Abstract Though political candidates, observers, and voters often express concern about partisan meddling in supposedly neutral elections, existing research has not directly studied partisan bias among election administrators. We report results from a field experiment conducted in Wisconsin during the 2014 general election. Local election clerks were sent an information request from a putative constituent, randomizing the sender’s partisanship. Our findings are mixed. Overall, partisan email-writers were somewhat more likely to receive responses from local election clerks than email-writers who provided no partisan signal, though these effects are driven mostly by greater responsiveness to Republican constituents. We also find some evidence of increased responsiveness to requests from copartisan constituents, particularly among Republican municipalities. However, we find no evidence that local institutional context moderates the effects of the partisan treatments. Our findings provide new evidence about the presence of partisan biases from administrators in ostensibly neutral settings and raise important questions about the capacity for insulating election administration from partisan influences. Potential voters confront a myriad of informational deficits. In addition to deciding which candidate to support, they must also assemble information about how, where, and when to cast their votes. As the “administrators of democracy” (Moynihan and Silva 2008), local election officials (LEOs) are well-positioned to mitigate such informational deficits. Whether they do so without bias—that is, whether election officials work on behalf of ordinary voters without taking into account the attributes of those voters—is central to contemporary debates over election administration and integrity. Understanding potential biases among election administrators is critical because their decisions “may affect whether some people are able to cast a vote” (Kimball, Kropf, and Battles 2006, 448). Scholars, policymakers, and international organizations widely agree about the importance of neutral election administration for democratic health. As the European Commission for Democracy through Law (2008, 38) argues, “Only transparency, impartiality, and independence from politically motivated manipulation will ensure proper administration of the election process.” Similarly, the Office for Security and Co-operation in Europe (2013, 29) advises member countries that “[n]o election-administration body should act in a partisan manner or exhibit partiality in the performance of its duties.” The nature of electoral administration has important implications for the success of democratic transitions (Pastor 1999) and perceptions of legitimacy (Anderson et al. 2005; Kropf and Kimball 2013). A report by the US Commission on Federal Election Reform (2005, 49) therefore concluded that “[t]o build confidence in the electoral process, it is important that elections be administered in a neutral and professional manner.” Election administrators’ success in doing so may also shape their professional reputations which help to secure bureaucratic autonomy from political principals (Carpenter and Krause 2012). Although it may be desirable for election officials to be immune to political biases and interested only in tallying the vote, history suggests that elections may not always be characterized by the neutral administration. A rich set of examples, including the post-2000 US presidential election battle in Florida and more recent controversies in Gabon, Kenya, and Nicaragua, suggests that partisan meddling in supposedly neutral elections is not uncommon. Recent literature documents tremendous heterogeneity in how election laws and procedures are implemented by local administrators and poll workers (Alvarez and Hall 2006; Atkeson 2010) and suggests that the administration of local elections tends to favor an official’s preferred political party (Bassi, Morton, and Trounstine 2008; Burden et al. 2013; Kimball, Kropf, and Battles 2006; Kropf, Vercellotti, and Kimball 2013; Stuart 2004). Similar discretion on the part of election officials can also lead to race-based bias in election administration (Atkeson et al. 2010; Cobb, Greiner, and Quinn 2010; White, Nathan, and Faller 2015). In this article, we report results from an audit study designed to learn about the neutrality of election administrators. More specifically, we study the relationship between LEOs’ responsiveness to requests for information about an upcoming election and the putative partisanship of voters. Though recent scholarship reports results from a number of audit studies conducted with public officials (for a review of this literature, see Costa 2017); few if any of these studies focus on identifying potential partisan biases. In doing so, we contribute to a conversation about partisan influence in election administration that has been long ongoing among academics, practitioners, and policymakers. Like most street-level bureaucrats, election administrators have wide discretion over the distribution of scarce resources, including their own time and energy, which may result in biased patterns of responsiveness (Alvarez and Hall 2006; Burden et al. 2012; Lipsky 1980). Suspicion of partisan bias in election administration (e.g., Burden et al. 2013; Kimball, Kropf, and Battles 2006; Kropf, Vercellotti, and Kimball 2013) has led voices from across the political spectrum to call for nonpartisan reforms.1 Building on the theory of administrative burden (Burden et al. 2012; Herd et al. 2013), we argue that changes to election laws and procedures can impose administration burden on local officials, who can then exercise discretion in determining whether to help mitigate learning costs for constituents or instead pass those burdens along to them. Building on a design used to study biases among street-level bureaucrats (Einstein and Glick 2017; Jilke, Van Dooren, and Rys Forthcoming; White, Nathan, and Faller 2015), we conducted an email-based experiment with LEOs in Wisconsin during the week before the 2014 midterm elections. Do election officials respond differently to requests for information based on voters’ putative partisan affiliations? Are they equally responsive to both co- and counter-partisans? And how does the potential for partisan bias vary across local partisan and institutional contexts? In 2014, Wisconsin offered an ideal laboratory to help answer these questions. We find some evidence of differential responsiveness based on the partisanship of the putative constituent. Overall, requests from constituents that were accompanied by a partisan cue received somewhat greater rates of response, and these patterns were driven primarily by greater responsiveness to requests from Republican constituents. We find some conditional evidence of partisan bias, with greater responsiveness to Republican email-writers among clerks serving Republican municipalities. However, we find no similar response patterns to requests from Democratic email-writers among clerks in Democratic municipalities. Our results are robust across a wide range of additional analyses, and we find no evidence that the effects of the partisan treatment were significantly conditioned by the mode of clerk selection, municipality size or administrative resources, or county partisan context. The findings suggest that partisanship can shape the behavior of local administrators in even ostensibly neutral contexts and raise new questions about how administrative institutions can be designed to mitigate against potential partisan influences. Partisanship and Election Administration Street-level bureaucrats such as LEOs have wide discretion over how they interpret and implement formal policies. According to Lipsky (1980, xii), “the decisions of street-level bureaucrats, the routines they establish, and the devices they invent to cope with uncertainties and work pressures effectively become the public policies they carry out.” Scholarship on representative bureaucracy advocates for public administrators such as street-level bureaucrats to hold attitudes and values that reflect the constituencies they serve and argues that these shared values will lead street-level bureaucrats to be responsive to the public (e.g., Kropf, Vercellotti, and Kimball 2013; Meier 1993). However, the discretion provided to street-level bureaucrats raises the possibility that they act upon their own biases absent mechanisms to mitigate potential discrimination (Lipsky 1980). In this citizen-state interaction, to extend the theory of administrative burden (e.g., Herd et al. 2013; Moynihan, Herd, and Harvey 2014), policy changes—or the potential for policy changes—may lead election administrators to subsidize the costs of learning the consequences of these changes for their copartisan constituents but pass down higher learning costs to constituents who identify with the other party. Introducing such biases would run counter to the goals of Progressives and other reformers who sought to infuse public administration with neutral competence. Voting irregularities in recent federal elections and a host of state and local election controversies have placed new attention on how partisanship may affect election administration. Partisanship shapes a wide range of political attitudes and behavior (e.g., Campbell 1960; Green, Palmquist, and Schickler 2002) and election administrators may not be immune to its influence. As the report of the Commission on Federal Election Reform (2005, 49) emphasized: Elections are contests for power and, as such, it is natural that politics will influence every part of the contest, including the administration of elections. In recent years, some partisan election officials have played roles that have weakened public confidence in the electoral process. Many other partisan election officials have tried to execute their responsibilities in a neutral manner, but the fact that they are partisan sometimes raises suspicions that they might favor their own party. Indeed, as Kimball, Kropf, and Battles (2006, 480), summarize, proponents of nonpartisan election administration argue that “partisan election officials may make administrative decisions intended to benefit their political party, while nonpartisan officials will administer elections in a more independent and neutral fashion.” Empirical scholarship has found that partisanship has influenced a range of behaviors among federal bureaucrats, including corruption prosecutions (Gordon 2009), IRS audits (Mete 2002), and the enforcement of federal election law (James and Lawson 1999). Scholars have also found that partisan administration of elections may contribute to differences in turnout rates (Burden et al. 2013) and uses of provisional ballots (Kimball, Kropf, and Battles 2006). These findings suggest that partisan election officials administer elections to benefit their party. As one observer put it, “Partisanship is thus a spectre haunting the making of election laws, as well as their implementation.”2 Though more than a half-century of scholarship debates the virtues of an independent public administration rather than one politicized by principals or parties (Heclo 1977; Kaufman 1956), existing research has not directly studied the possibility of partisan bias among local government officials or among election officials in particular. As partisanship is perhaps the most salient political identity in U.S. politics, partisan bias in local election administration could have tangible consequences for both voter turnout and election outcomes. Evidence of partisan bias could also be used to suggest administrative changes