Control, accountability, and constraints: rethinking perceptions of presidential responsibility for the economy.
I think the world vests too much power--certainly in the president, probably in Washington in general--for its influence on the economy, because most all of the economy has nothing to do with the government.
--Austan Goolsbee, former Chairman of President Obama's Council of Economic Advisers
A specter has long haunted the fortunes of American presidents--the specter of the national economy. Research has repeatedly demonstrated that the economy matters for presidential approval and election outcomes (Duch and Stevenson 2008; Erikson, Mackuen, and Stimson 2002; Fiorina 1981; Gronke and Newman 2009; Lewis-Beck and Stegmaier 2000; Mueller 1970; Tufte 1980; Vavreck 2009). Indeed, there are essentially no limits on the extent to which presidents can be publicly credited or blamed for economic conditions by other elites, nor are there any rules constraining the public claims political elites can make regarding presidents' lone ability to effect desired economic outcomes. The state of the national economy, in other words, is more than just fair game for evaluating presidents--it is quite often the game.
When it comes to discussing the economy in an electoral context, for example, presidential candidates regularly engage in glib, grandiose claims as a means of improving their respective electoral chances (e.g., see Vavreck 2009). In the opening moments of the second presidential debate of 2012, for example, Republican presidential candidate Mitt Romney was asked by a college student for reassurance that he will be able to find employment upon graduating. After noting the poor state of the economy under President Barack Obama's tenure, candidate Romney boldly stated, "When you [graduate] in 2014--I presume I'm going to be president--I'm going to make sure you get a job." (1) Despite their boldness, Romney's claims did not provoke much controversy, likely because such claims possess an all-too-familiar character. That is, they lay blame for present economic conditions squarely upon the incumbent president while boasting of the challenger's superior ability to singlehandedly manage the economy. And, what is a crucial point, such rhetoric deliberately omits any mention of the myriad constraints modern presidents inevitably face in trying to steer the course of the national economy.
Indeed, in the world's largest market economy, with multiple sources of power determining domestic policy making, presidents regularly encounter formidable limitations on their ability to influence economic outcomes (Blinder and Watson 2014; Franzese 2002; Golden and Poterba 1980). We thus find ourselves in a puzzling situation where the national economy matters far more for the fortunes of presidents than may actually be warranted by reality (e.g., see Thomas E. Mann's comments in Gross 2008; Newman 2013). In addressing this conundrum, scholars have typically concluded that this phenomenon occurs because the president is viewed as the "nation's chief economic manager" (quoted in Newman 2013, 870), and citizens simply act on this perception when answering opinion polls or come Election Day (Lewis-Beck et al. 2008, 381-82). (2) But in accepting this assessment as an immutable fact of American political life, scholars have heretofore been somewhat reluctant to ask the more fundamental question: where might this widely shared (mis)perception of presidential responsibility for the national economy come from? (3)
Cognizant of the importance of presidential campaigns as a major source of political and economic information for the general public (Gelman and King 1993; Sears and Valentino 1997; Vavreck 2009), this study is among the first to directly explore whether the candidate claims we commonly observe--as well as claims we fail to observe--in presidential campaigns may be responsible for citizens' perceptions of the extent to which the presidency is responsible for national economic outcomes. Using a survey experiment designed to circumvent partisan, ideological, and topical considerations, this study expands upon important recent scholarship (see Newman 2013; Sirin and Villalobos 2011) by directly measuring (1) the extent to which the public believes that presidents control national economic outcomes and (2) the extent to which the public believes that presidents should be credited or blamed (i.e., held accountable) for national economic conditions. Paradoxically, I find that the public appears inclined to credit or blame presidents for the economy to a significantly greater extent than it perceives presidents to actually control the economy. Second, the findings consistently suggest that, while the selfserving claims of presidential candidates do not themselves appear capable of significantly influencing perceptions of presidential responsibility for the economy, the omission of information regarding constraints presidents face enables perceptions of presidential responsibility to remain significantly higher than would otherwise be the case. A discussion of the implications of these findings and avenues for future research concludes the paper.
The President and the Economy
There exists a compelling, yet complex relationship between politics and economic performance, and this relationship has a variety of important consequences for politicians. As Gomez and Wilson (2006, 129) observe, "That citizens hold elected officials accountable for fluctuations in the economy has become a truism of political life." This "truism" manifests itself in myriad ways, particularly for presidents. Forecasting models of presidential elections, for example, commonly include a measure of economic performance to predict election outcomes (e.g., Abramowitz 1988; Erikson and Wlezien 2012), and a wealth of research highlights the disproportionate influence of recent economic conditions on presidential election outcomes (Achen and Bartels 2004b; Bartels 2008; Campbell 2000, 127-31; Eisenberg and Ketcham 2004; Erikson and Wlezien 2008; Erikson, Mackuen, and Stimson 2002; Hibbs 1987, 182-83). According to theories of retrospective economic voting, voters electorally reward or punish an incumbent administration for economic performance and, in so doing, they are in some sense communicating whether the administration has "performed poorly or well" (Fiorina 1981, 5; Peffley 1984), though some evidence suggests that dissatisfaction with the economy may weigh more heavily than satisfaction (Bloom and Price 1975). (4) Similarly, research has consistently found that public approval of presidents is heavily dependent upon the current state of the national economy (Erikson, Mackuen, and Stimson 2002; Hibbs, Rivers, and Vasilatos 1982; MacKuen, Erikson, and Stimson 1992; Mueller 1970; Norpoth 1985). Moreover, presidential approval is itself predictive of election outcomes (Abramowitz 1988; Brody and Sigelman 1983; Lewis-Beck and Rice 1982; MacKuen 1983; Sigelman 1979).
Thus, the essential consensus is that the public holds presidents accountable for their performance in effecting desired economic outcomes, which suggests that a vast portion of the general public operates on the assumption that presidents have the capacity to significantly control the country's economic performance (Gomez and Wilson 2001). Indeed, one may easily observe the frequency with which public opinion polls, media reports and even scholars habitually refer to a given president's "management" or "handling" of the economy.
While there is therefore little doubt that the economy is consequential for presidents' electoral fortunes, the effect that presidents actually have on the economy is far less certain. In other words, to what extent do presidents "control" the economy? In fact, this is a terrifically complicated question, one that Bartels (2008, 42) has suggested would require "a small army of economists" to begin to answer it, whereas McGraw (1991, 1134) implies that establishing such a causal link is "generally impossible." Perhaps as a symptom of this inherent complexity, accounts of the relationship between the presidency and the economy commonly leave the distinction between (1) the extent to which a president alone can actually effect desired economic outcomes and (2) the extent to which citizens happen to believe a president can effect desired economic outcomes, conspicuously underexamined. (5)
Presidents, with their own ideological predilections and respective judgment in selecting economic advisors, are unlikely to be of zero consequence for economic outcomes. A vast body of research on "political business cycles" has suggested that presidents may attempt to strategically manipulate the economy during election years in the interest of electoral gain, though evidence for this is mixed (Achen and Bartels 2004b; Franzese 2002; Golden and Poterba 1980; Tufte 1980). Moreover, various studies have argued that, due in part to differential prioritization of economic problems--for example, unemployment versus inflation--Democratic and Republican administrations have presided over significantly different rates of economic growth since World War II, with the former proving better at reducing overall economic inequality (Bartels 2008; Hibbs 1987). However, in a comprehensive study of presidential economic performance since WWII, economists Blinder and Watson (2014) find no evidence for the claim that superior macroeconomic performance under Democratic presidents vis-a-vis Republican presidents can be attributed to better macroeconomic policies--instead, favorable international economic climates and "good luck," among other factors, appear to better account for the Democratic advantage.
Moreover, certain conditions must, of course, be present in order for a president to bring about desired economic outcomes. For example, without the support of Congress--the institution primarily responsible for crafting and legislating fiscal policy--a president's power to influence economic outcomes may be reduced to the mere "power to persuade" (Edwards 1989; Neustadt 1991; Rudolph 2003). (6) Also consequential for national economic outcomes, of course, are the policies of individual state and local governments (e.g., McNichol 2012), the dynamics of the global economy (Hibbs 1987, 6; Rodrik 2011), various monetary policy decisions independently undertaken by the Federal Reserve (De Boef and Kellstedt 2004; Harwood 2013; Weatherford 2013), (7) and a multitude of other factors given the relatively free-market nature of our national economic system (Taylor 2012, 597).
Thus, though it would be an exaggeration to argue that presidents have no ability to deliberately effect desired economic outcomes (i.e., no independent effect on the national economy), it is not altogether unreasonable to conclude that this ability is generally quite limited and, at the very least, poorly described by terms such as "control" or "management" of the national economy (e.g., see Greenblat 2013; Ryssdal 2013). This is not a new insight, of course. Such a position is consistent with this study's opening quote from former Obama advisor Austan Goolsbee, as well as the conclusions of Tufte (1980). In his seminal book, Political Control of the Economy, Tufte (1980, 138-42) reminds us that government's ability to influence the national economy "often operates at the margins," in part because "much economic activity transpires completely outside the scope of public control or even review." And, if the entire federal government faces constraints, the constraints faced by a president are likely considerably more onerous insofar as "Legislatures, legislative committees, bureaucracies, agencies, central banks, and all others involved in shaping economic outcomes all have quite different constituencies from the executive, and those constituencies may not be at all sympathetic to the economic program of the executive" (Tufte 1980, 138-42). With these formidable constraints in mind, it is puzzling, to say the least, that the condition of the national economy remains so consequential for presidential elections and approval.
Economic Attribution in the United States
Given the aforementioned evidence, it becomes clear that much of the public, in all likelihood, has an exaggerated sense of the president's independent ability to effect desired economic outcomes. Research has found that this misperception is particularly characteristic of those low in political sophistication insofar as these citizens are more likely to attribute responsibility for economic conditions largely to the president (Gomez and Wilson 2003). Thus, given the relatively low levels of political sophistication in the United States (e.g., Delli Carpini and Keeter 1996), it follows that the problem of assigning too much economic responsibility to presidents is likely quite ubiquitous (see also Jenkins-Smith, Silva, and Waterman 2005).
