Judging politicians: The role of political attentiveness in shaping how people evaluate the ethical behaviour of their leaders

. Political theorists have developed well-deﬁned normative understandings of what constitutes ethical political conduct. Based on democratic theory as well as the demands of practical politics, these understandings prescribe certain types of behaviour and proscribe other types. However, it is unclear to what extent this normative framework has resonance for ordinary citizens. This article demonstrates that attention to politics tends to increase the resonance of this normative framework. The analysis identiﬁes three norms about the holding of public ofﬁce that are expected to structure citizens’ ethical judgments:the avoidance of conﬂicts of interest;conformity with the law or institutional rules;and the maximisation of the public good. The article assesses the importance of these norms in structuring judgments by means of an experiment embedded in a population survey conducted in Great Britain.The analysis ﬁnds that informational cues pertaining to conﬂict-of-interest avoidance only condition responses among the attentive, while information pertaining to law conformity has far wider resonance. This ﬁnding has implications for approaches to political ethics focusing on normative considerations that appear to have low salience for much of the general public.


Introduction
Maintaining the trust of the citizenry is crucial for politicians if they are to retain legitimacy, yet there is ample evidence to suggest that political leaders are very often believed by those they represent to behave unethically.There is also an extensive political theory literature that has elaborated normative frameworks for evaluating ethical standards among elected representatives -frameworks that are in many cases reflected in institutional codes of conduct and other practical instruments employed to regulate the behaviour of politicians.However, neither strain of research has succeeded in demonstrating adequately how members of the public actually go about forming their evaluations of elected leaders and whether their judgments reflect the norms established by political theory.This article seeks to fill this gap by drawing on recent developments in political psychology to examine how people respond to informational cues when evaluating the ethical conduct of those who represent them and the role of attentiveness to politics in shaping such responses.
In considering the normative frameworks that are likely to shape citizens' evaluations of political ethics, we focus on values that are deeply embedded in contemporary developed democracies as well as the normative political theory subtending such values.Our analysis thus serves to provide both empirical elucidation of ethical judgment and evidence of the relevance of normative theory conceptions of political ethics to such judgment.We argue that those who are most attentive to politics and public affairs are likely to respond to informational cues reflecting nuanced ethical norms, whereas those less well versed in politics will likely use simpler rules of thumb.Specifically, we propose a multidimensional conceptualisation of ethical judgments in politics based on three norms about the holding of public office that are expected to structure citizens' responses to information: the nuanced norms of (i) avoiding conflicts of interest and (ii) delivering outcomes that maximise the public benefit; and the simple norm of (iii) conformity with the law or institutional rules.We assess the impact of each normative frame on the basis of data from an experimental survey question fielded in Great Britain.In so doing, we examine the extent to which political attentiveness affects citizens' propensity to respond to each of the three normative frames when expressing opinions about political conduct.
The next section provides a theoretical and practical grounding for our analysis.We then introduce the data and the methodological approach employed in the empirical analysis before presenting our empirical results.The final section discusses these results and concludes.

Judging politics and politicians
Existing attitudinal research into citizens' judgments about political ethics and perceptions of corruption has shed light on questions of what, who, where and how much.In practice, most studies in the existing corruption perceptions literature tend to fall into one of two categories: the first involves judgments about the propriety or corruptness of given acts and what citizens regard as acceptable or unacceptable behaviour; the second is concerned with judgments about the prevalence of misconduct or corruption in a given institution, system, region or nation.We know that people's overall views of elected politicians are generally negative (though many citizens are favourably disposed toward individual politicians), and there is evidence as to the forms of misconduct citizens find most worrisome (e.g., Atkinson & Bierling 2005;Dolan et al. 1988;Jackson & Smith 1996;Johnston 1986Johnston , 1991;;Mancuso et al. 1998;McCann & Redlawsk 2006;Peters & Welch 1978) and how people react to politicians' justifications of their actions (Chanley et al., 1994;Gonzales et al., 1995).We also have some idea of which types of people are most likely to be critical of political elites, and which groups are generally more tolerant of ethically dubious behaviour (e.g., Allen & Birch 2011, 2012;Blais et al. 2005;Davis et al. 2004;Gonzales et al., 1995;Johnston 1986;Redlawsk & McCann 2005;Steenbergen & Ellis, 2006). 1 There has been relative neglect, however, of the question of how ethical evaluations of politicians are formed.This article addresses this 'how' question -understood in terms of the normative frameworks that inform ethical judgment -by drawing on recent research in political psychology.The political psychology perspective has not often been used in the study of corruption perceptions, 2 and we believe it sheds important new light on judgment formation that will be of relevance to normative theorists and to policy makers, as well as to students of public opinion.