intended to mitigate its influence. How Partisanship May Affect Administrative Responsiveness We study how partisanship affects responsiveness to information requests among LEOs.3 In this context, responsiveness amounts to the provision of information that helps constituents cast votes. As we will explain in greater detail, responsiveness by LEOs in our empirical case could have been especially useful to help potential voters understand the requirements for voting in an ambiguous and possibly confusing information environment. We posit that local officials’ incentives and administrative constraints shape whether and through what means partisanship affects responsiveness. Their incentives are structured largely by the nature of the officials’ principal, the principal’s goals, and the degree to which officials can be held accountable for their performance. Partisan principals could create incentives for LEOs to perform the duties of their office in a way that advantages their political party, perhaps by prioritizing information requests from copartisans and ignoring requests from counter-partisans. Alternatively, principals could place value on maximizing voter turnout by promoting effective and competent management of the office, in which case partisan bias may not be expected. These incentives must be balanced against local officials’ administrative burdens and capacities (Burden et al. 2012). In particular, given time and resource constraints, officials may not be able to respond to every request of them but instead, prioritize them according to the potential efficacy of an official’s response or their wish to minimize the possibility of a complaint from an unsatisfied constituent. Based on the nature of their incentives, partisan signals may affect responsiveness from LEOs in several ways. First, partisanship may serve as a cue by which officials infer other characteristics of the sender. For instance, if a local official were to encounter a known partisan constituent, the official may infer the constituent is especially interested in elections and voting, as partisans tend to be more active in electoral politics. The local official may thus be more likely to respond to a request from this individual because there is a greater chance that the official’s response will affect the individual’s behavior. Similarly, the official’s response may depend on whether they know the constituent is a Democrat or a Republican. If the official believes Republicans were more likely than Democrats to vote in 2014 (as they generally are in midterm elections), then the official may be more likely to respond to an information request from a Republican constituent because of the greater chance the response from the official would affect the constituent’s behavior. Another possibility is that a partisan signal—whether Democratic or Republican—could serve as an indicator that a local constituent is a high demander and would be more likely to complain to other authorities if the query were not answered. Each of these patterns would be consistent with statistical discrimination (e.g., Arrow 1973), wherein clerks would infer characteristics about the constituent that structure their responsiveness. Moreover, across each of these potential scenarios, clerks use statistical discrimination to contend with the administrative burden (Burden et al. 2012) and prioritize requests according to the inferences they can make about the sender based on the limited information that is provided. Second, a constituent’s partisanship communicates to local officials that the constituent has a political bias, which may or may not be aligned with the clerk’s own partisan views or the views of the community they serve. Consistent with the theory of representative bureaucracy (Kropf, Vercellotti, and Kimball 2013) local officials may employ taste-based discrimination (Becker 1957) to actively prioritize responding to copartisan constituents while ignoring counter-partisans. This form of taste-based discrimination could be based on local officials’ personal preferences for serving the requests of copartisan constituents. This pattern could also manifest due to electoral incentives. To the extent that constituents rely on information requested from LEOs, lower responsiveness to counter-partisan email-writers may enable LEOs to reduce turnout among residents who identify with the party opposite their own. Election officials could be motivated to produce this result either because of their personally-held partisan views or because they perceive career incentives to do so. Empirically, taste-based discrimination could manifest in increased [decreased] responsiveness to Democratic [Republican] constituents among Democratic [Republican] LEOs or LEOs who serve Democratic [Republican] constituencies. Third, clerks could exhibit different overall patterns of responsiveness based on their own partisan views or the partisanship of their constituencies. As Burden et al. (2012) and Kropf, Vercellotti, and Kimball (2013) argue, Democratic and Republican election officials (and officials who serve Democratic and Republican constituencies) may administer elections differently because they prioritize different values. Democratic officials, for instance, may place greater value on expanding ballot access while Republican officials may be more concerned with ensuring election integrity. These different emphases may generate increased rates of responsiveness from Democratic officials or officials who serve Democratic constituencies even in the absence of the clerk’s motivation to directly benefit his or her own party. Put in different terms, it is also possible that a clerk’s failure to respond to a request for information about voting could be an intentional attempt to reduce voting rates. Our account suggests that responsiveness from local government officials—in our case, election administrators—reflects the combination of administrative constraints, professional incentives, and personal political commitments. It is also possible, however, that the method of LEO selection may condition responsiveness to partisan cues. In particular, elected officials may be less responsive to partisanship than appointed LEOs. Elected officials must receive the approval of local voters, while appointed LEOs are chosen by ostensibly partisan municipal boards. Burden et al. (2013) provide suggestive evidence of such a relationship and document greater support for voter access among elected LEOs compared with LEOs who were appointed. On the other hand, to the degree local partisan constituencies desire election administration in a way that advantages their political party, elections may induce greater partisan responsiveness among LEOs. Thus, we compare clerks’ responsiveness to partisan requests for information among appointed LEOs and elected LEOs. Empirical Research on Partisanship And Public Officials Our focus on partisanship contributes to a growing literature that studies responsiveness among legislators (e.g., Butler and Broockman 2011) and public administrators (e.g., Einstein and Glick 2017; Grohs, Adam, and Knill 2016; White, Nathan, and Faller 2015). These studies generally employ email-based field experiments in which information is requested from public officials, where the key manipulation concerns some attribute of the email-writer. For instance, studies of racial bias have varied the name of the email-writer to signal the writer’s putative racial group (Butler and Broockman 2011) and typically find decreased responsiveness to constituents from different racial groups than the official. Related research has varied the religious affiliation (Distelhorst and Hou 2014) or ethnic group (Grohs, Adam, and Knill 2016) of letter-writers to examine how those characteristics affect responsiveness among local officials. Despite the centrality of partisanship for organizing government and political decision making, most research in this area does not directly investigate partisan bias in government responsiveness. Costa (2017) reports results from a meta-analysis of all recent research using audit studies to evaluate responsiveness from political officials. While several of the referenced studies focus on policy disagreement between the constituent contact and the political official, few studies have directly manipulated the partisanship of the contacting individual and none in the context of election administration. Two recent studies, both conducted outside the United States, do manipulate partisan alignment with the contacted official but find divergent results. In a study of Swedish politicians, Öhberg and Naurin (2016) report decreased responsiveness to constituents when they express disagreement on an issue with the politician’s political party, while Chen, Pan, and Xu (2016) find no difference in responsiveness among local government officials in China based on the sender’s membership in the Communist Party. The best evidence to date about the potential for partisan effects in election administration comes from several important observational studies. First, scholars have documented that election administrators sometimes evince attitudes toward the policies they are responsible for administering in ways that differ among partisan and ideological lines (e.g., Burden et al. 2012; Kimball et al. 2013; Kropf, Vercellotti, and Kimball 2013). To the extent, these attitudes are associated with clerks’ approaches to administering elections, as Kropf, Vercellotti, and Kimball (2013) show in the context of provisional voting, suggests the possibility that clerks respond differently to Republican and Democratic constituents. Second, existing research has found evidence consistent with the proposition that clerk partisanship conditions the association between constituency partisanship and election outcomes. For instance, Kimball and Kropf (2006) show that more provisional votes in the 2004 presidential election were cast and counted in jurisdictions where the election administrator’s partisanship matched the constituency’s partisanship; similarly, Burden et al. (2013) find that Wisconsin voter turnout among jurisdictions served by Republican election officials in the 2008 presidential election was lower in more heavily Democratic areas, but finds no evidence that turnout figures were associated with constituency partisanship in jurisdictions served by Democratic election administrators. We build upon this body of research and study how partisanship affects responsiveness to individual constituents using an experimental design to maximize internal validity. Study Context and Design We examined the effect of partisanship on responsiveness from LEOs by conducting an email-based audit study in Wisconsin in 2014. For three reasons, Wisconsin offered an opportune environment to study how partisanship affects local election administration. First, elections are more decentralized in Wisconsin than in virtually any other state. Wisconsin is one of eight states (the others being Michigan and the six New England states) to vest local responsibility for elections with municipalities. Each of the state’s 1,853 municipal clerks has primary responsibility for conducting local elections. The autonomy enjoyed by local administrators in such a decentralized system provides ample opportunity for individual administrators to make decisions based on their own discretion. The officially nonpartisan nature of Wisconsin election administration is a second important feature of our case. From 2008 to 2016, state elections were overseen by the General Accountability Board (GAB), which was comprised of six former state judges and served as