But what explains this exaggerated sense of presidents' responsibility for the economy? It may be tempting to first opine that citizenries always overestimate the ability of the country's leaders to manage the economy, and that little consideration is ever given to the real-world hindrances--personal, political, or economic--politicians inevitably face. Yet, we nevertheless observe meaningful variation, across institutional and economic contexts, in the extent to which different citizenries praise or punish incumbent leaders and governments for economic conditions. Powell and Whitten (1993), for example, find that, across 19 industrialized democracies, decreased "clarity of responsibility" (e.g., in situations where coalition governments exist) significantly attenuates the effect of economic conditions on incumbent vote share and that this occurs primarily among the politically sophisticated (see also Arceneaux 2006). Hellwig (2001) demonstrates that, across multiple democracies, a country's degree of exposure to the world economy significantly reduces the effect of domestic economic perceptions on electoral outcomes. Similarly, Kayser and Peress (2012) provide evidence for the notion that, across Organisation for Economic Co-operation and Development countries, voters exhibit a tendency to take into account the international economic context when deciding how to cast their votes.
It is not, therefore, a foregone conclusion that national leaders are universally regarded as omnipotent managers of the national economy. And yet, in the case of the United States specifically, research consistently demonstrates that voters are less sensitive to contextual factors that prevent presidents from moving the economy in a favorable direction. A case in point, instances of divided government are both readily apparent and likely to hinder a president from enacting his preferred economic program (or perhaps any program at all); and yet, numerous studies have found that the presence of divided government exerts no discernible tendency to disincline the U.S. electorate from engaging in economic voting or attributing responsibility for the economy to the president (Eisenberg and Ketcham 2004; Lewis-Beck et al. 2008; Nadeau and Lewis-Beck 2001; Norpoth 2001; Nicholson, Segura, and Woods 2002; cf. Rudolph 2003). Other factors that might affect the national economy, such as variation in state-level expenditures, global economic growth patterns, and Federal Reserve policy making, are therefore perhaps even less likely to impact "voting the economy" given their inherent complexity and relative obscurity from public view. Thus, with few exceptions (e.g., see D'Elia and Norpoth 2014), American voters appear to perceive an incumbent president as "the fundamental political leader of the economy, and act on that basis" (Lewis-Beck et al. 2008, 382).
What is not clear, though, is whether this tendency is purely a result of citizens not perceiving any constraints on presidential control over the economy and/or if citizens' beliefs about accountability might fundamentally differ from their perceptions of presidents' control over the economy. Certainly, a wealth of studies on presidential attribution imply that, in multiple policy-making areas, perceptions of accountability for presidential performance stem from perceptions of presidential control in these policy domains (Lewis-Beck and Paldam 2000; Peffley 1984; Sirin and Villalobos 2011); (8) however, the relationship between accountability and perceptions of control is unlikely to be perfectly correlated. Other factors, such as the desire to voice satisfaction or dissatisfaction with the status quo (via answering an opinion poll or voting; e.g., see Hibbing and Theiss-Morse 2002), may lead to citizens' normative beliefs about presidential accountability becoming decoupled from their beliefs about the control presidents actually wield over the national economy. (9) As Wattenberg (1992, 130) argues, "Regardless of whether their policies have an effect or not, presidents receive more of the credit and the blame for the country's immediate condition." Hamilton (1978) similarly argues for a distinction between causal responsibility and role responsibility (see also Peffley 1984). In other words, citizens may see a distinction between the empirical question of presidential control over the economy, and the normative question of whether presidents should be held accountable for the state of the national economy. As such, I propose to test the following hypothesis:
H1 Accountability over Control Hypothesis: Respondents will be inclined to hold presidents accountable for the national economy to a significantly greater extent than they perceive presidents to control the national economy.
Presidential Campaigns and Perceptions of Responsibility
The previous sections revealed that the American public appears all too willing to credit or blame the presidency for the economy, even if presidents are not fully deserving of such attributions. Campbell (2000, 137) puts this point succinctly when he acknowledges that, "Right or wrong, Americans hold the president responsible for the economy." But while much research seems to simply regard this latter point as an immutable fact of American political life, this study seeks to investigate why this inflated degree of responsibility for the economy is assigned to the presidency in the first place. Given the overwhelming importance of economic attribution for elections and presidential approval, it is surprising that more scholars have not asked the basic question: how do citizens come to believe that the president alone is largely responsible for national economic conditions?
One potential answer to this question is that these beliefs are, at least in part, the consequence of presidential campaigns. Because of their regular recurrence and salience in the media, there is good reason to suspect that presidential campaigns influence public beliefs about the ability of the presidency to "manage" the economy. Scholars have long demonstrated that presidential campaigns can serve as major sources of political and economic information to members of the public, particularly those who do not ordinarily follow politics or are being socialized into politics for the first time (Arceneaux 2006; Sears and Valentino 1997). Within these campaigns, economic conditions often function as a major--or, "fundamental" (e.g., Jacobson 2015)--issue, no doubt because the basic material needs of individuals, and the prospect of losing or enhancing the means by which to satisfy these needs, have a tendency to be prioritized over other, nonmaterial concerns (Inglehart 1981). Political candidates have good reason, therefore, to use lackluster economic outcomes as an issue on which to attack incumbents or, conversely, favorable economic outcomes as an issue on which to aggrandize themselves and their own policies.
A point often overlooked, of course, is that electoral incentives may compel elites to attribute credit or blame to presidents for the economy completely irrespective of the extent to which such credit/blame is warranted. Given that previous research has found that voters demonstrate a tendency to vote out incumbent administrations in response to events such as shark attacks (Achen and Bartels 2004a) and outcomes of football games (Healy, Malhotra, and Mo 2010)--that is, events whose relation to public policy is essentially absent--it is not difficult to understand why presidents, who most certainly do have at least a marginal influence on the economy (Tufte 1980), are so often wholeheartedly credited or blamed by their opponents for economic conditions regardless of whether such claims are legitimate. Moreover, given the especially complex nature of macroeconomics, much of the public is likely ill equipped to respond critically to spurious economic claims made by elites (e.g., see Zaller 1992, 311). Political elites, including presidents and presidential candidates, thus stand to reap substantial political gains from exploiting the inherent fuzziness of the line separating coincidence and causation.
In exploring the potential sources of citizens' perceptions of presidential responsibility for the economy, it therefore seems natural to first examine the information that citizens are likely to be exposed to during the course of presidential campaigns (Hillygus and Shields 2009; Vavreck 2009). This particular approach is novel in at least two key respects. First, extant theories of "responsibility attribution" have often focused exclusively on the characteristics of citizens and officials that facilitate citizens' attribution of responsibility for economic outcomes to various public officials (e.g., Gomez and Wilson 2001; Peffley and Williams 1985). But this approach neglects the possibility that such perceived responsibility exists because elites--and presidential candidates in particular--commonly advance claims that directly communicate to voters these very same attributions (Bartels 2006; Druckman and Holmes 2004; Vavreck 2009). This suspicion is consistent with a large body of research that identifies public opinion as being powerfully shaped by elite cues (Berinsky 2007; A. Campbell et al. 1960; Carmines and Stimson 1986; Converse 1964; Hayes and Guardino 2011; Levendusky 2009), particularly when such elite claims receive substantial media coverage (e.g., see Iyengar and Simon 1993), as is the case with presidential campaigns.
Second, and perhaps more importantly, those who study attribution of responsibility for the economy tend to be concerned with particular presidents (e.g., Newman 2013; Rudolph 2003), not beliefs about the extent to which the presidency itself is held responsible for the national economy. (10) Research on presidential campaign effects, for example, finds that "priming the economy" inclines voters to use current economic conditions as a more important criterion for candidate evaluation and vote choice (see Jacobson 2015). In this vein, analyses by Johnston, Hagen, and Jamieson (2004, 117) point to Al Gore's failure to prime the (strong) economy as a key reason for his loss in the 2000 presidential election, while Sides and Vavreck (2014) point to the Obama campaign's priming of the (weak) economy as a key reason for Obama's presidential victory in 2008. As these examples illustrate, scholars have typically been concerned with particular campaigns and particular presidential candidates rather than how campaign rhetoric may influence beliefs about the general power of the presidency to control the national economy. And, of course, by focusing on particular elements, such studies cannot speak to whether economic attribution occurs for context-specific reasons or, alternatively, because of more fundamental perceptions of presidential responsibility for the economy.
During any given presidential contest, citizens tend to be exposed to a particular combination of candidate claims regarding the economy. Contemporaneous economic indicators are especially important insofar as they may or may not permit both candidates to effectively engage in economic attribution of responsibility. In splendid economic times, the incumbent has good reason to "prime the economy" and take credit for it; the opponent, on the other hand, would be best advised to avoid discussing the economy as it might only serve to benefit the incumbent, and vice versa, for poor economic times (e.g., Bartels 2006; Johnston, Hagen, and Jamieson 2004; Vavreck 2009; cf. D'Elia and Norpoth 2014). Alternatively, when there exist mixed economic signals, the incumbent may take credit for positive economic developments (e.g., lower unemployment), while the opponent blames the incumbent for the negative economic developments (e.g., weak economic growth rates). So long as this attribution is perceived as plausible in citizens' minds (Rudolph 2003), the basic strategies of "claiming credit," "avoiding blame" (Weaver 1986), or attributing blame are regularly employed in presidential campaigns (Iyengar 1991, 8-9; Vavreck 2009, 31-35).
These three scenarios--(1) presidential incumbents claiming credit for the economy, (2) presidential challengers attributing blame to the incumbent for the economy, and (3) one (though generally not both) of the candidates avoiding discussion of the economy altogether--may potentially serve as informational cues to the general public about the power the presidency wields over the national economy. In other words, setting aside partisan, ideological, and other context-specific considerations, do presidential candidates' claims regarding their own--or their respective opponent's--responsibility for national economic outcomes impact how voters conceptualize presidential power vis-a-vis the economy? Relatedly, do such claims impact the extent to which citizens believe a president should be praised or punished for economic conditions? If we assume presidential contenders to be perceived as credible sources of information for voters (see Druckman 2001), we should expect such claims to indeed have a positive effect on the extent to which citizens perceive that presidents can generally control economic outcomes--a concept hereafter labeled Control. Insofar as respondents alter their perceptions of presidential control, it stands to reason that we should also observe a similar--though not necessarily perfectly commensurate--adjustment in the extent to which citizens believe presidents should be held accountable for economic outcomes--a concept hereafter labeled Accountability. This logic leads us to the following hypotheses:
H2 Credit-Claiming Hypothesis: Individuals who observe a presidential candidate claiming credit for economic outcomes should exhibit higher values for both Control and Accountability.
H3 Blaming Hypothesis: Individuals who observe a presidential candidate blaming an incumbent president for economic outcomes should exhibit higher values for both Control and Accountability. (11)
These hypotheses do not speak to the potential effects of a far less common, but potentially more consequential message--that is, an instance in which a candidate makes the case that presidents face many constraints on their respective abilities to effect economic outcomes. Given the frequency with which candidates from the two major parties engage in credit claiming and blaming for economic conditions, such a contrarian scenario is perhaps unlikely to be observed in a contest between candidates of the two major parties. Nevertheless, given the purpose of the study and the design of the survey experiment (see below), we would do well to explore the following question: if a "contrarian candidate" were to claim that presidents are quite limited in their ability to influence the national economy, would citizens readjust their perceptions of Control and/or Accountability?