In many cases, evaluations of politicians' conduct are better understood as political judgments rather than moral judgments per se.Yet, insofar as it is possible to distinguish clearly between them, we are interested in the latter and particularly in judgments about the rights and wrongs of conduct.Political theorists and those charged with regulating institutional political conduct, such as parliamentary ethics officers and committees, have elaborated understandings of political ethics in terms of norms that pertain to the role of elected office holders (e.g., Alexandra 2007: 89;Hampshire 1978: 48-52;Philp 2007: 152-163;Thompson 1987: 96-122;1995: 11-25).These norms almost always rest on a distinction between 'the public' and 'the private', according to which the actions of public officials are distinguished from those of individuals in their private lives or in the private sector by virtue of the fact that they are chosen to act in the public interest.
The key feature that marks out the performance of a public or political role, as opposed to a private role, is that it involves a public trust.Breaches of this public trust are often seen by theorists and regulators as among the most significant forms of political wrongdoing.And since such breaches commonly result from conflicts of interest between the actors' public duty and their personal or partisan interests, the avoidance of conflicts of interest is usually a central preoccupation of theorists' and regulators' conceptualisations of political ethics (Committee on Standards in Public Life 1995: 19-45;Thompson 1995: 49-76).
At the same time, there is reason to believe that most ordinary citizens have little appreciation of theoretical and elite conceptualisations such as these.Though theories about political conduct may make their way into the educational curriculum via classes in civics (or 'citizenship education' as it is called in the United Kingdom), relatively nuanced normative ideas are unlikely to figure prominently in the day-to-day judgments of those who take little interest in public affairs.Most people can be expected to take shortcuts of one sort or another in forming their ethical evaluations of political leaders.
Research in political psychology can help to shed light on the kinds of shortcuts that citizens might employ.In line with recent trends in moral psychology, we expect that most people will devote little effort to making ethical judgments about politicians, and that they will thus use less-demanding forms of 'low information' or 'peripheral processing' to make most evaluations (Kuklinski & Quirk 2000;Lieberman et al. 2003;Sears, 2001;cf. Marcus et al. 2000).Models of motivated reasoning further suggest that most people's judgments most of the time are determined by their initial intuitive responses, and that these intuitions are themselves a reflection of prevailing group norms and/or are shaped by prior reasoned persuasion (Cassino & Lodge 2007;Haidt 2001Haidt , 2012;;Lodge & Taber 2000, 2013;Taber et al. 2001;Wagner et al. 2014).Such group norms and beliefs can serve as frames, which simplify judgment making by giving 'meaning to key features of some topic or problem' (Lau & Schlesinger 2005: 80; emphasis in original).The literature on cognitive bias further suggests that citizens can be expected to react strongly to the substantive outcome of an action, and that loss-aversion will induce them to evaluate especially negatively actions whose outcomes result in a diminution of benefit (Arceneaux 2012;Kahneman & Tversky 2000;Tversky & Kahneman 2000).
