journal of Public Relations Research, 28(2), 102-119. https://doi.org/10.1080/1062726X.2016.1166367

37.

Marks

(2020, January 29). The cybersecurity 202: DNC heads to Iowa to help protect caucuses from digital attacks and disinformation. The Washington Post. https://www.washingtonpost.com/news/powerpost/paloma/the-cybersecurity-202/2020/01/29/the-cybersecurity-202-dnc-heads-to-iowa-to-help-protect-caucuses-from-digital-attacks-and-disinformation/5e309dc888e0fa6ea99d626b/

38.

Martin

G. J.

McCrain

(2019). Local news and national politics. American Political Science Review, 113(2), 372-384. https://doi.org/10.1017/S0003055418000965

39.

Merica

(2020, February 4). Iowa Dem chair faces criticism from campaigns on overnight call. CNN. https://www.cnn.com/politics/live-news/iowa-caucuses-live-results-coverage-2020/h_875da71434aa16dc652bc3fdbf0dd220

40.

Ogden

C. K.

Richards

I. A.

(1923). The meaning of meaning of meaning: A study of the influence of language upon thought and of the science of symbolism. Harcourt, Brace, & Company, Inc.

41.

O’Neill

O’Connor

(2008). The passive journalist: How sources dominate local news. Journalism Practice, 2(3), 487-500. https://doi.org/10.1080/17512780802281248

42.

Parks

Detrow

(2019, August 30). DNC recommends scrapping plans for virtual Iowa, Nevada caucuses over security. NPR. https://www.npr.org/2019/08/30/755819116/dnc-recommends-scrapping-plans-for-virtual-iowa-caucuses-over-security-concerns

43.

Patton

M. Q.

(2015). Qualitative research and evaluation methods (4th ed.). Sage.

44.

Pew Research Center. (2016). State of the news media 2016. [Technical Report]. https://assets.pewresearch.org/wp-content/uploads/sites/13/2016/06/30143308/state-of-the-news-media-report-2016-final.pdf

45.

Pfannenstiel

(2020, February 8). Iowa caucus 2020: Inside the Iowa Democratic Party’s “boiler room,” where “hell” preceded the results catastrophe. Des Moines Register. https://www.desmoinesregister.com/story/news/elections/presidential/caucus/2020/02/08/iowa-caucus-2020-inside-democrats-chaotic-call-center-boiler-room/4690263002/

46.

Redlawsk

D. P.

Tolbert

C. J.

Donovan

(2011). Why Iowa? How caucuses and sequential elections improve the presidential nominating process. University of Chicago Press. https://doi.org/10.7208/chicago/9780226706979.001.0001

47.

Richards

(2013). Beyond city limits: Regional journalism and social capital. Journalism: Theory, Practice & Criticism, 14(5), 627-642. https://doi.org/10.1177/1464884912453280

48.

Romm

Satija

Harwell

(2020, February 4). How tech firm shadow sought to revolutionize Democratic campaigns—but stumbled in Iowa. The Washington Post. https://www.washingtonpost.com/technology/2020/02/04/how-tech-firm-shadow-sought-revolutionize-democratic-campaigns-stumbled-iowa/

49.

Sarkar

(2016). Text analytics with python: A practitioner’s guide to natural language processing. Apress. https://doi.org/10.1007/978-1-4842-2388-8

50.

Scacco

J. M.

Coe

(2017). Talk this way: The ubiquitous presidency and expectations of presidential communication. American Behavioral Scientist, 61(3), 298-314. https://doi.org/10.1177/0002764217704321

51.

Schlesinger

J. A.

(1984). On the theory of party organization. Journal of Politics, 46(2), 369-400. https://doi.org/10.2307/2130967

52.

Schudson

(2011). Why democracies need an unlovable press. In Graber

D. A.

(Ed.), Media power in politics (6th ed, pp. 33-44). CQ Press.

53.

Smith

Ceni

Milic-Frayling

Shneiderman

Mendes Rodrigues

Leskovec

Dunne

(2012). NodeXL: A free and open network overview, discovery and exploration add-in for Excel 2007/2010/2013/2016. http://nodexl.codeplex.com/

54.

Sohn

Y. J.

Lariscy

R. W.

(2014). Understanding reputational crisis: Definition, properties, and consequences. Journal of Public Relations Research, 26(1), 23-43. https://doi.org/10.1080/1062726X.2013.795865

55.

Squire

(2008). The Iowa caucuses, 1972-2008: A eulogy. The Forum, 5(4), Article 1. https://doi.org/10.2202/1540-8884.1212

56.

Tavory

Timmermans

(2014). Abductive analysis: Theorizing qualitative research. University of Chicago Press. https://doi.org/10.7208/chicago/9780226180458.001.0001

57.

Thorson

Chen

Lacy

(2020). National and local news consumption and indices of community participation in an American civic-journalism news environment. Journalism Studies, 21(4), 443-458. https://doi.org/10.1080/1461670X.2019.1672576

58.

Twint. (2018). Twint project [GitHub repository]. https://github.com/twintproject/twint

59.

Waldman

(2011). The information needs of communities: The changing media landscape in a broadband age (Report prepared for the Federal Communication Commission). www.fcc.gov/infoneedsreport

60.

Warner

B. R.

Galarza

Coker

C. R.

Tschirhart

Hoeun

Jennings

F. J.

McKinney

M. S.

(2019). Comic agonism in the 2016 campaign: A study of Iowa caucus rallies. American Behavioral Scientist, 63(7), 836-855. https://doi.org/10.1177/0002764217704868

61.

Wasserman

Faust

(1994). Social network analysis: Methods and applications. Cambridge University Press. https://doi.org/10.1017/CBO9780511815478

62.

Weiner

(2006). Social motivation, justice, and the moral emotions: An attributional approach. Lawrence Erlbaum. https://doi.org/10.4324/9781410615749

63.

Wiemer

E. C.

Scacco

J. M.

(2018). Disruptor-in-chief? The networked influence of president Trump in building and setting the agenda. Agenda Setting Journal, 2(2), 191-213. https://doi.org/10.1075/asj.18020.wie

64.

Winfrey

K. L.

(2017). Crazy for caucusing: Media use and voter opinions in the 2016 Iowa caucus. American Behavioral Scientist, 61(9), 1002-1023. https://doi.org/10.1177/0002764217735624

65.

Zaller

(1994). Elite leadership of mass opinion: New evidence from the Gulf War. In Bennett

W. L.

Paletz

D. L.

(Eds.), Taken by storm: The media, public opinion, and U.S. foreign policy in the gulf war (pp. 186-209). University of Chicago Press.