To date, very little research exists on this type of question. Newman (2013), also using an experimental approach, finds that presenting respondents with similarly contrarian claims made by economists (rather than political candidates vying for the presidency) led to improved evaluations of a specific president--that is, President Obama--and attenuated perceptions of President Obama's "responsibility for national economic conditions," all else being equal. While Newman (2013) does not distinguish perceptions of control over the economy from attitudes toward accountability for it, the key implication is that such contrarian information should be capable of decreasing the extent to which the economy is a legitimate criterion by which to evaluate presidents in general because it communicates to voters that presidents have little control over the economy. We therefore have reason to suspect that such contrarian information would indeed reduce perceptions of presidential control of the economy (Control) and reduce the extent to which citizens believe presidents should be held accountable for the economy (Accountability). Stated formally,
H4 Contrarian Claim Hypothesis: Individuals who observe a presidential candidate arguing that presidents have relatively little control over economic outcomes should exhibit lower values for Control and Accountability.
Importantly, though the presence of such a contrarian candidate in an actual U.S. election is unlikely, evidence in support of H4 would more broadly suggest that, in the absence of credible information regarding the president's limitations on effecting desired economic outcomes, the public's perception of (1) a president's ability to control the economy and (2) the extent to which a president should be held accountable for economic outcomes, remain exaggerated and, thus, likely affect presidential elections and evaluations to a greater degree than should be the case.
Given the many real-world factors that could influence perceptions of presidential responsibility for the economy, as well as the limited real-world frequency with which citizens encounter information that decouples the vicissitudes of the macroeconomy from the powers of the presidency, an experimental approach represents a superior method of testing the aforementioned hypotheses (see also Newman 2013; Sirin and Villalobos 2011). Two experimental studies were conducted in July of 2014, which, taken together, permit us to test each of the aforementioned hypotheses. (12) Our total sample consists of 476 respondents and was collected using Amazon.com's Mechanical Turk (MTurk) service. This web-based service enables "requesters" to post "human intelligence tasks" (HITs), which MTurk "workers" can then complete. (13) In order to avoid any self-selection confounds, the HITs for the experiments did not provide workers with any inkling as to the nature and purpose of the study. Previous research has found that adult samples acquired via MTurk are, on most demographic variables (see Table 1), reasonably comparable to those acquired in more established national surveys (e.g., the American National Election Studies [ANES]) and are considerably more representative of the national population than student samples often used in political science experiments (Berinsky, Huber, and Lenz 2012). (14) Recent research on the external validity of MTurk finds that the results of framing experiments launched on MTurk were substantively similar to the effects observed in a nationally representative sample (Leeper and Mullinix 2014).
Design of the Experiment
The purpose of the experiment was to instrument, as treatments, the types of information a citizen is likely to encounter during a presidential campaign and then examine whether these various treatments are predictive of changes in perceptions of presidential responsibility for the economy. Importantly, these treatments closely resemble the distinct campaign strategies presidential candidates regularly employ (see Vavreck 2009, chap. 3). It must also be stressed that, in contrast to other experiments investigating attributions to the president (e.g., Newman 2013), this study was carefully designed so as to render irrelevant any considerations involving current perceptions of the economy, attitudes toward specific political figures (e.g., President Obama), partisanship, and ideology. This is important insofar as ideological and partisan orientations have been shown to moderate the degree to which individuals attribute economic performance to particular presidents (Lebo and Cassino 2007; Rudolph 2003). (15) Partisans, for example, will likely exhibit skepticism when their unpreferred candidate is attributed praise for, or tries to attribute blame to, the preferred candidate for economic outcomes (Taber and Lodge 2006) and will even discount factual economic information if it appears to undermine a co-partisan president (Ramirez and Erickson 2014). Similarly, context-specific factors, such as the perceptions of individual candidates (Peffley 1989) and contemporaneous economic conditions (Achen and Bartels 2004b; Bartels 2008; Campbell 2000; Erikson, Mackuen, and Stimson 2002; Erikson and Wlezien 2008), have also been shown to influence presidential fortunes (e.g., presidential election outcomes, presidential approval, etc.).
Given the role of debates as key focal points of presidential campaigns (Jamieson and Birdsell 1988), respondents were presented with (fabricated) debate transcript excerpts that, respondents were told, were taken from a previous presidential debate in the months just prior to a presidential election. To avoid any possibility that respondents might believe the debate to have involved the current president, respondents were informed that the debate took place after World War II--that is, during the post-Keynesian period in which governments "are seen to exert real influence over the course of the national economy" (Bartels 2008, 99)--but prior to the 2004 election. Respondents were explicitly instructed to not concern themselves with "which presidential election the transcript is taken from" (see Supporting Information for details).
Respondents were randomly assigned to view one of several different forms of presidential campaign claiming, forming a total of six conditions. (16) These conditions are listed and briefly described in Table 2. The first condition (Baseline) functions as a placebo control, featuring credit claiming for various educational outcomes by the incumbent president, while the challenger blames the incumbent for various educational outcomes. There is thus credit claiming and blaming by presidential candidates, but no reference to the economy whatsoever. The second condition (Claim/Avoid) was designed to test H2, and featured the incumbent president claiming credit for positive economic outcomes, while the challenger avoided discussing the economy and instead blamed the president for poor performance in the area of education. The third condition (Avoid/Blame), designed to test H3, is the exact inverse of the second condition: the presidential incumbent avoids discussing the economy, choosing instead to highlight his success in the area of education, while the challenger chooses to blame the lackluster economy on the president. The fourth condition (Claim/Blame) allows for credit claiming and blaming to occur at the same time, wherein the incumbent president claims credit for positive national economic outcomes, and the challenger blames the incumbent for negative economic outcomes.
The remaining two conditions were designed to test H4. The fifth condition is identical to the Claim/Blame condition in every respect except one: it also includes a "contrarian" third-party challenger who highlights the limits on a president's ability to manage the national economy (Contrarian/Claim/Blame). Admittedly, the presence of such a third-party candidate in a real-world presidential campaign is unlikely; however, given the nature of both our research question and the survey experiment itself, featuring claims by this contrarian candidate holds the potential to offer additional valuable insights concerning citizens' beliefs about presidential responsibility for the economy. Finally, in the interest of completely isolating the effects of the contrarian candidate's statements, the sixth condition simply added to Baseline the same third-party candidate arguing that the president has a limited ability to manage the national economy (Contrarian/Baseline) (17) Importantly, the order in which respondents observed each candidate's transcript excerpt was randomized to ensure that no set of candidate claims was systematically more influential than the other(s). (18)
Operationalization of Variables
Respondents answered two questions measuring their perceptions of presidential responsibility for the economy. Importantly, these two questions were presented in random order and both utilize a nine-point response scale. Responses to these questions function as our two dependent variables of interest: Control and Accountability. Control was measured using the following question:
Generally speaking, to what extent do you feel that a president has control over what happens in the economy during his time in office? A score of "1" would mean that you believe that a president has essentially no control over what happens in the economy. A score of "9" would mean that you believe that a president has essentially complete control over what happens the economy.
Similarly, Accountability was measured with the following question:
Generally speaking, to what extent do you feel that a president should be held accountable (that is, either praised or punished) for what happens in the economy during his time in office? A score of "1" would mean that you believe that a president should not be held accountable at all for what happens in the economy. A score of "9" would mean that you believe that a president should be held entirely accountable for what happens in the economy. (19)
Following the measures for Control and Accountability, respondents answered a series of demographic and political questions regarding their gender, age, and race. In addition, the survey measured annual income range (Income; "Below $20,000" = 1; "$90,000 or more" = 9), party identification and political ideology. Party identification (PID) and political ideology (Ideology) are measured using the standard questions from the National Election Studies (NES) survey and range from "Strong Democrat"/"Extremely Liberal" (= 1) to "Strong Republican"/"Extremely Conservative" (= 7). For the analyses, PID, Ideology, and Income were recoded to range from 0 to 1 for interpretive ease.
Finally, a variable that scholars have deemed to be particularly important in analyses of economic attribution is that of political knowledge (e.g., see Gomez and Wilson 2001, 2003). Those with greater political knowledge should, in theory, be more cognizant of the limits a president may face in trying to "manage" the economy (e.g., the presence of divided government) and, therefore, should attribute to the president less blame/ credit as a result. Thus, because substantial variability exists in citizens' knowledge of political matters (Delli Carpini and Keeter 1996), the variable Political Knowledge, recoded to range from 0 to 1, is included in several of the statistical models. (20)
In order to test H1, a t test was performed to determine whether, among those respondents who were randomly assigned to the Baseline condition, a significant difference exists between the mean values of the Control and Accountability measures. The results are summarized in Table 3. In clear support of H1, within the Baseline condition, the mean value of Accountability equaled 6.1 while the mean value of Control equaled 5.5, a substantively and statistically significant difference (p < .05, two-tailed test). Similarly, pooling together all conditions (the right-most column in Table 3), the mean values for Accountability and Control are 5.6 and 5.1, respectively--a highly significant difference (p < .001). On one hand, these differences may not appear all that large. Yet, because the presentation of the Control and Accountability measures was randomized and sequential, respondents should have been inclined to give equal values for both measures, either to appear consistent or because of response set bias. To the extent that this was the case, these results may in fact understate the true difference between Control and Accountability. (21) Figure 1 displays graphically the differences within and between each indicate that the individual's perceptions of control are greater than his/her beliefs about accountability, while positive values would indicate the opposite (a value of zero would indicate that is perceptions of control are perfectly in line with beliefs about accountability).) The results confirm that only 17.6% of the entire sample (19% of the Baseline respondents) exhibited negative values, while fully 45% of the entire sample (48% of the Baseline condition) exhibited positive values. This serves as additional evidence in support of H1. experimental condition, visually confirming that Accountability consistently rates higher than Control.
As one might expect given previous research (e.g., Lewis-Beck and Stegmaier 2000), the correlation between these two measures is positive and moderately strong (0.56 in the Baseline condition, 0.62 across all conditions). However, in contrast to previous scholarship, this study represents the first direct test of whether there exists a fundamental disconnect between perceptions of presidential power and attribution of responsibility. In finding support for H1, this study provides direct evidence that, ceteris paribus, citizens are willing to credit or blame the presidency for the national economy more so than they believe the presidency to be in control of the national economy.