The framing effects of elite discourse (Gamson & Modigliani 1989;Nelson & Kinder 1996;Tversky & Kahneman 2000;Zaller 1992) are thus particularly relevant to our conceptualisation of how people judge politicians' conduct, since such judgments are unlikely to be the product of belief change -that is, the acquisition of new positive or negative information about an attitude object that was not formerly part of an individual's belief structure and which causes a change in opinion.Rather, ethical judgments will in most cases be instant responses to the information presented, the details of which will activate 'information already at the recipients' disposal, stored in long-term memory' (Nelson et al. 1997: 225).
We are interested in identifying those normative cues to which the public is likely to respond when judging political acts and situations from an ethical point of view.Each cue will thus reflect a different understanding of what is proper behaviour for public officials.In seeking to develop hypotheses, we first recognise that the norms against which behaviour in public life is judged will reflect, at least in part, distinct democratic cultural traditions.Given that the data on which we draw are from Great Britain and pertain to Members of Parliament, we identify three sets of culturally embedded norms that are likely to be especially relevant in understanding the propriety (or otherwise) of their conduct. 3We do not claim that these are the only norms that are likely to frame citizens' ethical judgments; but they do represent particularly important ideas that structure expectations of elite conduct.Despite being drawn from British parliamentary practice, we also expect that, with minor cultural variations, these norms are also relevant in most other established democracies.
The first norm reflects a central preoccupation of most theorists and ethics regulators, and understands ethical conduct as the avoidance of conflicts of interest between an office holder's public duties and his or her private interests (Allen 2008;Mancuso 1995;Thompson 1995).In the British context, Members of Parliament have long been expected to represent their constituencies, individual constituents and wider interests, as well to act upon their own judgment.Yet it is well established that they should not profit from their position or influence parliamentary proceedings for personal, private gain (Allen 2011;Judge 1999;Williams 1985).As the current Code of Conduct for MPs states very clearly: 'Members shall base their conduct on a consideration of the public interest, avoid conflict between personal interest and the public interest and resolve any conflict between the two, at once, and in favour of the public interest' (House of Commons 2012: 3).The conflict-ofinterest avoidance norm takes us closest to the classic definition of political corruption as 'abuse of public office for private gain'.It is also the norm that is most specific to the evaluation of political ethics.The role of public service norms and related understandings of what is proper behaviour for public officials has been found to be prominent in various previous empirical studies of ethical reasoning (Dolan et al. 1988;Alvarez & Brehm 2002;Redlawsk & McCann 2005).In some cases there has been found to be a political-cultural element to such evaluations, which may vary from one cultural context to another (Frohlich & Oppenheimer 2000).
The second norm construes political ethics and conduct in act-utilitarian terms, specifically in terms of substantive outcomes and the maximisation of public benefit.An obvious feature of public office is that those in positions of authority are expected to deliver tangible benefits to those they serve.The importance of such considerations has been clearly demonstrated in MPs' own under understandings of political ethics (Allen 2008;Mancuso 1995), while the current Code of Conduct makes clear that MP 'have a general duty to act in the interests of the nation as a whole; and a special duty to their constituents' (House of Commons 2012: 5).Accordingly, if individuals can see clear positive benefits from certain behaviour or a specific act, conflicts of interest and breaches of the law, especially relatively minor conflicts and breaches, might be considered ethically tolerable if those who benefit are the socially marginalised.Conversely, if individuals perceive clear negative outcomes, ethical evaluations of dubious conduct may be more damning.These kinds of act-utility considerations may be tempered further by individuals' beliefs and values about social justice -for example, a preference for the disadvantaged to receive additional benefits as a matter of priority.Previous attitudinal research into political ethics has certainly found evidence that distributional considerations matter for some individuals.Popular ethical evaluations of certain acts can be shaped by the payoffs involved, by the identity and number of other individuals affected, by the consequences for those affected by the act and by the respondent's perception of the wider context (Chibnall & Saunders 1977;Johnston 1986Johnston , 1991;;Dolan et al. 1988;Frohlich & Oppenheimer 2000).Meanwhile, the literature on loss-aversion cited above also supports the notion that judgments may be affected by perceptions of outcomes.
