FORCES OF CHANGE: EXAMINING INFLUENCERS OF CONGRESSIONAL BEHAVIOR By Joshua Brendon Koss A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Political Science – Doctor of Philosophy 2024 ABSTRACT In an increasingly polarized political climate, understanding the sources that influence the behavior of institutional actors is essential to understanding the outcomes and dynamics that emerge from our institutions as a whole. While much has changed over the years, the core motives shared by members of Congress largely remain the same - though changes in the political environment force members to adapt their behavior to best satisfy those objectives. The secular belief shared by many social scientists and the public at large is that members are chiefly motivated by reelection, and they are willing to do any- thing to secure it. The increased rate of voluntary departures from the institution testify to the potency of other objectives that factor into the member career calculus. Examining the patterns of voluntary congressional retirement from 1973 to 2023, I find evidence that members consider their effectiveness and influence as legislators when determining if it is worth seeking reelection. Beyond intrinsic features that influence member behavior, external factors such as the President of the United States are also able to impact both behavior and the political fates of legislators. Presidents have long used their "bully pulpit" to engage with and ultimately influence Congress. President Trump’s use of Twitter, and its amplified media coverage, maximized his pulpit. I study how he used the platform over the length of his tenure in office as a political tool towards members of Congress - exploring who he talked about, how and when he did so, and finally, some of the consequences faced by his copartisans as a result. While the medium, scale, and tone departed from his predecessors in the Oval Office, Trump’s tactics still bear semblance to traditional executive-legislative relations. ACKNOWLEDGMENTS The winding road of graduate school and dissertation composition is one of the most arduous yet rewarding experiences of my life. With the completion of this dissertation there is no better time to express my deep and sincere gratitude towards the many individuals who helped me reach this point. My family and close friends have been my most important and essential support structure for me throughout my time in graduate school. I let my mother know how much she has meant to me every day but I must once again express my thanks for her endless sup- port for me throughout my life but especially these past six years when she has always ensured that I was comfortable and taken care of at my most anxious times. My brother and best friend, Jacob Koss, has been of vital assistance in providing a second set of eyes to help guide data collection and proofreading. Our close family friend Patrick Hayner, has also been tremendously supportive over the years and his confidence in me often out- stripping my own has helped propel me forward. I am grateful for my feline companions, Jaina and Kemba, for exclusively providing welcome distractions that helped alleviate my stress throughout the various stages of this project. My legions of aunts, uncles, and cousins whose company has been a welcome escape from the stresses of graduate school, with a special shout out for my cousin Brian who lent his graphical expertise to assist me with data visualization. I have to thank the Department of Political Science at Michigan State University which has been my home-away-from-home and a tremendous support system for me these past six years. The academic staff (especially Sarah Krause, Rhonda Burns, and Krista Zeig) and program coordinators (Ani Sarkissian, Sarah Reckhow, and Ian Ostrander) rarely get the credit they deserve for their tireless efforts to ensure the department functions and that we, as graduate students, are able to navigate any potential bureaucratic pitfalls along the way. Various faculty in the department, beyond my committee, played key and important roles in providing advice and encouragement throughout my MSU career, especially Marty Jordan, Eric Juenke, Andrew Kerner, Nazita Lajevardi, Mariana Medina, and Shahryar Minhas. I want to give special recognition and thanks to Matt Grossmann, whose mentorship during my time in IPPSR and assistance on the market helped mold my scholarship and find success in securing a job in academia. iii During my time in the Department, I was fortunate enough to meet several amazing friends and colleagues along the way. My institutionalist seniors, Jonathan King, Jessica Schoenherr, and Elizabeth Lane, helped me find my footing in grad school and East Lansing life and were always willing to offer excellent advice for my work and professionalization efforts. It has been a pleasure to share my journey with the final members of my cohort to remain from the beginning of my graduate school till the end, Ha Eun Choi and Tianhong Ying. I have also appreciated the support from many other fellow graduate students in the department including Benjamin Yoel, Erika Vallejo, Gerson Guevara, Raymundo Lopez, and Stanislav Skritki. Anthony Jesuale has been an incredibly supportive colleague, who will hopefully enjoy more of my focus on our collaborative research now that this milestone has been conquered. Having had the opportunity to teach multiple classes as a Ph.D. student at MSU, many of my students have provided an endless source of motivation and inspiration. I wish to provide special thanks to my undergraduate re- search assistants (Ellen Trease, Josh Taft, and Rebekah Batu) and learning assistants (Janey Salenbien and Allie Abramowski) for easing considerable burdens in data collection and course administration, without which completing this dissertation would have been far more difficult. I also wanted to give thanks to the various professors that mentored me as an undergraduate at Oakland University, David Dulio, Roger Larocca, Terry Towner, Alan Epstein, Nicole Matthews, Peter Trumbore, Paul Kubicek, and John Klemanski without whom I would not have gone on this journey in the first place. Many of these individuals continued to provide support that helped guide my development as a scholar and improved the production of this dissertation as well. Finally, there are the four individuals who made up my committee who have been instrumental in shaping my dissertation and academic career as a whole. All of them have provided incredible feedback on my research over the years along with sage advice for academia as a whole. Cory Smidt was instrumental in helping me find my footing within the discipline at the start of my graduate career, and has always been willing to provide excellent explanations for any of my methodological inquires. Ryan Black has been incredibly supportive of me since even before I joined the program and was another great reference to help me obtain gainful employment at the end of this journey. I owe a lot of my pedagogy and workflow habits to the lessons I learned as Ryan’s (frequent) TA. iv Jennifer Wolak joined our department about half-way through my graduate school tenure, and has been an exceptional mentor for my research and job market preparation ever since. When I’ve had my confidence waver in my research waver, Jenny has helped restore my faith in the validity and interest of my work. At last, my advisor Ian Ostrander, I’m not sure my words can fully express the extent of my gratitude. Your tireless support in reading countless drafts, bringing me on to other projects, securing research funding, aiding my navigation of the job market, and being an endless source of wisdom about academia and life in general. I will always be grateful for your constant encouragement and reassurances, especially while I was exploring options on the job market. I will miss our frequent (and often cynical) chats about politics and academia, but the wisdom you imparted will always be treasured. v TABLE OF CONTENTS CHAPTER 1: INTRODUCTION ............................................................................................................................ 1 CHAPTER 2: QUITTING CONGRESS: WHEN REELECTION LOSES ITS LUSTER ............................ 3 CHAPTER 3: TARGETING THE TWEETS: EXAMINING PRESIDENT TRUMP’S CONGRESSIONAL TARGETS ON TWITTER ............................................................................................... 29 CHAPTER 4: PUTTING THE "BULLY" IN BULLY PULPIT: PRESIDENT TRUMP’S CONGRESSIONAL PURGE THROUGH TWITTER ..................................................................................... 62 CHAPTER 5: CONCLUSION ............................................................................................................................... 84 BIBLIOGRAPHY .................................................................................................................................................... 86 APPENDIX A: CHAPTER 2 APPENDIX ...................................................................................................... 100 APPENDIX B: CHAPTER 3 APPENDIX ...................................................................................................... 102 vi CHAPTER 1: INTRODUCTION Conventional wisdom among social scientists, political pundits, and the public at large has long held that one of the greatest motivations for members of Congress is the continual quest for reelection, yet the willingness for many members to abandon their political office speaks to elusive ambitions within the institution which make their career increasingly unattractive. In particular, Chapter 2 advances a theory that members who are marginalized, as characterized by relative ineffectiveness in pursuing their policy agenda, are more likely to retire. This theory is empirically tested and supported using data on senators and representatives in the 93rd through 117th Congresses, and finds that ineffective legislators are more likely to voluntarily depart from the institution. In a world where incumbents enjoy low rates of electoral defeat, willful retirement becomes instrumental to member turnover and the composition of Congress in general, and better understanding of the motivations that lead members to leave has major consequences for the types of members we would expect to remain in Congress. Looking beyond internalized factors that influence the behavior of members of Congress, outside forces, particularly other political actors, can also prove influential. Chapters 3 and 4 more closely examine the role of the chief executive and how their actions towards Congress and it’s members can influence behaviors more broadly, especially among their copartisans. The highly public nature of Trump’s presidency, particularly through his prolific use of Twitter gave the American people a near-constant and unprecedented look at the not- so- inner thoughts of the president. While the medium President Trump used to vent his frustrations with members diverged from his predecessors due to the increased prevalence of social media and his own proclivities, the underlying nature of his actions are politically patterned and far from unique. Since the Trump-years, active efforts have been undertaken to emulate him within the Republican party, and his style may provide a road map for his successors. This paper investigates which members of Congress Trump used social media to target, how he talked about these members, and how this compares with historical trends in 1 presidential-congressional relations. The long-term use of this medium is uniquely suited to exploring inter-branch dynamics. Donald Trump’s victory in the 2016 presidential election produced a reckoning in the American political landscape, particularly within his own political party. The ascension of President Trump sparked debate regarding partisan realignment, and the extent to which he was a product of and/or contributor to the shifts occurring within the Republican party. This paper explores the active role Trump played, while president, to reshape the congressional delegation of the Republican party in his own image. Grievances be- tween the president and members of Congress, however, are far from new. Presidents have historically made overtures to discipline members to come in line with their political perspective and have also made attempts to oust members disadvantageous to their political vision, all to various degrees of success. Trump’s use of social media and his abrasive personality may have been a departure from previous Oval Office occupants, but his overarching objectives were far from new. Using data from Trump’s Twitter ac- count while he was in office, I find evidence that Trump was fairly successful in driving Republicans he did not like, out of office. 2 CHAPTER 2: QUITTING CONGRESS: WHEN REELECTION LOSES ITS LUSTER Why do members of Congress willingly choose to leave the institution on their own accord? The prevailing understanding of congresspersons is that they are first and fore- most interested in job security through their perpetual reelection. Indeed, being asked about voluntarily ended their congressional career has been known to leave some members speechless. Mayhew (1974a) has gone as far to say that some members "have no other interest" (16) than achieving reelection, and Fenno (1978) identifies it as one of their three most important goals, along with implementing good policy and achieving power and influence within the institution. Over the past several years, many high-profile members have left one of the two legislative bodies at ages most would consider to be premature for retirement, a strong indication that there is some credence to the notion that at least some members critically assess the utility they derive from their seats and determine if the time and effort exerted to maintaining that seat is worth it. Arguably, the primary appeal of reelection is that it is the most proximate goals and, by institutional design, is necessary in order to pursue the others outlined by Fenno (1978) - but failure to achieve those objectives may undermine the desire to continue seeking reelection. The implications for why members opt to leave are of paramount importance to the makeup and composition of Congress as a whole. The modern Congress is dominated by incumbency, with members serving longer in part due to institutional advantages that make reelection easier to secure (Carson, Engstrom and Roberts, 2007; Cox and Katz, 1995; Eckles et al., 2014; Erikson, 1971; King, 1991; Mayhew, 1974b). The incumbency advantage is realized through a variety of factors including the increased benefits such as "pork" they bring to the district, general familiar- ity with their member and his/her name, and the lack of collective responsibility which makes it difficult for constituents to blame their member of Congress for poor economic performance or other systemic struggles in the same capacity as they do for their exec- utives (Ansolabehere and Snyder Jr, 2002). These contribute to the classic paradox that 3 citizens hate Congress but love their individual member (Fenno, 1978). Not only has reelection become a more easily achievable objective, members have also opted to maintain careers within the legislature which has played a key role in the institutionalization of both chambers (Polsby, 1968; Price, 1971; White, 2000). While it may be true that the majority of members maintain this fixation on staying in office, and likely maintain this attitude throughout most of their congressional careers, the fact remains that many members willingly decide to retire from seats they would have otherwise kept. Bringing the consideration of the other accepted member objectives of influence and policy, sheds light on additional incentives to leave. Congress itself has been historically unproductive in recent sessions (Statistics and Historical Comparison, 2020), but there is considerable variation amongst members in their contributions to the bills that do still make it through the process. Even fake-out retirees such as Marco Rubio (R-FL) who initially planned to retire from the Senate regardless of the outcome of his 2016 Presidential bid, cited his general dissatisfaction with the lurching speed of the legislative process as the reason for his planned departure, but he ultimately made the decision to seek re- election, two months before the Senate primary (Press, 2016). This lack of power, or the prediction of a future diminished power, should make holding a Congressional seat less attractive if a member has interest in more than retaining a job and title. A key measure used to test this paper’s theoretical expectations of marginalized members being more likely to retire are Legislative Effectiveness Scores (LES) which are an aggregate variable consisting of a variety of factors that attribute to a members relative power in advancing legislation within their institution (Volden and Wiseman, 2009, 2014). LES provides an industry standard for legislator productivity and has even been used by members of Congress such as Senator Amy Klobuchar (D-MN) in the 2020 South Carolina Democratic Presidential Primary debate to tout her success in the chamber (South Carolina Democratic Debate Transcript: February 25 Democratic Debate, 2020). I find evidence that members in both chambers of the United States Congress are motivated to leave by their lack of personal productivity but there are some nuances that suggest some contrast between the House and Senate on this front. Members in both chambers are influenced by their institutional seniority, further supporting existing theories. These findings are highly consequential for our understanding of congressional 4 exits and member behavior as a whole. That less effective law makers are more prone to voluntary departures further reinforces that members of Congress have concerns and goals beyond simple reelection and that retirement’s important role in congressional turnover is more prevalent among the institution’s least productive members. 2.1 Congressional Retirement Why do members of Congress choose to retire? Some congressional retirements are, naturally, explained by age and health related reasons. The biological drivers of retirements are perhaps the most individualistic and difficult to predict as there is no standard cutoff age or standardized repository of member medical records. For the eldest members, out- side observers speculate each term whether or not that member will retire which, some- times cynically, turns into speculation of whether that member will die in office when they choose to pursue another term. Senator Dianne Feinstein (D-CA) was a lightning rod for this type of speculation for years before her death in the middle of the 118th Congress, with the 90- year-old’s failing health and mental state such that she was even unaware of her staff issuing a statement that she would not seek reelection in 2024 (Gans and Weaver, 2023). Other health factors are often kept private. While the deteriorating health conditions of some members become public knowledge well before a member leaves (through retirement or death) such as John McCain’s (R-AR) brain cancer diagnosis in 2017, many others either choose to keep these matters private until they announce that it has motivated their decision to retire from Congress or die from their ailment. The private nature of member health records and the more obvious implication poor health has on career decisions has largely minimized scholarly examination (Maltzman, Sigelman and Binder, 1996), which has generally accepted that it plays some role though not one that is as easily predicted. Another important category is political retirements such as those motivated by electoral vulnerabilities and scandals. Although the vast majority of members seeking re- election handily win their races, there are still members every cycle who are defeated and there are members with enough foresight to predict their likely loss and will ultimately choose to voluntarily call it quits rather than waste the time and financial resources needed to run a campaign (Ahuja et al., 1994; Bianco, Spence and Wilkerson, 1996; Brace, 1985; 5 Wolak, 2007). Similarly, members who are entrapped in scandal, particularly if that scandal has legal ramifications, are also likely to quit due to the vulnerable and embarrassing position it places them in (Shea, 1999). These strategic forms of retirement have been discussed in the literature but this project seeks to address an additional question that is not readily explained by electorally vulnerable members leaving - why do members who would most likely win their reelection efforts, and are in good health, willingly choose to leave Congress? This paper seeks to answer this question and contribute to the existing literature with a new focus on theoretically driven factors that influence retirement decisions, some of which have been under-examined. I forward a new theory that argues that marginalized members, as characterized by their relative ineffectiveness in the legislative process, are more likely to retire. The current literature on Congressional retirement has been narrowly focused in several capacities. It has been fairly typical for studies addressing other aspects of Congress more broadly to passively acknowledge that members retire strategically when they face reelection challenges (Jacobson and Kernell, 1983). This approach to Congressional retirement sets the stage for more deliberate focus on retiring members while acknowledging that their simple existence does not dispel Mayhew (1974a)’s work. In fact, retirement can very much be seen from the perspective of an electoral decision where members either elect to forfeit their reelection opportunity when they fear their electoral prospects or improve their political station by running for a higher office. A common approach to studying retired members has involved looking at large "waves" within single Congresses where particular mechanisms motivated a higher than average number of exits (Ahuja et al., 1994; Bianco, Spence and Wilkerson, 1996; Carson and Engstrom, 2005; Clarke et al., 1999; Groseclose and Krehbiel, 1994). For example, Carson and Engstrom (2005) provide a pre-modern example of constituents punishing legislators for their actions, and some legislators leaving before they were electorally ousted, during the 1826 Congressional elections as retribution for Congressmen who supported John Quincy Adams in the previous presidential election in districts that voted for Andrew Jackson. A more contemporary example of strategic retirement that received a considerable focus of scholarly analysis are the congressional elections of 1992 (Ahuja et al., 1994; Clarke et al., 1999; Shea, 1999). Ahuja et al. (1994) outline contextual factors that 6 contributed to retirement rates including a strong anti-incumbent mood caused by check- kiting scandals and the Clarence Thomas-Anita Hill Supreme Court hearings, 1992 marking the final opportunity for members to cash out their campaign funds into their personal assets, and the presence of newly drawn congressional districts. Further analysis of the following cycle by Clarke et al. (1999) suggests that members’ prospective income, as diminished by the campaign finance reform, was the most important factor in that cycle as it was an equal incentive for members of both parties to retire, unlike the scandals. Other studies also address members quitting rather than facing tough reelection bids, beyond the 1992 cycle. Stone et al. (2010) examine evidence from the 1998 cycle and conclude that even seemingly secure incumbents are still fearful about the prospects of electoral defeat and consider retirement accordingly. A recent analysis of the 2018 midterms reinforces previous findings that contextual factors can shape members of particular par- ties to more strongly consider their electoral vulnerability and make career decisions based on their perceived likelihood of defeat (Brant and Overby, 2021). More broadly, Brace (1985) and Kiewiet and Zeng (1993) offer frameworks that suggest members engage in risk-taking analyses and seek to avoid electoral defeat, especially in the wake of redistricting and that the existence, or lack thereof, of political opportunities to run for higher office shapes whether members choose to retire. These opportunities are most frequently characterized by members of the House making the decision to run for a vacant or vulnerable Senate seat but could also refer to members who ran for state-level offices rather than seeking reelection. Scandals are another angle retirements have been viewed from that are connected to the previous studies on strategic retirement due to the electoral costs and ramifications associated with scandals. Examining the aforementioned checking scandal from 1992, Shea (1999) finds that the media coverage or lack thereof, in a members district shaped the electoral ramifications of the scandal for those members, and in turn their decision to run again, with a positive relationship between media coverage and retirement. Paschall, Sulkin and Bernhard (2019) observed 253 cases of personal or professional scan- dals among members of congress from 1989-2012. They find that scandals have strong negative effects on members’ careers, harming their electoral prospects, reducing their legislative effectiveness, and harming their influence within the institution potentially 7 costing them committee chairmanships and other roles, important attributes of congressional service. Naturally those detrimental outcomes contribute to strategic decisions to not seek reelection or even to outright resign before the expiration of the member’s term. The effects of scandals can potentially linger well beyond the subsequent term after the scandal began, potentially affecting future bids. Taylor and Cobb (2019) finds that individual-level characteristics increase the likelihood for a member to become em- broiled in scandal such as those in positions of power and elected from districts that have traditionally been forgiving of scandals. They also find that marginalized racial minority members are more likely to be involved in scandal and subsequently be sanctioned by the chamber. Hamel and Miller (2019) examines congressional scandals from 1980 to 2010 and finds that their electoral impact has been mediated by the increased support those members receive from donors and voters becoming less punitive in the post-1994 period of nationalized electoral politics. The decreased costs of scandals suggested by these findings may mean that these events have grown to provide less incentive for members to willingly leave when their unethical actions have been exposed in recent years. A more niche perspective in the literature is the idea that members leave Congress because they have lost "enjoyment" of the institution and that the desirability of the job has lessened over time (Frantzich, 1978). Hibbing (1982) finds that a considerable spikes in Congressional retirement in the 1970s coincided with significant institutional reforms that reduced power among senior members, more elderly members reaching retirement age, and growing time-demands associated with the job. Lost enjoyment may disproportionately affect some members more than others, as Moore and Hibbing (1998) argues that situational dissatisfaction with Congress is strongly tied to a member holding an electorally marginal seat or being denied positions such as committee chairmanships. Displeasure with one’s current position in Congress is also closely tied to ambitions to run for higher office (Livingston and Friedman, 1993; Theriault, 1998). These findings tie into member’s ambitions to have influence in the institution and demonstrate a need for greater utility out of a congressional seat to merit retaining it. A gender component may also influence this frustration in the House, as Lawless and Theriault (2005) finds that women have greater rates of voluntary retirement as they are more susceptible to these factors that make Congress less enjoyable. 