an independent regulatory agency on elections, campaign finance, and ethics. The GAB was the only nonpartisan state entity to oversee elections in the United States and was “a worthy model for other states considering alternatives to partisan election administration” (Tokaji 2013, 577).4 Local election administration is performed by municipal clerks chosen through officially nonpartisan means.5 The GAB “consider[ed] each of these clerks to be a partner in the process of carrying out open, fair and transparent elections.”6 Third, there was considerable controversy and ambiguity in the state about voting requirements. A new voter identification law passed in 2011 generated widespread attention from the courts and voters. The law was struck down in April 2014 by the US District Court for the Eastern District of Wisconsin,7 but this decision was overturned by the Seventh Circuit Court of Appeals in September 2014.8 Less than a month before the election, the US Supreme Court issued a stay on October 9, which blocked its implementation.9 In addition, following the 2012 election, Governor Scott Walker, with the support of leading state legislators, called for an end to same-day registration, which the state had offered since 1976.10 Although this effort failed, more than 33 voting laws were implemented or changed during the Walker’s first term as governor and could have contributed to real uncertainty about voting procedures.11 The slew of recent court decisions and policy proposals about voting requirements in the state meant that voters were likely to be uncertain about voting procedures, enhancing the plausibility of our audit approach. Do nonpartisan election administrators maintain their partisan neutrality in communicating with constituents when enjoying local autonomy and in the presence of political and legal controversy over voting requirements? On the one hand, it would be not altogether unsurprising if the answer is yes, given the insulation of election administration from partisan politics. After all, the GAB is politically independent, and LEOs are selected through nonpartisan means, suggesting that LEOs have no partisan principals with whom to curry favor. On the other hand, though the means of selection may be officially nonpartisan, neutral offices alone may not remove the potential for partisan considerations to affect officer decision-making or the possibility that constituents attribute partisan motivations to putatively nonpartisan officials. Moreover, allegations of partisan behavior by clerks in Wisconsin are not infrequent. For instance, in 2016, Dane County clerk Scott McDonnell was challenged by Karen McKim, who alleged that McDonnell was “very skilled at politics” and “one of the most partisan clerks in the state.”12 If they wished, election administrators could act upon their own partisan preferences and attempt to help their preferred party. Thus, evidence of partisan bias from the Wisconsin context might suggest the potential for such biases to be even larger in contexts where election administrators are more overtly politicized. Wisconsin municipal clerks administer elections in their city, town, or village. In addition to recruiting poll workers and making decisions about polling locations and other facets of administration, municipal clerks themselves frequently engage directly with the public, particularly with questions about registration and voting requirements.13 Figure 1 displays the number of municipal clerks in each county. The mean number of LEOs per county is 25 and ranges from one (Menominee County in northeast Wisconsin) to 57 (Dane County in south-central Wisconsin). Figure 1. View largeDownload slide Number of Municipal Clerks Per County Figure 1. View largeDownload slide Number of Municipal Clerks Per County In the week before the 2014 election, we sent emails to municipal clerks asking for information about how to vote. The emails came from a putative resident named Michael Schmidt.14 Clerk contact information, including email addresses and phone numbers, is readily available and regularly updated on the state Government Accountability Board’s website at http://www.gab.wi.gov/clerks/directory.15 Before proceeding, we note that our intervention sought to avoid substantially contributing to clerks’ experiences of administrative burden (e.g., Burden et al. 2012). As we describe below, our emails were relatively short, and our queries could be answered with simple responses. While the timing of our survey corresponded with a period in which clerks were actively engaged with election activities, the potential confusion introduced by the conflicting court decisions in the month before the election,16 as our account above described, raises important practical questions about how local administrators implement policies under such conditions. Municipal clerks are our units of observation. Several clerks served multiple localities, however; to avoid potential SUTVA violations, we randomly selected one of the municipalities they served to remain in our data and dropped the other municipalities.17 We also dropped observations with bad email addresses (either the emails were returned, or we were informed that the email was addressed to an incorrect recipient)18 and omitted municipalities for whom no email addresses were provided. Our results are also robust to omitting municipalities whose borders crossed county lines.19 Our final sample consisted of 1,750 municipal clerks.20 Following practices in similar research (Butler and Broockman 2011; Jilke, Van Dooren, and Rys Forthcoming; White, Nathan, and Faller 2015; Worthy, John, and Vannoni 2017), we conducted our study by sending email to each clerk in our sample which requested information about voter identification and registration. The key manipulation concerned the presence of a partisan signal in the text of the email. Using simple random assignment, clerks were randomized with equal probability to one of three treatment conditions in which the email-writer signaled being either a Republican or a Democrat, or made no mention of either party. The text of the email read as follows: Hello, I am writing for some information about voting in the November election. I have two questions. I am not registered to vote. Can I register to vote on Election Day? I wanted to vote in the [Republican/Democratic/(neither)] primaries earlier this year but wasn’t sure I could because I wasn’t registered.21 I have heard that voters need to show ID to vote. Will I need to show an ID to vote in the election this year? What kinds of ID will you accept? Thank you very much for your help. Sincerely, Michael Schmidt22 Our queries sought to minimize the time required to respond. Based on Wisconsin laws at the time of the midterm election, the correct responses were that yes, voters could register on Election Day and that no identification was required to vote.23 Moreover, randomization provided excellent balance across treatment groups on a range of observable municipal and county characteristics.24 The clerks in the study displayed close engagement with the request, and many appeared exceedingly helpful. For instance, several clerks asked the email sender for his address so they could look up his polling location while others advised the sender to make sure he was a resident of the Town of X rather than the Village of X because those municipalities had different polling places. Other clerks wrote back having apparently found a resident in their municipality with the same name as the email-writer and confirmed the sender’s eligibility to vote in their community. In line with our theoretical account, we conducted two primary sets of analyses to evaluate the effect of partisanship on LEO responsiveness. First, we compared the overall response rates to our emails based on whether the municipal clerk received an email that contained a partisan signal. We created indicators Republican signal and Democratic signal and compared response rates for each to the response rates for messages that did not include a partisan signal. Greater responsiveness to signals from one party over another could provide evidence of systematic bias against the party whose signal generated the lower response rate. We also investigated whether partisan signals—whether Democratic or Republican—received greater response rates than email messages that did not contain a partisan signal, which could signal greater responsiveness to constituents deemed to be high-demanders or who might be more likely to utilize the information provided by the clerk.25 Second, we evaluated whether responsiveness to messages with partisan signals is conditioned by the municipal clerks’ partisan context. In particular, we studied whether clerks were more responsive to queries with Republican [Democratic] signals in increasingly Republican [Democratic] municipalities. We used the municipal-level Republican vote share in the 2012 presidential election as a measure of local partisan context.26 Larger values of this measure characterize more Republican communities; in these communities, there may also be an increased probability that the local clerk identifies as a Republican. We evaluated whether responsiveness to partisan signals depends on the local partisan context by interacting our indicators for assignment to the Republican signal and Democratic signal treatment conditions with the local Republican vote share. If local clerks privilege requests from constituents who share their or their community’s partisanship while ignoring requests from putative constituents who identify with the other party, we would expect a positive coefficient for the interaction between Republican signal and Republican vote share and a negative coefficient for the interaction between Democratic signal and Republican vote share.27 In additional analyses, we estimate model specifications that include a variety of controls for local contextual factors and also explore whether these factors moderate our estimated treatment effects. While we note that random assignment in our experimental design rules out the possibility of potential confounders, we will account for municipal population size, per capital general government expenditures (which serves as a measure of local municipal capacity), and whether municipal clerks were chosen through nonpartisan elections or were appointed by local municipal boards, as the means of selection defines the relevant principal and structures the incentives for the clerk’s behavior.28 The vast majority of municipalities are quite small, with approximately two-thirds serving 1,000 or fewer registered voters. On the one hand, clerks serving smaller municipalities may be more responsive since they are serving fewer voters; on the other hand, clerks in these municipalities may take on more responsibilities, even though they have fewer resources, and thus may exhibit less responsiveness. We expect that per capita general government expenditures are positively associated with responsiveness, as clerks in jurisdictions with more administrative funding may have the greater administrative capacity. We suspect that elected clerks may be more responsive than appointed clerks, as the former group could be responding to a potential future voter. However, given the compactness of many of these communities, clerks may likely know many of their local constituents such that means of appointment is not a significant predictor of responsiveness. We also estimate models which account for several characteristics of the counties in which municipalities reside, including the county’s vote for Romney in the 2012 election, county population density, and