The third norm reflects the much simpler criterion of legality.For many people, perhaps the most obvious way of understanding political ethics and especially misconduct may be in rule-utilitarian terms -in other words, as conformity with existing laws or institutional rules.Conduct that is in breach of either is 'wrong'; conduct that conforms to both is 'right' (Thompson 1995: 22-23).On this basis, we can hypothesise that many, if not all, individuals will judge behaviour at least in part in terms of its legality or its accordance with existing rules.This legalistic approach can in some sense be seen as minimalist, as those who employ it are relieved of the task of exercising their own independent ethical judgment; in order to evaluate a situation, they merely need to ascertain whether or not it is permitted under existing rules, regulations or laws.This approach is also potentially complicated by the fact that, in some jurisdictions, holders of certain political offices enjoy immunity from civil or criminal proceedings.In Italy and Germany, for example, constitutional provisions protect members of the national parliaments in this way, and a substantive vote is usually required by the legislature to lift this protection.In Britain, however, MPs enjoy no such immunity.As the current Code of Conduct makes clear, 'Members have a duty to uphold the law' (House of Commons 2012: 3).MPs, like other office holders, are always expected to adhere to the law, and we can anticipate that they are judged accordingly.
Our specific goal in this analysis is to examine the impact of political attentiveness on ethical judgments.We anticipate, first, that individuals who are attentive to and informed about public affairs are more likely to be able to apply the more nuanced norms to concrete political situations.This expectation chimes with evidence from neuropsychology that individuals' ability to generalise from everyday experience to specific political evaluations depends on attention to and knowledge of politics (Lieberman et al. 2003;Schreiber 2007).This is because when many mental acts are practiced repeatedly, they become habitual and require less conscious attention to accomplish.Thus for those who frequently read about politics, discuss politics with friends and pay attention to political information, the norms common to political discourse become familiar, and their application to concrete situations becomes automatic (Lieberman et al. 2003;Schreiber 2007).Working with survey data, Alvarez and Brehm also find that respondents who are better informed about politics find it easier to apply abstract values consistently to concrete situations, as 'information makes values and predispositions relevant for beliefs about public policy issues' (Alvarez & Brehm 2002: 50; emphasis in the original; cf.Chanley et al. 1994;Zaller 1992).
Accordingly, we hypothesise that the politically attentive are more likely than the inattentive to draw on norms pertaining to the distinctive duties of holding public office when making ethical judgments, and that they will be more sensitive to information about potential conflicts of interest, which is the primary focus of institutional ethics regulation in Britain.This is because those who absorb large amounts of political information on a regular basis are familiar with elite normative frameworks relevant to political life and apply these with relative ease to the political situations they encounter.For the same reason, we can also predict that responses to informational cues relating to the substantive outcome of politicians' behaviour will be more common among the politically attentive, as the above-cited evidence indicates that application of such a norm to concrete situations will be easier for the attentive as well.
Second, there is reason to believe that for most respondents, including the non-attentive, considerations of legality will have particular resonance.The rule that legal acts are acceptable and illegal acts are unacceptable has many of the characteristics of a heuristic or cognitive shortcut of the type generally employed in peripheral processing (Popkin 1991;Sniderman et al. 1991).If this is the case, then we might expect to see the non-attentive respond more readily to informational cues that tap legal norms than they would to cues tapping either of the other two norms.
A final prediction we can make is that, as Zaller (1992) and Alvarez and Brehm (2002) note, political sophisticates are more consistent in the responses they give to pollsters.Following this reasoning, we can anticipate that the politically attentive may be somewhat more adept at applying all three norms consistently to concrete political situations.

Experimental design and data
The role of norms in framing individuals' ethical judgments is examined here through the analysis of experimental data incorporated into a 2009 survey in the United Kingdom. 4  Embedded in the survey was a 2×2×2 factorial vignette designed to operationalise the three norms we predict to be associated with judgments of the behaviour of political elites.