8 A noticeable trend in the House of Representatives is the relatively higher rates of retirement among Republicans compared to their Democratic counterparts. Various explanations for this phenomenon have been discussed (Ang and Overby, 2008; Gilmour and Rothstein, 1993; Murakami, 2009). Ang and Overby (2008) attribute Republican retirement decisions to their status as the "permanent minority" in the forty-two years be- tween 1953 and 1995. Republicans in more recent years are also faced with an additional institutional consideration as their chairs and ranking members have term limits,1 at the end of which they return the rank and file and lose much of their policy influence (Deering and Wahlbeck, 2006). Reynolds (2017) suggests these term limits played a role in the retirement rates among House members in the Republican caucus. Most of the literature has focused on the House of Representatives with only very few studies giving attention to the Senate. Examinations of retirement behaviors in both chambers finds commonality in the importance of age and electoral vulnerabilities (Livingston and Friedman, 1993). There is some variation in the degree to which these are important for members in each chamber, with representatives more concerned about electoral security and senators with their age (Bernstein and Wolak, 2002). Masthay and Overby (2017) test many of the existing theories on retirement in the House in the context of the Senate. They find that the institutional features of the Senate, including electoral insulation and less dominant party control, change retirement dynamics. Notably, Re- publicans were far more likely to retire from the House during the era of Democratic dominance of the chamber, but that relationship is nonexistent in the Senate, while age, being up for reelection, and being appointed to the Senate increase retirement. The body of Senate retirement research has thus far been limited in quantity and scope, largely focusing on contrasts between the House and Senate while not necessarily forwarding new explanations for retirement decisions for members in the upper chamber. Legislative Effectiveness Scores which form a key component of this paper’s analysis of retirement have been used in various capacities in existing political science work. Volden and Wiseman (2009)’s examination of their measure has found that it is closely correlated 1 Members of the Republican caucus are limited to serving a combined six years, or three terms combined as the ranking member and/or chair of a single committee. 9 with a member’s seniority, position within the chamber (e.g. leadership, committee chair, etc.), and in the House, majority/minority partisan status. The LES measure has proven to be a significant and useful factor in addressing questions about members’ legislative activities, such as the difference that may exist between men and women in Congress (Jeydel and Taylor, 2003; Stanley and Blair, 1989; Volden, Wiseman and Wittmer, 2010, 2013). Volden and Wiseman (2018)’s examination of legislative effectiveness within the Senate finds that, compared to the House, senators have less disparity between their scores due to the Senate being a more egalitarian body with more individualistic law- making potential for its members. Additionally, Volden and Wiseman (2018) note that members well known for obstructing the legislative process, score fairly poorly on this metric. While constituents are generally unaware of how effective their members are, they are more approving of members who have an effective track record when presented with their record (Butler et al., 2023). 2.2 Theory and Expectations The existing congressional retirement literature has made avoiding presumptive electoral defeat a strong theme and has also dabbled in idiosyncratic and personal factors that influence a members decision to leave. This paper seeks to contribute to this literature with a new focus on theoretically driven and predictable institutional factors that attribute retirement decisions to congressional dynamics, some of which have been under-examined. At the core of this theory are marginalized members which are defined as members who are ineffective at accomplishing their policy goals. This theory assumes that members engage in a cost-benefit analysis when they considering running again for Congress. May- hew (1974a)’s framing of reelection as an instrumental goal should mean that members with decreased predicted policy benefits upon reelection, should place a lower value on the pursuit of that goal. The opportunity costs of maintaining a congressional career are too great for simple job security to be the primary objective – members of Congress are relatively underpaid compared to jobs that require similar levels of expertise (Hibbing, 1982). Empirically this theory is examined here primarily through the policy angle - looking at the passage or lack thereof of members bills - but it also extends into the hierarchical 10 realm of influence within the institution. It is hard to pierce the veil of members being denied a committee chairmanship or other prestigious positions when the modern rules ultimately leave these choices to the party with no guarantees that seniority or other easily deduced traits would secure those roles (U.S. Senate: About the Committee System | Committee Assignments 2023, 2023). However, we do have some insight into when members could be expected to lose these positions, on account of the House Republican Caucus rules placing term limits on their ranking-members/chairs (Reynolds, 2017). There are numerous examples of committee leaders in the Republican party leaving before the term they would be expected to limited out of their position such as Dave Camp (R-MI) (O’Keefe and Kane, 2014) and Mike Rogers (R-MI) (Bresnahan and Sherman, 2014) in 2014 who were the chairs of Ways and Means and Intelligence respectively, and Bob Goodlatte (R-VA), Jeb Hensarling (R-TX), and Lamar Smith (R-TX) in 2017 who were the chairs for Judiciary, Financial Services, and Science Committees (Reynolds, 2017). Members in this position are at the zenith of their influence over the policy-making process, but the recognition that this power will soon expire and they will rejoin the rank-and-file in relative ineffectiveness makes remaining in Congress far less attractive. The willful departure of the majority of recent Republican House leaders is also reflective of members taking future prospects of diminished power, as opposed to their current state, in considering the choice to leave. The Speaker of the House has long been understood as the most influential member of Congress and one of the most powerful individuals in the U.S. Government as a whole (Follett, 1896; Green, 2010; Taylor, 1908). Nonetheless, the Speaker’s role as a focal point of Congress makes it a vulnerable position with many speakers opting to leave when their party or they are otherwise likely to lose their position (Opsahl, 2018). Since the Speakership of John Boehner (R-OH), who served in that capacity from the start of the 112th Congress through October 2015, Re- publican Speakers have struggled to hold on to for considerable tenure due in large part to internal fractions within their party that threatened to forcefully end their speakership. John Boehner and Paul Ryan (R-WI) resigned and retired before dissident elements within their party could oust them or before their party lost an upcoming election, while most recently Kevin McCarthy (R-CA), who lacked the foresight of his predecessors, has re- signed following the successful vote to remove him from his position (Blood and Freking, 2023). All 11 of these examples are consistent with our main theory, as the fear of becoming politically marginalized within the institution, or the realization that they have become so in the case of McCarthy, serves a clear motivation to leave office. The revolving door phenomenon, wherein members easily find work as lobbyists following their departure, willing or otherwise, from the institution provides members a financially lucrative alternative to congressional service that undoubtedly plays a role in decisions to stay or leave. For those choosing to remain in Congress, they must rationalize that the utility of their seat outweighs the financial perks outside the institution which would include their influence over policy and perhaps long-term political leverage to seek a political "promotion" in the future. Members are acutely aware of these prospects irrespective of where they fall into the "stay" or "leave" category at the end of a term. During John McCain’s (R-AZ) funeral eulogy, Joe Lieberman (I-CT) recalled a conversation between the two after Lieberman had announce his retirement where McCain purportedly told him "I’ve been thinking if you go out into the private sector, you’re going to make some more money" and that he should use his new salary to buy them a home in Jerusalem (Hubbard, 2018). I examine evidence for the following hypotheses to determine their respective relationships with voluntary Congressional retirement. Hypothesis 1: Legislative Ineffectiveness. Members of the House and Senate who are ineffective legislators are more likely to voluntarily leave Congress. Fenno (1978) describes one of the major goals for members of Congress in both chambers as creating good public policy, with reelection being the most proximate goal to help achieve those ends. If this indeed is the ambition of most members, then a member’s level of success, or lack thereof, in getting bills through the legislative process should be part of their calculus when contemplating the future of their congressional career. Although there are important hierarchical differences between the House and Senate I anticipate that members of both chambers will be influenced to retire based on their success or lack thereof in the legislative arena. It is no secret Congress as a whole has been lamented in recent years, by both elites and the public at large, for its perceived inability to accomplish much in the legislative arena. Since the 107th Congress (2001-2003), the number of enacted laws has been 12 considerably lower than in pre-2001 sessions (Statistics and Historical Comparison, 2020). Members are acutely aware of the reputation their institution has earned but they undoubtedly have some optimism that at least some of the items on their policy agenda are accomplished, or at the very least minimize losses to the opposing party. Incoming members likely possess some degree of patience when it comes to their legislative objectives as freshmen – especially in the Senate – were expected to be fairly passive players according to the norms of the institution (Matthews, 1959). Although the draconian norms that constrained fresh- man in the past have eased, lower effectiveness is still observed among the junior ranks of each chamber on account of their lack of expertise and networks within the institution (Volden and Wiseman, 2009, 2018). While I expect that less effective congress persons are more likely to retire, it is generally expected that these retirements will not emerge among first term office holders, with perhaps occasional exceptions who may have held naive expectations of the institution. Because of the lack of predicted overlap between these two groups of members with low effectiveness, the predicted negative relationship between effectiveness and retirement is likely driven at least in part by the most effective members choosing to remain in office. Hypothesis 2: Senate Electoral Structure. Senators are more likely to retire when they are up for reelection. This prediction is rooted in a few assumptions. First, the decision of whether to retire should weigh more heavily on senators as they approach the end of their term as winning reelection adds another six years to a member’s career. This may encourage members who are older or simply wish to explore other options outside of Congress to choose to leave earlier than they may otherwise have in the House, where reelection would only occupy the next two years, a much briefer period to be stuck with a position that may be losing its appeal. While the option to resign in the middle of one’s term is a possibility, doing so outside of extraordinary circumstances (serious health issues and scandal for example) is going to be generally discouraged by the impact losing a member has for that member’s party, and by extension that members long-term policy objectives. This should be especially true when there is the possibility that the now vacant seat may fall into the hands of the opposing party via gubernatorial appointment. An early resignation may also adversely impact their state, reducing its representation in the Senate which could have consequences 13 for federal funding directed towards the state. 2.3 Operationalization and Investigation Before delving into the various variables at play in motivating congressional retirement, it is important to identify what qualifies as a retirement. I classify a retirement as any form of voluntary exit, including a member’s decision not to run for reelection or their resignation before the end of their term. This classification strategy is in line with existing literature (Masthay and Overby, 2017). I additionally consider a separate classification of retirement in the statistical analysis that factors whether political ambition played a role in that members retirement via the pursuit of a different political office. This is done to evaluate whether there are any meaningful differences when that subset is excluded as politically ambitious members, depending on the office they are pursuing, may be less jaded by the political process than someone who completely calling it quits. There is considerable variety in members who choose to resign from office before the end of their term. Such members may choose to resign because they are in the midst of scandal which is embarrassing and potentially threatening to their reelection efforts, because they are experiencing declining health, or because they have been appointed to a new position. Former President Barack Obama (D-IL) and former Vice-President Al Gore (D-TN), who vacated their Senate seats when they were elected to their higher offices, are also classified as retirements for the sake of consistency. Previous congressional rules granted greater seniority status to their successors if they entered office slightly earlier than the rest of their freshmen cohort by being appoint to fill the seat of their resigning predecessor, providing another potential incentive to resign, such as Wallace Bennett (R-UT), Alan Bible (D-NV), and Phil Gramm (R-TX), though the latter was seemingly unaware that the rule had been changed prior to his early exit. Even electoral losers concerned with their constituents’ interests have resigned to give their former opponent a head start in office, though these individuals are not treated as retirees in this data. Figure 2.1 presents the total number of all exits, by Congress, from the U.S. Senate over the observed period of 1973 through 2023, divided between retirements, defeats, and deaths. While the number of total departures per Congress varies drastically over this time, there is generally a clear pattern of voluntary retirements outnumbering exit via electoral 14 defeat and death, particularly from the 98th Congress (1983-1985) on. The most recently observed 117th Congress witnessed turnover exclusively through retirements. Despite the concerns some may have about members seeking to hold onto their seats seemingly forever, very few members die in Congress.2 191 Senate departures out of the 323 total (approximately 59%) are attributed to some type of retirement rather than being electorally forced out of the chamber. The only Congresses where more members were voted out than retired are the 95th (1977-1979), 96th (1979-1981), 106th (1999-2001), and 109th (2005- 2007) Congresses. There are no factors that would suggest an obvious pattern in these elections that stand out (96th and 106th Congresses are cycles in which new presidents were elected into office while the others are midterms), without delving into the particulars of those election cycles. Figure 2.1: Senate Departures 93rd-117th Congresses Curiously, the heavily examined 102nd (1991-1993) Congress features a relatively un- remarkable number of retirements in the Senate (8) – suggestive of the fact that, at the macro level, the Senate has retirement patterns differ from those of the House. This is 2 Members who die in office are considered separate from retirements and are not counted as such in the empirical models 15 consistent with general expectations of the Senate as it was designed to be the more stable chamber with more electoral isolation. When considering the factors that Ahuja et al. (1994); Clarke et al. (1999); Shea (1999) suggested were responsible for the upswing in House retirements in this year, it is likely that most Senators would not have benefited from the ability to liquidate their war chest as they were not running a campaign and similarly would avoid the electoral consequences of being involved with the checking scandal. Some of the peaks in Senate retirement, such as the one in the 111th (2009-2011) Congress are in part motivated by presidential cabinet picks leaving the Senate for their new offices (i.e. Hillary Clinton (D-NY) resigning to assume the position of Secretary of State) and governor appointees to fulfill the vacancies left by the new presidential administration that ultimately opted to not seek reelection when their term expired (i.e. Roland Burris (D-IL) who was appointed to fulfill the rest of Barack Obama’s term after he was elected president). Looking at the House from 1973-2023 in Figure 2.2, a slightly different pattern emerges. Predictably, there are far more exits from the House in general compared to the Senate, which is to be expected from a body with over four times the membership and where the entire body faces reelection each cycle as opposed to only a third. In contrast from the Senate, there is usually not as severe of a disparity between voluntary and involuntary exits. While retirements generally outnumber electoral defeats and member deaths, there are noteworthy exceptions in the 93rd, 96th, 103rd, 109th, and 111th congresses. Most of these sessions saw the results of notable wave elections including the 1974, 1994, 2008, and 2010 election cycles, where there were considerable swings in the partisan make up of the chamber. The 102nd Congress continues to remains noteworthy with it’s record number (59) of retirements, consistent with the level of interest it has received from House dedicated studies of voluntary exit. The asymmetry between the House and Senate in the prevalence of retirements and defeats (retirements make up approximately 55% of departures in the House) reflects the different electoral circumstances that exist in the two legislative bodies, with "wave" elections having greater potential to impact membership of the former. A potential source of contribution for why the House sees proportionally more defeats is a potential consequence of House seats receiving less quality challengers (Nice, 1984), and by extension, leaving many representatives with an inflated sense of security that may encourage them to run for office 16 even when the political tides do not favor them or their party. Figure 2.2: House Departures 93rd-117th Congresses To study the determinants of congressional retirement I expand upon Volden and Wiseman (2014)’s House and Senate Effectiveness data set providing additional variables of interest pertaining to retirement. This data set ranges from the 93rd to 117th Congresses (1973-2023) and includes all members that served in both chamber of Congress during the observed period along with several descriptive statistics of the members including their race, vote share in the prior election, gender, and their seniority. Most importantly, this dataset includes legislative effectiveness scores (LES) for all members who served in both chambers over this period of time. The LES measure captures the positive legislative activities of members of Congress, particularly their bill sponsorships. The Senate portion also adds variables for whether the member is up for reelection and, importantly for this research, if they retired at the end of the Congress. The retirement variable has been reviewed with corrections made to accurately reflect retired membership along with creating additional classifiers for members who chose to resign before their term came to an end, those who died in office, and those who lost reelection bids. Information on member 17 exit for the House has also been collected for this paper for the 93rd through 117th Congresses. Retirement and its derived subsets are dichotomous variables. Elaborating on the key LES measure further, it is important to consider how it is con- structed and what this means for legislator scores over time. Per Volden and Wiseman (2014), scores are normalized across legislators within each two-year congressional session to have an average of "1." As this average is maintained regardless of the productivity of particular legislative sessions at the Congress level, legislators across different sessions cannot be appropriately compared on a score-to-score basis. Within legislators, effectiveness can vary considerably from one Congress to the next, especially when that member transitions from their party to being in the majority to being in the minority, or vise versa, or they are elevated to more powerful positions such as committee chairs. As a robustness check for the role effectiveness plays more broadly, I also utilize Volden and Wiseman (2014)’s LES Benchmark scores, primarily through whether a member met, exceeded, or fell below them, and the independent components of total bills sponsored and bills that became law. The LES Benchmark scores predict a member’s effectiveness score based on their seniority, majority party status, and if they hold any committee chairs. Factoring expected effectiveness into the model and, more importantly, where members fall in regards to those expectations lends further credence to my theory in regards to members expectations for the institution not being met and opting to leave as a result. Separating two of the key LES components further clarifies the interest in participating in the policy-making process and if the tangible success members enjoy influences career choices. I control for several factors associated with higher LES. These factors include majority party status, whether a member is a committee chair, if they are on a "power" committee,3 seniority, and if the member is in their party’s leadership structure. Party leaders rarely propose bills themselves so tend to have low LES but are unquestionably tremendously influential in the actual legislative process. Because previous studies using LES have found differences between male and female members of the institution, with some evidence indicating that women are more effective in certain contexts (Volden, Wiseman and Wittmer, 2013), controls are included for gender. Similarly, racial identity, reduced to a 3 One of the top four highest ranked committees according to Groseclose-Stewart Scores. 18 simple white/non-white dichotomy due to the under-representation of African, Latino, Asian, and Native Americans in the Senate as a whole, is controlled for due to its potential consequence on other observed factors. Lastly, to mediate the issues of freshmen senators sharing the expectation of low effectiveness with retirees, this status is also controlled for in the logit model. I include two related, yet distinct ideological metrics. The first, a member’s ideological distance from their chamber median is a directionless measure, as an extreme conservative and extreme liberal would have similar positions, that is utilized due to the important role pivotal members play in passing legislation, wherein the median members of each chamber provide crucial votes to determine the passage or defeat of a bill (Krehbiel, 1998). Existing on either end of the ideological extreme is thus expected to have some bearing on member success and in turn have implications for a member’s likelihood to retire. I also incorporate the first dimension of NOMINATE scores obtained from Boche et al. (2018) to extrapolate if there is an ideological direction that would in- fluence retirement decisions, as conservatism has been theorized as a component of the greater levels of retirements among Republicans in the past (Murakami, 2009). My hypotheses on the causes of congressional retirements are tested with multivariate logistical regression models in each chamber. To account for the panel-like nature much of the dataset has due to the repetition of many members across several Congresses, I utilize clustered standard errors for each member and include congressional year fixed- effects to address the statistical issues that arise from modelling on so many closely related observations. 