county population. Perhaps most importantly, we include an indicator for counties with a Democratic county clerk. County clerks may have taken different approaches to share information among constituents about voting procedures, and as noted above Democratic clerks tend to prioritize voting access more than Republican clerks (Burden 2012). Thus, it is possible that municipal clerks could have taken their cues from the county clerks and exhibit greater responsiveness to information requests in counties with Democratic clerks. Results Overall, the municipal election clerks exhibited an extremely high level of responsiveness, particularly given that they likely were at their busiest when the audit study was conducted. We received responses from 1,400 of the 1,750 (80%) clerks in our sample. This figure is nearly identical to the 79.4% response rate elicited from LEOs in Wisconsin in a study conducted in September 2012 (White Nathan and Faller 2015). The high response rate provides some preliminary evidence, moreover, that LEOs were highly attuned to information requests from constituents in the lead-up to the 2014 election. We begin by examining whether partisan messages affected clerk responsiveness. We modeled clerk responsiveness with logistic regression and included an indicator for whether the clerk received a message that included a partisan signal. Standard errors are clustered on the county to account for potential interdependencies in clerk responsiveness within a county due to the local context and to the potential collaboration between municipal clerks and the county clerk.29 We estimate four models to study the relationship between partisanship and responsiveness. Column (1) shows results from a bivariate model in which responsiveness is regressed on an indicator for whether the email contained a partisan cue. In model (2), we account for local partisanship by including Republican vote share, centered so that a value of zero indicates a municipality that supported Barack Obama and Mitt Romney at equal rates in the 2012 presidential elections. In model (3), we include controls for a variety of local municipal characteristics that could influence responsiveness, including an indicator for whether the local clerk is Elected rather than appointed, 2014 municipal Population (logged),30 per capita General government expenditures (logged),31 and indicators for whether the municipality is a Town or Village where cities are the omitted category.32 In addition to these variables, model (4) contains a series of county-level controls, including an indicator for counties with an elected Democratic county clerk, the County Republican vote share in the 2012 presidential election (centered at zero), County population density (logged), and County population (logged). The results are shown in the first four columns of table 1. Across all four models, the coefficient estimate for Partisan message is consistently positive and ranges between 0.20 and 0.23, and reaches statistical significance at p < .09 with the inclusion of the control variables. The coefficient estimates suggest that a constituent who self-identified as a partisan was 3–4 percentage points more likely to receive a response than a constituent who did not provide a partisan cue. The results for the covariates are also of substantive interest. We find no significant differences in responsiveness based on municipal partisanship, whether the municipal clerk is elected or appointed, municipal expenditures on government administration, or county population. We do find, however, that municipal population and county population density is associated with significantly greater response rates, while the probability of response is greater in towns and villages than in cities.33 We also find that responsiveness is greater for municipalities in counties with a Democratic county clerk and that supported Romney at higher levels in 2012. Table 1. Effect of Partisan Messages on Election Clerk Response Rates (1) (2) (3) (4) (5) (6) (7) (8) Partisan message 0.20 (0.13) 0.20 (0.13) 0.22* (0.13) 0.23* (0.13) Republican message 0.30* (0.17) 0.29* (0.17) 0.32* (0.17) 0.32* (0.17) Democratic message 0.11 (0.14) 0.11 (0.14) 0.14 (0.15) 0.15 (0.14) Republican vote share 1.22 (0.75) 0.98* (0.54) −0.19 (0.80) 1.20 (0.75) 0.96* (0.53) −0.22 (0.82) Elected clerk −0.14 (0.17) −0.14 (0.17) −0.15 (0.16) −0.15 (0.17) Population 1.63** (0.26) 1.27** (0.30) 1.62** (0.26) 1.26** (0.30) General gov’t expenditures 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) Town 0.71** (0.28) 0.61** (0.30) 0.71** (0.28) 0.62** (0.30) Village 0.73** (0.27) 0.45 (0.29) 0.73** (0.27) 0.45 (0.29) Democratic county clerk 0.27** (0.13) 0.27** (0.13) County Republican vote share 1.74 (1.15) 1.75 (1.17) County density 0.62* (0.33) 0.61* (0.33) County population 0.03 (0.15) 0.03 (0.14) Constant 1.26** (0.12) 1.24** (0.13) 0.80** (0.24) −0.28 (1.81) 1.26** (0.12) 1.24** (0.13) 0.81** (0.24) 0.29 (1.82) N (clerks) 1,750 1,750 1,750 1,750 1,750 1,750 1,750 1,750 N (counties) 72 72 72 72 72 72 72 72 (1) (2) (3) (4) (5) (6) (7) (8) Partisan message 0.20 (0.13) 0.20 (0.13) 0.22* (0.13) 0.23* (0.13) Republican message 0.30* (0.17) 0.29* (0.17) 0.32* (0.17) 0.32* (0.17) Democratic message 0.11 (0.14) 0.11 (0.14) 0.14 (0.15) 0.15 (0.14) Republican vote share 1.22 (0.75) 0.98* (0.54) −0.19 (0.80) 1.20 (0.75) 0.96* (0.53) −0.22 (0.82) Elected clerk −0.14 (0.17) −0.14 (0.17) −0.15 (0.16) −0.15 (0.17) Population 1.63** (0.26) 1.27** (0.30) 1.62** (0.26) 1.26** (0.30) General gov’t expenditures 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) Town 0.71** (0.28) 0.61** (0.30) 0.71** (0.28) 0.62** (0.30) Village 0.73** (0.27) 0.45 (0.29) 0.73** (0.27) 0.45 (0.29) Democratic county clerk 0.27** (0.13) 0.27** (0.13) County Republican vote share 1.74 (1.15) 1.75 (1.17) County density 0.62* (0.33) 0.61* (0.33) County population 0.03 (0.15) 0.03 (0.14) Constant 1.26** (0.12) 1.24** (0.13) 0.80** (0.24) −0.28 (1.81) 1.26** (0.12) 1.24** (0.13) 0.81** (0.24) 0.29 (1.82) N (clerks) 1,750 1,750 1,750 1,750 1,750 1,750 1,750 1,750 N (counties) 72 72 72 72 72 72 72 72 Note: Table entries are logistic regression coefficients with standard errors clustered on county in parentheses. The outcome variable is an indicator for whether clerk in municipality i provided a response. **p < .05; *p < .10 (two-tailed tests). View Large Table 1. Effect of Partisan Messages on Election Clerk Response Rates (1) (2) (3) (4) (5) (6) (7) (8) Partisan message 0.20 (0.13) 0.20 (0.13) 0.22* (0.13) 0.23* (0.13) Republican message 0.30* (0.17) 0.29* (0.17) 0.32* (0.17) 0.32* (0.17) Democratic message 0.11 (0.14) 0.11 (0.14) 0.14 (0.15) 0.15 (0.14) Republican vote share 1.22 (0.75) 0.98* (0.54) −0.19 (0.80) 1.20 (0.75) 0.96* (0.53) −0.22 (0.82) Elected clerk −0.14 (0.17) −0.14 (0.17) −0.15 (0.16) −0.15 (0.17) Population 1.63** (0.26) 1.27** (0.30) 1.62** (0.26) 1.26** (0.30) General gov’t expenditures 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) Town 0.71** (0.28) 0.61** (0.30) 0.71** (0.28) 0.62** (0.30) Village 0.73** (0.27) 0.45 (0.29) 0.73** (0.27) 0.45 (0.29) Democratic county clerk 0.27** (0.13) 0.27** (0.13) County Republican vote share 1.74 (1.15) 1.75 (1.17) County density 0.62* (0.33) 0.61* (0.33) County population 0.03 (0.15) 0.03 (0.14) Constant 1.26** (0.12) 1.24** (0.13) 0.80** (0.24) −0.28 (1.81) 1.26** (0.12) 1.24** (0.13) 0.81** (0.24) 0.29 (1.82) N (clerks) 1,750 1,750 1,750 1,750 1,750 1,750 1,750 1,750 N (counties) 72 72 72 72 72 72 72 72 (1) (2) (3) (4) (5) (6) (7) (8) Partisan message 0.20 (0.13) 0.20 (0.13) 0.22* (0.13) 0.23* (0.13) Republican message 0.30* (0.17) 0.29* (0.17) 0.32* (0.17) 0.32* (0.17) Democratic message 0.11 (0.14) 0.11 (0.14) 0.14 (0.15) 0.15 (0.14) Republican vote share 1.22 (0.75) 0.98* (0.54) −0.19 (0.80) 1.20 (0.75) 0.96* (0.53) −0.22 (0.82) Elected clerk −0.14 (0.17) −0.14 (0.17) −0.15 (0.16) −0.15 (0.17) Population 1.63** (0.26) 1.27** (0.30) 1.62** (0.26) 1.26** (0.30) General gov’t expenditures 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) 0.27 (0.17) Town 0.71** (0.28) 0.61** (0.30) 0.71** (0.28) 0.62** (0.30) Village 0.73** (0.27) 0.45 (0.29) 0.73** (0.27) 0.45 (0.29) Democratic county clerk 0.27** (0.13) 0.27** (0.13) County Republican vote share 1.74 (1.15) 1.75 (1.17) County density 0.62* (0.33) 0.61* (0.33) County population 0.03 (0.15) 0.03 (0.14) Constant 1.26** (0.12) 1.24** (0.13) 0.80** (0.24) −0.28 (1.81) 1.26** (0.12) 1.24** (0.13) 0.81** (0.24) 0.29 (1.82) N (clerks) 1,750 1,750 1,750 1,750 1,750 1,750 1,750 1,750 N (counties) 72 72 72 72 72 72 72 72 Note: Table entries are logistic regression coefficients with standard errors clustered on county in parentheses. The outcome variable is an indicator for whether clerk in municipality i provided a response. **p < .05; *p < .10 (two-tailed tests). View Large Because our indicator of Partisan message does not distinguish messages with Republican signals from Democratic signals, we estimated the same four models described above but with indicators to distinguish clerks who received a Republican message from those who received a Democratic message. Messages sent without a partisan signal are the omitted category. The results are shown in columns 5–8 of table 1. The models provide the strongest evidence of bias in the context of responsiveness to requests from Republican constituents. The coefficients for Republican message range from 0.29 to 0.32 and each is statistically significant at p < .10. The coefficients for Democratic message are considerably smaller in magnitude, ranging between 0.11 and 0.15, but are estimated less precisely and with standard errors similar in magnitude to the coefficient estimate. Using the coefficients shown in column (8), the results imply that a Republican constituent was about 4.5 percentage points more likely to receive a response than a constituent who did not provide a partisan cue, roughly double the magnitude of the increased responsiveness to a Democratic message relative to a nonpartisan request (2.2 percentage points). These findings suggest that the increased responsiveness to partisan messages relative to messages that did not contain a partisan cue was due largely to preferential treatment to Republican requests. These results could suggest some systemic level of partisan favoritism. Yet we must be careful in interpreting this finding. First, we cannot statistically distinguish the magnitude of the treatment effect for Republicans from that for Democrats, and the substantive magnitudes are relatively modest.34 Second, we cannot say for sure whether clerks were more responsive to the Republican signal because they wished to advance the Republican Party, or because they inferred that putative Republican voters were more likely to be more politically interested or “high demanders,” as we discussed above. Clerks may have been working to serve their own or their community’s preferred party, or they may simply have been responding to putative constituents for whom they believed a response would be more efficacious or might react more negatively if they did not receive a response. The results for the other municipal and county covariates parallel the findings shown in the first four columns of table 1. In addition, our findings are robust to alternative model specifications, as supplementary table A.6 shows nearly identical patterns of results when estimating models with county fixed effects instead of county covariates and multilevel models with varying county-level coefficients. We now investigate whether responsiveness to partisan signals is conditioned by the local partisan context. We focus on examining whether