Survey experiments based on vignettes combine the internal validity inherent in experiments with the external validity of sample surveys (Gibson 2008;Sniderman & Grob 1996).They are particularly useful for the analysis of social judgments, which is the context in which the technique was originally developed by sociologists (Rossi & Anderson 1982).Moreover, this technique has been used in previous research to analyse framing effects (Kinder & Sanders 1990;Sniderman & Grob 1996).The vignette-based experiment is therefore an ideal tool to analyse how informational cues condition evaluations of political elites and thus which normative frameworks respondents draw on when making their judgments.
The experimental vignette employed in this analysis took nine versions, one of which was a control treatment describing the following scenario involving a fictitious politician Susan Barnes: 'MP Susan Barnes helps a firm whose headquarters is in her constituency.'The eight other versions of the vignette were treatments reflecting permutations of the three norms set out above.Each of these was embodied in two versions of a piece of qualifying information that served as a cue.In one version, Barnes acts in a manner consonant with the established norm in question -avoiding conflicts of interest (the conflict-of-interest norm), acting in accordance with the law (the legality norm) and delivering a tangible benefit for constituents (the public-benefit norm) -and in the other version she violates the norm in question. 5 The conflict-of-interest statements were crafted so that one version of the vignette (version A in Table 1) reflects the absence of any obvious conflict of interest: 'The firm has never made any donations either to Barnes or to her party', whereas the other (version B) implied conflict of interest: 'The firm has for several years made regular donations to Barnes's party.' The law-conformity statements were designed so that one version (A) describes conventional representative behaviour by British MPs, which suggests that the action in question conforms to rules to which politicians are subject:'Barnes writes a letter to the Minister The first version (A) of the public-benefit statement described an outcome that suggested the action taken might potentially be justified on the basis of its social consequences: 'Closure of the firm would result in the loss of 1,200 jobs.'The second version (B) suggested that the action would result in minimal social benefit: 'Closure of the firm would result in the loss of a dozen jobs.' The details of each version of the vignette are summarized in Table 1. 6Each respondent was randomly assigned to receive one of the nine versions of the vignette.
Respondents were then asked to judge Barnes's behaviour by means of the following question: Please indicate your opinion about Barnes's help for the firm using this 0 to 10 scale.The wording of this question was designed to elicit what might be termed an 'everyday' ethical response to the action in question.A higher score on this question reflects as straightforwardly as possible the perception of normative unacceptability or disapprobation.Many previous studies that have asked respondents to judge the propriety of politicians' actions and certain types of conduct have used questions that explicitly refer to 'corruption' (Allen 2008;Atkinson & Mancuso 1985;Jackson & Smith 1996;Johnston 1986).Corruption is a morally loaded concept, however, which means different things to different people.Since levels of political attentiveness may well affect how the term is understood, we avoid it in the present study. 7

Results
The mean normative approbation scores for each version of the vignette are displayed in Table 2. 8 The figures in this table demonstrate that all three of the manipulations had the expected effect.Respondents who received 'negative' versions of the vignette were more likely to judge the actions of Barnes to have been unacceptable, and the differences in means are in all cases significant at the 0.01 level or higher.
The results suggest that all three of the norms -avoiding conflicts of interest, abiding by the law and delivering a public benefit -play a role in shaping ethical evaluations of politicians' conduct.Interestingly, even those respondents given the 'positive' versions of the vignettes had fairly jaundiced views of the behaviour of the fictional Barnes, indicating that people are generally inclined to judge critically situations with even a whiff of ethical dubiety.Those harsh judgments are clearly exacerbated, however, by cues which tend to confirm the notion that the MP had behaved wrongly.Most important from our perspective is the fact that respondents' views appear to respond systematically to the norms embodied in the manipulations.
The impact of the manipulations can be probed in greater depth through multivariate analysis, which takes acceptability perceptions as a dependent variable (see Table 3).
Respondents were coded as to whether they received a 'positive' or 'negative' version of each vignette, and these codings were used to create independent variables.Specifically, the 'negative' versions of each manipulation were entered into an ordinary least squares (OLS) regression model, together with the interactions between them (as is conventional in analyses of factorial vignettes).The basic model is presented in the second column of Table 3.