Table 4 lists the results of the remaining hypothesis tests. The independent effects of the treatment conditions alone on Control and Accountability were estimated using ordinary least squares (OLS) multiple regression analyses and can be observed in the columns labeled Model 1 and Model 3, respectively. (22) In both cases, we find the most striking results within the conditions in which the "contrarian" candidate discussed the president's limited ability to influence the national economy. Compared to those assigned to the Baseline condition, for example, those assigned to the Contrarian/Baseline condition exhibited a 0.81 reduction in Control and 0.93 reduction in Accountability (p < .01, two-tailed). Similar results are found among those assigned to the Contrarian/Claim/Blame condition, though the effect sizes are somewhat smaller and, in the case of Accountability, fall narrowly short of marginal statistical significance. (23) It is also worth noting that Contrarian/Claim/Blame shows a far larger substantive effect on Control than on Accountability, again suggesting that these two concepts are distinct. Finally, in the Control and Accountability models in which covariates are included (Models 2 and 4, respectively), the effect sizes of the pure contrarian condition (Contrarian/Baseline) remain virtually unaltered, easily attaining statistical significance.
In all, these results provide clear evidence in support of H4: upon observing evidence that presidents face considerable constraints when trying to improve the economy, citizens are likely to reduce the extent to which they believe presidents control the economy and the degree to which they believe presidents should be credited or blamed for national economic conditions. (24)
The effects of incumbent claiming and challenger blaming, in contrast, do not provide any evidence for H2 or H3; in fact, because the coefficients on the first three treatment conditions are consistently negative in sign, the results actually point in the opposite direction (though generally not significantly so). That is, compared to those in the Baseline condition, those in the Claim/Avoid condition exhibited a marginally significant lower value for Control (p < .10) and a significantly lower value for Accountability (p < .05). The effects of the Claim/Blame and Avoid/Blame conditions never meet conventional levels of statistical significance. These null findings are important insofar as they suggest that, absent other cues, the claiming/blaming messaging citizens tend to hear from presidential candidates may be unlikely to exert much influence on perceptions of presidential power over the economy (if anything, these results suggest it will tend to reduce the extent to which citizens believe presidents are responsible for the economy). However, two points are worth emphasizing. First, the effect sizes of these treatment conditions tend to be quite small relative to the effect sizes observed for the "contrarian" treatments. Second, it must again be stressed that these results say nothing about how the inclusion of partisan and ideological cues might moderate these treatment effects. Future research would do well to examine how various forms of economy-related elite messaging impacts perceptions of presidential responsibility when citizens are made aware of the candidates' respective parties and ideological inclinations as well as whether a general suspicion of political candidates might limit the perceived veracity of their claims (e.g., see McGraw, Lodge, and Jones 2002).
Political Characteristics and Perceptions of Presidential Responsibility
While Models 1 and 3 report only the effects of the treatment conditions, Models 2 and 4 also include the effects of the control variables on Control and Accountability, respectively. First, note that the substantive and statistical significance of the treatment conditions remain essentially intact, which is to be expected given that randomization was largely successful. Second, in examining the effects of the political covariates in Models 2 and 4, perhaps the most impressive result is the effect of Ideology. Holding all else constant, moving from "Extremely Liberal" to "Extremely Conservative" is predictive of more than a two-point increase in Control (p < .001) and a 1.85-point increase in Accountability (p < .001). As an effect for Ideology was not hypothesized, future research may wish to examine whether factors underlying conservative ideological self-identification are also predictive of a tendency to assign greater economic responsibility to the presidency.
Finally, it is instructive to examine the effects of our Political Knowledge measure. Previous research has suggested that those higher in political knowledge should be less inclined to attribute responsibility for economic conditions to the president alone (Gomez and Wilson 2003). Again, this is because more knowledgeable citizens should be more aware of the president's limitations in shaping national economic outcomes and, as a result, attribute to him less credit and blame for the economy. The results listed in Table 4, while partially corroborating this finding, suggest a somewhat more nuanced story. The size of, and negative sign on, the coefficient for Political Knowledge in Model 2 does indeed suggest that, relative to those lowest in knowledge of political affairs, those highest in knowledge of political affairs tend to believe that presidents wield less control over the economy, all else being equal. (25)
And yet, the near-zero and statistically insignificant coefficient on Political Knowledge obtained in Model 4 suggests that moving from lowest to highest Political Knowledge has virtually no effect on beliefs about the extent to which a president should be credited or blamed for national economic conditions. This result again reinforces the logic underlying H1--that is, that Control and Accountability are actually distinct constructs. Some citizens may be knowledgeable of a given president's limited ability to effect positive national economic outcomes, but this does not necessarily portend that such citizens will be any less willing to credit or blame the president for national economic conditions when a pollster contacts them or when Election Day arrives.
Lastly, additional models (see Supporting Information for details) tested for the presence of heterogeneous treatment effects. Specifically, Models 1 and 3 (above) were expanded to include interactions between each of the treatment conditions and Political Knowledge, PID, and Ideology. Regarding the Political Knowledge interactions, the results of these analyses reveal that, regardless of whether Control or Accountability was specified as the dependent variable, nearly all of the interaction terms in these models failed to attain conventional levels of statistical significance (i.e., p < .05, two-tailed tests). Regarding our most robust treatment condition (Contrarian/Baseline), the failure of the Contrarian/ Baseline X Political Knowledge interaction to attain statistical significance suggests that, regardless of one's level of political knowledge, simply being exposed to contrarian information concerning the president's ability to influence the economy significantly reduces perceptions of presidential responsibility for the economy.
Regarding the interactions involving PID and Ideology, the results offer little evidence for the notion that Democrats/liberals responded differently to the treatment conditions than did Republicans/conservatives. Given the design of the experiment, which deliberately omitted partisan and ideological cues, these results were not unexpected (i.e., there existed no reason, a priori, to suspect that partisans/ideologues would differentially respond to the treatment conditions). In all, despite individual differences in political knowledge, party identification, and ideological self-placement, the treatment conditions influenced perceptions of presidential control over the national economy, and beliefs about the extent to which the presidency should be held accountable for it, across individuals in a fairly similar fashion.
Discussion and Conclusion
This study set out to investigate a basic set of questions that have heretofore been somewhat neglected by scholars interested in public perceptions of presidential power: to what extent and why do citizens hold presidents so responsible for the national economy? It was theorized that citizens' conceptualization of responsibility comprises two related, but distinct components: (1) the degree of control presidents exercise over the national economy and (2) the extent to which presidents should be held accountable for what happens in the economy. In addition, it was theorized that citizens' exaggerated sense of the presidency's responsibility for the economy are, in part, the result of presidential campaigns, which regularly feature candidates (1) engaging in credit claiming and blaming for the national economy as well as (2) giving little attention to the many constraints on presidential control of the economy. The ideal test, therefore, is one that examines how the candidate claims one typically hears during a presidential election contest impact perceptions of presidential control of, and accountability for, the national economy, while also neutralizing partisan, ideological, and contextual considerations (e.g., candidate qualities, current economic indicators, etc.). Given that observational data are limited in their ability to meet these criteria, a survey experiment was implemented using a sample of adult U.S. citizens.
Several key findings emerged from the experimental analyses. First, citizens appear willing to hold a president--that is, any president--accountable for economic conditions to a greater extent than they believe him to be in control of the economy. This finding is consistent with--and sheds new light upon--extant literature suggesting that citizens will punish incumbents for events over which the incumbents possess little or no control (Achen and Bartels 2004a; Healy, Malhotra, and Mo 2010). The finding also makes an important contribution to a wealth of literature demonstrating that variation in the president's ability to enact his preferred economic policies has little effect on the extent to which the public is willing to praise or punish him for economic conditions. Because perceived control of the economy and accountability for the economy are actually distinct concepts (the former an empirical question; the latter a normative question), shifts in perceptions of presidential control (if and when they do occur) may not always translate into commensurate shifts in Americans' willingness to hold presidents accountable for the economy when responding to opinion polls or come Election Day.
Second, though somewhat small in effect size (and rarely attaining statistical significance), the outcomes of the H2 and H3 hypothesis tests are important to consider. These results suggest that, without the inclusion of political cues, the self-serving candidate messaging involving credit claiming and blaming for the economy, often integral to electoral politics, is largely ineffective in achieving its ostensible purpose--that is, increasing the extent to which citizens believe presidents to be responsible for national economic conditions. How can we reconcile these findings with extant literature on campaign effects, which generally argues that presidential candidates' priming of the economy increases its importance relative to other issues, thereby inclining citizens to "vote the economy" (for a review, see Jacobson 2015)? One possibility is that while priming the economy may affect how voters weigh the current state of the economy in their evaluations of the president, it may exert less influence on voters' perceptions of presidential responsibility for the economy. Testing this possibility with real-world campaigns would be tremendously valuable, though it would require more direct measures of citizens' perceptions of presidential responsibility for the economy than are currently available in most national surveys. Second, and from a more theoretical standpoint, the results suggest that scholars may benefit from distinguishing priming (i.e., mentioning the state of the economy) from credit claiming and blaming. In other words, it may be the case that priming the economy assists presidential candidates, while claiming/blaming for the economy per se is far less effectual. This potentially represents an important distinction insofar as much scholarship uses "priming" and "credit claiming" somewhat interchangeably (e.g., Johnston, Hagen, and Jamieson 2004; Vavreck 2009).
The analyses also revealed strong and consistent support in favor of H4, which hypothesized that a credible candidate who discusses presidents' limited control over economic outcomes should be capable of diminishing perceptions of presidential control of, and accountability for, the national economy. One might have anticipated that public opinion about presidential responsibility for the economy would be fairly resistant to manipulation by brief statements made by unknown politicians in an online survey; and yet, the results of the "contrarian" treatment conditions consistently demonstrate otherwise. One tentative possibility, particularly when viewed in conjunction with the results of Newman (2013), is that citizens may actually be learning from such contrarian information and that its effects on attitudes and behavior should be quite durable, though more research is certainly needed on this question. Moreover, because treatment effects were statistically indistinguishable across those low and high in political knowledge, it suggests that if candidates and/or other elites were to invoke such information more regularly, a vast majority of the public might respond in kind. Conversely, in the absence of this contrarian information, the public's sense of presidential responsibility for the economy remains higher than it probably should, given the array of formidable constraints on presidential power. This finding represents an important contribution to our understanding of how the American public views the presidency in relation to the national economy.