2.4 Retirement in the Senate Table 2.1 and Figure 2.3 presents the results of the main model for retirement from the United States Senate. Model 1 shows the results for all retirees, while Model 2 removes those who retired with political ambition from the subset. There is support for both hypotheses that members with lower legislative effectiveness are more likely to retire and Senators are more likely to retiree when they are up for reelection as opposed to the middle of their term. That LES is significant even in the Senate where legislators have access to far more negative power to shape the process, further speaks to the value of legislative 19 productivity to retirement calculations.4 For many senators the capacity to obstruct the other side is simply an inadequate substitute for implementing desirable policy. While there is a range of retirement probabilities at each level of effectiveness, there is a noted decrease in a member’s likelihood to willingly leave office the more effective they are. This holds true even when factoring out politically ambitious types as in the Senate this is frequently appointees who are forced to resign early-to-mid term and naturally have a lower level of effectiveness as a consequence. Examining the effects of electoral status on retirement, there is a strong positive relationship between retirement and senators being up for reelection. The role of the electoral cycle in shaping the retirement decisions of senators – even those who choose to resign before their time in office naturally expires – suggests that there are strong norms that encourage even those contemplating not serving a full six years to closely align their retirement with the end of the term. This should come to relatively little surprise considering individuals such as Dianne Feinstein (D-CA) have faced incredible public pressure to resign due to failing health and it’s debilitating impact on her service in the Senate, but there was no movement in that direction before her death. Seniority is also positively associated with retirement, as consistent with existing the- ories. As a variable highly correlated with age, it is natural that seniority would motivate a member to consider retirement eventually, as holds true in most professions, and may also allude to diminishing returns derived from continued congressional service. Sena- tors are also influence by their prior electoral performance in considering their retirement calculations. This is interesting as the institution was designed to be more insulated from electoral pressures with its six year terms, though clearly this is not quite lengthy enough to completely diminish electoral-vulnerability concerns. It is noteworthy that freshman status does not have a statistically significant negative relationship with retirement. This suggests that there is a non-trivial population of senators who choose to leave at the end of only a single term. As even freshmen senators tend to be older, and terms last for 6 years, time related concerns may potentially influence many members to not seek reelection after the first term. Connecting this with LES, it is also possible that retiring 4 These results are consistent with the models that evaluate actual LES versus predicted Benchmark LES featured in Appendix Table 1 20 freshmen had more optimistic expectations for their legislative productivity when they entered office – naivete that was shattered at some point within their first six years. Another interesting null is that majority party status seemingly plays no role in retirement calculations in the Senate. The Senate has lacked a lengthy period of uni-party dominance such as Democrats enjoyed for over forty years in the House (Ang and Overby, 2008; Gilmour and Rothstein, 1993; Murakami, 2009), and is generally seen as more egalitarian in how legislators are allowed to participate in the process, even as a member of the minority party. It may also be the case, that legislative (in)effectiveness is the underlying feature that makes being in the minority party so unattractive. The gender difference in the results, women being less likely to retire than men, is noteworthy in that it contrasts with prior findings in the House (Lawless and Theriault, 2005). However, the most likely explanation for this occurrence is the nature of the data, and more pointedly, the relatively low number of women who have served in the chamber over this period and the even smaller number who have retired. The null results in Model 2 which ignore those who have retired for the sake of pursuing an alternative political office reinforces this. Figure 2.3: Senate Retirement Coefficient Plot Model 1 Breaking down LES into two of it’s major components, bills proposed and bills that 21 were enacted into law, in Table 2.2 I find that the stronger effect emerges from members sticking around if they have sponsored several bills, whereas the passage of those bills has null results. While this may be indicative of some final term shirking in terms of actually getting a bill past the finish line, this is inconsistent with the types of shirking that have Table 2.1: Retirements in the Senate Over Time, 93rd-117th Congresses previously been identified by scholarship which has found evidence in votes cast and ideological scores (Rothenberg and Sanders, 2000; Tien, 2001). The rules of the Senate do provide ample opportunities for legislators to accomplish their policy objectives through means other than the passage of their specifically sponsored piece of legislation, such as 22 offering non-germane amendments to other bills. This makes the null results on bills passed less surprising, especially as unorthodox lawmaking becomes increasingly the norm for passing legislation (Gluck, 2021). Table 2.2: Alternative Senate Retirement Models with Effectiveness Components 23 2.5 Retirement in the House Examining the House in Table 2.3 and Figure 2.4 I also find support for my hypothesis regarding legislator effectiveness.5 The commonality of legislative productivity in influencing retirement decisions in the House and Senate is a noteworthy similarity that has previously evaded comparisons between the two chambers. Some divergence between the chambers is present in this regard as the successful passage of legislation is a con- sideration for House members weighing retirement, though not so when factoring out politically ambitious types as seen in Table 2.4. It’s harder, though far from impossible, to get policy provisions incorporated into other bills in the House compared to the Senate. That bill passage has null results when factoring out ambitious retirements suggests this result is carried more by members who are genuinely interested in implementing policy and are looking for better opportunities to do so in other offices. Hierarchical features such as being on a power committee play a far more significant role in the House than in the Senate. Curiously majority-party membership similarly has null results just as in the Senate. As discussed before, the Senate has more features on the positive and negative power front that make being a minority party member less daunting than in the House. However, in addition to the possibility that effectiveness is the primary characteristic of majority party status that interests legislators, representatives recognize that control of the chamber is fleeting and the loathsome conditions of being the minority are temporary (Lee, 2016). The ideological measures present interesting results, with a negative correlation with general ideological extremity, as measured by distance from the chamber median, and the positive correlation with NOMINATE scores. The positive effect of NOMINATE, meaning more conservative members are more likely to retire, is consistent with earlier findings that Republicans are more prone to voluntary exit than their Democratic colleagues, particularly Murakami (2009) who attributed this greater rate to members’ conservative ideology. The effect of ideology on House retirement decisions is a notable asymmetry between member exit behavior in the two chambers, and aligns with existing findings (Ang and Overby, 2008). 5 This is also backed up with the comparison with Benchmark LES in the Appendix, see Table 2 24 Table 2.3: Retirement from the U.S. House of Representatives in the 93rd -117th Congresses 25 Table 2.4: Alternative House Retirement Models with Effectiveness Components 26 Figure 2.4: House Retirement Coefficient Plot Model 1 2.6 Conclusion The empirical findings here provide support for both of my hypotheses. Members in both chambers are found to be more likely to retire when they are less effective legislators. The commonality in findings for the LES measure in both chambers suggests that member motivations are not too dissimilar between the two, and fulfilling legislative objectives is important for members to justify continuing their congressional career. Our expectation that less effective legislators are more prone to retire – a key component of this paper’s main theory – is supported and passes several robustness checks when examining multiple measures of effectiveness. In the Senate, being up for reelection at the end of a given term is positively correlated with a senator’s decision to leave the chamber as is predicted by hypothesis 2. This finding is normatively desirable, as it suggests that once elected, senators will generally honor their commitment to a full six-year term. Seniority, as unveiled in previous literature, has a strong positive relationship with congressional retirement in both chambers. Other findings produced by this empirical analysis that were not previously predicted are worthy of further examination, particularly the negative relationship ideological extremity in the House and gender in the Senate have on retirement in those 27 chambers. The correlation between ineffectiveness and retirement has many potential consequences on Congress as an institution and for representation. As retirement is the premier form of membership turnover, less effective members opting to leave the institution filters out bad legislators which may benefit their former chamber and their constituents at home who have an opportunity to elect a potentially more productive member. This trend presents related benefits in that the most effective legislators are more likely to stick around which means their clout may be used to further aid their state/district. The normative benefits of the self-filtering out of poor legislators are clear, but it is less clear if the findings in this paper will persist into the future. Congress as a whole continues to become less productive, and there is an apparent class of congresspersons who see this phenomenon as a desirable feature rather than a bug. This is best exemplified by the backing of failed Speaker candidate Jim Jordan (R-OH) whose 16 years in Congress have netted a grand total of 0 pieces of proposed legislation becoming law and a mere 36 bills even proposed, an applause line for some of his supporters when referenced on the floor (Blake, 2023). My findings suggest the majority of members of Congress are still at least nominally interested in using their time in office to produce legislation, but it is worth monitoring whether this remains the case in the future. 28 CHAPTER 3: TARGETING THE TWEETS: EXAMINING PRESIDENT TRUMP’S CONGRESSIONAL TARGETS ON TWITTER In a negative ad released by the Biden campaign in May 2024 entitled "Snapped," a dark recollection of the Trump presidency is painted, focusing in on some of the lowest points of Donald Trump’s time in office (Gangitano, 2024). That this ad would open it’s run- down of the former president by zeroing-in on "the midnight Tweets" speaks to the defining, and at times despised, role Trump’s use of the social media platform then known as Twitter played in his presidency. The implicit suggestion of Robert De Niro’s narration in this ad is that Trump’s tweets were the ravings of a mad man and the window into a self-evidently depraved mind, a sentiment that likely resonates with many in the public (Bowler, Carreras and Merolla, 2023; Nelson, 2017; Thomsen, 2017). Indeed, journalistic (Karl, 2021) and academic (Altheide, 2022; Monahan and Maratea, 2021) observations of the Trump administration frame many facets of his time in office, especially his tweets, as akin to a sequel of his reality show, The Apprentice. While there are undisputed examples of Donald Trump using his platform in precisely that manner, this framing undersells that Trump also used it for political communication and, although his combination of medium and rhetoric may have been less than conventional, the overall pattern of engagement is reminiscent of many of his predecessors in the Oval Office. For the President of the United States, the ability to communicate with and persuade members of Congress is essential to the pursuit of their legislative agenda (Bond and Fleisher, 1990; Fleisher and Bond, 1996; Spitzer, 2012). Various presidents have favored different styles to this approach, but modern presidents have consistently employed some level of public engagement with members in their interactions with Congress as a means of leveraging constituent support (Kernell, 2006). President Donald Trump’s constant online presence through social media, particularly on X/Twitter,1 offered unique insights into the day-to-day musings of the president in a manner never previously on public display at such 1 Since the platform was branded as Twitter for the entirety of the Trump administration, Twitter will be used to refer to the platform throughout this piece. 29 a scale. With over 26,000 messages posted during his time in office, all aspects of the Trump presidency were colored in some way by his Twitter use, and his relationship with Congress and its members was no exception. Although he was not the first president to take advantage of the platform, Trump’s utilization of Twitter was a near constant source of intrigue both as a candidate for the White House and as the chief executive. Several tweets, trivial2 and consequential alike, came to dominate the news cycle at various points during Trump’s political career (Morales, Schultz and Landreville, 2021). Twitter use was a frequent point of contention among campaign and White House staffers. Trump’s former social media manager, Justin Mc- Conney, recalled the revelation of Trump using his own Twitter account as "comparable to the moment in ‘Jurassic Park’ when Dr. Grant realized that velociraptors could open doors" - being a major "oh no" moment (Schreckinger, 2018). During the first Trump campaign, staffers worked to plant favorable stories in right wing media outlets in order to sooth the president and dissuade him from tweeting against various media outlets and people (Palmeri, 2017). Ultimately, the staffers attempting to curb this access lost that battle, and White House staff largely gave Trump free reign even if it meant covering for a tweet over the course of a given media cycle (Karni, 2017). The significance of Trump’s Twitter account in defining his presidency is not merely a media creation. During his presidency, the official position of the White House was that the tweets from Trump’s personal account were "official statements" (Landers, 2017), a position which was further cemented in court with Knight First Amendment Inst. at Columbia Univ. v. Trump, No. 1:17-cv-5205 (S.D.N.Y.), No. 18-1691 (2d Cir.), No. 20-197 (Supreme Court) (2019) which ruled that the account was a "public forum" under the First Amendment. Consequently, Trump’s tweets are covered under the Presidential Records Act of 1978 (Treisman, 2019). While the prolific nature of Trump’s Twitter account is undeniable, the actual con- tent of his tweets remain divisive. Trump’s candidacy was frequently cast in controversy due to his proclivity to launch personal attacks on his rivals and various other political 2 Trump, Donald J. "Twitter/@RealDonaldTrump: Despite the constant negative press covefe" May 30, 2017, 9:06 PM. https://twitter.com/realDonaldTrump/status/8697669948 30 figures – attacks that were either echoed or had sometimes originated on his Twitter ac- count (Gross, 2018; Kelly, 2019; Oprysko, 2019). These attacks could stem from policy differences or personal attributes that had earned Trump’s disdain, and members of both parties could easily find themselves on the receiving end of a digital tirade. Contrary to the expectations of countless political pundits, this behavior did not subside once Trump reached the Oval Office. Surveys expressed that the content of Trump’s tweets in particular were a major factor in shaping their view that he was "unpresidential" (Nelson, 2017; Thomsen, 2017), a view Trump would sardonically appropriate when he tweeted "My use of social media is not Presidential – it’s MODERN DAY PRESIDENTIAL."3 Though his heightened political station may have failed to adjust Trump’s character, it did grant increased weight to his political attacks, especially when directed towards members of Congress. His precise public-facing demeanor may have been a departure from prior presidents, but Trump’s flamboyant approach to social media interactions with his Washington fellows blends some of the private conflicts of his predecessors with exercise of the bully pulpit. As extraordinary as Trump’s often crass embrace of social media is often framed, it is not entirely dissimilar from how other presidents engaged with inter-branch relations – possibly marking more of an (d)evolution rather than a full deviation. Twitter’s role as a "public forum" nests its use within the broader context of the "going public" strategy (Kernell, 2006). Several prior presidents were also known for having a similarly "larger than life" persona such as Theodore Roosevelt who first coined the term "bully pulpit" in reference to the public-facing persuasive power enhanced by the station of the president (Roosevelt and Taft, 1892). Theodore Roosevelt’s exercise of this power was constrained by the instruments of his time, taking advantage of public speeches and what passed for mass media in his era, while his successors were able to enhance their reach through the technological innovations of their eras. Franklin D. Roosevelt enjoyed more direct access to the public through improvements in radio technology, most notably through his fire side chats (Kiewe, 2007) and John F. Kennedy embraced further technological evolution with the 3 Trump, Donald J. "Twitter/@RealDonaldTrump: My use of social media is not Presidential - it’s MODERN DAY PRESIDENTIAL. Make America Great Again!"July 1, 2017, 6:41 PM. https://x.com/realDonaldTrump/status/881281755017355264 31 use of televised press conferences (Hoover, 1988). Trump’s immediate predecessor, Barack Obama, was on the cutting edge in pioneering political usage of many of the same tools and platforms that came to define a significant portion of the Trump presidency (Aharony, 2012). Tonally, Trump in many ways echoes the interpersonal approach of Lyndon Baines Johnson, whose infamous "Johnson treatment" brought his crass and threatening demeanor into the fray to domineer presidential-congressional relations (Beckmann, Chaturvedi and Garcia, 2017; Zelizer, 2015). 3.1 Tweets in Political Context Trump is far from being the first or sole political actor on Twitter. Many members of Congress utilize the platform for various political communiqué. Republicans and younger members were the most common early adopters of Twitter (Lassen and Brown, 2011; Peterson, 2012), with some already setting the stage for the more polarizing rhetoric that became the norm under Donald Trump, particularly when they were comparatively free from electoral pressure (Ballard et al., 2022). Not all Twitter use by political officials facilitates harsh rhetoric. From the onset, members of Congress have used Twitter for self- promotion and campaigning (Golbeck, Grimes and Rogers, 2010; Hemphill, Otterbacher and Shapiro, 2013; LaMarre and Suzuki-Lambrecht, 2013). This has developed to encompass more policy discussion over time (Hemphill, Russell and Schöpke-Gonzalez, 2021), with usage being predictive of voting patterns (Hemphill, Culotta and Heston, 2013), and has played a role in agenda-setting with the media in some issue areas (Shapiro and Hemphill, 2017). Donald Trump’s active and near constant use of Twitter during his candidacy and presidency caught the eyes of journalists and politicians alike (Wells et al., 2020), and his oftentimes inflammatory rhetoric has been cited by Republicans in Congress as a sign that they have been left behind by the party (Krause and Byers, 2022). Prior presidents have enjoyed the capacity to shift the media narrative through press conferences and the like (Miles, 2014). And, Trump’s extension of this into the low-cost tool of social media posts created an impact that extended well beyond the individual tweet itself and ensured even the most social media illiterate were aware of them (Lazarus and Thornton, 2021). With the Trump presidency taking place within an era of increased partisan fervor, the timing was 32 perfect for his hardened brand of politics (Tucker et al., 2019). Communication can be utilized in numerous ways by the president against members of Congress. Long before the era of social media, presidents such as Lyndon Baines John- son would directly lobby and coerce members – typically those more moderate and less ideologically committed to his projects – into supporting them (Beckmann, Chaturvedi and Garcia, 2017). Beckmann, Chaturvedi and Garcia (2017)’s exploration of President Johnson’s phone records uncovered which members he engaged with the most and found that congressional leaders and chairs occupied most of his attention. Trump’s Twitter rants about members may be somewhat akin to the harshness of the fabled "Johnson Treatment" exercised by LBJ against members of Congress in tone (Zelizer, 2015), if not necessarily the same in purpose. The treatment has been described as "an incredible, potent mixture of persuasion, badgering, flattery, threats, reminders of past favors and future advantages" (Coleman, 2018) and has been seen as a hallmark of presidential- congressional coalition building. Johnson’s approach worked with the congressional tools and members available, and it is worth examining the extent to which Trump’s tweeting behavior resembled LBJ’s interactions. Trump’s predecessor, Barack Obama, trailblazed the media tools that played a major role in defining the Trump-era. Through innovation and some level of growing necessity in an evolving media landscape brought about by technological advancement and the economic recession, Obama used the internet to bypass the press and directly engage with voters (Pfeiffer, 2018). Towards the end of his time in office, Obama’s communication team launched the @POTUS account which served as an effective and direct tool for Obama to engage with the evolving political conversation. The effort to create this account came after overcoming reservations regarding the security of installing the app on his BlackBerry and general suspicions over the validity of the platform (Pfeiffer, 2018). While dwarfed by the catalog of Trump Tweets, Obama did engage with the platform more frequently than other contemporary heads of state (Aharony, 2012). Prior scholarship on Trump’s Twitter use has focused on a variety of political and eco- nomic consequences. It has been argued that his social media presence has normalized right-wing populist discourse (Kreis, 2017; Varis, 2020) and that it has spread misinformation and conspiracies, with noted impacts on vaccine hesitancy during the 33 COVID-19 pandemic (Hornsey et al., 2020; Ross and Rivers, 2018). This is a departure both in con- tent and tone from his immediate predecessor, Barack Obama (Ott, 2017; Wignell et al., 2021). Others have examined the consequences the Tweets have had on various financial markets in their aftermath (Filippou et al., 2020; Gjerstad et al., 2021; Huynh, 2021; Klaus and Koser, 2021). Trump’s Twitter activity has also been identified as a mechanism to influence political change through less-formal channels. The rollback of Obama-era environmental regulations and reorientation of immigration policy are two such examples of this (Sahin, Johnson and Korkut, 2021). Ouyang et al. (2020) broadly examined Trump’s Twitter posts from the creation of his account in 2009 through November 27, 2019 and identified many of the core topics Trump discussed through the account while in office. These ranged from offers of support and praises for friends and political allies, commentary on policies and other broad issues, as well as promotion of himself and his properties, merchandise, tweets, books, and other accomplishments. Additionally, there were the verbal attacks he levied against other public figures and groups (Ouyang et al., 2020; Kelly, 2019; Oprysko, 2019). 