The regression results refine the understanding gained through bivariate analysis above; the conflict-of-interest and legality manipulations have statistically significant impacts on judgments of political behaviour, with the substantive-outcome manipulation having no perceptible influence.Tellingly, the legality manipulation has by far the largest coefficient and is most strongly significant.This suggests that when an act is framed in terms of law-breaking, this sends a very strong signal to members of the public that it is unacceptable, regardless of the other contextual information with which they are presented.The impact of the informational cue based on conflict-of-interest norms is also significant, but of considerably lesser magnitude and lower significance.The non-significance of the publicbenefit variable is somewhat puzzling given the strong tradition of constituency service in Britain; this could potentially be due to the peculiarities of vignette design (the difference between the loss of 1,200 and 12 jobs might not have had sufficient resonance for most people), or it could be that the substantive outcomes of politicians' behaviour are genuinely not as relevant to people as is conformity to procedural norms, as suggested by some literature on legitimacy (Major & Schmader 2001;Tyler 2001).
A caveat is in order here, however.The nature of vignette-based survey experiments is such that it is not possible to compare directly across manipulations and thus compare the impacts of the different norms they are designed to operationalise.The wording of vignettes is by definition highly idiosyncratic, such that it is virtually impossible to design manipulations with precisely equivalent impacts.Further research employing a different research design would be necessary to probe the weakness of the public-benefit frame in comparison to the framing effects of the legality and conflict-of-interest norms.The research design employed here does allow us to compare the impact of each norm across subcategories of respondent, however, and it is to this that we now turn.As noted above, we are primarily interested in investigating the impact of political attentiveness on how culturally embedded norms impact on moral judgments.Political attentiveness was operationalised for this purpose by means of a variable designating attentiveness to 'what is going on in government and public affairs', which we expect to be associated with a greater propensity to understand impropriety in terms of conflict-of-interest and public-benefit norms.To test this and the other hypotheses outlined above, the sample was divided into highly attentive and less attentive groups, 9 and the regression was run on the two sub-groups.The third and fourth columns of Table 3 present these models.
The main effects in this model show that when judging the acceptability of Barnes's behaviour, the minority of the sample (approximately a third of the total) who claimed to be highly attentive to public affairs responded to cues related to legality as well as those designed to reflect the presence or absence of a conflict of interest, as expected by our understanding of the role of political attentiveness in conditioning ethical judgments.In contrast, the non-attentive responded to legal norms alone.This finding suggests that, as expected, normative concerns about conflicts of interest have resonance mainly for that portion of the public which is highly attuned to politics.Considerations about legality or rule-compliance, by contrast, would appear to serve as a convenient rule of thumb that enables even the non-attentive to judge a situation.The results for the public-benefit norm were again not significant for either group.
Analysis of the interaction term in this model enables us further to refine our understanding of the relative importance of the different norms.The negatively signed coefficient on the interaction between conflict of interest and legality demonstrates that when an action is illegal, the impact of conflict-of-interest considerations is virtually obliterated.In other words, legality 'trumps' conflict of interest, suggesting a hierarchical ordering of norms in the minds of the attentive: the most important consideration is whether or not an action is legal, and only when it is legal (or there is no indication that it is not legal) does the conflict-of-interest norm come into play. 10 In order to probe the data further, we also segmented them according to actual (as opposed to reported) attentiveness.For this purpose we chose an indicator of newspaper readership, from which we constructed a dummy variable designating readership of what are commonly referred to in Britain as 'quality' newspapers (see online supplementary material for details).These are newspapers that are held to adhere to relatively high journalistic standards and to report news stories in depth (Kuhn 2007).We assume that most people derive most of their knowledge of political elites from news sources.The specific sources into which they tap can thus be expected to be one of the main determinants of their degree of awareness of the details of political life.In addition, newspaper consumption is far more likely to reflect a conscious choice than the consumption of broadcast news, which may well depend on fortuitous factors.