How presidents' actual responsibility for policy outcomes corresponds to the degree to which the public holds presidents accountable for policy outcomes is a topic that has received greater scholarly interest more recently (e.g., Ruder 2014; Sirin and Villalobos 2011). The evidence in support of H4 suggests that citizens are indeed attentive--and responsive--to information about the real-world constraints that presidents encounter in the domain of economic policy making, which opens the possibility that such information may be effective for other policy domains (e.g., health care, national security, etc.). While Sirin and Villalobos (2011), for example, find evidence suggesting that citizens perceive greater presidential control over U.S. military action than the national economy, the results of the present study suggest that such perceptions can nevertheless be adjusted in line with nonpartisan information about constraints on the presidency. But again, adjustments in citizens' perceived control of presidential power may not translate into commensurate adjustments in perceptions of presidential accountability. Thus, in addition, examining the relationship between perceived control and accountability across multiple policy domains, particularly in response to varying amounts of information about constraints on presidential power, also represents a potentially fruitful area of research.
Importantly, this study is among the first to explicitly seek out a causal explanation for the American public's widely held perception that presidents reside firmly at the helm of the national economy. Nevertheless, it is a single study, using an adult convenience sample, and it must therefore necessarily be left to future research to determine whether the effects we observed in the survey experiment are comparable to what we would observe in the "real world" (e.g., see Barabas and Jerit 2010). Moreover, the study examined only one aspect of a potential cause--candidate claims in a presidential campaign--thus leaving unexplored a variety of other potential sources such as media discussion of presidential campaigns or media coverage of the presidency outside of a campaign context. An enormous literature on media agenda setting and framing of issues suggests that media can powerfully influence perceptions of economic performance (Hetherington 1996) and the extent to which citizens believe the economy to be an important issue that is relevant to the election (e.g., Iyengar, Peters, and Kinder 1982; Miller and Krosnick 2000); media, in other words, often communicate to citizens a connection between the state of the economy and the presidency. Setting aside partisanship and selectivity into partisan media (Iyengar and Hahn 2009), it stands to reason that, particularly when viewed as reliable sources (Druckman 2001), media outlets should be more persuasive in convincing citizens of this connection than the candidates themselves (again, potentially because the claims of the latter may be viewed by citizens as self-serving and with some suspicion [McGraw, Lodge, and Jones 2002]). Given the regularity with which the state of the national economy is invoked during media coverage of presidential elections, such a research agenda holds great potential to meaningfully impact campaign politics as we know it.
However, the importance of both citizens' own partisan attachments, as well as the perceived partisan leanings of media sources, are difficult to overstate and should be accounted for by researchers interested in further exploring this topic. For any particular president, cues that he is or is not responsible for economic conditions are likely to be uncritically accepted or quickly dismissed depending on the partisan leanings of the recipient (Lodge and Taber 2013; cf. Green, Palmquist, and Schickler 2002, chap. 5) and/ or the source (Iyengar and Hahn 2009; Levendusky 2013), even if such information is entirely factual and from a neutral source (Ramirez and Erickson 2014).
Nevertheless, even after accounting for partisan biases, media cues regarding presidential responsibility for the economy hold the power to affect the fortunes of presidents in meaningful ways. For example, these cues should exert a substantial effect upon political independents--a group whose attitudes are particularly consequential for presidential elections (see Klar 2014), for example. Moreover, though somewhat mixed, some research finds that the state of the economy and perceptions of responsibility also matter for turnout in presidential elections (e.g., Arceneaux 2003). In addition, though partisan identification certainly correlates with economic perceptions, it does not determine them entirely, and, even among Democrats/Republicans, perceptions of the economy have been shown to predict "defection" to the other party's candidate in presidential elections (e.g., Lewis-Beck et al. 2008, 374-77). Thus, the fortunes of presidents will likely remain inextricably linked with public perceptions of the economy for the foreseeable future, and the media play an important role in informing the latter (Hetherington 1996).
Regarding the potential effects of contrarian information specifically, the results of this study demonstrate that when the source is not perceived as aiding/undermining the current president, but rather making a larger point about the presidency in general, perceptions of presidential responsibility shift relatively evenly among all partisans and independents. Thus, all else being equal, the more frequently and prominently such information is communicated to the public, the more we should expect the public as a whole to rein in its belief that presidents control the national economy and the extent to which it uses the economy as a major criterion for evaluating presidents. Needless to say, the consequences of such a scenario are potentially enormous, likely impacting future presidential candidates' strategies, elections, and approval--that is, all political outcomes that are predicated upon the public's belief that presidents are responsible for the economy--and, thus, downstream outcomes such as the degree of presidential influence in Congress (Edwards 2009) and results of congressional elections (Lebo and O'Geen 2011).
Finally, because this study finds that information about presidential constraints on economic policy making tends to attenuate perceptions of presidential control of, and accountability for, the economy, important normative questions arise. For example, should journalists, national polling firms, and even scholars continue to entertain and/or reinforce the notion that presidents are largely responsible for the national economy? Do national media have any responsibility to dedicate more coverage to the limits on presidential power over the economy? Given the well-documented importance of the national economy for presidential fortunes, in conjunction with the surprisingly meager evidence for presidents' actual influence on economic outcomes, it is hoped that such questions will receive greater attention and consideration in the near future.
In conclusion, V. O. Key, writing of citizens' use of their recent experience to evaluate candidates, famously insisted that "voters are not fools." The problem, one might counter, is that candidates tend not to be fools either. By omitting information concerning the president's limited ability to influence national economic outcomes, they are perhaps improving their chances of electoral victory, but at the significant cost of serving to perpetuate the widespread perception that presidents are akin to omnipotent "managers" of the national economy. Until they are confronted with information to the contrary-- whether from presidential candidates or elsewhere--the results of this study suggest that Americans will continue to have unrealistic expectations of presidents and will continue to judge the economy to be a far more important criterion for evaluating presidents than is likely warranted by reality.
Abramowitz, Alan I. 1988. "An Improved Model for Predicting Presidential Election Outcomes." PS: Political Science and Politics 21 (4): 843-47.
Achen, Christopher H., and Larry M. Bartels. 2004a. "Blind Retrospection: Electoral Responses to Drought, Flu, and Shark Attacks." Centro de Estudios Avanzados en Ciencias Sociales Working Paper.
--. 2004b. "Musical Chairs: Pocketbook Voting and the Limits of Democratic
Accountability." Presented at the Annual Meeting of the American Political Science Association, Chicago.
Arceneaux, Kevin. 2003. "The Conditional Impact of Blame Attribution on the Relationship Between Economic Adversity and Turnout." Political Research Quarterly 56 (1): 67-75.
--. 2006. "Do Campaigns Help Voters Learn? A Cross-National Analysis." British Journal of Political Science 36 (1): 159-73.
Barabas, Jason, and Jennifer Jerit. 2010. "Are Survey Experiments Externally Valid?" American Political Science Review 104 (2): 226-42.
Baron, Reuben M., and David A. Kenny. 1986. "The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations." Journal of Personality and Social Psychology 51 (6): 1173-82.
Bartels, Larry M. 2006. "Priming and Persuasion in Presidential Campaigns." In Capturing Campaign Effects, ed. Henry E. Brady and Richard Johnston. Ann Arbor: University of Michigan Press, 78-112.
--. 2008. Unequal Democracy: The Political Economy of the New Gilded Age. Princeton, NJ: Princeton University Press.
Berinsky, Adam J. 2007. "Assuming the Costs of War: Events, Elites, and American Public Support for Military Conflict." The Journal of Politics 69 (4): 975-97.
Berinsky, Adam J., Gregory A. Huber, and Gabriel S. Lenz. 2012. "Evaluating Online Labor Markets for Experimental Research: Amazon.com's Mechanical Turk." Political Analysis 20 (3): 351-68.
Blinder, Alan S., and Mark W. Watson. 2014. "Presidents and the U.S. Economy: An Econometric Exploration." National Bureau of Economic Research Working Paper 20324. http://www.nber. org/papers/w20324 (accessed August 1, 2015).
Bloom, Howard S., and H. Douglas Price. 1975. "Voter Response to Short-Run Economic Conditions: The Asymmetric Effect of Prosperity and Recession." American Political Science Review 69 (4): 1240-54.
Brody, Richard, and Lee Sigelman. 1983. "Presidential Popularity and Presidential Election: An Update and Extension." Public Opinion Quarterly 47 (3): 325-28.
Bullock, John G., Donald P. Green, and Shang E. Ha. 2010. "Yes, but What's the Mechanism? (Don't Expect an Easy Answer)." Journal of Personality and Social Psychology 98 (4): 550-58.
Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The American Voter. Chicago: University of Chicago Press.
Campbell, James E. 2000. The American Campaign: U.S. Presidential Campaigns and the National Vote. College Station: Texas A&M University Press.
Carmines, Edward G., and James A. Stimson. 1986. "On the Structure and Sequence of Issue Evolution." American Political Science Review 80 (3): 901-20.
Converse, Philip E. 1964. "The Nature of Belief Systems in Mass Publics (1964)." Critical Review 18 (1-3): 1-74.
De Boef, Suzanna, and Paul M. Kellstedt. 2004. "The Political (and Economic) Origins of Consumer Confidence." American Journal of Political Science 48 (4): 633-49.
D'Elia, Justine, and Helmut Norpoth. 2014. "Winning with a Bad Economy." Presidential Studies Quarterly 44 (3): 467-83.
Delli Carpini, Michael X., and Scott Keeter. 1996. What Americans Know about Politics and Why It Matters. New Haven, CT: Yale University Press.
Druckman, James N. 2001. "On the Limits of Framing Effects: Who Can Frame?" The Journal of Politics 63 (4): 1041-66.
Druckman, James N., and Justin W. Holmes. 2004. "Does Presidential Rhetoric Matter? Priming and Presidential Approval." Presidential Studies Quarterly 34 (4): 755-78.
Duch, Raymond M., and Randolph T. Stevenson. 2008. The Economic Vote: How Political and Economic Institutions Condition Election Results. New York: Cambridge University Press.
Edwards, George C. 1989. At the Margins: Presidential Leadership of Congress. New Haven, CT: Yale University Press.
--. 2009. "Presidential Approval as a Source of Influence in Congress." In The Oxford Handbook of the American Presidency, ed. George C. Edwards III and William Howell. New York: Oxford University Press, 338-61.
Eisenberg, Daniel, and Jonathan Ketcham. 2004. "Economic Voting in U.S. Presidential Elections: Who Blames Whom for What." Topics in Economic Analysis & Policy 4 (1): 1-23.
Erikson, Robert S., and Christopher Wlezien. 2008. "Leading Economic Indicators, the Polls, and the Presidential Vote." PS: Political Science & Politics 41 (4): 703-07.