3.2 Expectations People paid attention to the president’s Twitter account, especially once he was in office, and this granted considerable leverage and influence in many areas (Pérez-Curiel and Limón-Naharro, 2019). This influence was exercised by the former president towards non- incidental political ends. Focusing on Trump’s comments on and towards individuals, specifically members of Congress, this paper expands on the literature by cataloging and categorizing who Trump talked about, when, and how. Not all of these messages were attacks – President Trump had allies (some more temporary than others) who received boosting from the president as well – and the level and source of insult could vary tremendously. The focus on Tweets aimed at members allows for a deeper dive into these disparities and more closely observes his attitude towards individual members over the course of his tenure. To evaluate these trends, I examine the following expectations for President Trump’s Twitter communications towards members of Congress: Hypothesis 1: Targeting Leadership. Trump is more likely to tweet about Congressional leaders than rank-and-file. 34 Working with congressional leaders is essential to pushing and setting a president’s agenda in Congress (Beckmann, 2010). Donald Trump’s experience as a political out- sider may have made him more dependent on these figures than the average president in terms of navigating the particulars of the legislative process, but this strategy would also be consistent with the approach of a seasoned political veteran. Beckmann, Chaturvedi and Garcia (2017)’s analysis of President Johnson’s phone records show that leaders were his favorite lobbying targets, and the more public-facing nature of the Trump presidency should make it likely this tradition extended to the more public-facing tweets. Additionally, when both the era of hyper-polarized parties with and the role of negative partisan- ship in the electorate are taken into consideration, out-party leaders are natural avatars to attack for electoral and political gain (Phipps and Montgomery, 2022). Hypothesis 2: Partisan/Ideological Hostility. Trump is more likely to negatively tweet about Democrats/Liberals than Republicans/Conservatives. The president has assumed a role at the head of their political party (Klinghard, 2005; Seligman, 1956), and this role has become increasingly partisan in their approach to politics (Skinner, 2008). These biases associated with the modern presidency, naturally suggest that the highly public tweets from President Trump, should express similar biases. A partisan/ideological disparity in the tone Trump uses to talk about political opponents would also follow established trends in polarization and the former president’s role in exacerbating that process (Abramowitz and McCoy, 2019). Hypothesis 3: Political Timing. Trump is more likely to mention members of Congress during heightened political contexts. Assuming President Trump’s tweets are politically patterned, it would stand to rea- son that the timing of their usage would vary depending on the political contexts. Prior presidents have used various types of media appearances to help reinforce their agenda (Baum, 2013). In the congressional context, one of the defining events during Trump’s time in office was his first impeachment. As this was an event entirely initiated by Congress and Trump’s influence over the process would be entirely indirect through his allies, I anticipate he would have engaged with members of Congress at an increased rate during that process. Also, presidents are often active participants in Congressional races to help bolster key allies and attempt to harm political foes (Herrnson, Morris and Mc- Tague, 2011), so it should 35 similarly be expected that Trump would amp up his Twitter engagement with members during election years. 3.3 Data The analyses of this paper use data collected from all tweets from the former president’s Twitter account, @RealDonaldTrump as archived by Brown (N.d.), from November 8, 2016 through January 8, 2021, capturing every message posted by Donald Trump as president-elect and President of the United States of America. This choice in time frame reflects the intentions of this paper’s focus on presidential-congressional relations, with that early president-elect period being included due to the equivalent influence Trump held rhetorically at that point as he would hold following inauguration. These data include a total of 26621 tweets, including retweets and deleted posts during the observed time period. The tweets were then hand-coded for whether they mentioned members of Congress, groups of members, and the level of endorsement/threat expressed towards them. Important nuances had to be considered with the identification of different members within Trump’s tweets as he was not always as explicit in referring to particular members. Trump was infamous for employing the use of various degrading nicknames for the sub- jects of his Tweets. Many of these were far from enigmatic such as Head Clown/Cryin’ Chuck Schumer (D-NY), Crazy Bernie Sanders (I-VT), Sleazy/Shifty Adam Schiff/Schitt (D-CA), Jeff Flake(y) (R-AZ), Al Frankenstein (D-MN), or Liddle’ Bob Corker (R-SC). Others require more awareness of Trump’s vernacular or rely on an introductory post for that alias’s debut to properly ascertain the referenced individual, such as "Pocahontas" for Senator Elizabeth Warren (D-MA), "Da Nang Dick" for Richard Blumenthal (D-NY), and "King of Baltimore" for Elijah Cummings (D-MD). Other times Trump would allude to a member through their circumstances and state such as his remarks towards Representative Justin Amash (I-MI) after leaving the Republican Party where he celebrated the news while declaring him "one of the dumbest and most disloyal men in Congress."4 These cases are taken into account 4 Trump, Donald J. "Twitter/@RealDonaldTrump: Great news for the Republican Party as one of the dumbest amp, most disloyal men in Congress is “quitting” the Party. No Collusion, No Obstruction! Knew he couldn’t get the nomination to run again in the Great State of Michigan. Already being challenged for his seat. A total loser!" July 4, 2019, 9:05 AM. https://x.com/realDonaldTrump/status/1146766931573301248 36 with the construction of a thesaurus of notable nicknames along with searches of other keywords that may be associated with a given member such as their state, but some may lack enough specificity to take into account. Some cases are admittedly far too indirect or discreet to properly catalogue as referring to a particular member. A noteworthy incident reflective of this phenomenon occurred early in Trump’s presidency where he promoted an episode of Fox News’ “Justice With Judge Jeanine”5 where the top of the show opened with a scathing rebuke of Speaker Paul Ryan failure in leading the effort to repeal the Affordable Care Act (Wagner, 2017). Both President Trump and Speaker Ryan denied that the tweet promoting the show in-and- of- itself was an attack against the latter, highlighting the challenge of not only identifying these uncharacteristically subtle jabs from the president but also the shield of plausible deniability that makes judgement calls on those messages overly presumptuous. Key to the analyses in this paper was the creation of hand-coded endorsement/threat scores to evaluate Trump’s expressed sentiment towards members of Congress. To develop these scores a five-point (-2 to 2) scale was created which factors in the general tone of the tweet along with a level of commitment to politically support or threaten the member. These scores are recorded at the member-tweet level, with tweets that reference multiple members having individual, sometimes contrasting, scores for each. Retweets were coded in this capacity in the same manner as Trump’s original tweets. While there are competing "models" for how users of X/Twitter view their retweets, Trump never indicated that they were not an endorsement of their content, and much of the content of his retweets when it came to member sentiments reflect similar views and positions stated in his original posts. Starting on the negative spectrum of the scale, a score of "-2" indicates that Trump’s tweet had a negative tone and posed electoral and/or political threats against the member(s) discussed in the tweet. Examples of -2 scored tweets include a variety of post genres that are generally aimed at pursuing punitive actions against members of Congress (regardless of the feasibility or sensibility of following through on such a threat), explicitly 5 Trump, Donald J. “Twitter/@RealDonaldTrump: Watch @JudgeJeanine on @FoxNews tonight at 9:00 P.M.” March 25, 2017, 10:41 AM. https://x.com/realDonaldTrump/status/845646761704243200 37 express the desire to electorally oust a member, and/or undermine the political support for a member through stating that they have done a disservice to their constituents, in often less flattering terms. Calls for punitive actions have included the expressed desire to investigate members for alleged wrongdoings such as his counter-Russiagate requests to investigate ties between Democratic leaders Chuck Schumer6 and Nancy Pelosi,7 and repeated calls to "impeach"8 or censure9 Adam Schiff for his role in the impeachment process. Several tweets tied to electoral action will deride the candidate and endorse their op- ponent in one fell swoop such as with Debbie Stabenow and her opponent John James,10 or in the case of senators not up for reelection during Trump’s tenure in office he’d have to play a longer game to threaten their electoral status such as Lisa Murkowski.11 Other lines of attack that fall under this extreme category had Trump directly invoke the member’s electorate to highlight their failings such as Thomas Massie12 and Mitt Romney.13 Less severe on the negative spectrum, "-1" scores fail to meet the severity of the prior 6 Trump, Donald J. "Twitter/@RealDonaldTrump: We should start an immediate investigation into @Sen- Schumer and his ties to Russia and Putin. A total hypocrite! https://t.co/Ik3yqjHzsA" March 3, 2017, 12:54 PM. https://x.com/realDonaldTrump/status/837722869106880517 7 Trump, Donald J. "Twitter/@RealDonaldTrump: I hearby demand a second investigation, after Schumer, of Pelosi for her close ties to Russia, and lying about it. https://t.co/qCDljfF3wN" March 3, 2017, 4:02 PM. https://x.com/realDonaldTrump/status/837770149767827456 8 Trump, Donald J. "Twitter/@RealDonaldTrump: The only crimes in the Impeachment Hoax were com- mitted by Shifty Adam Schiff, when he totally made up my phone conversation with the Ukrainian Presi- dent and read it to Congress, together with numerous others on Shifty’s side. Schiff should be Impeached, and worse!" October 29, 2019, 11:58 PM. https://x.com/realDonaldTrump/status/1189028602052042752 9 Trump, Donald J. "Twitter/@RealDonaldTrump: Hope all House Republicans, and honest House Democrats, will vote to CENSURE Rep. Adam Schiff tomorrow for his brazen and un lawful act of fabricating (making up) a totally phony conversation with the Ukraine President and U.S. President, me. Most have never seen such a thing!" October 17, 2019, 9:14 PM. https://x.com/realDonaldTrump/status/1184638812150009858 10 Trump, Donald J. "Twitter/@RealDonaldTrump:John James, running as a Republican for the Senate from Michigan, is a spectacular young star of the future. We should make him a star of the present. A distinguished West Point Grad and Vet, people should Vote Out Schumer Puppet Debbie Stabenow, who does nothing for Michigan!" November 5, 2018, 12:04 AM. https://x.com/i/web/status/1059310562771984390 11 Trump, Donald J. "Twitter/@RealDonaldTrump: Few people know where they’ll be in two years from now, but I do, in the Great State of Alaska (which I love) campaigning against Senator Lisa Murkowski. She voted against HealthCare, Justice Kavanaugh, and much else..." June 4, 2020, 7:36 PM. https://x.com/realDonaldTrump/status/1268688013791506450 12 Trump, Donald J. "Twitter/@RealDonaldTrump: By empowering the Radical Left Democrats, do nothing Kentucky politician @RepThomasMassie is making their War on the 2nd Amendment more and more difficult to win (But don’t worry, we will win anyway!). He is a disaster for America, and for the Great State of Kentucky!" March 27, 2020, 10:13 AM. https://x.com/realdonaldtrump/status/1243541556424048643 13 Trump, Donald J. "Twitter/@RealDonaldTrump: I’m hearing that the Great People of Utah are consid- ering their vote for their Pompous Senator, Mitt Romney, to be a big mistake. I agree! He is a fool who is playing right into the hands of the Do Nothing Democrats! IMPEACHMITTROMNEY" October 5, 2019, 3:06 PM. https://x.com/realDonaldTrump/status/1180559858699030529 38 category but still display clear misgivings Trump holds towards the member. The lack of any firm commitments to oppose or threaten the member is the key distinction be- tween a threat score of -2 and -1. Perhaps as a result of this distinction, some of these posts may come across as more petty such as chastising Mitch McConnell over failing to repeal Obamacare,14 or attacks on Bob Corker for his support, or lack thereof, for given policies.15 Another common set of messages are ones where President Trump punched down on specific policy proposals or votes such as attacks on Bernie Sander’s advocacy for single- payer healthcare16 or his reaction John McCain’s infamous Obamacare vote.17 The middle of the scale, "0" scores are for posts where Trump is tonally neutral in referring to the member. Many retweets fall into this classification, with them largely serving as announcements of events of political or more personal nature such as a retweet of Rand Paul’s COVID diagnosis.18 Other cases that receive a 0 score are ones where the referenced members are not the real focus of the message such as Trump’s prediction that Alexandria Ocasio-Cortez being unlikely to support Biden’s candidacy.19 Another cate- gory of tweets with 0 scores are ones where Trump advocates for particular members20 to pursue a particular policy objective without any prior context of their existing positions or making a particularly enticing appeal for the members to act. 14 Trump, Donald J. "Twitter/@RealDonaldTrump: "Can you believe that Mitch McConnell, who has screamed Repeal amp, Replace for 7 years, couldn’t get it done. Must Repeal amp, Replace ObamaCare!" August 10, 2017, 6:54 AM. https://x.com/realDonaldTrump/status/895599179522650112?lang=en 15 Trump, Donald J. "Twitter/@RealDonaldTrump: Bob Corker, who helped President O give us the bad Iran Deal amp, couldn’t get elected dog catcher in Tennessee, is now fighting Tax Cuts..." October 24, 2017, 8:31 AM. https://x.com/realDonaldTrump/status/922798321739161600?lang=en 16 Trump, Donald J. "Twitter/@RealDonaldTrump: Bernie Sanders is pushing hard for a single payer healthcare plan - a curse on the U.S. amp, its people..." September 14, 2017, 3:31 PM. https://x.com/realDonaldTrump/status/908413019050463232 17 Trump, Donald J. "Twitter/@RealDonaldTrump: John McCain never had any intention of voting for this Bill, which his Governor loves. He campaigned on Repeal amp, Replace. Let Arizona down!" September 23, 2017, 6:42 AM. https://x.com/realDonaldTrump/status/911541328013676544 18 Trump, Donald J. "Twitter/@RealDonaldTrump: RT @RandPaul: Senator Rand Paul has tested positive for COVID-19. He is feeling fine and is in quarantine. He is asymptomatic and was teste. . . " March 23, 2020, 1:36 PM. https://x.com/RandPaul/status/1241780756617273345 19 Trump, Donald J. "Twitter/@RealDonaldTrump: "Can’t see AOC plus 3 supporting Sleepy Joe!" April 8, 2020, 12:12 PM. https://x.com/realDonaldTrump/status/1247920239045681159 20 Trump, Donald J. "Twitter/@RealDonaldTrump: Why does Twitter leave phony pictures like this up, but take down Republican/Conservative pictures and statements that are true? Mitch must fight back and repeal Section 230, immediately. Stop biased Big Tech before they stop you! @HawleyMO @MarshaBlackburn https://t.co/ah0nMeQdM0" September 8, 2020, 7:58 AM. https://x.com/realDonaldTrump/status/1303301563868479490 39 On the positive end of the spectrum, a score of "2" is obviously the opposite extreme from the -2 scored messages, being attributed to tweets where Trump used a positive tone to express direct electoral and/or political support for a member of Congress. Direct endorsements of members are some of the most obvious tweets that fall within this classification, some of which had some rather surprising recipients21 when considering the members Trump grew less than fond of over time. Other examples include more classical instances of "going public" with Trump using the platform to promote shared policy objectives such as Rand Pual’s theoretical replacement for the Affordable Care Act.22 Finally, a score of "1," naturally mirrors its negative counterpart as expressing generally positive sentiments without the firm commitments or tonal extremity of a 2 endorsement/threat score. Tweets the fall within this scoring are messages that were used to express gratitude towards members for specific actions such as Richard Burr’s over- sight of the Senate Intelligence committee’s review of Trump’s ties to Russia,23 or the passage of legislation including more regionally specific bills.24 Some tweets in this cate- gory also may reflect a twisting of a member’s comments to reflect better on himself such as his warping of Alexandria Ocasio-Crotez’s remark that the VA is not broken due to his work.25 Others reflect a mixed sentiment where one member is credited to contrast with another, such as Trump’s assessment of Kevin McCarthy’s performance compared to his predecessor at the 21 Trump, Donald J. "Twitter/@RealDonaldTrump: .@MittRomney has announced he is running for the Senate from the wonderful State of Utah. He will make a great Senator and worthy successor to @OrrinHatch, and has my full support and endorsement!" February 20, 2018, 9:21. https://x.com/realDonaldTrump/status/965773283554668544 22 Trump, Donald J. "Twitter/@RealDonaldTrump: I feel sure that my friend @RandPaul will come along with the new and great health care program because he knows Obamacare is a disaster!" March 8, 2017, 7:14 PM. https://x.com/realDonaldTrump/status/839268048313929729 23 Trump, Donald J. "Twitter/@RealDonaldTrump: "Highly respected Senator Richard Burr, head of Senate Intelligence, said, after interviewing over 200 witnesses and studying over 2 million pages of documents, “WE HAVE FOUND NO COLLUSION BETWEEN THE TRUMP CAMPAIGN AND RUSSIA.” The Witch Hunt, so bad for our Country, must end!" February 22, 2019, 9:11 AM. https://x.com/realdonaldtrump/status/1099996565551439872 24 Trump, Donald J. "Twitter/@RealDonaldTrump: "Just signed a critical bill to formalize drought contin- gency plans for the Colorado River. Thanks to @SenMcSallyAZ for getting it done. Big deal for Arizona!" April 16, 2019, 6:39 PM. https://x.com/realdonaldtrump/status/1118282780633382912 25 Trump, Donald J. "Twitter/@RealDonaldTrump: Rep. Alexandria Ocasio-Cortez is correct, the VA is not broken, it is doing great. But that is only because of the Trump Administration. We got Veterans Choice amp, Accountability passed. “President Trump deserves a lot of credit.” Dan Caldwell, Concerned Veterans of America" April 19, 2019, 12:54 PM. https://x.com/realDonaldTrump/status/1121095181690974208 40 helm of House Republicans, Paul Ryan.26 Beyond the baseline scores at the member-tweet level, aggregate scores were created at the member level to account for the average and collective total of scores they received from Trump across all mentions. The bulk of analysis here explores these composite sentiments towards members as they are more reflective of Trump’s general attitudes towards the member and also increased salience those attitudes have for Trump and, by extension, his base. Trump discussed members of Congress in over 4700 tweets. Figure 4.1 shows the break down of endorsement/threat scores for the primary member discussed within those tweets. The largest category of is the tonally neutral "0" scores. When considering that a significant volume of these are retweets, announcements, and personal musings that are often times more self-serving in highlighting Trump’s own thoughts and accomplishments, this perhaps not too shocking. Although the plurality of tweets convey a more neutral attitude, this is generally not reflected in his net views of individual members. While there is divergence on the extreme ends, Trump is overall more negative in addressing individual members than he is positive. Reputationally, this is hardly a surprise, as President Trump’s general negativity has been a focus of media and scholarly analysis alike, as discussed before. The more plentiful 2s over -2s is somewhat surprising but it may be explained by the electoral focus that is plentiful across both sets of tweets. For that, I will look at more individual members and their endorsement/threat scores. 26 Trump, Donald J. "Twitter/@RealDonaldTrump: House Minority Leader Kevin McCarthy is a far supe- rior leader than was Lame Duck Speaker Paul Ryan. Tougher, smarter and a far better fundraiser, Kevin is already closing in on 44 Million Dollars. Paul’s final year numbers were, according to Breitbart, “abysmal.” People like. " July 13, 2019, 5:19 PM. https://x.com/realDonaldTrump/status/1150152760362438657 41 Figure 3.1: Composition of Endorsement/Threat Scores Across Members 3.4 Analysis and Discussion Figure 3.2 shows the volume of Trump’s tweets each month, along with the number of tweets which mentioned one or more members of Congress. From the onset of his presidency, Trump’s volume of posts – while less impressive than during his candidacy – has been described as "absurd" (Gallucci, 2017). That being said, there is still an apparent radical increase in tweet output year over year. This is especially apparent in Trump’s final two years, which peaked with 1,415 tweets posted during October 2020. The explosion of activity leading into the 2020 election is naturally a reflection of his desire for reelection and his electioneering ambitions down the ballot as well. The other major spikes in activity line up with the first Trump impeachment and the summer of 2020, a period marked with the second mass wave of COVID-19 infections,27 primary elections wrapping up, and mass protests in the wake of the murder of George Floyd, among other political occurrences (Lee and Alba, 2020). Narrowing in on the subset of tweets that mentioned at least one member of 27 The Corona Virus was not only a defining issue in 2020, it was also a major obstacle to holding Trump’s signature rallies, frustrating the president and limiting his modes of public engagement. 42 Congress and evaluating my expectations for the timing of political communications, Figure 3.2 shows Trump’s maximal engagement with members perfectly aligns with the first impeachment with comparatively minimal fluctuations during periods where his signature legislation was under consideration. Running a t-test on member mentions with the first impeachment (December 18, 2019 through February 5, 2020) similarly returns significant results in support of hypothesis 3.28 The congressional mention peak in December 2019 encapsulates this well as during his first impeachment, Trump used his platform to rain vitriol on the members pushing the effort forward while retweeting and boosting messages from members who towed his line that the process was nonsense and unfair. During the first congress of his term, Trump mostly discussed members of Congress in the lead up to and during the 35 day government shutdown in 2018/2019. Related tweets primarily cast blame towards Democratic opposition. Considering these were the only years during which Trump held a Republican trifecta in government, this tips the hat to how Trump tended to more often criticize than cajole members of Congress. And, with regard to my expectations concerning mention volume, there are generally increased mentions during election years. This is further reflected in t-test statistics.29 28 t = 47.008, df = 1557, p-value < 2.2e-16 29 t = 276.26, df = 15803, p-value < 2.2e-16 43 Figure 3.2: @RealDonaldTrump Tweets and Member mentions Over Time Figure Note: The numbered points on the graph signify the timing of the following significant congressional events: 1) The failed Affordable Care Act repeal attempt in July 2017. 2) Tax Cuts and Jobs Act in November 2017. 3) Start of the first Trump impeachment in December 2019. 4) The Coronavirus Aid, Relief, and Economic Security (CARES) Act in March 2020. 5) The second Trump impeachment in January 2021. Unpacking these mentions a bit more for the big four Congressional leaders in Figure 3.4, we can examine this dynamic in greater detail. As these are amongst the most mentioned members in Congress, the scale between the 115th and 116th Congress varies considerably as is expected from the already observed dynamics. Considering their role in shaping the legislative process, some of their mention-rates are fairly surprising. During the 115th Congress, Mitch McConnell is more frequently discussed after the Obamacare/ACA repeal attempt has failed in an effort to deflect blame from Trump. All the leaders – except for the then lame-duck Paul Ryan – are also mentioned fairly frequently during the shutdown in the transition from the 115th into the 116th Congress, though Speaker Nancy Pelosi is the most mentioned. This completely warps the scale of the 116th Congress, especially during impeachment where she accounts for nearly a quarter of all Congressional mentions during the month of December 2019. Besides institutional activities, the months leading up to both election cycles also see considerable bumps in Democratic leadership engagement as Trump often used Pelosi and Chuck Schumer as mascots for general Democratic dissatisfaction. This is consistent with my expectations. 