Given the considerable diversity in news coverage among newspapers, it stands to reason that those who read 'quality' newspapers should have a more nuanced appreciation of elite-level ethical norms, including preoccupations with conflicts of interest, than those who derive their information from other sources.And this is precisely what we find.The fifth column in Table 3 reports a model based on readers of 'quality' papers alone.Though this is a small sample -representing approximately 17 per cent of the total -we nevertheless obtain strongly significant results on our core variables of interest.The impact of the legality manipulation remains the strongest of the three informational cues, but we see that the conflict-of-interest cue is also of considerable magnitude and significance in this model.The public-benefit cue remains non-significant.As in the model of self-identified 'attentive' respondents, the significant negative coefficient on the interaction terms between conflict of interest and legality suggests that the presence of a conflict of interest only becomes a relevant consideration for respondents when legality is not at stake.The final column of this table contains a model based on those respondents who did not report reading 'quality' newspapers.As can be readily ascertained, the conflict-of-interest cue is far from being significant in this model.
These findings suggest that the information to which people are exposed plays an important role in conditioning their attitudes toward the actions of politicians.Though the legality of an action appears to have universal resonance, conformity to norms specific to the public role of politicians conditions reactions only among the small minority of the population that is highly attentive to politics and consumes quality newspapers.And even among the attentive this norm only kicks in when legality is not at stake.That norms commonly deployed in elite discourse should be an important consideration for such a small sector of the population poses serious questions for the assumptions subtending a range of theoretical and practical approaches that have been taken to the question of political ethics.

Discussion and conclusion
Our findings indicate that people are critical of politicians, but they are not blindly critical.When judging behaviour they respond to contextual information in ways that are consistent with norms embodied in political culture, and they respond largely in the ways we would expect on the basis of common understandings of political ethics.However, most do not respond in the ways prescribed by theoretical accounts of public ethics.In particular, concerns with avoiding conflicts of interest may under certain circumstances shape the evaluations of those who are attentive to politics, but simple considerations of legality appear to weigh more heavily for the less attentive.And even among the attentive, legality 'trumps' the more nuanced conflict-of-interest norm.
These findings have significant implications for both theoretical and practical approaches to political ethics, which have largely ignored research in political psychology.Normative accounts of political ethics and the institutional regulation of political ethics are conducted largely on the basis of role-specific understandings of proper behaviour by those in public life.Politicians have a public trust, which means that they are bound to put the public good above their own private interests and to avoid conflict-of-interest situations where possible in the execution of their public duties.This sort of argument appears to have very little resonance with large section of the general public, however, when responding to information about political conduct.Our findings suggest that the way in which elites understand and deal with ethical matters pertaining to politics is at some remove from the way in which the majority of the public approaches these issues.This gap could go some way toward explaining why the solutions proposed by elites to declining trust in politicians are very often unsuccessful.
These findings are also relevant for the debate as to whether rules or culture are better means of ensuring high ethical standards among politicians (Atkinson & Mancuso 1985;Committee on Standards in Public Life 1995;Flinders 2012;Riddell 2011).Our results suggest that statutory regulation is a more robust means of keeping politicians on the perceived straight-and-narrow than non-binding codes of conduct or general principles of appropriate behaviour.If most people employ the evaluative short-cut of rule-compliance as a means of judging politicians, then it follows that relevant norms need to be embodied in rules, and these rules need to be enforced.
The results of this article are of particular relevance to Britain, which is the empirical focus of the analysis.Given the nature of contemporary media coverage of British politicians' real and alleged indiscretions, it is not surprising that so many sections of the British public take a generally dim view of their politicians' integrity (Allen & Birch 2011;Birch & Allen 2010).Widespread ethics reforms such as those undertaken in Britain at several points since the 1990s have done little to improve confidence in elected politicians.There is even some evidence to suggest that recent increases in transparency may actually have fuelled suspicion rather than alleviating it (Newell 2008).The analysis presented here sheds light on why this might have been the case, as it identifies a worrying gap between elite and mass understandings of political ethics.This finding suggests that greater emphasis could profitably be placed on role-specific political norms in citizenship education and other educational initiatives designed to instill democratic values in young people.