--. 2012. "The Objective and Subjective Economy and the Presidential Vote." SSRN eLibrary. http://papers.ssrn.com/sol3/papers.cfm?abstract_id = 2110791 (accessed August 1, 2015).
Erikson, Robert S., Michael B. Mackuen, and James A. Stimson. 2002. The Macro Polity. New York: Cambridge University Press.
Fiorina, Morris P. 1981. Retrospective Voting in American National Elections. New Haven, CT: Yale University Press.
Franzese, Robert J. 2002. "Electoral and Partisan Cycles in Economic Policies and Outcomes." Annual Review of Political Science 5 (1): 369-421.
Gelman, Andrew, and Gary King. 1993. "Why Are American Presidential Election Campaign Polls So Variable When Votes Are So Predictable?" British Journal of Political Science 23 (4): 409-51.
Golden, David G., and James M. Poterba. 1980. "The Price of Popularity: The Political Business Cycle Reexamined." American Journal of Political Science 24 (4): 696-714.
Gomez, Brad T., and J. Matthew Wilson. 2001. "Political Sophistication and Economic Voting in the American Electorate: A Theory of Heterogeneous Attribution." American Journal of Political Science 45 (4): 899-914.
--. 2003. "Causal Attribution and Economic Voting in American Congressional Elections." Political Research Quarterly 56 (3): 271-82.
--. 2006. "Cognitive Heterogeneity and Economic Voting: A Comparative Analysis of Four Democratic Electorates." American Journal of Political Science 50 (1): 127-45.
Greenblat, Alan. 2013. "Can A President Really Fix A Bad Economy?" NPR.org. http://www.npr. org/2011/11/16/141762700/can-a-president-really-fix-a-bad-economy (accessed May 4, 2015).
Green, Donald P., Shang E. Ha, and John G. Bullock. 2010. "Enough Already about 'Black Box' Experiments: Studying Mediation Is More Difficult than Most Scholars Suppose." ANNALS of the American Academy of Political and Social Science 628 (1): 200-208.
Green, Donald P., Bradley Palmquist, and Eric Schickler. 2002. Partisan Hearts and Minds: Political Parties and the Social Identities of Voters. New Haven, CT: Yale University Press.
Gronke, Paul, and Brian Newman. 2009. "Public Evaluations of Presidents." In The Oxford Handbook of the American Presidency, ed. George C. Edwards III and William Howell. New York: Oxford University Press, 232-53. http://works.bepress.com/brian_newman/3/ (accessed January 26, 2016).
Gross, Daniel. 2008. "The Power to Fix the Economy Rests With the Next President." Newsweek Magazine, June 27. http://www.thedailybeast.com/newsweek/2008/06/27/the-power-to-fix-theeconomy-rests-with-the-next- president.html (accessed August 1, 2015).
Hamilton, V. Lee. 1978. "Who Is Responsible? Toward a Social Psychology of Responsibility Attribution." Social Psychology 4 (1): 316-28.
Harwood, John. 2013. "Fed Chairman's Departure Casts a New Light on the Bush Legacy." New York Times, July 26. http://www.nytimes.com/2013/07/27/us/politics/fed-chairman-changecasts-new-light-on-bush-era.html (accessed August 1, 2015).
Hayes, Danny, and Matt Guardino. 2011. "The Influence of Foreign Voices on U.S. Public Opinion." American Journal of Political Science 55 (4): 831-51.
Healy, Andrew J., Neil Malhotra, and Cecilia Hyunjung Mo. 2010. "Irrelevant Events Affect Voters' Evaluations of Government Performance." Proceedings of the National Academy of Sciences 107 (29): 12804-09.
Hellwig, Timothy T. 2001. "Interdependence, Government Constraints, and Economic Voting." The Journal of Politics 63 (4): 1141-62.
Hetherington, Marc J. 1996. "The Media's Role in Forming Voters' National Economic Evaluations in 1992." American Journal of Political Science 40 (2): 372-95.
Hibbing, John R., and Elizabeth Theiss-Morse. 2002. Stealth Democracy: Americans' Beliefs About How Government Should Work. New York: Cambridge University Press.
Hibbs, Douglas A. 1987. The American Political Economy: Macroeconomics and Electoral Politics. Cambridge, MA: Harvard University Press.
Hibbs, Douglas A., R. Douglas Rivers, and Nicholas Vasilatos. 1982. "The Dynamics of Political Support for American Presidents among Occupational and Partisan Groups." American Journal of Political Science 26 (2): 312-32.
Hillygus, D. Sunshine, and Todd G. Shields. 2009. The Persuadable Voter: Wedge Issues in Presidential Campaigns. Princeton, NJ: Princeton University Press.
Imai, Kosuke, Luke Keele, and Dustin Tingley. 2010. "A General Approach to Causal Mediation Analysis." Psychological Methods 15 (4): 309-34.
Inglehart, Ronald. 1981. "Post-Materialism in an Environment of Insecurity." American Political Science Review 75 (4): 880-900.
Iyengar, Shanto. 1991. Is Anyone Responsible?: How Television Frames Political Issues. Chicago: University of Chicago Press.
Iyengar, Shanto, and Kyu S Hahn. 2009. "Red Media, Blue Media: Evidence of Ideological Selectivity in Media Use." Journal of Communication 59 (1): 19-39.
Iyengar, Shanto, Mark D. Peters, and Donald R. Kinder. 1982. "Experimental Demonstrations of the 'Not-So-Minimal' Consequences of Television News Programs." American Political Science Review 76 (4): 848-58.
Iyengar, Shanto, and Adam Simon. 1993. "News Coverage of the Gulf Crisis and Public Opinion A Study of Agenda-Setting, Priming, and Framing." Communication Research 20 (3): 365-83.
Jacobson, Gary C. 2015. "How Do Campaigns Matter?" Annual Review of Political Science 18 (1): 31-47.
Jamieson, Kathleen Hall, and David S. Birdsell. 1988. Presidential Debates: The Challenge of Creating an Informed Electorate. New York: Oxford University Press.
Jenkins-Smith, Hank C., Carol L. Silva, and Richard W. Waterman. 2005. "Micro- and Macrolevel Models of the Presidential Expectations Gap." The Journal of Politics 67 (3): 690-715.
Johnston, Richard, Michael G. Hagen, and Kathleen Hall Jamieson. 2004. The 2000 Presidential Election and the Foundations of Party Politics. New York: Cambridge University Press.
Kayser, Mark Andreas, and Michael Peress. 2012. "Benchmarking across Borders: Electoral Accountability and the Necessity of Comparison." American Political Science Review 106 (3): 661-84.
Klar, Samara. 2014. "Identity and Engagement among Political Independents in America." Political Psychology 35 (4): 577-91.
Lebo, Matthew J., and Daniel Cassino. 2007. "The Aggregated Consequences of Motivated Reasoning and the Dynamics of Partisan Presidential Approval." Political Psychology 28 (6): 719-46.
Lebo, Matthew J., and Andrew J. O'Geen. 2011. "The President's Role in the Partisan Congressional Arena." The Journal of Politics 73 (3): 718-34.
Leeper, Thomas J., and Kevin J. Mullinix. 2014. "What If You Had Done Things Differently?: Testing the Generalizability of Framing Effects with Parallel Experiments." Working Paper. https://dl.dropboxusercontent.com/u/414906/ParallelExperiments0nFraming.pdf.
Levendusky, Matthew. 2009. The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans. Chicago: University of Chicago Press.
--. 2013. How Partisan Media Polarize America. Chicago: University of Chicago Press.
Lewis-Beck, Michael S., and Martin Paldam. 2000. "Economic Voting: An Introduction." Electoral Studies 19 (2-3): 113-21.
Lewis-Beck, Michael S., and Mary Stegmaier. 2000. "Economic Determinants of Electoral Outcomes." Annual Review of Political Science 3 (1): 183-219.
Lewis-Beck, Michael S., and Tom W. Rice. 1982. "Presidential Popularity and Presidential Vote." Public Opinion Quarterly 46 (4): 534-37.
Lewis-Beck, Michael S., William G. Jacoby, Helmut Norpoth, and Herbert F. Weisberg. 2008. The American Voter Revisited. Ann Arbor: University of Michigan Press.
Lodge, Milton, and Charles S. Taber. 2013. The Rationalizing Voter. New York: Cambridge University Press.
MacKuen, Michael B. 1983. "Political Drama, Economic Conditions, and the Dynamics of Presidential Popularity." American Journal of Political Science 27 (2): 165-92.
MacKuen, Michael B., Robert S. Erikson, and James A. Stimson. 1992. "Peasants or Bankers? The American Electorate and the U.S. Economy." American Political Science Review 86 (3): 597-611.
McGraw, Kathleen M. 1991. "Managing Blame: An Experimental Test of the Effects of Political Accounts." American Political Science Review 85 (4): 1133-57.
McGraw, Kathleen M., Milton Lodge, and Jeffrey M. Jones. 2002. "The Pandering Politicians of Suspicious Minds." The Journal of Politics 64 (2): 362-83.
McNichol, Elizabeth. 2012. "Out of Balance: Cuts in Service Have Been States' Primary Response to Budget Gaps, Harming the Nation's Economy." Center for Budget and Policy Priorities. http://www.cbpp.org/research/out-of-balance (accessed February 4, 2016).
Miller, Joanne M., and Jon A. Krosnick. 2000. "News Media Impact on the Ingredients of Presidential Evaluations: Politically Knowledgeable Citizens Are Guided by a Trusted Source." American Journal of Political Science 44 (2): 301-15.
Mueller, John E. 1970. "Presidential Popularity from Truman to Johnson." American Political Science Review 64 (1): 18-34.
Nadeau, Richard, and Michael S. Lewis-Beck. 2001. "National Economic Voting in U.S. Presidential Elections." The Journal of Politics 63 (1): 159-81.
Nakamura, David. 2012. "Obama Blames Congress for Inaction on Jobs While Romney Calls President 'out of Touch.'" Washington Post, June 8. http://articles.washingtonpost.com/2012-06-08/ politics/35462171_1_romney-campaign-obama-campaign-obama-aides (accessed August 1, 2015).
Neustadt, Richard E. 1991. Presidential Power and the Modern Presidents: The Politics of Leadership from Roosevelt to Reagan. New York: Simon and Schuster.
Newman, Brian. 2013. "Polls and Elections: Decreasing the Economy's Impact on Evaluations of the President: An Experiment on Attribution Framing." Presidential Studies Quarterly 43 (4): 866-82.
Nicholson, Stephen P., Gary M. Segura, and Nathan D. Woods. 2002. "Presidential Approval and the Mixed Blessing of Divided Government." The Journal of Politics 64 (3): 701-20.