44 Figure 3.3: Mentions of Congressional Leaders in 115th and 116th Congresses Figure 3.5 shows President Trump’s 10 most referenced members of Congress30 – all of whom were a subject of over 100 posts – along with the number of times they were mentioned as the primary subject of a post. Overwhelmingly, the most discussed members were Nancy Pelosi and Adam Schiff, both of whom were the recipients of demonstrably negative attention. In these cases, much of Trump’s vitriol was rooted in their actions against him. Democrat chamber leaders Nancy Pelosi and Chuck Schumer (the most discussed Senator by far), were a consistent proxy for Trump’s frustration with opposition to his agenda. And, they frequently caught stray attacks aimed at other members that Trump had labelled as "puppets" of either individual. Adam Schiff, too, earned condemnation primarily due to his congressional role as the top Democrat on the House Intelligence Committee.31 In this position, he was not only a key player for investigating collusion between Russia and the Trump campaign in the 2016 election but later the lead 30 Nancy Pelosi, Adam Schiff, Chuck Schumer, Jim Jordan, Mitch McConnell, Lindsey Graham, Bernie Sanders, Mark Meadows, Kevin McCarthy, and Ted Cruz. 31 Ranking member during the 115th Congress, then Chair for the 116th. 45 impeachment manager for Donald Trump. The latter of which was what undoubtedly incited the most animus towards Schiff. By contrast, Jim Jordan was the most referenced Republican in either chamber, and he largely secured his spot through personal loyalty to President Trump. Jordan frequently posted messages of his own that praised Trump and/or attempted to undermine Democratic attacks against him – many of which Trump retweeted. Because of this, Jordan and members like him, have increased notoriety on Trump’s social media because they effectively double as self-promotion for the president. Figure 3.4: Top 10 Most Referenced Members on @RealDonaldTrump Table 3.1 displays the total number of times Trump referenced a Republican Senator who served in the 115th and/or 116th Congresses on social media along with two composite measures that convey his expressed attitude towards them through their endorsement/threat scores, listed in order of mentions. Table 3.2 does the same for Democratic (and Independent) Senators with a full list of all mentioned Senators in alphabetical order being included in the Appendix with tables 3 and 4. These tables only include members that Trump mentioned by name, or nickname, at least once on his @RealDonaldTrump Twitter account, and ignores the 1732 that served during Trump’s tenure that were not referenced through this medium. Interestingly, all members of the Senate ignored by Trump on social media are Democrats, meaning that every Republican that served in the Senate received at least one acknowledgement from the former president on the platform. Another broad observation is that President Trump had polarized views of nearly every member he talked about. The only net-neutral member, Pat Toomey, earned a collective 0 score due to tweets where Trump reversed track on his perspective, initially being positive towards him before having a negative post. Looking at the Republicans in more detail on Table 3.1, the overall sentimentality from Trump is quite positive with only eight members33 having a negative collective score. 32 See Table 15 for the full list of senators Trump never mentioned. 33 Mitt Romney, Richard Burr, John Thune, Bob Corker, John McCain, Jeff Flake, Ben Sasse, and Lisa Murkowski 46 However, this is missing some elements of the whole picture as even the members with the highest positive endorsement/threat scores could and did find themselves on Trump’s bad side on an occasion. As can be inferred from the slightly under 1,000 tweet count to collective score disparity. Many of the messages Trump issues towards these members are tonally neutral, many being pseudo-self-promoting retweets similar to what was discussed with Jim Jordan. 47 Table 3.1: President Trump’s Attitudes Towards Republican Senators Other Republican members that received considerable levels of Trump’s attention on Twitter demonstrated a mixture of utility Trump derived from engaging with Congress. 48 Majority Leader Mitch McConnell topped the list. This was both reflective of his status as Trump’s go-to member to advance his own agenda through the Senate and as an occasional punching bag when that agenda was stalled or objectives failed. The next four most frequently mentioned members are all former presidential rivals and noted critics- turned- allies: Lindsey Graham, Ted Cruz, Marco Rubio, and Rand Paul. The disparity between the tweet count and collective score for each of these members demonstrates both Trump’s mixed perspective on these members, and that he often used them – in a similar capacity to Jim Jordan – as a a source to farm personal praise through retweets. Lindsey Graham and Ted Cruz, along with other members such as Josh Hawley, David Perdue, and Marsha Blackburn, also demonstrate Trump’s efforts to directly endorse and promote loyal members of his party in their (re)election bids. Compared to his copartisans, Trump discussed specific Democratic Senators far less and far more negatively. As shown in Table 3.2, only a single Democrat barely ekes out a positive score: Angus King. King, who is technically an independent that caucuses with the Democrats was awarded one retweet to praise a bipartisan bill he was a part of.34 Excluding this exception, Trump was overwhelmingly negative towards his Democratic rivals, so much so that the total collective endorsement/threat score slightly exceeds the total tweet count where he referenced Democratic members and nearly mirrors the positive score of their Republican colleagues despite being mentioned far less. A similar pattern is apparent in terms of which Democratic Senators Trump talked about the most when compared with their Republican counterparts. Minority Leader Chuck Schumer (again, the most mentioned senator as a whole) is Trump’s continual source of frustration over Democratic opposition in the Senate. Despite some early praise for his fellow New Yorker,35 Trump quickly transitioned into turning Chuck Schumer into the Democratic boogeyman that was at the heart of every perceived policy failure instituted by the Democrats. He was also a frequent secondary subject when Trump discussed Schumer’s 34 Trump, Donald J. "Twitter/@RealDonaldTrump: RT @SteveDaines: Glad to see my bipartisan bill with @SenAngusKing pass the Senate today to increase flexibility for small businesses and w. . . " June 5, 2020, 12:14 PM. 35 Trump, Donald J. "Twitter/@RealDonaldTrump: I have always had a good relationship with Chuck Schumer. He is far smarter than Harry R and has the ability to get things done. Good news!" November 20, 2016, 9:05 AM. https://x.com/realDonaldTrump/status/800339165829283840 49 so-called "puppets."36 The next most mentioned Democrats were leading 2020 presidential contenders that served in the Senate: Bernie Sanders, Elizabeth Warren, and Kamala Harris. The latter also received a lot of attention from Trump on account of her position as Joe Biden’s running mate. The focus on these individuals unquestionably reflect Trump’s own electoral efforts of beating down potential rivals to improve his odds at remaining in the White House. But, his attacks on Sanders and Warren are also emblematic of the broader Republican strategy of increasing voter perception of a radicalization of the Democratic party. 36 Trump, Donald J. "Twitter/@RealDonaldTrump: The last thing we need in Alabama and the U.S. Senate is a Schumer/Pelosi puppet who is WEAK on Crime, WEAK on the Border, Bad for our Military and our great Vets, Bad for our 2nd Amendment, AND WANTS TO RAISES TAXES TO THE SKY. Jones would be a disaster!" November 26, 2017, 8:52 AM. https://x.com/realDonaldTrump/status/934781939088629761? 50 Table 3.2: President Trump’s Attitudes Towards Democratic Senators Table 3.3 shows the results of a Poisson count model examining factors associated with Trump’s propensity to mention a member and an OLS model to explore explanations for his expressed sentiment towards those members in the Senate. Mentions use a Poisson Count model as it best accounts for the structure of the mentions variable, which is a count of the number of tweets that made reference to a particular member, and the presence of 51 several cases with 0 mentions. OLS is utilized for the endorsement/threat scores because the scores themselves can take on negative values, thus making the count approach ineligible. These models utilize year fixed effects and are clustered on the in- dividual member to account for repeated member entries. The first model supports the described findings and hypothesis 1 that Trump is more likely to reference Congressional leaders. It also shows that Trump was more focused on Republicans and, similar to Beckmann, Chaturvedi and Garcia (2017)’s findings with LBJ, disinterested in median pivotal members and more focused on ideologically extreme ones. Trump is also more likely to mention committee chairs. As these are powerful copartisans for his tenure in office, they are critical to his agenda so it makes sense he would at least make passing mention of members in that position. 52 Table 3.3: President Trump’s Twitter Engagement with U.S. Senators When it comes to the endorsement/threat scores of senators, Trump is more negative towards leaders (carried by Chuck Schumer), and there is support for hypothesis 2 with Trump expressing more negativity towards Democrats. A member’s ideological proximity to the median Senator seemingly played little role in Trump’s evaluations of them, and this further reinforces the idea that he did not prioritize heavily catering to pivotal 37 37 The drop off between mentions and endorsement/threats is due to members without mentions not having these scores as they are functionally different from "0"s which represent a collectively neutral sentiment. 53 voters on the regular. Trump was more positive to members who were up for reelection and demonstrated this with frequent endorsements of Republican allies leading up to their elections. The negative relationship between a member’s prior vote share and Trump’s sentimentality suggests that Trump was more likely to target safe members with negative tweets rather than pursuing more vulnerable ones. It’s unlikely Trump felt these attacks would reasonably sway enough voters against these members to make a difference, but it’s possible he felt these attacks against more notable Democrats would have a cascading effect on their copartisans. It is also indicative of Trump perhaps not playing defensive enough with rallying support for marginal Republicans. When it comes to the House, Trump mentions members more frequently than in the Senate. This comes with little surprise considering the membership scale of the House of Representatives compared to the Senate. However, mentions are far more spread-out on the lower end, and 283 members38 are never brought up on the Twitter timeline. House Republicans, seen in Table 3.4,39 received more positive scores overall but have lower averages on the high-mention end. This is primarily on account of them having a high number of "neutral" endorsements when mentioned on Twitter. As has been previously noted with Jim Jordan, many Republican members show up on Trump’s timeline through retweeting glowing praise of the president, which is why many of the upper ranks of mentions consist of firm Trump allies while touting low average endorsements. A key difference between the House and Senate on this front is that the House GOP leaders are discussed far less frequently than Senate Majority Leader Mitch McConnell. This may be a product of the switch between Paul Ryan and Kevin McCarthy for the top position. Furthermore, the lost majority may have spared McCarthy from the tumultuous relationship Trump had with Ryan – characterized by its rapid swings between praise and condemnation in accordance with the trajectory of Republican policy initiatives. 38 See Appendix Tables B3.8, B3.9, B3.10, B3.11, B3.12, B3.13 39 Abridged for space, cutting positive/neutral scoring Republican representatives with less than 6 mentions. 54 Table 3.4: President Trump’s Attitudes Towards Republican Representatives 40 Mark Meadows goes on to serve as Trump’s final Chief of Staff but the vast majority of mentions he has predate his ascension to the position on March 31, 2020. Jeff Van Drew is elected as a Democrat in 2018 but defects to the Republican Party in 2020. Paul Ryan’s neutral score is the result of a shift in attitude from 55 40 Unlike in the Senate, House members are up for reelection every cycle so electoral status fails to explain any variation in mentions or endorsement. There are some examples such as Dan Bishop (who boasts the highest endorsement score in the House), Troy Balderson, and Mike Garcia where Trump’s frequent endorsements of a candidate correlate with their victory in a special election, or Ron DeSantis who successfully ran from the House for Florida governor. Trump identified slightly fewer objectionable co- partisans than in the Senate with five House Republicans earning net negative endorsement/threat scores including Liz Cheney, Mark Sanford, Thomas Massie, Justin Amash, Bradley Byrne, and Raul Labrador. For Democratic references (as seen in Table 3.5),41 Trump’s tweeting patterns are similar to the Senate – being demonstrably negative and skewed towards leadership. Com- pared with the Senate, chairs/ranking-members42 and popular freshmen members43 take the places of presidential candidates44 in the upper ranks. There are two exceptions to the Democratic negativity rule in the form of Jeff Van Drew, who defected to the Republican party in 2020 (his sentiments were already positive before the formal party switch), and Tulsi Gabbard, former presidential candidate turned Democratic pariah following her "present" vote on impeachment and on the Trump 2024 vice presidential shortlist (Hernández and Carman, 2024). Democratic representatives are the most discussed group in Congress by Trump. This fits with his generally more vindictive style as the House was the one chamber controlled by Democrats during his time in office, and it is consistent with the uptick in Trump Twitter activity during the impeachment proceedings. Reinforcing the vindictiveness is that the collective sentiment for House Democrats is very dense in negativity, especially when compared to the positive score of their Republican counterparts. With 91 more tweets than House Republicans, Democratic scores are so negative that the collective total for the chamber is -994,45 a testament to Trump’s frustrations with them. positive to negative, rather than tonally neutral statements. 41 All members mentioned only a single time have been cut for space 42 Adam Schiff, Jerrold Nadler, Maxine Waters, and Elijah Cummings 43 "The Squad," more specifically Alexandria Ocasio-Cortez, Ilhan Omar, and Rashida Tlaib. 44 Though Beto O’Rourke, Tulsi Gabbard, and Eric Swalwell do have multiple mentions 45 See Appendix Table 9 56 Table 3.5: President Trump’s Attitudes Towards Democratic Representatives 57 Table 3.6: President Trump’s Twitter Engagement with U.S. Representatives 46 Table 3.6 shows the results of Poisson count and OLS models47 similar to those used for the Senate but using data from House members instead. These models also support hypothesis 1, with similar effects present for leadership in terms of being mentioned and having negative sentiments (again carried by the Democratic leadership). Chairmanship status is influential in driving the likelihood of mention and negative sentiment in the House as opposed to the Senate where chairmen had null results. This makes sense when 46 The drop off between mentions and endorsement/threats is due to members without mentions not having these scores as they are functionally different from "0"s which represent a collectively neutral sentiment. 47 These use year fixed-effects and have the data clustered on the member to account for overlaps in membership and maintain the same specifications outside of dropping the functionally meaningless "up for reelection" variable. 58 considering many of his top Democratic mentions fell into these ranks. Partisanship reflects a Democratic bias in Trump’s mentions and negative endorsement/threat scores, thus supporting the expectations for hypothesis 2. Ideologically, Trump is more likely to mention ideologically extreme members but the null results as it pertains to sentiment are likely a product of the shift in control from Republicans to Democrats in the 116th Congress. There are also mixed results for electoral vulnerability in shaping when and how Trump talks about a member, where he is more likely to mention less secure members but has a more positive attitude towards electorally safe members. 3.5 Conclusion As a whole, Donald Trump’s usage of Twitter during his presidency synthesized member and public communication in a manner which provides unique insight into his attitudes towards other elected officials that largely support my expectations regarding their political nature. Leaders in both parties were typically at the forefront of the former president’s commentary, though the House was more egalitarian with key allies and enemies outpacing the heads of the Republican caucus in each term. Across Congress as a whole, Trump was generally more negative than positive in his public commentary about members. In no small part was this the result of his heavily negative attitude towards Democrats – with his top targets being major members from the opposition party. The timing of his messages are also reflective of the consideration of political context. The election years and the impeachment process acted as peak seasons for his Twitter activity as is made evident by the volume of his discussion of members at these times. Examining Republicans more closely unveils a considerable divide in how he talked about members by party. While Republicans received far more attention when combining both chambers, they are referenced much of the time only through their own discussion of and praise for President Trump. Some did, however, receive a bit more direct support through promotion of their policies and upcoming elections. Contrarily, Democrats were talked about far more negatively and intensely, nearly mirroring Republican endorsement/threat scores while having a considerably lower group tweet count. These messages highlighted (and exaggerated) policy contrasts in addition to providing numerous calls to action to mobilize voters against the president’s opposition party. 59 The general pattern of congressional-engagement on Twitter from Trump suggests that he was more focused on agenda setting than persuading members that might be pivotal to the passage of key bills. There are several examples of Trump expressing frustration with members for slow-rolling their support or failing to get legislation across the finish-line.48 Just as well, there are attempts to "grease the wheels" with certain members in hopes they vote a certain way.49 However, these are fairly low-effort compared to something like the Johnson Treatment. While congressional leaders are frequent targets, other pivotal members – namely centrists – lacked salience for Trump. Trump did not use Twitter to engage with the same level of lobbying that Johnson did, but his choices of more top-loaded engagement are more closely aligned with who Johnson contacted most regularly (Beckmann, Chaturvedi and Garcia, 2017). Social media was not the exclusive domain of communication from the Trump ad- ministration. There were certainly behind-the-scenes meetings and efforts to try to gain support for various administration initiatives. Not to mention, there were abundant political rallies throughout his time in office. These engagements are worthy of further ex- amination but are beyond the scope of this paper. Twitter was a consistent and relatively unfiltered tool in the Trump arsenal that – because of its ease of access – provides a com- prehensive collection of the subjects Trump publicly prioritized throughout his administration. This opens the door for using his presidential tweets as a window into a wide array of political subjects outside of Congress as well. While Twitter’s conversion to X and many of the subsequent policies adopted by owner Elon Musk have depreciated several facets of that particular platform (Reed, 2024), social media as a whole shows no sign of vanishing any time soon. And, the degree to which Donald Trump utilized this powerful communication tool remains an available and potentially viable strategy for future presidents. Trump’s immediate successor, Joe Biden, is hardly the prolific tweeter that previously occupied the oval office. However, Biden and his 48 Trump, Donald J. "Twitter/@realDonaldTrump: "The Freedom Caucus will hurt the entire Republican agenda if they don’t get on the team, amp, fast. We must fight them, amp, Dems, in 2018!" March 30, 2017, 1:07:00 PM. https://x.com/realDonaldTrump/status/847435163143454723 49 Trump, Donald J. "Twitter/@realDonaldTrump: @SenJohnMcCain-Thank you for coming to D.C. for such a vital vote. Congrats to all Rep. We can now deliver grt healthcare to all Americans!" July 25, 2017, 3:24 PM. https://x.com/realDonaldTrump/status/889929412539543552 60 reelection campaign are acutely aware of happenings on the social plat- form (Stokols, 2024; Weigel, 2024) and seem interested in advancing more elements of social media engagement into their strategy (Titcomb, 2024). With Trump himself mounting another presidential bid in 2024 as well as the lifting of his Twitter/X suspension, time will tell if such tactics will again be put to use very soon. 61 CHAPTER 4: PUTTING THE "BULLY" IN BULLY PULPIT: PRESIDENT TRUMP’S CONGRESSIONAL PURGE THROUGH TWITTER At the Faith and Freedom Coalition conference on July 17, 2022, Senator Lindsay Graham (R-SC) pined for the yesteryears of the Trump presidency, stating the following: "You know what I liked about Trump? Everybody was afraid of him. Including me" (Shoaib, 2022). Senator Graham’s confession of fearing the former president expresses the power Trump wielded over his fellow Republicans. In Graham’s case it has been well documented by journalists and politicos how transformative of an effect his fear of the president was. He went from being one of then-candidate Trump’s sharpest and most outspoken critics with noteworthy statements referring to Trump as "a race-baiting, xenophobic religious bigot," to a stalwart ally and sycophant for President Trump, professing that he “deserves the Nobel Peace Prize and then some" (Ring, 2020). Graham is but one example of how President Trump used his influence to remold Congress and the Republican party to fit his personal brand. It is not difficult to imagine why Senator Graham held some level of fear for the former president. Several cases during the Trump era demonstrate the political costs Republican congressmen endured for crossing the president. Although Trump’s feud with John Mc- Cain (R-AZ) was one of his most prolific, his animosity towards fellow Arizonan Senator Jeff Flake had perhaps more direct political implications. Unlike McCain who wouldn’t live to run for another election during Trump’s time in office, Flake had to face a primary challenger who became the vessel of Trump’s retribution. Trump quickly endorsed Flake’s challenger, Dr. Kelli Ward. He then took to defaming Flake on a number of key is- sues, referred to him as "toxic" (Brown, N.d.),1 and would later celebrate his drop from the 1 Trump, Donald J. "Twitter/@RealDonaldTrump: Great to see that Dr. Kelli Ward is running against Flake Jeff Flake, who is WEAK on borders, crime and a non-factor in Senate. He’s toxic!" August 17, 2017, 6:56 AM. https://x.com/realDonaldTrump/status/898136462385979392 62 race. Trump continued piling disparaging remarks upon Flake throughout the remainder of the latter’s term by referring to him as "unelectable... weak and ineffective" (Brown, N.d.).2 Would Flake have met this fate had his critiques of the president escaped Trump’s notice? Or, was Trump’s enumerated disdain for the senator a key factor in shaping his political fortunes? The answer to these questions has important repercussions for the relationship between Congress and the president not only during the Trump era but also beyond. The president has long been recognized as the predominant leader of their political party (Seligman, 1956), with this role taking shape during the presidencies of Grover Cleveland and William McKinley (Klinghard, 2005). This position at the top of the party hierarchy typically manifests in setting the party’s political agenda. But, some presidents have not been fully satisfied in this capacity, especially when they perceive members of their own party as a consistent obstacle to their policy goals. A particularly noteworthy example is Franklin Delano Roosevelt (FDR), who sought to remove disloyal Democrats from Congress when he found his domestic agenda strained (Dunn, 2012). Beyond seeking systemic purges of copartisan apostates, the overarching strategy of "going public," identified by Kernell (2006), is an adversarial tactic that presidents have utilized to place members of Congress directly in the line of fire with their constituents and bypassing traditional bargaining and negotiation lines. Compared with prior presidents, Trump may have been ideally positioned to instigate and carry out a purge of his copartisans. His campaign as a political outsider firmly identified his candidacy as a rejection of the Republican candidates that preceded him. And, in turn, their rejection of him cemented a clear rift in the party (Burns and Barbaro, 2016). Trump’s electoral success where McCain and Romney had failed was lorded over the GOP (especially the individuals3 on those campaigns),4 nearly until the end of his presidency. The 2 Trump, Donald J. "Twitter/@RealDonaldTrump: "Jeff Flake(y) doesn’t want to protect the Non- Senate confirmed Special Counsel, he wants to protect his future after being unelectable in Arizona for the “crime” of doing a terrible job! A weak and ineffective guy!" November 9, 2018, 12:10 PM. https://x.com/realdonaldtrump/status/1060942600994050048 3 Trump, Donald J. "Twitter/@RealDonaldTrump: Paul Ryan, the failed V.P. candidate former Speaker of the House, whose record of achievement was atrocious (except during my first two years as President), ultimately became a long running lame duck failure, leaving his Party in the lurch both as a fundraiser leader " June 11, 2019, 11:10 PM. https://x.com/realDonaldTrump/status/1149516400341348354 4 Trump, Donald J. "Twitter/@RealDonaldTrump: Obama is lucky he ran against @MittRomney, a man with very little talent or political skill, as opposed to someone who knows how to fight and win!" September 10, 63 battle for the "soul" of the GOP was a struggle throughout Trump’s term. Many, including some of Trump’s once sharpest critics like Lindsey Graham and Marco Rubio (R-FL), quickly fell in line with the president. Some "never-Trump" Republicans, however, coalesced around a political action committee, the Lincoln Project, to oppose the president. Notably, though, its failure to meaningfully chip away support from the Republican base is a testimony to Trump’s role in the party (McLean, N.d.). President Trump’s evergreen utilization of X/Twitter5 as a communication tool synthesized personal communication with members and public visibility. And with the high level of media coverage his tweets garnered, many messages of his were significantly amplified. (Lazarus and Thornton, 2021; Wells et al., 2020). From his time as president-elect until his banning from Twitter in the final days of his presidency, Trump posted more than 26,600 tweets. Of those, approximately 18% of them were about one or more mem-bers of Congress. While Twitter was hardly the exclusive method Trump used to pursue his bully pulpit, his near-constant engagement with the platform makes his account an excellent catalog of the political issues, individuals, and other subjects that were salient with the president at any given moment in addition to being his "loudest" microphone. Trump’s eventual banning from Twitter in the closing days of his lame-duck presidency came with the tacit acknowledgement that his usage of the platform had consider- able and dangerous influence on his followers (X, 2021). The rhetoric on and leading to the January 6th insurrection can be viewed as the maximal discontent Trump expressed towards Congress and a testament to his capacity to use the lever of the public to – in this case quite literally – threaten members of Congress. It should therefore come as little surprise that Trump was able to exercise his position to influence the departure of several members that did not fit his vision for the party. 