clerks were more responsive to queries with Republican [Democratic] signals in increasingly Republican [Democratic] municipalities. To do so, we estimated model (4) while also interacting the partisan signal provided by the email-writer with Republican vote share, recalling that the latter measure is centered at zero. If LEOs respond to information requests in a manner consistent with taste-based discrimination, we would expect a positive coefficient for the interaction between Republican vote share and the indicator for Republican message and a negative coefficient for its interaction with Democratic message. We note that our investigation of conditional effects is necessarily more tentative because the proposed moderating variables were not randomly assigned.35 The results are shown below in table 2.36 Because the measure of Republican vote share is centered at zero, the coefficient estimates for Republican message and Democratic message indicate the effect of partisan signals in municipalities that were evenly split in their support for Romney and Obama in 2012. While both coefficients are positive, neither is statistically significant. More importantly, the table reveals an asymmetry in how local partisan context affected responsiveness to an email containing partisan signals. The interaction between Republican message and Republican vote share is positive, large in magnitude, and statistically significant, and indicates that the probability of receiving a response to the message with the Republican signal was greater in more Republican municipalities. The interaction between Democratic message and Republican vote share is also positive (though not statistically significant), suggesting that response rates to Democratic messages were higher in more Republican municipalities, in the direction opposite from what we might expect if clerks in Democratic communities were exhibiting bias against Republican email-senders. To the extent that clerks were more responsive to information requests from copartisans than to requests from other email-writers, the data in table 2 show that this relationship is found among Republicans but not Democrats.37 Table 2. Partisan Context and Responsiveness among Local Election Officials (1) Republican message 0.29 (0.18) Democratic message 0.13 (0.15) Republican vote share −1.44 (1.05) Republican message × Republican vote share 2.75** (1.26) Democratic message × Republican vote share 1.02 (1.24) Constant −0.29 (1.79) Municipal controls ✓ County controls ✓ N (clerks) 1,750 N (counties) 72 (1) Republican message 0.29 (0.18) Democratic message 0.13 (0.15) Republican vote share −1.44 (1.05) Republican message × Republican vote share 2.75** (1.26) Democratic message × Republican vote share 1.02 (1.24) Constant −0.29 (1.79) Municipal controls ✓ County controls ✓ N (clerks) 1,750 N (counties) 72 Note: Table entries are logistic regression coefficients with standard errors clustered on county in parentheses. The outcome variable is an indicator for whether clerk in municipality i provided a response. Municipal controls include Population, General government expenditures, Town, and Village. County controls include Democratic county clerk, County Republican vote, County density, and County population. **p < .05; *p < .10 (two-tailed tests). View Large Table 2. Partisan Context and Responsiveness among Local Election Officials (1) Republican message 0.29 (0.18) Democratic message 0.13 (0.15) Republican vote share −1.44 (1.05) Republican message × Republican vote share 2.75** (1.26) Democratic message × Republican vote share 1.02 (1.24) Constant −0.29 (1.79) Municipal controls ✓ County controls ✓ N (clerks) 1,750 N (counties) 72 (1) Republican message 0.29 (0.18) Democratic message 0.13 (0.15) Republican vote share −1.44 (1.05) Republican message × Republican vote share 2.75** (1.26) Democratic message × Republican vote share 1.02 (1.24) Constant −0.29 (1.79) Municipal controls ✓ County controls ✓ N (clerks) 1,750 N (counties) 72 Note: Table entries are logistic regression coefficients with standard errors clustered on county in parentheses. The outcome variable is an indicator for whether clerk in municipality i provided a response. Municipal controls include Population, General government expenditures, Town, and Village. County controls include Democratic county clerk, County Republican vote, County density, and County population. **p < .05; *p < .10 (two-tailed tests). View Large Figure 2 graphically displays the results from table 2. Each plot shows the marginal effect of partisan messages, calculated as the difference in the predicted probability of a response for the two messages noted at the top of the plots. The predicted probabilities were calculated while holding the other continuous variables at their mean values and the indicators variables at their modal values. The marginal effects are shown across the range of values of Republican vote share. The shaded areas represent the 90% confidence intervals, and the horizontal dashed lines indicate no differences in responsiveness. The vertical lines across the top and bottom of the plots show the distribution of the Romney vote for the group of clerks indicated by the text. Figure 2. View largeDownload slide Responsiveness to Partisan Information Requests. Note: Plots show the marginal effects between the two messages indicated at the top of each plot. Marginal effects are calculated as the difference in predicted probabilities estimated from the model shown in table 2. Shaded regions are the 90% confidence intervals, and the dashed horizontal lines indicate the null hypothesis of no difference between the predicted probabilities. The vertical lines across the top and bottom of each plot show the distribution of the Romney vote across the two treatment groups being compared. Figure 2. View largeDownload slide Responsiveness to Partisan Information Requests. Note: Plots show the marginal effects between the two messages indicated at the top of each plot. Marginal effects are calculated as the difference in predicted probabilities estimated from the model shown in table 2. Shaded regions are the 90% confidence intervals, and the dashed horizontal lines indicate the null hypothesis of no difference between the predicted probabilities. The vertical lines across the top and bottom of each plot show the distribution of the Romney vote across the two treatment groups being compared. The plot on the left shows the difference in the predicted probability of response between messages with a Republican signal and messages that contained no partisan signal. Negative [positive] values indicate that Republican messages were less [more] likely to receive a response compared with messages that did not contain a partisan signal. In municipalities that provided less than 40% support for Romney, the plot shows that responsiveness was lower to the Republican signal than to messages that contained no partisan signal, though these differences are also not statistically distinguishable from zero. However, for municipalities that provided more than 50% support for Romney, messages with Republican signals were significantly more likely to receive responses than messages that did not contain partisan signals. Moreover, the positive slope indicates that the magnitude of this difference increases with the degree of local support for Romney. In more substantive terms, the predicted probability (holding all other variables at their mean or modal values) of a response to the Republican message relative to the nonpartisan message increased by about 8 percentage points as the municipality’s support for Romney increased from 40% to 60% (which corresponds to a nearly two standard deviation increase in the values of Romney vote share). The plot in the middle conducts a similar comparison for response rates to Republican versus Democratic signals. The slope is positive, suggesting that response rates were higher to Democratic messages in Democratic areas and higher to Republican messages in Republican areas. However, at no point along the x-axis are these differences statistically significant. The slope is also considerably less steep than the one shown in the left plot. Finally, the plot on the right compares responsiveness to Democratic signals and messages with no partisan signal across the range of values of Republican vote share. The slope is again positive, opposite from what we would expect if clerks in Democratic areas are disproportionately responsive to information requests from putatively Democratic constituents. However, the confidence intervals overlap the dashed line at zero for all values along the x-axis, again indicating that none of the marginal effects are statistically significant. Our data provide some support for the hypothesis that LEOs prioritize information requests from respondents who share the clerk’s partisanship or the partisanship of the community, but the evidence is limited to Republican messages that were received by LEOs in Republican municipalities. We uncover no evidence that clerks in Democratic municipalities are more responsive to requests from Democratic email-writers than they are to Republican email-writers or emails that contain no partisan signals. It is possible that these results reflect the greater propensity of Republicans to turn out in midterm elections and which was illustrated in Republican candidates’ success in the 2014 election, including Governor Scott Walker’s third successful state-wide victory in 4 years. As with any such analysis, we also cannot rule out the possibility that our analyses of conditional responsiveness to partisan cues are subject to unobserved confounding, as the partisan composition of local municipalities was not randomly assigned, and Democratic areas are likely to differ from Republican areas on dimensions beyond partisanship. Some features of the substance of the responses are also worth noting. As discussed, before this election Wisconsin had been the site of intense partisan controversy about what forms of identification were required to vote. A law was passed; a court overturned the law, but a higher court reversed that decision. For this reason, some inaccurate responses may be understandable, for which the whiplash pace of policy change, rather than outright nefariousness, may have been to blame. Yet among all our responding clerks, none wrote responses that could be characterized as misleading. While some clerks were more verbose than others, and some took longer to respond than others, we observed no clerk who attempted to outright mislead the putative voter. Therefore, our results can be interpreted as providing some evidence of LEOs’ disproportionate efforts to encourage voting (or help make it easier to vote) among copartisan constituents. We do not wish to overstate the substantive implications of our findings given the potential asymmetries we found in how LEOs responded to Democratic and Republican email-writers; however, our results do suggest that partisanship may affect not only whether LEOs respond to requests for information but the degree to which they encourage or facilitate voter turnout among their constituencies. Local Institutions, Context and the Possibility of Interactive Effects As our theoretical discussion above indicated, local clerk selection mechanisms have important implications for defining a clerk’s relevant principal and structuring their behavioral incentives. Accordingly, we evaluated whether partisan messages sent to elected clerks elicit different responses compared with the response patterns of appointed clerks. To do so, we estimated three supplementary models. In the