The results will also be of interest to political psychologists in that virtually all of the political psychology research on which we draw in developing our theoretical expectations has been carried out in the United States on American subjects.It will be of value to political psychologists to see that their findings hold also in the United Kingdom, despite the cultural differences between the two settings.
Finally, this article has shown that survey experiments provide a fruitful means of probing patterns of ethical judgments of politicians.Further research could usefully explore the stability of such evaluations over time, and their sensitivity to other types of contextual cues.
3. For a more detailed exposition of these norms, see Allen and Birch (2015).4. The survey formed part of the British Co-operative Campaign Analysis (B/CCAP) multi-wave survey project led by Raymond Duch, Simon Jackman and Lynn Vavreck.The wave of the survey used in this analysis was administered online in September 2009 by the YouGov polling agency.The total sample size was 1,978 and the achieved sample was 809, or 40.1%.The sample has been weighted to the demographic profile of the population.Recent analyses have established the validity of online surveys (Sanders et al. 2007;Twyman 2008;Vavreck & Rivers 2008). 5.The precise wordings of the different vignette versions are included in the supplementary material available online.6. See also the online supplementary material for the full wording of each version of the vignette.7. Respondents were also asked two comprehension questions in order to test that the vignettes were understood in the intended way.We asked first whether Barnes had broken the law when she helped the firm.The mean evaluation was significantly more unacceptable at the 0.0001 level for those responding that the law had been broken (the t-statistic for difference of means tests was 25.19), suggesting that the respondents were correctly influenced by the experimental manipulation (in other words, they identified a breach of the Official Secrets Act as a violation of the law but not the act of letter-writing).A total of 70.39% answered correctly, 7.82% answered incorrectly and 21.79% said they didn't know.We also tested the appropriateness of the public interest manipulation by asking respondents whether the fate of the firm would greatly affect the local economy or not.Those who thought the fate of the firm would affect the local economy were significantly less likely at the 0.0001 level to judge Barnes's behaviour to be unacceptable (t-statistic, −6.60), which again indicates that respondents were being appropriately influenced by the experimental manipulation.A total of 48.6 per cent answered correctly, 20.11 per cent answered incorrectly and 31.28 per cent said they didn't know.The high number of 'don't knows' suggests that respondents struggled with this question somewhat, but we are reassured by the proportion providing correct answers is more than double the proportion providing incorrect answers.8.More detailed results can be found in Appendix Table 1. 9. The highly attentive group included respondents who claimed to follow what is going on in government and public affairs 'most of the time' or 'some of the time', whereas the less attentive group included those who claimed to follow such events 'only now and then' or 'hardly at all' (see the online supplementary material for precise question wording).10.The frequencies presented in Appendix Table 1 confirm this point; the mean evaluation of those respondents given the baseline version of the vignette -which included no indication of the legality/ illegality of the action -yield a more positive rating of the politician than those versions of the vignette where her action was specified as being illegal.It is worth noting that the coefficient could also be interpreted to mean that legality is less salient to respondents when there is a perceived conflict of interest, but given the overwhelming importance of the legality norm in the models presented here, the former interpretation of the coefficient is undoubtedly of greater relevance to the argument under consideration.

Table 1 .
Structure of vignette experimental manipulations (MP Susan Barnes helps a firm whose headquarters is in her constituency . ..)Industry asking the ministry to intervene to save the firm.'The other version (B) signals clear law-breaking: 'Barnes shows the firm confidential government documents that give it a competitive advantage.In so doing she breaches the Official Secrets Act.'

Table 2 .
Effects of vignette manipulations on corruption perceptions

Table 3 .
Variation in judgments of acceptability/unacceptability: Ordinary least squares regressions Variable