4.1 Purges, Presidents, and Tweets From historical perspectives, prior presidents utilized various tools to influence the public on various elements of their political objectives. The presidency of FDR in many 2010. 8:56 AM. https://x.com/realDonaldTrump/status/1304040990550958081 5 While the platform has been re-branded as "X" as of 2023, "Twitter," and related vernacular, will be used to identify the website throughout this manuscript as that was the name it carried throughout the entirety of the Trump administration. 64 aspects revolutionized the office of the president by strengthening and building many of the institutional features that continue to define the presidency to this day (Burke, 2000). One such notable contribution is FDR’s efforts to bring the president closer to the public with his Fireside Chat radio addresses. FDR’s usage of the mass communication technology of his era has been interpreted as a credit to his leadership, and it helped calm and reassure the public in the throes of the Great Depression (Kiewe, 2007). In the broader scheme of his presidency, the Fireside Chats were used to set expectations and persuade the public of the merits of the administration’s various policy initiatives (Brockman, 2014; Pedemonte, 2019). Similar to Trump’s use of Twitter, FDR was comfortable shifting this mechanism into an offensive tool to attack obstacles to his administration’s agenda. In a Fireside Chat on June 24th 1938, FDR maligned the "Copperheads" in his party and spoke directly to the public about the upcoming Democratic primaries. He urged listeners to prevent candidates who impeded his objectives in carrying out the "liberal declaration of principles" he had set forth in 1936 from securing Democratic nominations (Roosevelt, 1938). Although Roosevelt’s disdain for disloyal Democrats was newly made apparent, divisions in the Democratic coalition had existed for years and intraparty disputes over the New Deal existed from the onset, and the ranks of obstructionist Democratic Senators had expanded before FDR even secured reelection (Dunn, 2012).6 FDR did more than simply encourage voters to carefully consider whether Democrats in their primaries were sufficiently committed to the New Deal’s liberal ideals; he directly intervened in their primaries and backed liberal challengers (Dunn, 2012). The efforts to exile conservative members of his own party were consistent with Roosevelt’s beliefs that the party system in America required a more cohesive conservative/liberal divide (Milkis, 1985), a notion largely taken for granted in far-more-polarized contemporary politics. FDR’s primary moves against members of his own party quickly earned the derisive label of "purge," likening it to more drastic efforts undertaken in the Soviet Union, but the label helped undermine these efforts (Dunn, 2012; Shogan, 2006). One of the purge’s 6 Dunn (2012) identifies Carter Glass (VA), Harry Byrd (VA), Millard Tydings (MD), Thomas Gore (OK), Josiah Baily (NC), Walter George (GA), Ellison "Cotton Ed" Smith (SC), Robert Bulkley (OH), Bennett Champ Clark (MO), Burton Wheeler (MT), Tom Connally (TX), Pat McCarran (NV), and Guy Gillette (IO) as early members of the Democratic party to oppose the New Deal. 65 targets, Walter George (GA), electorally slaughtered his Roosevelt-anointed opponent (Telfeyan, 2002) which set a clear tone for the elections of the members FDR targeted, with nine of the ten Democratic members he sought to replace winning reelection. Paired with a high number of electoral defeats for Democrats in both chambers and the bolstering of the Conservative coalition in Congress, FDR’s attempted purge is historically viewed as one of the biggest failures of his presidency (Dunn, 2012; Milkis, 2004; Wolf, 2019). Roosevelt’s attempt to forge a party more broadly loyal to him and his agenda is a precursor in many ways for the intraparty dynamics of the Republican party under Donald Trump. A testament to the overall influence Donald Trump and his character on the Republican party is the media narratives frequently positing the former president as a fac- tor for member exit from his party throughout his presidency. The framing of this issue has taken a few forms. Some, including the former president himself, have interpreted Trump- era exits within a broader pattern of Republican retirement, particularly the departure of committee chairmen who have maxed out their time in the position (Reynolds, 2017). Others have framed it as reflecting negative reactions among Republicans to Trump and his brand of politics, suggesting members either willfully departed on their own so as to dissociate from the party or felt forced out due to their criticism of the president being met with retaliation (Bade, 2019; Krause and Byers, 2022). Trump’s general impact on the Republican party has given rise to questions over whether or not he has realigned the party’s voting base, with pointed losses among suburban voters and shoring up of rural and evangelical voters being characteristic shifts of this potential seventh party system (Vance, 2021). Regardless of whether Trump’s role in the Republican party and its overall impact warrant the evaluation of a new American party system, it is clear that his grip on the party has had an important impact on the party’s base and – by electoral extension – its congressional branch. Trump’s influence in the party seems to go beyond the traditional role of president by bringing radical elements of the Republican party into the forefront (Baker and Bader, 2022; Espinoza, 2022), tying public perceptions of the party to himself (Jacobson, 2018, 2020), and controlling the RNC (Heersink, 2018). While few presidents have attempted systemic purges of poor team players in the party, the parallels between some of the departures from the Republican party and the Democrats FDR attempted to oust highlight 66 an effort not so unfamiliar to American politics. Trump, in contrast with his modern Re- publican predecessors, has tossed aside efforts to broaden the party’s voting coalition in favor of doubling down on a devoted base (Galvin, 2020). It is worth more deeply examining Trump’s efforts to purge Republicans he did not like from Congress. A concerted effort to electorally purge members is likely conducted through multiple avenues. For Trump, none likely stand as pronounced as his use of social media as a mode of communicating to the party base. Indeed, Trump’s mastery of digital communication has been deemed to be something of a spiritual successor to the Fireside Chats of FDR and "The Radio Priest" Charles Coughlin (Oborne and Roberts, 2017). The lack of coherently committed conservative ideology in Trump’s policies raised questions over whether his initiatives would see support from his party, but surveys of Republicans have shown a willingness to embrace Trump over their self-identified conservative ideals (Barber and Pope, 2019). This contrasts with the reception of his rhetorical style which was seen as a source of division within the party’s ranks (Krause and Byers, 2022) and had mixed reception among voters (Nelson, 2017; Thomsen, 2017). In utilizing Twitter, Trump melded the harsh interpersonal bargaining posture of Lyn- don Baines Johnson’s interpersonal communication style (Zelizer, 2015) with highly visible mass communication and kept the electorate in on the conversation – potentially to the detriment of the other party (Kernell, 2006). As a candidate and president, Donald Trump commanded attention in the public sphere (Wells et al., 2020) and was reinforced through the widespread proliferation of his tweets (Pérez-Curiel and Limón-Naharro, 2019). Presidents have long been able to reorient media narratives in a more favorable capacity through press conferences and other events (Miles, 2014), and Trump’s immediate predecessor, Barack Obama, pioneered presidential use of the Twitter platform (Aharony, 2012; Pfeiffer, 2018). However, the media was uniquely captivated by Trump and guar- anteed that all but the most disengaged citizens were aware of the posts on @RealDonaldTrump (Lazarus and Thornton, 2021). Beyond the potential congressional influence, prior scholarship has identified several consequences that emerged as a result of Trump’s tweets. Several tweets have been found to influence financial markets in their immediate aftermath (Filippou et al., 2020; Gjerstad et al., 2021; Huynh, 2021; Klaus and Koser, 2021). It has also been generally used to further 67 right-wing political projects through the normalization of discourse (Kreis, 2017; Varis, 2020), spread conspiracy theories (Hornsey et al., 2020; Ross and Rivers, 2018), and shift the political discussion surrounding different policy areas (Sahin, Johnson and Korkut, 2021). 4.2 Expectations People paid attention to the president’s Twitter account, especially once he was in of- fice, and this granted considerable leverage and influence in many areas (Pérez-Curiel and Limón-Naharro, 2019). Therefore, it is not far-fetched to speculate that Trump’s social media use could mobilize the Republican base to revolt against Republicans Trump regularly lamented in political primaries. Or, it could even be used against Democrats in the general election in a similar though less physically-threatening manner than he had in rallying the efforts to interfere with his failed reelection results (X, 2021). Indeed, the wider- reaching and more-frequent nature of Trump’s Twitter feed may have enabled Trump to carry out a far more successful (in terms of departing members) purge than FDR or any other prior presidents. I examine the following expectations for President Trump’s Twitter communications towards congressional Republicans. Hypothesis 1a: Ideological Preference. Trump is more likely to tweet positively/negatively about members of his own party who are more conservative/liberal. Even if Trump himself may have lacked an ideological core personally (Barber and Pope, 2019), his presidency still rhetorically supported conservatism (Varis, 2020). These proclivities in the partisan sphere should lead to preferential treatment towards copartisans in-part based on where they fall on the ideological spectrum. Hypothesis 1b: Disloyalty Dissatisfaction. Trump is more likely to tweet negatively about members of his own party who vote less in line with his policy agenda. Donald Trump’s status as a political outsider may have molded a less-than-typical president (Schier and Eberly, 2017; Worthy and Bennister, 2017), but he should have concrete motivations to go after members of his own party he did not like. The "going public" approach has been used as a vehicle for politicians to get member votes, which should extend toward more negative tweets directed at members who deviated from 68 Trump’s agenda and policy preferences more frequently. Hypothesis 2: Dissident Purge. Members of the Republican party which received negative attention from Trump were likely to leave or be removed from office. Political dynamics have shifted tremendously since Franklin Roosevelt attempted to purge conservative southern Democrats from his party. The partisans are far more ideo- logically cohesive, removing the rift which attributed to the uphill battle FDR faced when trying to oust those members he felt jeopardized his agenda. Donald Trump’s ascension to the central figure of his party (Heersink, 2018; Jacobson, 2020) arrived at a time that should have enabled him to more successfully target and remove members who he found objectionable than FDR. Social media is also a far larger and more sweeping tool than what FDR had at his disposal, thus giving Trump the capacity to engage in more relent- less rhetoric against purge targets than FDR would have been able to. 4.3 Data This paper uses data collected from Donald Trump’s Twitter account, @RealDonaldTrump as archived by Brown (N.d.). As the analyses of this paper are concerned with presidential communication and it’s influence, the data field is limited to Trump’s time as president- elect starting November 8, 2016 through his account’s original suspension on January 8, 2021. This time period captures a total of 26621 posts, including retweets and deleted messages posted by Donald Trump during this time frame. Tweets were hand-coded for references to members of Congress, and the level of endorsement/threat Trump posited towards them in those messages. Trump’s proclivities posed some interesting cases which required extra levels of at- tention to properly code. While his tendency towards giving nicknames for members of Congress was primarily aimed at Democrats, Trump occasionally developed one for his copartisans. Many of these are readily identifiable such as "Little" Ben Sasse (NE) or "Big" Luther Strange (AL). Less specific ones such as "RINO" (used for several members) and "Mitch’s Boy" (John Thune (SD)) are helpfully linked with a proper noun reference to the member Trump is referencing. Occasionally some additional awareness of the member’s state and political circumstances to correctly identify the member. Cases that were far too indirect are ignored in terms of their implications for members. An infamous example of 69 such a tweet involved Paul Ryan (WI) where Trump’s recommendation to watch Judge Jeanine’s show on Fox News7 treated viewers with an admonishment of Speaker Ryan as a negligent political shepherd who was actively harming Trump’s policy agenda in Congress (Wagner, 2017). Paul Ryan and Donald Trump denied any connection between the tweet and the attack on Jeanine’s show, adding plausible deniability to the list of rea- sons to ignore hundreds of hours of tweet-promoted Fox News content as potentially member- related messaging. Each tweet with one or more identified members also includes a corresponding endorsement/threat score for each of those members to reflect Trump’s expressed attitude towards that member. Endorsement/threat scores are on a five-point (-2 to 2) scale which considers the general tone and level of commitment to politically support or threaten the assigned member. A tweet that references multiple members has a unique score for each, which may be the same or vary depending on the construction of the tweet and how each member is treated in it (i.e. an endorsement of a candidate challenging an incumbent is a positive score for the candidate and a negative one for the incumbent). These scores are also compiled for composite measures at the member level. Breaking down the scale, the positive spectrum is differentiated between "2" and "1" through the expression of direct electoral and/or political support. Electoral endorsements of members as well as elevating awareness and support of shared policy objectives are typical tweets within the "2" genre, while expressing gratitude, congratulatory messages, and general pleasant remarks about a member fall into the categorization of "1." "0s" reflect tonally neutral expressions towards the member with examples being basic statements of fact, messages where the member is not a primary subject, and most frequently, retweets from a member. The negative spectrum is of greater focus for this paper and its analyses. "-1s" consist of attacks on members that focus more on specific policy actions or failures and milder attacks without going fully scorched-earth on the recipient. The failure to repeal the Affordable Care Act was a particularly noteworthy theme in these types of tweets, 7 Trump, Donald J. “Twitter/@RealDonaldTrump: Watch @JudgeJeanine on @FoxNews tonight at 9:00 P.M.” March 25, 2017, 10:41 AM. https://x.com/realDonaldTrump/status/845646761704243200 70 particularly taking aim at congressional leaders for their self-inflicted embarassment.8 Some of these can be fairly petty attacks as well, such as a complimenting John Kerry for making a joke at Thomas Massie’s (KY) expense9 after Trump previously expressed his frustration over his lack of support for COVID relief policies, or kicking the departing Bob Corker (TN) while he was down with several snide remarks about his inelectibility.10 In contrast, "-2s" pose a directed political and/or electoral threat to the receiving member. The most obvious instances of this are Trump promising to campaign against someone, such as he had with Lisa Murkowski, fully aware she would not be up for reelection for another two years.11 Maximal pursuits of "going public" (Kernell, 2006), wherein Trump directly encouraged the member’s constituency to reevaluate their member12 also fall into the -2 classification. Trump referenced at least one member of Congress in over 4700 tweets. Figure 4.1 shows the breakdown of endorsement/threat scores for all Republicans Trump discussed via tweet during his time in office. On the whole, Trump was very positive towards his copartisans on the platform with a combined 43% of scores being a "2" or "1." Neutral "0" scores make up the largest categorization of member sentiments with negative "-1" and "-2" scores making up 5.5% of the remainder. This breakdown is fairly unsurprising, considering how consolidated Trump has made the Republican party behind him (Espinoza, 2022; Owens, 2021). Taking to Twitter to more frequently prop up Congressional allies in his own party, rather than relentlessly pursuing the perceived "bad seeds," casts Trump as generally more of a team player in this capacity. Trump’s endurance of a fairly low level of dissension Never knew John Kerry had such a good sense 8 Trump, Donald J. "Twitter/@RealDonaldTrump: The only problem I have with Mitch McConnell is that, after hearing Repeal & Replace for 7 years, he failed!That should NEVER have happened!" August 24, 2017, 9:42 AM. https://x.com/realdonaldtrump/status/900714982823821313?lang=bg 9 Trump, Donald J. "Twitter/@RealDonaldTrump: of humor! Very impressed!" March 27, 2020, 12:16 PM. https://x.com/realDonaldTrump/status/1243572588200280064 10 Trump, Donald J. "Twitter/@RealDonaldTrump: Isn’t it sad that lightweight Senator Bob Corker, who couldn’t get re-elected in the Great State of Tennessee, will now fight Tax Cuts plus!" October 24, 2017, 9:30 AM. https://x.com/realDonaldTrump/status/922817539251556353 11 Trump, Donald J. "Twitter/@RealDonaldTrump: Few people know where they’ll be in two years from now, but I do, in the Great State of Alaska (which I love) campaigning against Senator Lisa Murkowski. She voted against HealthCare, Justice Kavanaugh, and much else..." June 4, 2020, 7:36 PM. https://x.com/realDonaldTrump/status/1268688013791506450 12 Trump, Donald J. "Twitter/@RealDonaldTrump: By empowering the Radical Left Democrats, do nothing Kentucky politician @RepThomasMassie is making their War on the 2nd Amendment more and more difficult to win (But don’t worry, we will win anyway!). He is a disaster for America, and for the Great State of Kentucky!" March 27, 2020, 10:13 AM. https://x.com/realdonaldtrump/status/1243541556424048643 71 in the ranks translates into a seemingly proportionate level of grievances in response. In addition to the tweet related variables, several control variables related to member biographical data including party, ideological Scores (NOMINATE), Bycoffe (2021) Trump Loyalty Scores, leadership status, committee ranks, gender, and vote shares in their prior elections were also collected to run analyses. Figure 4.1: Composition of Endorsement/Threat Scores Across Republican Members 4.4 Analysis and Discussion Table 4.1 displays the results of a Poisson Count model for variables associated with Trump’s likelihood of tweeting about a Republican member of Congress with fixed effects for year and standard errors clustered on the individual member to account for repeated member observations. A Poisson Count model is used as it best accounts for the structure of the Twitter mentions variable, which is a count of the number of tweets that contained a member’s name, and the presence of several cases with 0 mentions. Trump was far more likely to mention Republican leaders and committee chairs than rank-and-file members. Trump was less likely to discuss those who were less supportive of his agenda13 and those who were more electorally secure, as measured by their prior vote share. Gender seemingly played no role in his proclivity to discuss Republicans in Congress. It is somewhat surprising that Trump would be less likely to discuss members who deviate from his agenda on social 13 As derived from Trump support scores from Bycoffe (2021), with the <75% support variable referring to 70 out of 1092 member-year observations. 72 media, considering his vindictive reputation. However, deeper analysis of Trump’s endorsement/threat scores and who was discussed will provide some clarity on this front. 73 Table 4.1: Determinants of Trump’s Member-Tweet Count Figure 4.2 shows the result of an OLS model for the drivers of Trump’s endorsement/threat scores for his copartisans. Leadership and ideology, as measured by first dimension nominate scores, are the only statistically significant variable which is positively associated with higher endorsement scores. Trump was more likely to post positively about more conservative members of the caucus, supporting my ideological expectations for his sentiments about copartisans from hypothesis 1a. The lack of significance that decreased loyalty would be associated with more negative tweets against a member fails to support my expectations from hypothesis 1b. There are only eight observations of Republicans with a support score below 75% that have a net negative sentiment across their congressional career: Justin Amash in 2017-2019,14 Steve King in 2020, Raul Labrador in 2018, Thomas Massie in 2018, John McCain in 2017, and Mark 14 Ironically, Amash’s highest score, 81.2% was in 2020, the year he defected from the Republican party and voted to impeach Trump which consequently was the only year Trump tweeted about him. 74 Sanford in 2018. The disconnect between loyalty scores and expressed threats against members is perhaps the product of general disinterest from Trump in policy and that only defections on key votes earned a sufficient level of ire to publicly complain about a member. It is also likely that Trump was motivated by more personal animus related to member’s statements and critiques of him which was not reflected in their support for his legislative agenda. This will be explored as I more closely examine the members Trump primarily discussed in a negative light. Table 4.3 shows all Republicans that Trump posted at least one negative tweet about while both were in office, ordered in ascending order from lowest to highest collective sentiments. As previously inferred from Figure 4.1, as Trump was mostly positive and neutral towards his copartisans on Twitter, fairly few Republican members received negative posts, and even fewer (14) received enough for it to reflect Trump’s prevailing sentiment of that member. Some of the members Trump expressed the most positive sentiments towards overall were not entirely immune from receiving occasional criticism. Mitch McConnell’s position as Senate majority leader – the shepherd of much of Trump’s congressional agenda – placed him in the line of fire when Trump’s agenda was impeded legislatively15 or electorally.16 Even some of Trump’s most steadfast allies such as Jim Jordan and future chief of staff Mark Meadows would earn ire when they strayed from his chosen path such as slow-rolling support on key votes.17 15 Trump, Donald J. "Twitter/@RealDonaldTrump: The only problem I have with Mitch McConnell is that, after hearing Repeal Replace for 7 years, he failed!That should NEVER have happened!" August 24, 2017, 9:42 AM. https://x.com/realdonaldtrump/status/900714982823821313?lang=bg 16 Trump, Donald J. "Twitter/@RealDonaldTrump: I hope the Democrats, and even more importantly, the weak and ineffective RINO section of the Republican Party, are looking at the thousands of people pouring into D.C. They won’t stand for a landslide election victory to be stolen. @senatemajldr @JohnCornyn @Sen- JohnThune" January 5, 2021, 5:12 PM. https://x.com/realDonaldTrump/status/1346580318745206785 17 Trump, Donald J. "Twitter/@RealDonaldTrump: If @RepMarkMeadows, @Jim_Jordan and @Raul_Labrador would get on board we would have both great healthcare and massive tax cuts & reform." March 30, 2017, 5:20 PM. https://x.com/realDonaldTrump/status/847559239149158401 75 Table 4.2: Determinants of Trump’s Expressed Sentiments Towards Republicans 76 Table 4.3: President Trump’s Attitudes Towards Republicans Trump’s Twitter account was suspended prior to his second impeachment which took place between January 26th and February 13th 2021. Had his account been active he undoubtedly would have had plenty of attacks towards the Republicans who voted against him with the precedents set by Mitt Romney and Justin Amash, the only Republicans18 who voted in favor of his first impeachment and conviction. As a result, only Liz Cheney of the ten19 House Republicans that voted in favor of impeachment has any negative mentions going in, and only Bill Cassidy (LA) of the seven20 Senate Republicans who voted to convict him escaped any negative attention. Exploring the members Trump expressed collectively negative views towards, his success in purging undesirable members of the party can be better accessed. Mitt Romney, who has by far the most negative score, was the subject of one of Trump’s most public feuds. 