first, we interacted our indicator for clerks that are Elected with our indicators for Republican message and Democratic message to account for the possibility that elected clerks respond differently to partisan messages. In the second model, we interacted Elected with Republican vote share to evaluate whether local partisan context is differently associated with responsiveness depending on whether a clerk is elected or appointed. Third, we examined whether the relationship between local partisan context and responsiveness to partisan messages varies across municipalities in which clerks are elected or appointed. In general, none of these three models provides systematic evidence that elected clerks responded any differently to the partisan messages compared with appointed clerks.38 Although LEOs may indeed have different principals based on whether they are appointed by municipal boards or elected by their peers, our data indicate they respond in similar ways to potential partisan incentives.39 We also evaluated the possibility that the effects of partisan messages could depend on other local factors. In particular, we considered whether the size of local municipalities and municipal expenditures affected responsiveness to partisan messages. Larger municipalities could indicate local clerks with more significant workloads and change how clerks prioritize responding to partisan messages. On the other hand, there may be greater specialization in larger offices while clerks in small offices are responsible for a wider range of election-related tasks. Clerks in smaller offices may thus experience greater administrative burden and exhibit lower responsiveness to our treatments. It is also possible that clerks serving smaller jurisdictions may have had easier access to tax and other records such that they could have more easily verified the email-writer’s place of residence, or would have been especially concerned about the possibility of a complaint from a counter-partisan. Similarly, municipalities with larger per capita budgets for general government expenditures could have greater administrative resources for clerks and could also affect the ways clerks respond to partisan messages. We also considered the possibility that responsiveness to partisan messages varied across cities, towns, and villages, each of which has its own administrative structure and local government culture. Across all these supplementary analyses, however, we again uncover no evidence that population size, administrative budgets, or the nature of administrative units affected the relationship between partisan messages and response rates from local clerks.40 Finally, we explored two features of Wisconsin counties that could condition how local clerks respond to partisan messages. We note, however, that there is no formal relationship between county administrative structures and local clerk behavior, although in practice municipal clerks often work together with county clerks on matters related to election administration. First, we estimated a model that included interactions between Republican message and Democratic message and whether the county-level clerk was a Democrat. We find a positive and significant interaction between Democratic message and county Democratic clerks, indicating that responsiveness was greater to the Democratic message among municipal clerks working in counties with an elected Democratic clerk. However, the interaction between Republican message and Democratic county clerk is also positive (though not statistically significant), and thus provides little evidence of systematic bias against counter-partisans on the basis of the partisanship of the county clerk. We also interacted Republican message and Democratic message with the county’s Republican vote share for Romney. Here we find no evidence that constituent partisanship in the county is associated with different response patterns to partisan messages.41 On the whole, therefore, our analysis provides some evidence that contextual factors affect how local government officials respond to partisan constituents. Clerks in increasingly Republican jurisdictions are more responsive to information requests from copartisans, though our data provide more evidence of similar patterns among clerks in Democratic jurisdictions. We do not find, however, that these patterns are strongly conditioned by the means through which local clerks are chosen, nor do other features of local institutions moderate the effects of partisan messages. Conclusion Using a study design well-established in recent research on racial bias among bureaucrats and legislators, we evaluate the presence of partisan bias among election administrators. Our results uncover some evidence of partisan bias, with putative Republicans more likely to receive a response than other constituents, particularly in Republican areas. To our knowledge, our study is the first to document how partisanship may shape election officials’ responses to individual constituents, particularly in the context of providing information about how to vote. That local election administrators evince some degree of partisan bias is not especially surprising given the importance of partisanship in the contemporary United States. At the same time, we do not wish to overstate our findings. Its effects were relatively limited, and its substantive magnitude was relatively small in comparison the hyper-partisan environment that has characterized US politics generally and has spilled over to election-related policies in particular. It may be more surprising that partisan bias was not wider in both scope and scale. In Wisconsin, as in many other states around the country, the battle over voter identification requirements was intensely partisan. Yet clerks did not respond to our seemingly innocuous requests for information in a manner commensurate to that battle. The number of elections and elected officials combine to make election administration in the United States a daunting task. Our evidence leads us to think that the task is made less daunting by the general probity of election officials, similar to what we observed. Civic virtue, which prizes fealty to the rule of law (Walzer 1974), may be more prevalent among election officials than might be assumed. We also found some evidence that election officials were more responsive to information requests from partisans than they were to nonpartisans. These findings raise the possibility that local officials prioritized requests from putative “high demanders” who may be more likely to complain if their request were not answered. Yet we must be cautious in interpreting these results, as our audit design does not allow us to distinguish greater responsiveness to supposed high-demanders from greater responsiveness to one’s preferred party. Local officials may want to be responsive to their copartisans while also not risking the ire of their counter-partisans. Further studies could investigate the extent to which this kind of behavior shapes bureaucratic responsiveness in nonelectoral contexts and in cases outside election administration. While our experimental design allowed us to maximize internal validity within a fixed institutional setting, our ability to extrapolate from our results is limited by our laboratory of choice. It is possible that, in another state, and in another year, Democratic letter-writers might have yielded higher response levels than Republicans, for example, or that nonpartisans would be treated more favorably than their partisan counterparts. One possible explanation for the greater responsiveness to Republican email-writers in our study is that state government had been under unified party control for the preceding several years, motivating clerks to respond at greater rates to these putative constituents due to the ample number of other government officials to whom Republican constituents could complain. It is also possible that both the effect size and direction could differ in a state where party affiliation is more formally integrated into election administration. We speculate, however, that the bias against non-partisans we observe might be magnified in such an environment. We also acknowledge that the limited evidence we uncover of partisan bias could be due to ceiling effects from the high overall response rate. Even if Wisconsin LEOs were prone to exhibit stronger partisan biases, local civic culture and norms could prevail upon them to respond to requests from constituents of all partisan stripes. Thus, it is possible that a similar experiment in other contexts where the norms of responsiveness are less strong would reveal more substantial evidence of partisan bias. The results from our study provide some support for arguments presented by political observers and academics in favor of nonpartisan election administration. Though election administrators may indeed have intense partisan views of their own, the election administration system in place in Wisconsin—no stranger to polarizing partisan politics—appears to help insulate the conduct of elections from the passions that animate electoral campaigns. At the same time, even nonpartisan LEOs appear not to be completely immune to the influence of partisanship, whether based on their own personal preferences or those of the constituency they serve. While there may not necessarily be partisan ways of fighting crime, constructing highways, protecting public health or assigning patents, virtually every decision made by election administrators has potential partisan implications, and the politicization of election administration could contribute to a less effective and more contested administrative apparatus. As current events would have it, Wisconsin lawmakers may allow future researchers to directly investigate how politicization affects local election administration. In 2015, Governor Scott Walker signed Wisconsin Act 118, which eliminated the Government Accountability Board and replaced it with separate Elections and Ethics Commissions. One of the most important changes that accompanied this restructuring is a provision that allows the governor and state legislative leaders to make appointments to the commissions. In so doing, the state disbanded its nonpartisan approach to election administration by politicizing administrative appointments, and thus changed the institutional context in which local election administrators do their work. It is unclear how this development might affect the behavior of LEOs, who remain responsible to their municipal constituencies. However, this policy change opens new opportunities for researchers to study how partisanship affects behavior from LEOs in a more politicized setting. Scholars should remain attentive to the possibility that the politicization of election administration may result in election administrators imposing differential learning and compliance costs, and thereby adjusting administrative burden, based on voters’ perceived partisanship (Moynihan, Herd, and Harvey 2014). Such behavior would corroborate claims that the disbursement of administrative burden can have large social consequences (Heinrich 2016). Partisan bias among public administrators and bureaucrats—street-level and otherwise—deserves further scrutiny. To our knowledge, the present study is the first to test whether partisanship affects responsiveness from street-level bureaucrats in an electoral context. Though partisan bias may be especially troubling in election administration, in our study, LEOs were not the only source of voting information for email-writers. Examining potential biases in other settings, such as when a bureaucrat determines whether an application for government services or a grant is approved, is an important goal for future research. We conclude by commenting on the method we used to investigate clerk responsiveness. Our use of an audit study helps maximize both internal validity, through random assignment, and external validity, by studying clerk behavior as they might typically interact with constituents. While these attributes are desirable from a design perspective, our intervention also consumed valuable time from local election administrators at a moment when demands on their attention were especially high. While we attempted to balance these costs against the potential benefits of our approach, we acknowledge the public service provided by LEOs and encourage researchers to evaluate these tradeoffs carefully. Supplementary Material Supplementary data are available at Journal of Public Administration Research and Theory online. We thank Barry Burden, Dan Butler, Jonathan Homola, Will Howell, Betsy Sinclair, Michelle Torres, Patrick Tucker, Ariel White, three anonymous reviews, and the Editor for helpful comments. We gratefully acknowledge Barry Burden for sharing some of the data used in this project and Stephanie Langella and Enrique Rodriguez for providing excellent research assistance. Footnotes 1 See, for example, Linda Killian, 2014, “The Case for Nonpartisan Election Overseers,” The Wall Street Journal, May 16, 2014; http://blogs.wsj.com/washwire/2014/05/16/the-case-for-nonpartisan-election-overseers/ (accessed April 4, 2016). 