Norpoth, Helmut. 1985. "Politics, Economics and the Cycle of Presidential Popularity." In Economics and Electoral Outcomes, ed. H. Eulau and Michael S. Lewis-Beck. New York: Agathon, 167-86.
--. 2001. "Divided Government and Economic Voting." The Journal of Politics 63 (2): 41435.
Peffley, Mark. 1984. "The Voter as Juror: Attributing Responsibility for Economic Conditions." Political Behavior 6 (3): 275-94.
--. 1989. "Presidential Image and Economic Performance: A Dynamic Analysis." Political Behavior 11 (4): 309-33.
Peffley, Mark, and John T. Williams. 1985. "Attributing Presidential Responsibility for National Economic Problems." American Politics Research 13 (4): 393-425.
Pfiffner, James P. 2011. "Decision Making in the Obama White House." Presidential Studies Quarterly 41 (2): 244-62.
Poole, Keith T., and Howard L. Rosenthal. 2007. Ideology and Congress. New Brunswick, NJ: Transaction Publishers.
Powell, G. Bingham, and Guy D. Whitten. 1993. "A Cross-National Analysis of Economic Voting: Taking Account of the Political Context." American Journal of Political Science 37 (2): 391-414.
Ramirez, Mark D., and Nathan Erickson. 2014. "Partisan Bias and Information Discounting in Economic Judgments." Political Psychology 35 (3): 401-15.
Rodrik, Dani. 2011. The Globalization Paradox: Why Global Markets, States, and Democracy Can't Coexist. Oxford, UK: Oxford University Press.
Ruder, Alex I. 2014. "Institutional Design and the Attribution of Presidential Control: Insulating the President from Blame." Quarterly Journal of Political Science 9 (3): 301-35.
Rudolph, Thomas J. 2003. "Who's Responsible for the Economy? The Formation and Consequences of Responsibility Attributions." American Journal of Political Science 47 (4): 698-713.
Ryssdal, Kai. 2013. "Does the President Actually Influence the Economy?" Marketplace.org. http://www.marketplace.org/topics/economy/freakonomics-radio/does-president-actually-influence-economy (accessed May 4, 2015).
Sears, David O., and Nicholas A. Valentino. 1997. "Politics Matters: Political Events as Catalysts for Preadult Socialization." American Political Science Review 91 (1): 45-65.
Sides, John, and Lynn Vavreck. 2014. The Gamble: Choice and Chance in the 2012 Presidential Election. Updated ed. with a new preface by the authors. Princeton, NJ: Princeton University Press.
Sigelman, Lee. 1979. "Presidential Popularity and Presidential Elections." Public Opinion Quarterly 43 (4): 532-34.
Sirin, Cigdem V., and Jose D. Villalobos. 2011. "Where Does the Buck Stop? Applying Attribution Theory to Examine Public Appraisals of the President." Presidential Studies Quarterly 41 (2): 334-57.
Taber, Charles S., and Milton Lodge. 2006. "Motivated Skepticism in the Evaluation of Political Beliefs." American Journal of Political Science 50 (3): 755-69.
Taylor, Mark Zachary. 2012. "An Economic Ranking of the US Presidents, 1789-2009: A DataBased Approach." PS: Political Science & Politics 45 (4): 596-604.
Tufte, Edward R. 1980. Political Control of the Economy. Princeton, NJ: Princeton University Press.
Vavreck, Lynn. 2009. The Message Matters: The Economy and Presidential Campaigns. Princeton, NJ: Princeton University Press.
Wattenberg, Martin P. 1992. The Rise of Candidate-Centered Politics: Presidential Elections of the 1980s. Cambridge, MA: Harvard University Press.
Weatherford, M. Stephen. 2013. "The President, the Fed, and the Financial Crisis." Presidential Studies Quarterly 43 (2): 299-327.
Weaver, R. Kent. 1986. "The Politics of Blame Avoidance." Journal of Public Policy 6 (4): 371-98.
Zaller, John. 1992. The Nature and Origins of Mass Opinion. Cambridge, UK: Cambridge University Press.
Additional supporting information can be found in the supporting information tab for this article.
Supplemental Online Information. Control, Accountability, and Constraints: Rethinking Perceptions of Presidential Responsibility for the Economy
JOHN V. KANE
Stony Brook University
John V. Kane is a Ph.D. candidate of political science at Stony Brook University. His research primarily focuses on political behavior, partisanship and ideology, and quantitative research methods.
AUTHOR'S NOTE: I wish to give special thanks to Jason Barabas for his many insights and guidance on this project. I would also like to thank Reuben Kline, Matthew Lebo, Helmut Norpoth, and the anonymous reviewers for their assistance and thoughtful comments.
(1.) Debate transcript available at http://www.debates.org/index.phpPpage-october-1-2012-the-second-obama-romney- presidential-debate (accessed September 26, 2014).
(2.) In a similar vein, Erikson, MacKuen and Stimson (2002, 42) go so far as to argue that such behavior should not be considered "irrational" because, "Even if the economy is only marginally influenced by presidential behavior, the economy is one of the few measures of performance upon which most can judge and most can agree."
(3.) Throughout this study, the phrase "responsibility for the economy" is treated as comprising two concepts that are both theoretically and methodologically distinct (discussed below): (1) the degree to which presidents alone are thought to control what transpires in the national economy and (2) the degree to which voters believe presidents should receive credit or blame (i.e., be held accountable) for what transpires in the national economy.
(4.) Importantly, perceptions of economic conditions--which can be significantly shaped by media reporting--may at times matter more for economic voting than actual economic conditions (Hetherington 1996).
(5.) A recent exception is Taylor's (2012) effort to assign each of the previous U.S. presidents an economic "grade," thereby allowing for an "economic ranking of the presidents." The author indeed raises the question of the extent to which presidents actually influence the economy; but the discussion is limited, ultimately falling back on historical accounts of "transformational" presidencies and the fact that "voters seem to think" that presidents do have substantial influence on the economy (Taylor 2012, 597).
(6.) This constraint on presidential power is perhaps all the more formidable in an era of intense partisan polarization in Congress (Poole and Rosenthal 2007). Indeed, many have likened the Obama administration's post-2010 efforts to revive the economy as being symptomatic of precisely this dynamic (e.g., see Nakamura 2012). In his discussion of the Obama administration's crafting of the 2009 stimulus program, for example, Pfiffner (2011, 254) notes that the president's proposal of $787 billion was not larger, at least in part, because of a strong sense that Congress (and the Republican Party in particular) would roundly reject any larger sum.
(7.) In fact, a recent article in the New York Times asserts that the Federal Reserve's "vast influence over the economy rival[s] or exceed[s] that of presidents" (Harwood 2013).
(8.) Ruder (2014), for example, finds that as individuals perceive the president to have less control over bureaucratic agencies, they are less willing to attribute to the president responsibility for the agency's actions.
(9.) In a similar vein, a potential distinction exists wherein citizens may not believe a president alone has much control over the economy but might nevertheless expect him to force the hand of other existing institutions--for example, Congress, the Federal Reserve, foreign governments, etc.--to improve the economy; in the event that he succeeds (fails) in accomplishing the latter, citizens may be inclined to reward (punish) him, all else being equal.
(10.) Sirin and Villalobos (2011) represents a notable exception. Here, the authors use an innovative experimental design on a student sample to discern the independent effects of party cues, policy domain (domestic vs. foreign), and policy condition (deteriorating vs. improving) on attribution of responsibility to the (hypothetical) president. However, as the article does not explicitly focus on informational sources that might influence the extent to which citizens think the president (1) is actually responsible or (2) should be held responsible, for the state of the national economy, its research question is distinct from that of this study.
(11.) It is unclear, however, what effect observing credit claiming by the incumbent and blaming by the challenger would have on Control and Accountability. On the one hand, each claim may signal that the president is responsible for aspects of the economy, and thus generate a kind of "mainstream effect" (Zaller 1992) wherein elites appear to agree that the president's policies are consequential for economic outcomes. Alternatively, the noticeable discrepancy between the two sets of claims may make individuals less inclined to believe that the president has much responsibility for the economy after all, thus resulting in lower values for Control and Accountability. As such, no hypothesis is specified for instances in which an individual experiences both credit claiming and blaming.
(12.) Both experiments featured the exact same Baseline condition, thus the following analyses feature more respondents in the Baseline condition. Because all other conditions were distinct, and because respondents were prevented from taking both survey experiments, the two studies were combined into a single data set. In order to determine whether the two studies could be combined into a single analysis, a t test was conducted to examine whether the Baseline groups in each of the two studies exhibited significantly different mean values on either of the two dependent variable measures. The results of this test confirmed that the groups were statistically indistinguishable on both Control (p - 0.70, two tailed) and Accountability (p - 0.48, two-tailed). As such, the results of the two experiments were combined so as to maximize statistical power and to allow for easier presentation of results, and will hereafter be referred to as a single experiment.
(13.) Survey controls were put in place to ensure that respondents were located within the United States and were unable to complete the survey more than once.
(14.) Berinsky, Huber, and Lenz (2012) identify three key variables wherein MTurk respondents noticeably differ from those of ANES surveys: political knowledge, partisanship, and political ideology. First, MTurk users appear slightly more knowledgeable about politics than ANES respondents, though Berinsky et al. report that these differences are substantively rather small. Nevertheless, this study will control for (and examine the moderating effects of) political knowledge in the following analyses. Secondly, Berinsky et al. find that MTurk is significantly more liberal and Democratic than respondents in the ANES surveys. The figures in Table 1 are consistent with this observation, but, again, these variables (among others) will be controlled for in the following analyses so as to capture the independent effects of the treatments. Moreover, the experimental vignettes used in this study deliberately avoided using any partisan or ideological cues. As such, partisan and ideological leanings of our sample are less of a concern.
(15.) Using ANES data, Rudolph (2003), for example, finds that, in line with a large body of research on motivated reasoning and in-group biases, Democrats increasingly assigned responsibility for the economy to the (Democratic) president as their perceptions of the economy improved; Republicans, though somewhat more likely to attribute responsibility to the president as their economic perceptions improved, overwhelmingly attributed responsibility for the economy to businesspeople.
(16.) Again, because the purpose of this study is to examine beliefs about presidential responsibility for the economy apart from partisan and ideological biases and considerations, neither of the experiments featured any information regarding the candidates' respective parties, nor did the candidates make any statements that could be construed as ideologically motivated (e.g., the candidates said nothing as to whether the federal government should take a more or less interventionist role in the economy).
(17.) Careful effort was made within the design of the experiment to ensure that this third-party candidate would be perceived as credible (see Supporting Information for details).