18 Justin Amash had technically switched his party affiliation to "Independent" prior to casting his vote in favor of impeachment. 19 The others being: Anthony Gonzalez (OH), Jaime Herrera Beutler (WA), John Katko (NY), Adam Kinzinger (IL), Peter Meijer (MI), Dan Newhouse (WA), Tom Rice (SC), Fred Upton (MI), and David Valadao (CA). 20 The others being: Richard Burr, Susan Collins, Lisa Murkowski, Mitt Romney, Ben Sasse, and Pat Toomey. 77 Donald Trump actually endorsed candidate Mitt Romney,21 but Trump almost im- mediately turned on Senator Romney after he won his seat questioning whether he would be "a Flake" (in reference to the former Senator from Arizona).22 Romney remained in Trump’s bad graces through his continued public criticism of the president and being the only Republican senator to vote in favor of the first impeachment, an act which Trump pulled a reversal for insisted that Romney should instead be "impeached."23 Although he survived the Trump years, due to not facing reelection during them, Romney did announce he would not seek reelection in 2024. This is an outcome unquestionably satisfactory for the former president as he announced his explicit intention to rid the party "of the Romneys of the world" (Bates, 2024). Senators Jeff Flake and Bob Corker were early casualties of the Trump era, both up for reelection in the 2018 midterm and opting to retire rather than face voters with their mutual public animosity with their copartisan president. Corker24 and Flake25 opposed Trump in signature policy areas driving further negativity. Additional Senators Trump singled out to attack endured similar fates. Richard Burr fell into Trump’s bad graces when, as head of the Senate Intelligence Committee, he subpoenaed Donald Trump Jr. over potential election interference involvement with Russia (Collins, 2019). Burr would retire when he was next up for reelection in 2022. Ben Sasse saw a turn in Trump’s perspective on him following the release of leaked audio of him criticizing the president over various policies, his character, and "rage tweeting" (Axelrod, 21 Trump, Donald J. "Twitter/@RealDonaldTrump: .@MittRomney has announced he is running for the Senate from the wonderful State of Utah. He will make a great Senator and worthy successor to @OrrinHatch, and has my full support and endorsement!" February 20, 2018, 9:21 AM. https://x.com/realDonaldTrump/status/965773283554668544 22 Trump, Donald J. "Twitter/@RealDonaldTrump: Here we go with Mitt Romney, but so fast! Question will be, is he a Flake? I hope not. Would much prefer that Mitt focus on Border Security and so many other things where he can be helpful. I won big, and he didn’t. He should be happy for all Republicans. Be a TEAM player WIN!" January 2, 2019, 7:53 AM. https://x.com/realDonaldTrump/status/1080447092882112512 23 Trump, Donald J. "Twitter/@RealDonaldTrump: .@MittRomney has announced he is running for the Senate from the wonderful State of Utah. He will make a great Senator and worthy successor to @OrrinHatch, and has my full support and endorsement!" February 20, 2018, 9:21. https://x.com/realDonaldTrump/status/965773283554668544 24 Trump, Donald J. "Twitter/@RealDonaldTrump: Bob Corker, who helped President O give us the badIran Deal amp, couldn’t get elected dog catcher in Tennessee, is now fighting Tax Cuts..." October 24, 2017,8:31 AM. https://x.com/realDonaldTrump/status/922798321739161600?lang=en 25 Trump, Donald J. "Twitter/@RealDonaldTrump: Great to see that Dr. Kelli Ward is running against Flake Jeff Flake, who is WEAK on borders, crime and a non-factor in Senate. He’s toxic!" August 17, 2017, 6:56 AM. https://x.com/realDonaldTrump/status/898136462385979392 78 2020). Trump quickly lashed out at the "stupid and obnoxious"26 senator and directly compared him to previously ousted critics Bob Corker and Jeff Flake.27 He further suggested that Sasse should retire or be replaced with a more viable Republican.28 Fortunately for Sasse’s immediate political prospects, two weeks from the general election was far too late to be replaced by another Republican. He did ultimately resign from his office prematurely, though, to assume a position as the President of the University of Florida before having the chance to face a GOP primary again. John McCain was a member Trump entered office with hostility towards. These flames were only further fueled through McCain’s pivotal vote that killed the attempt to repeal the Affordable Care Act29 despite dubious efforts from Trump to grease the wheels for McCain on that vote.30 Two of the eight Senators trump targeted negatively did manage to survive their next election – a far more successful sweep than that of Franklin Roosevelt’s purge which saw all ten of his targets remain. John Thune received retaliation from Trump following his rejection of election denialism during the lame duck session. The president not only made the call for action for Thune to be primaried31 but also directly challenged his and others’ assertions that the 2020 election was fair.32 Thune held on in a competitive Republican 26 Trump, Donald J. "Twitter/@RealDonaldTrump: ...Nomination to run for a second term. Then he went back to his rather stupid and obnoxious ways. Must feel he can’t lose to a Dem. Little Ben is a liability to the Republican Party, and an embarrassment to the Great State of Nebraska. Other than that, he’s just a won- derful guy!" October 17, 2020, 9:39 AM. https://x.com/realDonaldTrump/status/1317460179223498753 27 Trump, Donald J. "Twitter/@RealDonaldTrump: Senator Little Ben Sasse of the Great State of Nebraska seems to be heading down the same inglorious path as former Senators Lid- dle’ Bob Corker, whose approval rating in Tennessee went from 55% to 4%, Jeff “the Flake” Flake, whose approval rating in Arizona went from 56% to...." October 17, 2020, 12:00 PM. https://x.com/realDonaldTrump/status/1317495722133065728 28 Trump, Donald J. "Twitter/@RealDonaldTrump: ....practically nothing. Both Senators be- came totally unelectable, couldn’t come even close to winning their primaries, and decided to drop out of politics and gracefully “RETIRE”. @SenSasse could be next, or perhaps the Republicans should find a new and more viable candidate?" October 17, 2020, 12:00 PM. https://x.com/realDonaldTrump/status/1317495725786353670 29 Trump, Donald J. "Twitter/@RealDonaldTrump: Democrats are laughingly saying that McCain had a "moment of courage." Tell that to the people of Arizona who were deceived. 116% increase!" September 23, 2017, 6:20 PM. https://x.com/realDonaldTrump/status/911717004222091264 30 Trump, Donald J. "Twitter/@RealDonaldTrump: So great that John McCain is com- ing back to vote. Brave - American hero! Thank you John." July 25, 2017, 6:44 AM. https://x.com/realDonaldTrump/status/889798407228989441 31 @RealDonaldTrump tweet from December 23, 2020: "Republicans in the Senate so quickly forget. Right now they would be down 8 seats without my backing them in the last Election. RINO John Thune, “Mitch’s boy”, should just let it play out. South Dakota doesn’t like weakness. He will be primaried in 2022, political career over!!!" 32 @RealDonaldTrump tweet from January 5, 2021: "I hope the Democrats, and even more importantly, the 79 primary in 2022. Trump followed through on his previous threat to campaign against Lisa Murkowski in 2022 (Ruskin, 2022) due to her repeated defections on marque votes, but she managed to hold on to her seat in a highly competitive election decided in Round 3 of Alaska’s ranked choice voting election. In the House, Trump was similarly successful in seeing his co-partisan opponents removed from office with only one survivor holding out among the seven who received negative attention. Thomas Massie, the sole survivor, earned considerable hostility on ac- count of his opposition towards COVID relief33 but would go unpunished by his Conservative constituency. His fellow derided colleagues were not as fortunate. Justin Amash’s renouncement of the Republican party in the lead up to his vote in favor of impeaching Donald Trump was met with mockery by the president who was gleeful at the knowledge he was being challenged for his seat,34 though Amash would opt not to seek reelection in 2020. Raul Labrador received relatively mild criticism for slow-rolling support for health care reform with Mark Meadows and Jim Jordan. But, unlike those two, he never received counteracting positive remarks, and he was defeated in a gubernatorial primary in 2018. Mark Sanford was similarly ousted in 2018, though he was the recipient of greater hostility as Trump considered him "unhelpful" to the MAGA agenda.35 Bradley Byrne’s negativity is more so the result of collateral damage as he was not Trump’s preferred candidate for the GOP primary for the Alabama Senate seat in 2020. Finally, Liz Cheney, well known for her role in the January 6th Committee following Trump’s 2020 defeat, earned ridicule from Trump for foreign policy disagreements36 before Trump grew frustrated with weak and ineffective RINO section of the Republican Party, are looking at the thousands of people pouring into D.C. They won’t stand for a landslide election victory to be stolen. @senatemajldr @JohnCornyn @SenJohnThune" 33 Trump, Donald J. "Twitter/@RealDonaldTrump: ... costly. Workers small businesses need money now in order to survive. Virus wasn’t their fault. It is “HELL” dealing with the Dems, had to give up some stupid things in order to get the “big picture” done. 90% GREAT! WIN BACK HOUSE, but throw Massie out of Re- publican Party!" March 27, 2020, 9:44 AM. https://x.com/realDonaldTrump/status/1243534445367492608 34 Trump, Donald J. "Twitter/@RealDonaldTrump: Great news for the Republican Party as one of the dumbest amp, most disloyal men in Congress is “quitting” the Party. No Collusion, No Obstruction! Knew he couldn’t get the nomination to run again in the Great State of Michigan. Already being challenged for his seat. A total loser!" July 4, 2019, 9:05 AM. https://x.com/realDonaldTrump/status/1146766931573301248 35 Trump, Donald J. "Twitter/@realDonaldTrump: Mark Sanford has been very unhelpful to me in my campaign to MAGA. He is MIA and nothing but trouble. He is better off in Argentina. I fully endorse Katie Arrington for Congress in SC, a state I love. She is tough on crime and will continue our fight to lower taxes. VOTE Katie!" June 12, 2018, 4:12 PM. https://x.com/realdonaldtrump/status/1006630395067039744 36 Trump, Donald J. "Twitter/@RealDonaldTrump: Liz Cheney is only upset because I have been actively 80 her failure to embrace his "stolen victory" narrative. Due to timing, Cheney would face the consequences in the next electoral cycle, losing the Republican primary in 2022. Table 4.4 shows the full breakdown of the negatively targeted members and when they were "purged" from Congress in the Trump-era. Ben Sasse has the "*" on account of him technically resigning in early 2023, but he aligns most closely with the 2022 cycle despite not being up for reelection during it. John McCain has been excluded on account of his death to brain cancer being divorced from any causality link between Trump’s attitude towards him and his departure from Congress. This shows the overall 10/13 success rate Trump had in eliminating less desirable Republicans from the party through identification and chastisement on Twitter. Comparing t-tests for pre-general election departures through retirement or primary election loss (for those with negative scores (t= 4.2682, df = 558.39, p-value = 2.315e-05) and those with positive scores (t = -16.304, df = 619.93, p-value < 2.2e- 16)) there is further support for my expectations regarding Trump’s purge successes and that his influence extends beyond mere coincidence. This is apparent when comparing the same type of exits in the party. Table 4.4: Negative Republican Departures in the Trump Era Beyond those with a net-negative score, there are other members who Trump changed his attitude towards for the worse. Former Speaker of the House Paul Ryan was generally a recipient of positive endorsements and even praise37 during his tenure. Following his departure from Congress, however, Trump would unleash a flurry of criticism his way.38 getting our great and beautiful Country out of the ridiculous and costly Endless Wars. I am also making our so-called allies pay tens of billions of dollars in delinquent military costs. They must, at least, treat us fairly!!!" July 23, 2020, 7:45 AM. https://x.com/realDonaldTrump/status/1286266178789310465 37 Trump, Donald J. "Twitter/@RealDonaldTrump: Speaker Paul Ryan is a truly good man, and while he will not be seeking re-election, he will leave a legacy of achievement that nobody can question. We are with you Paul!" April 11, 2018, 9:50 AM. https://x.com/realDonaldTrump/status/984066131303583746 38 Trump, Donald J. "Twitter/@RealDonaldTrump: ....Paul Ryan almost killed the Republican Party. Weak, ineffective stupid are not exactly the qualities that Republicans, or the CITIZENS of our Country, were looking 81 Trump’s online rhetoric, at least towards Paul Ryan, certainly did not play a role in the latter’s decision to leave Congress, but his bullying of the former Speaker may have helped keep him politically sidelined for the remainder of the Trump-era. 4.5 Conclusion The small cohort of Republicans Trump was negative towards in the Senate and House reinforce many of the expectations for the types of members Trump has spoken about removing from the party. This small sample indicates considerable success from the former president on this front. Trump did not make the same concerted effort FDR did in terms of selecting a specific cycle, so his critiques are scattered and – in the case of the Senate – not necessarily immediately actionable. Similar to FDR, he found ideology to be important in reflecting his expressed sentiment towards members. But, he diverges from his distant predecessor in that legislative support for his agenda was not a clear criterion for earning his wrath. Nonetheless, only two39 senators and one representative40 Trump opposed on Twitter remain in office after facing reelection and all of the representatives he negatively targeted have retired from Congress or lost primaries. Compared with the single representative, John O’Connor (NY), that FDR was able to knock off of the 11 conservative Democrats he tried to beat, Trump’s 10/1341 record is far more impressive. This even ignores other avenues through which he may have identified members beyond this group for removal. The volume and access of the tweets likely helped Trump succeed where Roosevelt failed. Despite that success, Trump largely landed in the same position post- primary with a reduced Republican coalition in both chambers and middling general electoral performances – even as far as the 2022 midterms among his preferred candidates (Bowman, 2022). This ultimately places the tangible policy outcomes in the same position as Roosevelt’s failed purge in spite of Trump’s accomplishing more of its immediate objective. Trump’s success in purging members he didn’t like from Congress pre-general election (and keeping many of them out of political office since) demonstrates the powerful for. Right now our spirit is at an all time high, far better than the Radical Left Dems. You’ll see next year!" July 13, 2019, 5:19 PM. https://x.com/realDonaldTrump/status/1150152762044440576 39 Lisa Murkowski and John Thune 40 Thomas Massie 41 John McCain would elevate this to 11/14 but considering his death presumably had little to do with Trump’s opposition to him, he’s excluded from this final count. 82 grip he holds over his party and has important implications for his successors. Com- pared with many prior occupants of the Oval Office, Trump has remained a political force out of office. He has helped engineer the ousting of other Republicans members of Congress who avoided adverse attention while he was in office. For example, his rescinded endorsement of Mo Brooks in favor of Katie Britt for the Alabama Senate seat in 2022 resulted in the former’s defeat (Koplowitz, 2022). Trump’s 2024 campaign has made no secret out of his desire to broadly purge disloyal, non-"MAGA," Republicans from the party entirely (Bates, 2024). And, this notion has sweeping implications that extend far beyond just presidential- congressional relations. Even during his time in office he certainly had plenty to say about candidates who never made it into Congress along with down ballot Republicans (Ouyang et al., 2020). Assessing the potential political ramifications of Trump’s rhetoric towards politicians outside of Congress is worthy of future examination. Beyond Trump, the tools to use the presidential bully pulpit to take out perceivably disloyal partisans remain largely intact. All that is required is an executive willing to employ them and less-than-willing targets. Though a radical departure from Trump in temperament, Joe Biden has been similarly presented with copartisans who have served as hindrances to his agenda. Some debate has emerged over whether Biden should con- sider purging centrist Democrats, such as the since-turned Independents Joe Manchin and Krysten Sinema. However, the fear of reliving FDR’s mistake (Rozsa, 2021) and a likely general lack of will has kept this tactic as a theoretical. Nonetheless, Biden has taken to X/Twitter to passive-aggressively chastise Democratic obstructionists with repeated calls for "two more Democratic senators."42 It is also possible that the particulars of the Republican base in the modern political climate, paired with X/Twitter’s right-wing pivot under Elon Musk’s ownership (Anderson, 2023) and absence of a consensus alternative platform, may limit the viability of these tactics to future Republican presidents. 42 Biden, Joeseph R. "X/@JoeBiden: If you give me two more Democratic senators, and Democrats keep the House, I promise you we will codify Roe v. Wade. We will once again make Roe the law of the land. We will once again protect a woman’s right to choose." September 23, 2022, 1:18 PM. https://x.com/JoeBiden/status/1573361187491921920 83 CHAPTER 5: CONCLUSION Technological innovations have added a new veneer to the dynamics between political ac- tors in Congress and beyond, yet beneath the surface we still find commonalities between these individuals and those who served before them. Much of our understandings about what motivates members of Congress (Fenno, 1978; Mayhew, 1974a) are still true, but the costs and benefits of serving have to be considered more broadly than simply seeking reelection for job security’s sake. Many key members, from committee chairs to chamber leaders, have left office in recent years which is indicative of broader motivations tied to the utility of their seats that are not being satisfied in the present political context. If these figures are reevaluating what they are getting out of remaining in Congress, rank-in- file members undoubtedly question if the probability of achieving their political ambitions within Congress is worth running for another term or not. The relationship between Congress, its members, and the President of the United States is also of paramount importance to the functioning of our institutions and the be- haviors of individuals therein. The advent of social media, and a president willing to engage with one such platform to its fullest, has opened a new realm through which the chief executive may fuse public and personal engagements with members of Congress. During Donald Trump’s presidency, his use of Twitter was on the cutting edge and granted him arguably the loudest megaphone of any American President to spread his message, especially when paired with the proliferation of his tweets as leading news stories (Wells et al., 2020). Despite the radical evolution Twitter use was over prior forms of political communication, and the former president’s divergence in tone compared to his predecessors, Trump’s engagement with members of Congress on the platform followed general patterns that resembled communication strategies from many prior presidents. Even Donald Trump’s copartisan feuds and efforts to reshape the Republican party were not without precedent. While Franklin Roosevelt could scarcely imagine the scale and reach of Twitter, he too identified a number of copartisan Democratic congressmen who he tried, and failed, to remove from office. Greater partisan-ideological cohesion and the 84 further reaching tools at his literal fingertips likely gave Trump an edge that made him more successful than FDR in this regard, as several targeted Republicans lost primaries or saw the writing on the wall and retired. Trump’s stated objectives since the end of his administration demonstrate a desire to continue purging partisan dissidents making the understanding of his nascent efforts all the more important for the future of American politics. 85 BIBLIOGRAPHY Abramowitz, Alan and Jennifer McCoy. 2019. “United States: Racial resentment, negative partisanship, and polarization in Trump’s America.” The ANNALS of the American Academy of Political and Social Science 681(1):137–156. Aharony, Noa. 2012. “Twitter use by three political leaders: an exploratory analysis.” Online information review 36(4):587–603. Ahuja, Sunil, Staci L Beavers, Cynthia Berreau, Anthony Dodson, Patrick Hourigan, Steven Showalter, Jeff Walz and John R Hibbing. 1994. “Modern congressional election theory meets the 1992 house elections.” Political Research Quarterly 47(4):909–921. Altheide, David L. 2022. Gonzo governance: The media logic of Donald Trump. Routledge. Anderson, Monica. 2023. “After Musk’s takeover, big shifts in how Republican and Democratic Twitter users view the platform.” Pew Research Center. URL: https://www.pewresearch.org/short-reads/2023/05/01/after-musks-takeover-big- shifts-in-how-republican-and-democratic-twitter-users-view-the-platform/ Ang, Adrian U-Jin and L Marvin Overby. 2008. “Retirements, retentions, and the balance of partisan power in contemporary Congressional politics.” The Journal of Legislative Studies 14(3):339–352. Ansolabehere, Stephen and James M Snyder Jr. 2002. “The incumbency advantage in US elections: An analysis of state and federal offices, 1942–2000.” Election law journal 1(3):315–338. Axelrod, Tal. 2020. “Trump excoriates Sasse over leaked audio.” The Hill. URL: https://thehill.com/homenews/administration/521529-trump-excoriates-sasse- over- leaked-audio/ Bade, Rachael. 2019. “Trump’s takeover of GOP forces many House Republicans to head for the exits.” The Washington Post. URL: https://www.washingtonpost.com/politics/trumps-takeover-of-gop-forces many- house-republicans-to-head-for-the-exits/2019/09/22/d89f99fc-d4bd-11e9-ab26 e6dbebac45d3story.html Baker, Joseph O and Christopher D Bader. 2022. “Xenophobia, partisanship, and support for Donald Trump and the Republican Party.” Race and Social Problems 14(1):69–83. Ballard, Andrew O, Ryan DeTamble, Spencer Dorsey, Michael Heseltine and Marcus Johnson. 2022. “Dynamics of polarizing rhetoric in congressional tweets.” Legislative Studies Quarterly. Barber, Michael and Jeremy C Pope. 2019. “Does party trump ideology? Disentangling party and ideology in America.” American Political Science Review 113(1):38–54. 86 Bates, Suzanne. 2024. “Trump says Republican Party is ‘getting rid of the Romneys’.” Yahoo! News . URL: https://news.yahoo.com/trump-says-republican-party-getting- 010236439.html Baum, Matthew A. 2013. “Presidential communication strategies in the age of three me- dias.” iPolitics: Citizens, elections, and governing in the new media era pp. 183–205. Beckmann, Matthew N. 2010. Pushing the agenda: Presidential leadership in US lawmaking, 1953–2004. Cambridge University Press. Beckmann, Matthew N, Neilan S Chaturvedi and Jennifer Rosa Garcia. 2017. “Targeting the Treatment: The Strategy behind Lyndon Johnson’s Lobbying.” Legislative Studies Quarterly 42(2):211–234. Bernstein, Jeffrey L and Jennifer Wolak. 2002. “A bicameral perspective on legislative retirement: The case of the Senate.” Political Research Quarterly 55(2):375–390. Bianco, William T, David B Spence and John D Wilkerson. 1996. “The electoral connection in the early Congress: The case of the Compensation Act of 1816.” American Journal of Political Science 40(1):145. Blake, Aaron. 2023. “Jim Jordan’s remarkably thin legislative track record.” The Washing- ton Post. Blood, Michael R. and Kevin Freking. 2023. “Kevin McCarthy was booted as House speaker two months ago. now he’s leaving Congress by year’s End.” AP News . Boche, Adam, Jeffrey B Lewis, Aaron Rudkin and Luke Sonnet. 2018. “The new Voteview. com: preserving and continuing Keith Poole’s infrastructure for scholars, students and observers of Congress.” Public Choice 176(1-2):17–32. Bond, Jon R and Richard Fleisher. 1990. The president in the legislative arena. University of Chicago Press. Bowler, Shaun, Miguel Carreras and Jennifer L Merolla. 2023. “Trump tweets and democratic attitudes: Evidence from a survey experiment.” Political Research Quarterly 76(3):1325–1339. Bowman, Bridget, Ben Kamisar Alexandra Marquez. 2022.“By the numbers: How Trump- backed candidates fared in the midterms.” NBC. URL: https://www.nbcnews.com/meet-the-press/meetthepressblog/numbers-trump- backed-candidates-fared-midterms-rcna61524 Brace, Paul. 1985. “A probabilistic approach to retirement from the US Congress.” Legislative Studies Quarterly pp. 107–123. Brant, Hanna K. and L. Marvin Overby. 2021. “Congressional Career Decisions in the 2018 87 Congressional Midterm Elections.” Congress & the Presidency 48(1):8–24. URL: https://doi.org/10.1080/07343469.2020.1811424 Bresnahan, John and Jake Sherman. 