2 Daniel P. Tokaji, October 1, 2010, “The Persistence of Partisan Election Administration,” The American Constitution Society; available at https://www.acslaw.org/acsblog/the-persistence-of-partisan-election-administration (accessed July 12, 2016). 3 Note that our use of responsiveness contrasts with other areas of research that evaluate how government policies respond to constituent preferences. 4 The GAB was replaced by separate Elections and Ethics Commissions starting in 2016, whose members were appointed by the governor and state legislative leaders. 5 Approximately 60% of the municipal clerks are elected through nonpartisan elections held in the spring of odd-numbered years while the others are appointed by local governing boards. 6 See http://www.gab.wi.gov/clerks (accessed July 9, 2016). 7 Monica Davey and Steve Yaccino, “Federal Judge Strikes Down Wisconsin Law Requiring Photo ID at Polls”, April 29, 2014, The New York Times; available at http://www.nytimes.com/2014/04/30/us/federal-judge-strikes-down-wisconsin-law-requiring-photo-id-at-polls.html (accessed April 4, 2016). 8 Monica Davey, “Federal Appeals Court Permits Wisconsin Voter ID Law”, September 12, 2014, The New York Times; available at http://www.nytimes.com/2014/09/13/us/voter-id-law-in-wisconsin-is-permitted-by-us-court.html (accessed April 4, 2016). 9 Adam Liptak, “Courts Strike Down Voter ID Laws in Wisconsin and Texas”, October 9, 2014, The New York Times; available at http://www.nytimes.com/2014/10/10/us/politics/supreme-court-blocks-wisconsin-voter-id-law.html?_r=0 (accessed April 4, 2016). 10 Jason Stein, “Walker calls for changes to same-day voter registration rules,” November 19, 2012, Milwaukee Journal-Sentinel; available at http://www.jsonline.com/news/statepolitics/walker-calls-for-changes-to-sameday-voter-registration-rules-hk7n9e8-180010171.html (accessed April 4, 2016). 11 See, for example, Katelyn Ferral, “Timeline: Changes Scott Walker has made to voting and elections in Wisconsin since 2011”, The Capitol Times, December 31, 2015; available at http://host.madison.com/ct/news/local/govt-and-politics/timeline-changes-scott-walker-has-made-to-voting-and-elections/article_8a716d48-af12-11e5-af9a-f762b14ea70e.html (accessed July 6, 2016). 12 See http://www.hngnews.com/deforest_times/news/government/article_9513ab40-3ed1 (accessed October 4, 2016). 13 Local news stories document appeals from local clerks for voters to contact them directly for voting-related questions. See, for example, http://www.wiscnews.com/portagedailyregister/news/local/article_0ddda1d9-6e08-56b4-87bb-4a1b58b69d4d.html (accessed October 4, 2016); and http://lacrossetribune.com/news/local/la-crosse-city-clerk-urges-voter-registration/article_60bb6945-c496-549e-aef0-aa87de7ced31.html (accessed October 4, 2016). 14 Both are common names. The surname Schmidt is found more frequently in Wisconsin than in any other state, as more than 9 percent of all US residents with the Schmidt surname live in Wisconsin. See http://mypage.siu.edu/lhartman/wisnamestable.html (accessed April 7, 2016). 15 The vast majority of clerks’ email addresses appeared to be private accounts which were based on the clerks’ names, so we believe it is reasonable to assume that clerks themselves—rather than administrative assistants—were the recipients of our emails. 16 As an example of the state’s efforts to address voters’ potential confusion, the state GAB website for In-Person Early Voting in 2014, posted on October 20, 2014 (the week before our study), specifically pointed out that “Wisconsin’s voter photo ID law is not in effect for this election due to a recent U.S. Supreme Court ruling, so voters do not have to show photo ID to receive their ballots.” See http://elections.wi.gov/node/3422 (accessed April 4, 2016). 17 For instance, if a clerk served multiple jurisdictions and received an email for each, the clerk’s response to the email received first would likely affect her response to the additional email and thus create spillover effects. 18 We performed additional analyses based only on assignment to treatment. This more conservative estimation strategy does not change any of our inferences. 19 See supplementary tables A.1 and A.2. 20 Supplementary table A.3 shows that we find no evidence that observable characteristics at either the municipal or county level predict inclusion in our sample. 21 A sizable proportion of Wisconsin primary voters does not identify with either of the two major political parties. Though Wisconsin does not register voters by party, our calculations from the 2016 Wisconsin presidential primary exit poll indicate that 27% of primary voters were Independents, compared with 30% who identified as Republicans and 38% who identified as Democrats. (Unfortunately, exit poll data from the 2014 state primary elections do not exist.) These figures are nearly identical to the distribution of partisanship in the 2012 general election exit polls (37% D, 32% R, and 31% I) and suggest that the expression of interest in primary voting does not serve as an indication that an email-writer is a partisan over and above what one would expect based on the composition of the state electorate. 22 While the text of the email contains multiple questions, all clerks received the same questions so we are unable to distinguish whether response rates would have varied depending on the nature of the queries. To avoid order effects, we randomized whether emails mentioned registration or identification concerns first 23 Reflecting the rather minimal time required to respond, many of the clerks’ responses were quite short while others directed the email-writer to state-issued guidelines on the identification requirement such as http://www.gab.wi.gov/node/3417 (accessed November 4, 2014). 24 These comparisons are shown in supplementary table A.4. 25 Inthis design, nonresponse can be understood as the administrator imposing additional learning costs upon the constituent; if the administrator cannot answer the putative query we sent, presumably the putative constituent would have to go elsewhere for accurate information (Moynihan et al. 2014). 26 These data were obtained from http://www.gab.wi.gov/sites/default/files/CanvassResults_Presidential_by_Assembly_Senate_0.xls (accessed July 10, 2016). 27 Municipality vote share is likely to be an imperfect measure of clerk partisanship, particularly in more competitive jurisdictions. Two potential consequences follow. First, local Republican vote shares can be interpreted as a measure of the clerk’s constituency-induced partisanship. Alternatively, local Republican vote share could be interpreted as a noisy measure of clerk partisanship which could lead to downward bias in the estimates for the interaction terms. However, we have also performed our analysis using the measures of municipal clerk partisanship reported in Burden (2013); though we were able to obtain data on individual clerks’ self-reported partisanship due to confidentiality issues, we note that our substantive results are unchanged when aggregating municipal clerks’ partisanship on a 7-point scale and interacting it with our treatment indicators. See supplementary table A.5. The county-level correlation between average clerk partisanship and Romney vote share is reasonably strong (r = 0.43), suggesting these measures are capturing similar quantities. Finally, we note that we attempted to secure information about LEOs’ partisanship using their donation patterns to political candidates from state contribution records. However, our search revealed contribution records for only about 6% of the clerks in our sample, limiting the usefulness of these data as a proxy for clerk partisanship. 28 Data on the means of appointment come from Burden (2013). We note that the data on selection means are current as of 2009, but we cannot definitively rule out the possibility that it may have changed in a small number of places. However, local referenda are required to change this provision and it is somewhat unlikely that many such referenda would have occurred between 2009 and 2014. 29 This accounts for the possibility that county clerks could adopt specific policies regarding communication with constituents and thus lead to intracounty correlations in response patterns that vary across counties. While this strategy may be conservative, it does not meaningfully change any of our inferences. The statistical significance of our results does not change in more than minimal ways when estimating conventional standard errors. 30 These data were obtained from http://www.doa.state.wi.us/Documents/DIR/Demographic\%20Services\%20Center/Estimates/MCD_Time_Series_2015.xlsx (accessed July 11, 2016). 31 These data were obtained from https://www.revenue.wi.gov/slf/cotvc/cmreb14.xlsx (accessed July 10, 2016) and may serve as an indicator of local government capacity. 32 Each of these units is a general government administrative division. Administrative unit is not determined by population or area but instead reflects the form of government chosen by local residents and approved by the state legislature. Cities and villages are both incorporated areas and have home rule but have different governance structures. Towns are unincorporated and have less authority than cities and villages. The data contain 182 cities, 385 villages, and 1,183 towns. 33 While the positive relationship between responsiveness and population would seem to be at odds with the finding that responsiveness was higher in smaller jurisdictions (towns and villages) than in large (cities), we suspect this is due to the uneven distribution of different municipality types across counties. 34 However, the magnitudes are not dissimilar from the effect sizes reported for bias against email-writers with a Latino alias (White, Nathan, and Faller 2015). 35 In an ideal scenario, we would also randomly assign values of local clerk partisanship so that we could be confident that any interaction effects were due to partisanship per se rather than any potential observed or unobserved confounders that are correlated with partisanship. 