(18.) Following exposure to the treatment conditions, respondents were asked to briefly describe the exchange between the candidates. This question functioned as a manipulation check, thus enabling the researcher to be confident that respondents had properly read and considered the debate transcript excerpts. Careful analysis of these open-ended answers found absolutely no evidence that respondents perceived the transcript excerpts, candidates, or election to be fabricated.
(19.) In accordance with the underlying purpose of this study, the language in these questions deliberately asks about presidential responsibility for the economy in a general sense, thus mitigating the possibility that citizens will only consider the power of the current or recent presidents vis-a-vis the national economy.
(20.) Political knowledge was measured prior to randomization into treatment conditions. It is operationalized similarly to that of Gomez and Wilson (2003) and Sirin and Villalobos (2011), wherein respondents were asked a series of six open-ended, factual questions pertaining to the offices held by Joe Biden, John Boehner, and John G. Roberts, as well as the political party of President Obama, and the party that currently holds a majority in both the U.S. House of Representatives and the U.S. Senate. These questions scaled well (Cronbach's alpha = 0.79), and, thus, subjects received a score of "1" for each correct response and a score of "0" for each incorrect response, resulting in scores ranging from 0 to 6 (mean = 4.32; SD = 1.77).
(21.) It must also be noted that 71.83% (68.19%) of respondents entered values between 4 and 7 for Control (Accountability), which limits the size of the difference we should expect to find. Moreover, these significant differences are not simply an artifact involving a handful of large, outlying values for either Control or Accountability. Subsequent analyses subtracted each individual's value for Control ([Control.sub.i]) from his or her respective value for Accountability ([Accountability.sub.i]). (Thus, a negative value for [Accountability.sub.i]- [Control.sub.i] would
(22.) Prior to any of the main analyses, randomization checks were performed so as to ensure that the experimental conditions were balanced on all observable covariates. These checks (using OLS and logistic regression models) demonstrated that nearly all covariates were not significantly predicted by treatment assignment at conventional levels of statistical significance (p < .05), which indicates that randomization was largely successful (regressing the treatment assignment variable onto all covariates yielded an insignificant joint test (prob > [chi square] = 0.24). The only exception involved those identifying as African American--these individuals appear to have been slightly overrepresented in the Claim/Avoid and Avoid/Blame conditions. As such, the regression models that include covariates (see below) also controlled for African American identification. Ultimately, this variable did not significantly predict either of the dependent variables of interest.
(23.) Given the data at hand, one can only speculate as to the reason for this weaker treatment effect. Of course, it may be that, with a different sample, we would obtain a stronger and more reliable effect. On the other hand, it might suggest that contrarian information, when transmitted in the midst of major party candidates engaging in claiming/blaming, influences citizens' perceptions of Control to a far greater extent than perceptions of Accountability--that is, that the presence of economic claiming/blaming between candidates perhaps mitigates the steeper declines in perceived presidential accountability we observed in the Contrarian/ Baseline condition. In either case, given that any contrarian information during a presidential campaign would likely occur in conjunction with the major party candidates engaging in economic claiming/blaming, additional research would do well to further explore the effects of this type of treatment condition.
(24.) Additional analyses revealed that a smaller, but still significant portion (p = 0.05) of the effect of the Contrarian/Baseline on Accountability remains even after accounting for Control (plus all other treatment conditions and covariates), suggesting that some of the effect of this condition on Accountability is likely mediated by Control (see Baron and Kenny 1986). Using the mediate package developed by Imai, Keele and Tingley (2010) for Stata , I find that, of the total effect of Contrarian/Baseline on Accountability (-0.89), an estimated 50.10% of this effect is mediated by a decrease in Control. While the design of this study is not entirely conducive to such mediation analyses (see Bullock, Green, and Ha 2010; Green, Ha, and Bullock 2010), and further investigation is still needed, the results are nevertheless broadly consistent with the long-standing theory that, as citizens believe the president to be less capable of controlling what happens in the national economy, they will consequently attribute to him less credit/blame for national economic conditions (Lewis-Beck and Stegmaier 2000; Newman 2013; Sirin and Villalobos 2011). Importantly, however, it must be stressed that the effect falls well short of being fully mediated. Consistent with the logic underlying H1, there is thus reason to suspect that the willingness to attribute credit or blame does not always prove perfectly responsive to shifts in perceptions of presidential control of the economy.
(25.) This effect only attained marginal statistically significance (p = 0.12, two-tailed; p =0.06, one tailed).
Caption: FIGURE 1. Difference Between Perceptions of Presidential Control and Beliefs about Accountability across Experimental Conditions. Note: Bars represent the mean values for Control and Accountability, respectively, and are displayed for each of the six experimental conditions.
TABLE 1 MTurk Sample Characteristics MTurk Sample 2012 ANES Sample Median Income $20-30K $40-45K Mean Age (SD) 35.44 yrs. (1.35) 49.44 yrs. (16.82) Female 45.38% 51.89% White 60.99% 59.1% Black 7.84% 17.18% Hispanic 4.78% 16.99% Democrat 48.95% 39.92% Independent 28.15% 31.20% Republican 22.89% 23.49% Liberal 49.58% 27.81% Moderate 25.63% 34.45% Conservative 24.78% 37.7% Note: A total of 476 respondents were surveyed. TABLE 2 Treatment Conditions and Descriptions Treatment Conditions Treatment Descriptions Baseline The incumbent president claims credit for improve- ments in education; the challenger criticizes the president for poor performance on education and claims that his own policies will be more effective. Claim/ Avoid The incumbent president claims credit for improve- ments in the national economy; the challenger criticizes the president for poor performance on education and claims that his own policies will be more effective. Avoid/Blame The incumbent president claims credit for improve- ments in education; the challenger criticizes the president for poor performance on the national economy and claims that his own policies will be more effective. Claim/Blame The incumbent president claims credit for improve- ments in the national economy; the challenger criticizes the president for poor performance on the national economy and claims that his own policies will be more effective. Contrarian Claim Blame A third-party candidate argues that, as a result of multiple factors, presidents do not actually pos- sess much control over the national economy; the incumbent president claims credit for improve- ments in the national economy; the challenger criticizes the president for poor performance on the national economy and claims that his own policies will be more effective. Contrarian Baseline A third-party candidate argues that, as a result of multiple factors, presidents do not actually pos- sess much control over the national economy; the incumbent president claims credit for improve- ments in education; the challenger criticizes the president for poor performance on education and claims that his own policies will be more effective Notes: The presentation of each (fictitious) candidate's statements was randomized so as to avoid any "recency" and-or "primacy" effects. Care was taken to ensure that statements were of approximately equal length, and that the main candidates' claims only varied with respect to the issue of interest (i.e., the national economy vis-a-vis a noneconomic issue). In addition, the candidates' party labels were not revealed to respondents, nor were the claims of the candidates in any way indicative of their ideological leanings (e.g., the candidates did not express support for "free market" vis-a-vis Keynesian policies, or vice versa). See Supporting Information for verbatim instruction and treatment wordings. TABLE 3 Differences between Perceived Presidential Control and Accountability for the Economy Baseline Condition All Conditions Presidential Control Mean (SD) 5.53 (1.85) 5.12 (1.88) Presidential Accountability 6.06 (1.82) 5.64 (1.86) Mean (SD) Difference between Control .53 .52 and Accountability p-value .03 < .001 N 110 476 Note: p values were estimated using t tests allowing for unequal variances (two-tailed hypothesis tests). Presidential Control (Control) is a nine-point measure capturing the extent to which respondents perceive presidents to have control over what happens in the economy (1 = No Control; 9 = Complete Control). Presidential Accountability (Accountability) is also a nine-point measure, but captures the extent to which respondents believe presidents should be held accountable for what happens in the economy (1 = Should not be held accountable at all; 9 = Should be held entirely accountable). The Baseline condition featured no discussion of the economy. The column labeled "All Conditions" contains estimates across all six experimental conditions pooled together. TABLE 4 Effects on Beliefs about Presidential Economic Control and Accountability Presidential Control (1) (2) Treatment Conditions Claim/Avoid -0.49 ([dagger]) -0.44 (0.28) (0.27) Avoid/Blame -0.39 -0.55 * (0.28) (0.27) Claim/Blame -0.28 -0.18 (0.28) (0.27) Contrarian/Claim/Blame -0.75 ** -0.72 ** (0.28) (0.27) Contrarian/Baseline -0.81 ** -0.79** (0.29) (0.28) Covariates Ideology -- 2.11 *** (0.47) PID -- -0.28 (0.48) Income -- 0.16 (0.28) Political Knowledge -- -0.47 (0.30) Constant 5.53 *** 4.88 *** (0.18) (0.36) N 476 473 Adjusted [R.sup.2] .01 .09 Presidential Accountability (3) (4) Treatment Conditions Claim/Avoid -0.62 * -0.54 * (0.28) (0.27) Avoid/Blame -0.28 -0.45 ([dagger]) (0.28) (0.27) Claim/Blame - 0.50 ([dagger]) -0.35 (0.27) (0.26) Contrarian/Claim/Blame -0.44 -0.33 (0.28) (0.27) Contrarian/Baseline -0.93 *** -0.89 *** (0.29) (0.28) Covariates Ideology -- 1.85 *** (0.47) PID -- 0.17 (0.47) Income -- 0.28 (0.28) Political Knowledge -- -0.08 (0.30) Constant 6.06 *** 4.60 *** (0.18) (0.35) N 476 473 Adjusted [R.sup.2] .02 .12 Notes: Entries are OLS unstandardized regression coefficients (standard errors in parentheses). ([dagger]) significant at p < .10 * significant at p < .05 ** significant at p < .01 *** significant at p < .001 (two-tailed hypothesis tests). Presidential Control (Control) is a nine-point measure capturing the extent to which respondents perceive presidents to have control over what happens in the economy (1 = No Control; 9 = Complete Con-trol). Presidential Accountability (Accountability) is also a nine-point measure, but captures the extent to which respondents believe presidents should be held accountable for what happens in the economy (1 = Should not be held accountable at all; 9 = Should be held entirely accountable). The Baseline condition serves as the omitted category to which all treatment condition effects should be compared. Models 2 and 4 also control for effects of Age, Gender, and Race, none of which exhibited a statistically significant effect on either the Control or Accountability dependent variables. All covariates included in the table were recoded to range from 0 to 1 for ease of interpretation.
|Printer friendly Cite/link Email Feedback|
|Author:||Kane, John V.|
|Publication:||Presidential Studies Quarterly|
|Date:||Jun 1, 2016|
|Previous Article:||The coevolution of affect toward presidents and their parties.|
|Next Article:||Polls and elections: character and political time as sources of presidential greatness.|