2014. “Rogers to retire, heading to Radio.” Politico . URL: https://www.politico.com/story/2014/03/mike-rogers-michigan-house- intelligence- committee-105127 Brockman, Christopher. 2014. Scaffolding the New Deal: Exploring the Legislative Roots of Franklin Delano Roosevelt’s Fireside Chats. Lehigh University. Brown, Brendan. N.d. “Trump Twitter Archive.” Forthcoming. URL: https://www.thetrumparchive.com/ Burke, John P. 2000. The institutional presidency: Organizing and managing the White House from FDR to Clinton. Johns Hopkins University Press. Burns, Alexander and Michael Barbaro. 2016. “Mitt Romney and John McCain Denounce Donald Trump as a Danger to Democracy.” New York Times. URL: www.nytimes.com/politics/first-draft/2016/03/03/mitt-romney-donald-trump/ Butler, Daniel M, Adam G Hughes, Craig Volden and Alan E Wiseman. 2023. “Do constituents know (or care) about the lawmaking effectiveness of their representatives?” Political Science Research and Methods 11(2):419–428. Bycoffe, Aaron, Nate Silver. 2021. “Tracking Congress In The Age Of Trump.” FiveThirtyEight. URL: https://projects.fivethirtyeight.com/congress-trump-score/ Carson, Jamie L and Erik J Engstrom. 2005. “Assessing the electoral connection: Evidence from the early United States.” American Journal of Political Science 49(4):746–757. Carson, Jamie L, Erik J Engstrom and Jason M Roberts. 2007. “Candidate quality, the personal vote, and the incumbency advantage in congress.” American Political Science Review 101(2):289–301. Clarke, Harold D, Frank B Feigert, Barry J Seldon and Marianne C Stewart. 1999. “More time with my money: leaving the house and going home in 1992 and 1994.” Political Research Quarterly 52(1):67–85. Coleman, John. 2018. “The Johnson Treatment: Pushing And Persuading Like LBJ.” Forbes. URL: https://www.forbes.com/sites/johncoleman/2018/07/30/the-johnson- treatment-pushing-and-persuading-like-lbj/?sh=2706ef614201 Collins, Eliza. 2019. “Firestorm erupts around Sen. Richard Burr, man behind Donald Trump Jr. subpoena. He’s not backing down.” USA Today. URL: https://www.usatoday.com/story/news/politics/2019/05/11/trump-jr-subpoena- richard-burr-faces-pressure/1169275001/ 88 Cox, Gary W and Jonathan N Katz. 1995. “Why did the incumbency advantage in US House elections grow?”. Deering, Christopher J and Paul J Wahlbeck. 2006. “US House committee chair selection: Republicans play musical chairs in the 107th Congress.” American Politics Research 34(2):223–242. Dunn, Susan. 2012. Roosevelt’s purge: How FDR fought to change the Democratic Party. Harvard University Press. Eckles, David L, Cindy D Kam, Cherie L Maestas and Brian F Schaffner. 2014. “Risk attitudes and the incumbency advantage.” Political Behavior 36(4):731–749. Erikson, Robert S. 1971. “The advantage of incumbency in congressional elections.” Polity 3(3):395–405. Espinoza, Michael. 2022. Donald Trump’s impact on the Republican Party. In The Trump Administration. Routledge pp. 134–150. Fenno, Richard F. 1978. Home style: House members in their districts. Harper Collins. Filippou, Ilias, Arie Eskenazi Gozluklu, My T Nguyen and Ganesh Viswanath-Natraj. 2020. “The information content of Trump tweets and the currency market.” Available at SSRN 3754991 . Fleisher, Richard and Jon R Bond. 1996. “The president in a more partisan legislative arena.” Political Research Quarterly 49(4):729–748. Follett, Mary Parker. 1896. The speaker of the House of Representatives. London and Bombay, Longmans, Green, and Company. Frantzich, Stephen E. 1978. “Opting Out: Retirement from House of Representative, 1968- 1974.” American Politics Quarterly 6(3):251–273. Gallucci, Nicole. 2017. “The number of tweets Donald Trump has sent since becoming president is downright shocking.” Mashable. URL: https://mashable.com/article/donald-trump-tweets-six-months Galvin, Daniel J. 2020. Party domination and base mobilization: Donald Trump and Re- publican Party building in a polarized era. In The forum. Vol. 18 De Gruyter pp. 135– 168. Gangitano, Alex. 2024. “Robert De Niro argues Trump has ‘snapped’ in new Biden ad.” The Hill. URL: https://thehill.com/homenews/campaign/4683887-robert-de-niro-argues- trump-has- snapped-in-new-biden-ad/ Gans, Jared and Al Weaver. 2023. “Feinstein corrected by staffer about retirement announcement.” The Hill. URL: https://thehill.com/homenews/senate/3858125- 89 feinstein-corrected-by-staffer-about-retirement-announcement/ Gilmour, John B and Paul Rothstein. 1993. “Early republican retirement: A cause of democratic dominance in the House of Representatives.” Legislative Studies Quarterly pp. 345–365. Gjerstad, Peder, Peter Filip Meyn, Peter Molnár and Thomas Dowling Næss. 2021. “Do President Trump’s tweets affect financial markets?” Decision Support Systems 147:113577. Gluck, Abbe R. 2021. Unorthodox lawmaking and legislative complexity in American statutory interpretation. In Comparative Multidisciplinary Perspectives on Omnibus Legislation. Springer pp. 195–225. Golbeck, Jennifer, Justin M Grimes and Anthony Rogers. 2010. “Twitter use by the US Congress.” Journal of the American society for information science and technology 61(8):1612– 1621. Green, Matthew N. 2010. The Speaker of the House: A study of leadership. Yale University Press. Groseclose, Timothy and Keith Krehbiel. 1994. “Golden parachutes, rubber checks, and strategic retirements from the 102d House.” American Journal of Political Science pp. 75– 99. Gross, Terry. 2018. “Who Is The Mystery Man Behind @realDonaldTrump? (Besides The President).” NPR. URL: https://www.npr.org/2018/04/19/603904858/who-is-the- mystery-man-behind-realdonaldtrump-besides-the-president Hamel, Brian T and Michael G Miller. 2019. “How voters punish and donors protect legislators embroiled in scandal.” Political Research Quarterly 72(1):117–131. Heersink, Boris. 2018. “Trump and the party-in-organization: Presidential control of national party organizations.” The Journal of Politics 80(4):1474–1482. Hemphill, Libby, Annelise Russell and Angela M Schöpke-Gonzalez. 2021. “What drives US congressional members’ policy attention on Twitter?” Policy & Internet 13(2):233– 256. Hemphill, Libby, Aron Culotta and Matthew Heston. 2013. “Framing in social media: How the US Congress uses Twitter hashtags to frame political issues.” Available at SSRN 2317335. Hemphill, Libby, Jahna Otterbacher and Matthew Shapiro. 2013. What’s congress doing on twitter? In Proceedings of the 2013 conference on Computer supported cooperative work. pp. 877–886. 90 Hernández, Alec, Scott Bland and Joella Carman. 2024. “Who could be Trump’s vice president?” NBC News. URL: https://www.nbcnews.com/politics/2024-election/who- could-be-trumps-vice-president-rcna150403 Herrnson, Paul S, Irwin L Morris and John McTague. 2011. “The impact of presidential campaigning for congress on presidential support in the US House of representatives.” Legislative Studies Quarterly 36(1):99–122. Hibbing, John R. 1982. “Voluntary retirement from the US House: The costs of congressional service.” Legislative Studies Quarterly pp. 57–74. Hoover, Judith. 1988. “An Early Use of Television as a Political Tool: The 1961 News Conferences of President John F. Kennedy and the Republican Opposition.” Journal of Popular Film and Television 16(1):41–48. Hornsey, Matthew J, Matthew Finlayson, Gabrielle Chatwood and Christopher T Begeny. 2020. “Donald Trump and vaccination: The effect of political identity, conspiracist ideation and presidential tweets on vaccine hesitancy.” Journal of Experimental Social Psychology 88:103947. Hubbard, Lauren. 2018. “Read the Full Transcript of Former Senator Joe Lieberman’s Speech at John McCain’s Memorial Service.” Town Country. URL: https://www.townandcountrymag.com/society/politics/a22892540/joe-lieberman- john-mccain-memorial-speech-full-transcript/ Huynh, Toan Luu Duc. 2021. “Does bitcoin react to Trump’s tweets?” Journal of Behavioral and Experimental Finance 31:100546. Jacobson, Gary C. 2018. “The effects of the early Trump presidency on public attitudes toward the Republican Party.” Presidential Studies Quarterly 48(3):404–435. Jacobson, Gary C. 2020. “Donald Trump and the parties: Impeachment, pandemic, protest, and electoral politics in 2020.” Presidential Studies Quarterly 50(4):762–795. Jacobson, Gary C and Samuel Kernell. 1983. “Strategy and choice in congressional elections.”. Jeydel, Alana and Andrew J Taylor. 2003. “Are women legislators less effective? Evidence from the US House in the 103rd-105th Congress.” Political Research Quarterly 56(1):19– 27. Karl, Jonathan. 2021. Front Row at the Trump Show. Penguin. Karni, Annie. 2017. “Aides give up on trying to control Trump’s tweets.” Politic . URL: https://www.politico.com/story/2017/11/17/trump-al-franken-tweets-twitter- 247662 Kelly, Caroline. 2019. “Trump attacks McChrystal after retired general called Trump im- 91 moral.” CNN. URL: https://www.cnn.com/2019/01/01/politics/trump-tweet- mcchrystal/index.html Kernell, Samuel. 2006. Going public: New strategies of presidential leadership. Cq Press. Kiewe, Amos. 2007. FDR’s first fireside chat: public confidence and the banking crisis. Texas A&M University Press. Kiewiet, D Roderick and Langche Zeng. 1993. “An analysis of congressional career decisions, 1947–1986.” American Political Science Review 87(4):928–941. King, Gary. 1991. “Constituency service and incumbency advantage.” British Journal of Political Science 21(1):119–128. Klaus, Jürgen and Christoph Koser. 2021. “Measuring Trump: The Volfefe index and its impact on European financial markets.” Finance Research Letters 38:101447. Klinghard, Daniel P. 2005. “Grover Cleveland, William McKinley, and the emergence of the president as party leader.” Presidential Studies Quarterly 35(4):736–760. Knight First Amendment Inst. at Columbia Univ. v. Trump, No. 1:17-cv-5205 (S.D.N.Y.), No. 18- 1691 (2d Cir.), No. 20-197 (Supreme Court). 2019. Koplowitz, Howard. 2022. “Donald Trump endorses ‘fearless America First warrior’ Katie Britt in Alabama Senate race.” Advance Local. URL: https://www.al.com/politics/2022/06/donald-trump-endorses-fearless-america-first- warrior-katie-britt-in-alabama-senate-race.html Krause, George A and Jason S Byers. 2022. “Confirmation Dynamics: Differential Vetting in the Appointment of US Federal Agency Leaders.” The Journal of Politics 84(2):1189– 1201. Krehbiel, Keith. 1998. Pivotal politics: A theory of US lawmaking. University of Chicago Press. Kreis, Ramona. 2017. “The “tweet politics” of President Trump.” Journal of language and politics 16(4):607–618. LaMarre, Heather L and Yoshikazu Suzuki-Lambrecht. 2013. “Tweeting democracy? Examining Twitter as an online public relations strategy for congressional campaigns’.” Public relations review 39(4):360–368. Landers, Elizabeth. 2017. “White House: Trump’s tweets are ‘official statements’.” CNN Politics. URL: https://www.cnn.com/2017/06/06/politics/trump-tweets-official- statements/index.html Lassen, David S and Adam R Brown. 2011. “Twitter: The electoral connection?” Social science computer review 29(4):419–436. 92 Lawless, Jennifer L and Sean M Theriault. 2005. “Will she stay or will she go? Career ceilings and women’s retirement from the US Congress.” Legislative Studies Quarterly 30(4):581–596. Lazarus, Jeffrey and Judd R Thornton. 2021. “Bully pulpit? Twitter users’ engagement with President Trump’s tweets.” Social science computer review 39(5):961–980. Lee, Carol, Kristen Welker and Monica Alba. 2020. “For Trump, June was the darkest of months.” NBC News. URL: https://www.nbcnews.com/politics/white-house/trump- june-was-darkest-months-n1232619 Lee, Frances E. 2016. Insecure majorities: Congress and the perpetual campaign. University of Chicago Press. Livingston, Steven G and Sally Friedman. 1993. “Reexamining theories of congressional retirement: Evidence from the 1980s.” Legislative Studies Quarterly pp. 231–253. Maltzman, Forrest, Lee Sigelman and Sarah Binder. 1996. “Leaving Office Feet First: Death in Congress.” PS: Political Science and Politics 29(4):665–671. Masthay, Theodore J and L Marvin Overby. 2017. “Dynamics of senate retirements.” Po- litical Research Quarterly 70(1):184–193. Matthews, Donald R. 1959. “The folkways of the United States Senate: Conformity to group norms and legislative effectiveness.” American Political Science Review 53(4):1064– 1089. Mayhew, David R. 1974a. Congress: The electoral connection. Vol. 26 Yale University Press. Mayhew, David R. 1974b. “Congressional elections: The case of the vanishing marginals.” Polity 6(3):295–317. McLean, Caroline. N.d. “The Lincoln Project–Success Story or Hype.” Biden vs Trump. Forthcoming. Miles, Matthew R. 2014. “The bully pulpit and media coverage: Power without persuasion.” The international journal of press/politics 19(1):66–84. Milkis, Sidney M. 1985. “Franklin D. Roosevelt and the Transcendence of Partisan Poli- tics.” Political Science Quarterly 100(3):479–504. Milkis, Sidney M. 2004. “FRANKLIN DELANO ROOSEVELT 1933–.” The American Pres- idents: Critical Essays p. 1945407. Monahan, Brian and RJ Maratea. 2021. “The Art of the Spiel: Analyzing Donald Trump’s Tweets as Gonzo Storytelling.” Symbolic Interaction 44(4):699–727. Moore, Michael K and John R Hibbing. 1998. “Situational dissatisfaction in Congress: 93 explaining voluntary departures.” The Journal of Politics 60(4):1088–1107. Morales, Erendira Abigail, Cindy J Price Schultz and Kristen D Landreville. 2021. “The impact of 280 characters: an analysis of Trump’s tweets and television news through the lens of agenda building.” Electronic News 15(1-2):21–37. Murakami, Michael H. 2009. “Minority status, ideology, or opportunity: Explaining the greater retirement of House Republicans.” Legislative Studies Quarterly 34(2):219– 244. Nelson, Louis. 2017. “Poll: 70 percent say Trump acts unpresidential.” POLITICO. URL: https://www.politico.com/story/2017/07/17/poll-trump-unpresidential-240622 Nice, David. 1984. “Competitiveness in House and Senate Elections with Identical Constituencies.” Political Behavior 6(1):95–102. URL: http://www.jstor.org/stable/586049 Oborne, Peter and Tom Roberts. 2017. How Trump thinks: His tweets and the birth of a new political language. Head of Zeus Ltd. Oprysko, Caitlin. 2019. “Trump tells intel chiefs to ‘go back to school’ after they break with him.” Politico. URL: https://www.politico.com/story/2019/01/30/trump-national- security-1136433?cid=apn Opsahl, Robin. 2018. “Why all the speakers left, 1935-2018.” Roll Call. URL: https://rollcall.com/2018/04/11/why-all-the-speakers-left-1935-2018/ Ott, Brian L. 2017. “The age of Twitter: Donald J. Trump and the politics of debasement.” Critical studies in media communication 34(1):59–68. Ouyang, Yu, Richard W Waterman, Yu Ouyang and Richard W Waterman. 2020. Trump, Twitter, and the American democracy. Springer. Owens, John E. 2021. “The Trump Party: Republican in Name Only.” Political Insight 12(2):32–35. O’Keefe, Ed and Paul Kane. 2014. “Dave Camp to retire after his current term.” The Washington Post. URL: https://www.washingtonpost.com/news/post- politics/wp/2014/03/31/dave-camp-to-retire-after-his-current-term/ Palmeri, Tara. 2017. “How Trump’s campaign staffers tried to keep him off Twitter.” Politico. URL: https://www.politico.com/story/2017/02/trump-twitter-staffer-235263 Paschall, Collin, Tracy Sulkin and William Bernhard. 2019. “The Legislative Consequences of Congressional Scandals.” Political Research Quarterly p. 1065912919826224. Pedemonte, Mathieu. 2019. “Fireside Chats: Communication and Consumers’ Expectations in the Great Depression.” The Review of Economics and Statistics pp. 1–46. 94 Pérez-Curiel, Concha and Pilar Limón-Naharro. 2019. “Political influencers. A study of Donald Trump’s personal brand on Twitter and its impact on the media and users.” Communication & Society pp. 57–75. Peterson, Rolfe Daus. 2012. “To tweet or not to tweet: Exploring the determinants of early adoption of Twitter by House members in the 111th Congress.” The social science journal 49(4):430–438. Pfeiffer, Dan. 2018. Yes we (still) can: Politics in the age of Obama, Twitter, and Trump. Hachette UK. Phipps, E Brooke and Fielding Montgomery. 2022. ““Only YOU Can Prevent This Night- mare, America”: Nancy Pelosi As the Monstrous-Feminine in Donald Trump’s YouTube Attacks.” Women’s Studies in Communication 45(3):316–337. Polsby, Nelson W. 1968. “The institutionalization of the US House of Representatives.” American Political Science Review 62(1):144–168. Press, Associated. 2016. “Marco Rubio changes his mind about retirement from Senate.” New York Post. URL: https://nypost.com/2016/06/22/marco-rubio-ditches-retirement-plans- will-seek-re- election-to-senate/ Price, H Douglas. 1971. “The congressional career: Then and now.” Congressional behavior 27. Reed, Betsy. 2024. “Twitter usage in US ‘fallen by a fifth’ since Elon Musk’s takeover.” The Guardian. URL: https://www.theguardian.com/technology/2024/mar/26/twitter- usage-in-us-fallen-by-a-fifth-since-elon-musks-takeover Reynolds, Molly E. 2017. “Retirement from Congress may be driven by term limits on committee chairs.” Brookings Institution. Ring, Trudy. 2020. “Lindsey Graham’s stunning transformation from Trump critic to Lap Dog.” Advocate. URL: https://www.advocate.com/election/2020/5/22/lindsey- grahams-stunning-transformation-trump-critic-lap-dog Roosevelt, Franklin D. 1938. “"Fireside chat" on "purging" the Democratic Party.” Teaching American History. URL: https://teachingamericanhistory.org/document/fireside-chat- on-purging-the-democratic- party/?fbclid=IwAR0QjC9pomZKZsq9ERzNs85ulwPFpo8PuClqE7IaCB−yGGElxLWnZIfQ DA Roosevelt, Theodore and William Howard Taft. 1892. “The Progressive Era and the Nadir.”. Ross, Andrew S and Damian J Rivers. 2018. “Discursive deflection: Accusation of “fake news” 95 and the spread of mis-and disinformation in the tweets of President Trump.” Social Media+ Society 4(2):2056305118776010. Rothenberg, Lawrence S and Mitchell S Sanders. 2000. “Severing the electoral connection: Shirking in the contemporary Congress.” American Journal of Political Science pp. 316– 325. Rozsa, Matthew. 2021. “FDR once tried to purge disloyal Democrats — would it work for Biden to do the same?” Salon. URL: https://www.salon.com/2021/09/26/fdr-once- tried-to-purge-disloyal-democrats–would-it-work-for-biden-to-do-the-same/ Ruskin, Liz. 2022. “Trump rallies his Alaska faithful against Murkowski, for Tshibaka and Palin.” Alaska Public Media. URL: https://alaskapublic.org/2022/07/09/trump-rallies- alaska-faithful-against-murkowski-for-tshibaka-and-palin/ S¸ahin, Osman, Richard Johnson and Umut Korkut. 2021. “Policy-making by tweets: dis- cursive governance, populism, and Trump Presidency.” Contemporary Politics 27(5):591– 610. Schier, Steven E and Todd E Eberly. 2017. The Trump presidency: Outsider in the oval office. Rowman & Littlefield. Schreckinger, Ben. 2018. “‘Oh, no’: The day Trump learned to tweet.” Politico. URL: https://www.politico.eu/article/oh-no-the-day-us-president-donald-trump-learned-to- tweet-twitter-social-media/ Seligman, Lester G. 1956. “The Presidential Office and the President as Party Leader.” Law & Contemp. Probs. 21:724. Shapiro, Matthew A and Libby Hemphill. 2017. “Politicians and the policy agenda: Does use of Twitter by the US Congress direct New York Times content?” Policy & internet 9(1):109–132. Shea, Daniel M. 1999. “All scandal politics is local: Ethical lapses, the media, and congressional elections.” Harvard International Journal of Press/Politics 4(2):45–62. Shoaib, Alia. 2022. “Sen. Lindsey Graham said he liked former president Donald Trump because he feared him.” Business Insider. URL: https://www.businessinsider.com/lindsey-graham-liked-trump-afraid-of-him-2022-6 Shogan, Colleen J. 2006. “The contemporary presidency: The sixth year curse.” Presidential Studies Quarterly 36(1):89–101. Skinner, Richard M. 2008. “George W. Bush and the partisan presidency.” Pol. Sci. Q. 123:605. South Carolina Democratic Debate Transcript: February 25 Democratic Debate. 2020. Rev. 96 URL: https://www.rev.com/blog/transcripts/south-carolina-democratic-debate- transcript-february-democratic-debate Spitzer, Robert. 2012. The presidency and public policy: The four arenas of presidential power. University of Alabama Press. Stanley, Jeanie R and Diane D Blair. 1989. “Gender Differences in Legislative Effectiveness.” Statistics and Historical Comparison. 2020. GovTrack.us. URL: https://www.govtrack.us/congress/bills/statistics Stokols, Eli. 2024. “The Petty Feud Between the NYT and the White House.” Politico. URL:https://www.politico.com/news/magazine/2024/04/25/new-york-times-biden- white-house-00154219 Stone, Walter J, Sarah A Fulton, Cherie D Maestas and L Sandy Maisel. 2010. “Incumbency reconsidered: Prospects, strategic retirement, and incumbent quality in US House elections.” The Journal of Politics 72(1):178–190. Taylor, Andrew J and Michael D Cobb. 2019. “The Individual-Level Origins of Congressional Corruption Scandals.” American Politics Research p. 1532673X19850093. Taylor, Hannis. 1908. “The Speaker and His Powers.” The North American Review 188(635):495–503. Telfeyan, Brad. 2002. The 1938 Georgia Democratic Senatorial Primary: The Repudiation of President Franklin D. Roosevelt’s’ Purge Campaign’ PhD thesis Vanderbilt University. Dept. of History. Theriault, Sean M. 1998. “Moving up or moving out: career ceilings and congressional retirement.” Legislative Studies Quarterly pp. 419–433. Thomsen, Jacqueline. 2017. “Poll: 70 percent say Trump’s actions unpresidential.” The Hill. URL: https://thehill.com/homenews/administration/342299-poll-majority-believes- trump-acts-unpresidentially/ Tien, Charles. 2001. “Representation, voluntary retirement, and shirking in the last term.” Public Choice 106(1-2):117–130. Titcomb, James. 2024. “Biden seeks 85k‘ Meme Lord in Race for Reelection.′′ TheTelegraph. URL:https://www.telegraph.co.uk/business/2024/05/23/biden-seeks-85k-meme-lord- social-media-battle-trump Treisman, Rachel. 2019. “As President Trump Tweets And Deletes, The Historical Record Takes Shape.” NPR. URL:https://www.npr.org/2019/10/25/772325133/as-president- trump-tweets-and-deletes-the-historical-record-takes-shape Tucker, Patrick D, Michelle Torres, Betsy Sinclair and Steven S Smith. 2019. “Pathways to 97 trump: Republican voters in 2016.” Electoral Studies 61:102035. U.S. Senate: About the Committee System | Committee Assignments 2023. 2023. U.S. Senate: About the Committee System | Committee Assignments. URL: https://www.senate.gov/about/origins-foundations/committee-system/committee- assignments.htm Vance, Chris. 2021. “The Seventh Party System.” Niskanen Center. URL: https://www.niskanencenter.org/the-seventh-party-system/ Varis, Piia. 2020. “Trump tweets the truth: metric populism and media conspiracy.” Trabalhos em Linguística Aplicada 59:428–443. Volden, Craig and Alan E Wiseman. 2009. “Legislative effectiveness in Congress.” Manuscript, The Ohio State University . Volden, Craig and Alan E Wiseman. 2014. Legislative effectiveness in the United States congress: The lawmakers. Cambridge University Press. Volden, Craig and Alan E Wiseman. 2018. “Legislative Effectiveness in the United States Senate.” The Journal of Politics 80(2):731–735. Volden, Craig, Alan E Wiseman and Dana E Wittmer. 2010. “The legislative effectiveness of women in Congress.” Manuscrito, The Ohio State University. Volden, Craig, Alan E Wiseman and Dana E Wittmer. 2013. “When are women more effective lawmakers than men?” American Journal of Political Science 57(2):326–341. Wagner, John. 2017. “Fox News host calls on Ryan to step down, hours after Trump tweets about her show.” Washington Post. URL: https://www.washingtonpost.com/news/post-politics/wp/2017/03/25/trump-asks- his-supporters-to-watch-fox-news-host-who-calls-on-ryan-to-step-down/ Weigel, David. 2024. “The viral ‘bloodbath’ clip and the rise of the liberal video influencer.” Semafor. URL: https://www.semafor.com/article/03/19/2024/the-viral-bloodbath- clip-and-the-rise-of-the-liberal-video-influencer Wells, Chris, Dhavan Shah, Josephine Lukito, Ayellet Pelled, Jon CW Pevehouse and JungHwan Yang. 2020. “Trump, Twitter, and news media responsiveness: A media systems approach.” New media & society 22(4):659–682. White, David Richard. 2000. “The institutionalization of the United States Senate, 1789- 1996.”. Wignell, Peter, Sabine Tan, Kay L O’Halloran and Kevin Chai. 2021. “The Twittering Pres- idents: An analysis of tweets from@BarackObama and @realDonaldTrump.” Journal of Language and Politics 20(2):197–225. 98 Wolak, Jennifer. 2007. “Strategic retirements: The influence of public preferences on voluntary departures from Congress.” Legislative Studies Quarterly 32(2):285–308. Wolf, Thomas Phillip. 2019. The 1938 Purge: A Re-Examination. In Franklin D. Roosevelt and Congress. Routledge pp. 108–121. Worthy, Ben and Mark Bennister. 2017. “The Outsider Inside? The Leadership Capital of Donald J Trump.” The Leadership Capital of Donald J Trump (April 25, 2017). X. 2021. “Permanent suspension of @realDonaldTrump.” X Blog. URL: https://blog.twitter.com/enus/topics/company/2020/suspension Zelizer, Julian E. 2015. The fierce urgency of now: Lyndon Johnson, Congress, and the battle for the Great Society. Penguin Books. 99 APPENDIX A: CHAPTER 2 APPENDIX Table A2.1: Alternative Senate Models 100 Table A2.2: Alternative House Models 101 APPENDIX B: CHAPTER 3 APPENDIX Table B3.1: President Trump’s Attitudes Towards U.S. Senators (A-KI) 102 Table B3.2: President Trump’s Attitudes Towards U.S. Senators (Ky-Y) 103 Table B3.3: President Trump’s Attitudes Towards U.S. Representatives A-C 104 Table B3.4: President Trump’s Attitudes Towards U.S. Representatives D-Je 105 Table B3.5: President Trump’s Attitudes Towards U.S. Representatives Jo-N 106 Table B3.6: President Trump’s Attitudes Towards U.S. Representatives O-Steu 107 Table B3.7: President Trump’s Attitudes Towards U.S. Representatives Stev-Z 108 Table B3.8: Members Trump Never Mentioned On Twitter (A-C) 109 Table B3.9: Members Trump Never Mentioned On Twitter (D-He) 110 Table B3.10: Members Trump Never Mentioned On Twitter (Hi-L) 111 Table B3.11: Members Trump Never Mentioned On Twitter (M-Ros) 112 Table B3.12: Members Trump Never Mentioned On Twitter (Rou-Web) 113 Table B3.13: Members Trump Never Mentioned On Twitter (Wes-Y + Senate) 114