36 We have also performed this analysis while including a quadratic expression of Republican vote share to capture the possibility that the treatment effects may be stronger in more competitive municipalities rather than in municipalities with heavy concentrations of Republicans or Democrats. The results are generally consistent with those described above: for municipalities where the Romney vote in 2012 was between 30% and 70%, where the vast majority of the observations fall, the Republican treatment effects continue to be positive and stronger than those for Democrats. 37 One possible explanation is that clerks in Republican areas inferred that Republican email-writers were more likely to be from their municipality than Democratic email-writers. Though we cannot rule out this possibility, the distribution of Romney support is relatively normally distributed with very thin tails and suggests Democrats and Republicans are found in roughly equal numbers in most municipalities. 38 These results are shown in supplementary table A.7. 39 The results from the third model merit some discussion. To the extent local Republican vote share is a proxy for a clerk’s personal partisanship, we would expect this proxy measure to be more accurate in places where clerks are elected rather than appointed. To the extent Republican vote share is a better measure of partisanship for elected clerks, we would expect the magnitude of the marginal effects to be larger in these places. The results are generally consistent with this intuition. Among appointed clerks, the predicted probability (holding all other variables at their mean or modal values) of a response to the Republican message relative to the nonpartisan message increased by about 5 percentage points as the municipality’s support for Romney increased from 40% to 60%. Among elected clerks, however, the predicted probability increased by around 11 percentage points. These differences are not themselves statistically distinguishable, and so we do not wish to overinterpret them. However, they do provide evidence to suggest that measurement error in our indicator for local clerk partisanship may be depressing the magnitude of our estimated results. 40 These results are shown in supplementary table A.8. 41 These results are shown in supplementary table A.9. References Alvarez, R. Michael, and Thad E. Hall. 2006. Controlling democracy: The principal-agent problems in election administration. Policy Studies Journal 34: 491– 510. Google Scholar CrossRef Search ADS Anderson, Christopher J., Andre Blais, Shaun Bowler, Todd Donovan, and Ola Listhaug. 2005. Losers’ consent: Elections and democratic legitimacy . New York: Oxford Univ. Press. Google Scholar CrossRef Search ADS Arrow, Kenneth J. 1973. The theory of discrimination. In Discrimination in labor markets , ed. Orley Ashenfelter and Albert Rees, 3– 33 Princeton: Princeton Univ. Press. Google Scholar CrossRef Search ADS Atkeson, Lonna Rae, Lisa Ann Bryant, Kyle Saunders, Thad E. Hall, and Michael Alvarez. 2010. A new barrier to participation: Heterogeneous application of voter identification policies. Electoral Studies 29: 66– 73. Google Scholar CrossRef Search ADS Bassi, Anna, Rebecca B. Morton, and Jessica Trounstine. 2008. Reaping political benefits: Local implementation of state and federal election law. Working paper. https://www.researchgate.net/profile/Anna_Bassi3/publication/253264433_Reaping_Political_Benefits_Local_Implementation_of_State_and_Federal_Election_Law/links/0c960522f7bad8ff9f000000/Reaping-Political-Benefits-Local-Implementation-of-State-and-Federal-Election-Law.pdf?origin=publication_detail ( accessed April 8, 2018). Becker, Gary S. 1957. The economics of discrimination . Chicago: University of Chicago Press. Burden, Barry C., David T. Canon, Stephane Lavertu, Kenneth R. Mayer, and Donald P. Moynihan. 2013. Selection method, partisanship, and the administration of elections. American Politics Research 41: 903– 36. Google Scholar CrossRef Search ADS Burden, Barry C., David T. Canon, Kenneth R. Mayer, and Donald P. Moynihan. 2012. The effect of administrative burden on bureaucratic perception of policies: Evidence from election administration. Public Administration Review 72: 741– 51. Google Scholar CrossRef Search ADS Butler, Daniel M., and David E. Broockman. 2011. Do politicians racially discriminate against constituents? A field experiment on state legislators. American Journal of Political Science 55: 463– 77. Google Scholar CrossRef Search ADS Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The American voter . New York: Wiley. Carpenter, Daniel P., and George A. Krause. 2012. “ Reputation and public administration.” Public Administration Review 72: 26– 32. Google Scholar CrossRef Search ADS Chen, Jidong, Jennifer Pan, and Yiqing Xu. 2016. Sources of authoritarian responsiveness: A field experiment in China. American Journal of Political Science 60: 383– 400. Google Scholar CrossRef Search ADS Cobb, Rachael V., D. James Greiner, and Kevin M. Quinn. 2010. Can voter ID laws be administered in a race-neutral manner? Evidence from the city of Boston in 2008. Quarterly Journal of Political Science 7: 1– 33. Google Scholar CrossRef Search ADS Commission on Federal Election Reform. 2005. Building confidence in U.S. elections. https://www.eac.gov/assets/1/6/Exhibit%20M.PDF (accessed May 16, 2018) Costa, Mia. 2017. How responsive are political elites? A meta-analysis of experiments on public officials. Journal of Experimental Political Science 4: 241– 54. Google Scholar CrossRef Search ADS Distelhorst, Greg, and Yue Hou. 2014. Ingroup bias in official behavior: A national field experiment in China. Quarterly Journal of Political Science 9: 203– 30. Google Scholar CrossRef Search ADS Einstein, Katherine Levine, and David M. Glick. 2017. Does race affect access to government services? An experiment exploring street-level bureaucrats and access to public housing. American Journal of Political Science 61( 1): 100– 116. Google Scholar CrossRef Search ADS European Commission for Democracy through Law. 2003. Code of good practice in electoral matters . Strasburg: Council of Europe Publishing. Gordon, Sanford C. 2009. Assessing partisan bias in federal public corruption prosecutions. American Political Science Review 103: 534– 54. Google Scholar CrossRef Search ADS Green, Donald, Bradley Palmquist, and Eric Schickler. 2002. Partisan hearts and minds . New Haven, CT: Yale Univ. Press. Grohs, Stephan, Christian Adam, and Christoph Knill. 2016. Are some citizens more equal than others? Evidence from a field experiment. Public Administration Review 76: 155– 64. Google Scholar CrossRef Search ADS Heclo, Hugh. 1977. A Government of strangers: Executive politics in Washington . Washington, DC: Brookings Institution Press. Heinrich, Carolyn J. 2016. The bite of administrative burden: A theoretical and empirical investigation. Journal of Public Administration Research and Theory 26: 403– 20. Google Scholar CrossRef Search ADS Herd, Pamela, Thomas DeLeire, Hope Harvey, and Donald P. Moynihan. 2013. Shifting administrative burden to the state: The case of Medicaid take-up. Public Administration Review 73: 69– 83. Google Scholar CrossRef Search ADS James, Scott C., and Brian L. Lawson. 1999. The political economy of voting rights enforcement in America’s gilded age: Electoral college competition, partisan commitment, and the federal election law. American Political Science Review 93: 115– 31. Google Scholar CrossRef Search ADS Jilke, Sebastian, Wouter Van Dooren, and Sabine Rys. Forthcoming. Discrimination and administrative burden in public service markets: Does a public-private difference exist? Journal of Public Administration Research and Theory . Kaufman, Herbert A. 1956. Emerging conflicts in the doctrines of public administration. American Political Science Review 50: 1057– 73. Google Scholar CrossRef Search ADS Kimball, David C., and Martha Kropf. 2006. The street-level bureaucrats of elections: Selection methods for local election officials. Review of Policy Research 23: 1257– 68. Google Scholar CrossRef Search ADS Kimball, David C., Martha Kropf, and Lindsay Battles. 2006. Helping America vote? Election administration, partisanship, and provisional voting in the 2004 election. Election Law Journal 5: 447– 61. Google Scholar CrossRef Search ADS Kimball, David C., Martha Kropf, Donald Moynihan, Carol L. Silva, and Brady Baybeck. 2013. The policy views of partisan election officials. University of California–Irvine Law Review 3: 551– 74. Kropf, Martha, and David C. Kimball. 2013. Helping America vote: The limits of election reform . New York: Routledge. Kropf, Martha, Timothy Vercellotti, and David C. Kimball. 2013. Representative bureaucracy and partisanship: The implementation of election law. Public Administration Review 73: 242– 52. Google Scholar CrossRef Search ADS Lipsky, Michael. 1980. Street-level bureaucracy: Dilemmas of the individual in public services . New York: Russell Sage Foundation. Meier, Kenneth J. 1993. Latinos and representative bureaucracy: Testing the Thompson and Henderson hypotheses. Journal of Public Administration Research and Theory 3: 393– 414. Mete, Mihriye. 2002. Bureaucratic behavior in strategic environments: Politicians, taxpayers, and the IRS. Journal of Politics 64: 384– 407. Google Scholar CrossRef Search ADS Moynihan, Donald, Pamela Herd, and Hope Harvey. 2014. Administrative burden: Learning, psychological, and compliance costs in citizen–state interactions. Journal of Public Administration Research and Theory 25: 43– 69. Google Scholar CrossRef Search ADS Moynihan, Donald P., and Carol L. Silva. 2008. The administrators of democracy: A research note on local election officials. Public Administration Review 68: 816– 27. Google Scholar CrossRef Search ADS Office for Security and Co-operation in Europe. 2013. Guidelines for reviewing a legal framework for elections . Warsaw: OSCE Office for Democratic Institutions and Human Rights. Öhberg, Patrik, and Elin Naurin. 2016. Party-constrained policy responsiveness: A survey experiment on politicians’ response to citizen-initiated contacts. British Journal of Political Science 46: 785– 97. Google Scholar CrossRef Search ADS Pastor, Robert A. 1999. The role of electoral administration in democratic transitions: Implications for policy and research. Democratization 6: 1– 27. Google Scholar CrossRef Search ADS Stuart, Guy. 2004. Databases, felons, and voting: Bias and partisanship of the Florida felons list in the 2000 elections. Political Science Quarterly 119: 453– 75. Google Scholar CrossRef Search ADS Tokaji, Daniel P. 2013. America’s top model: The Wisconsin government accountability board. UC Irvine Law Review 3: 575– 608. Walzer, Michael. 1974. Civility and civic virtue in contemporary America. Social Research 41: 593– 611. White, Ariel R., Noah L. Nathan, and Julie K. Faller. 2015. What do I need to vote? Bureaucratic discretion and discrimination by local election officials. American Political Science Review 109: 129– 42. Google Scholar CrossRef Search ADS Worthy, Ben, Peter John, and Matia Vannoni. 2017. Transparency at the parish pump: A field experiment to measure the effectiveness of freedom of information requests in England. Journal of Public Administration Research and Theory 27: 485– 500. © The Author(s) 2018. Published by Oxford University Press on behalf of the Public Management Research Association. All rights reserved. For permissions, please e-mail: email@example.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Journal of Public Administration Research and Theory – Oxford University Press
Published: May 24, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera