HOW SHOULD WE MEASURE POLITICAL TRUST?

Ben Seyd, University of Kent

Political trust represents a key indicator of how closely citizens relate to political actors; trust

also has significant effects on citizens behaviour in ways that affect the effectiveness of democratic governance. Yet in spite of the profile and importance of the concept, little

scholarly attention has been paid to the way political trust should be measured. This would

not matter if existing indicators of the concept were valid and reliable. However, this is by

no means clear. Most empirical studies of political trust rely on simple – arguably simplistic

– ways of measuring the concept, based on indicators whose validity is potentially limited. If

scholars are to improve the way they measure political trust, more rigorous and robust

indicators should be developed. This article identifies various such measures, drawing on

analyses conducted by scholars in the field of organisational studies. However, the task of

identifying suitable indicators is complicated and involves potential trade-offs. These must

be faced and resolved if new indicators – capable of providing more robust measures of

political trust – are to be developed.

Paper for PSA annual conference, Brighton, 21st-23rd March 2016

1

In understanding how citizens relate to political authorities, trust is seen to play a central

role. Political trust provides an important indicator of how positively citizens regard

decision-making actors and institutions. Trust also shapes how citizens behave in ways that

affect the health of democratic governance. Thus, people s feelings of trust towards

government have been found to shape their propensity to pay taxes (Scholz and Lubell,

1998) and to comply with collective obligations (Marien and Hooghe, 2011). Given these

effects, it is troubling that levels of trust among citizens in advanced democracies appear to

be low and, in many cases, to have declined in recent decades (Dalton, 2004). Yet our

grounds for concern are largely based on accepting that the conventional empirical

indicators of political trust, based on questions fielded on opinion surveys, provide an

accurate way of measuring the concept. In fact, the profile and salience of political trust as a

topic has not been matched by considered and careful empirical measurement of the

concept. Attempts made up to fifty years ago to develop robust measures of support for the

political system (eg. Agger et al, 1961; Finifter, 1970; Craig et al, 1990) have generally not

been followed up. While considerable effort has been devoted to devising measures of trust

in other fields – most notably in organisational studies (McEvily and Tortoriello, 2011) and

risk management (Earle, 2010) – studies of trust in the political sphere have tended to rely,

usually uncritically, on a set of primitive and well-worn indicators. However, these

indicators suffer from well-known methodological weaknesses, which undermine their

ability to serve as robust measurement instruments. There is thus a substantial gap between,

on the one hand, the contemporary interest in, and importance attached to, political trust

and, on the other hand, the rigour in the way the concept is measured.

The weakness of existing indicators of political trust has been noted by a number of scholars.

For example, having identified the propensity of researchers to gauge trust through simple

single-item survey questions, Gershtenson and Plane (2011: 132) observe:

… there is a dearth of serious evaluations of alternative trust measures. Without empirical evidence that one measure is better or worse than others, political science research on trust is

likely to remain fragmented, and thereby limited in its ability to truly understand the

relationship between political trust and other variables.

Given the importance of political trust, and its relevance in shaping patterns of citizen

behaviour, it is perplexing that scholars have based their research so unquestioningly on a

limited and simple set of indicators, and have devoted so little energy to considering

whether alternative indicators might do a better job in measuring the concept. This article is

intended to provide an initial foray into that task. It begins by considering how political

trust has conventionally been assessed, and reviews the main shortcomings of these

indicators. It then identifies some basic principles by which political trust might more

accurately be measured. This task is, however, far from straightforward and involves

various trade-offs depending on the purpose being pursued by the analyst. These different

purposes, and the trade-offs they entail, are laid out. At this point, the discussion draws

heavily on the measurement of trust within the organisational science literature. I suggest

that attention to the work of scholars in this field might help in formulating ways of gauging

trust more effectively in the political realm. However, comparing relations in the contexts of

the workplace and the political sphere also highlights various differences in the nature, and

thus in the measurement, of trust across the two fields. While political scientists can learn

2

much from the principles underpinning the measurement of trust in the organisational field,

not all of these principles are readily transferable to the political field. The analysis

concludes by highlighting alternative ways of measuring political trust, and considering

how far these constitute an improvement on existing practice.

1. Conventional methods of measuring political trust

Political trust was first measured via a battery of survey questions fielded on the US National

Election Study (NES) in 1958 (item 2 from 1964). This battery comprised the following items:1

1. How much of the time do you think you can trust the government in Washington to

do what is right?

2. Would you say the government is pretty much run by a few big interests looking out

for themselves, or that it is run for the benefit of all people?

3. Do you think that the people in government waste a lot of the money we pay in

taxes, wastes some of it, or don t waste very much of it?

4. Do you think that quite a few of the people running the government are a little

crooked, not very many are, or do you think hardly any of them are crooked?

These items are still used by the NES to measure trust, albeit that the wording of the fourth

item has been altered, with corrupt replacing crooked . However, although widely taken

to measure political trust, the items were originally designed to tap people s basic

evaluative orientations toward the national government (Stokes, 1962: 64), and in particular

their assessments of officials honesty, competence and procedural correctness. There was no

justification of the questions status as measures of political trust, and indeed no reference to

the concept, in Stokes (1962) account.

Hence, in spite of the NES items longevity, there are doubts over how far the indicators

genuinely and accurately tap political trust. Indeed, rather than tapping attitudes towards

the political system in general (what is normally assumed by the notion of political trust;

Easton, 1975), the NES items have instead been found to tap evaluations of incumbent

political actors (Citrin, 1974; Hill, 1981). Moreover, the relationship of the individual

indicators to one another has been found to be inconsistent (Poznyak et al, 2014), again

highlighting the tenuous relationship between the individual indicators and the background

concept.

However, while the NES persists with these measures, their potential flaws are at least well-

documented. Far less attention, however, has been paid to the ways in which subsequent

measures of political trust have developed. While various surveys outside the US have

replicated the NES items (eg. Muller, 1979: 281-2), most national and international surveys

have eschewed them, largely in favour of simpler indicators. These indicators tend to

manifest three features: they comprise a single item, they explicitly mention trust, and they

1 A fifth question was initially used to measure political trust, this time gauging whether respondents

thought that the people running the government are smart people who usually know what they are doing … or that quite a few of them don t seem to know what they are doing? . However, this question was fielded on the NES only until 1980. See http://www.electionstudies.org/.

3

contain no referent against which trust is to be gauged.2 Typically, trust is measured by a

single survey item worded along the following lines: How much do you trust [institution]

in [country]? Given the controversy over the NES items and what they measure, this

departure from the multi-item NES model is hardly surprising. Moreover, since single-item

measures take up less time within crowded survey schedules than do multi-item measures,

this practice frees up survey space which can be used to pose questions on trust about

various different political institutions.3 Yet while researchers now routinely resort to single-

item direct indicators to gauge political trust, there is no evidence that this practice provides

a more robust form of measurement than the widely criticised NES indicators. In particular,

single-item direct indicators may (i) only weakly tap the background concept, (ii) trigger

variations in understanding among survey respondents, (iii) fail to capture ambivalence in

attitudes among respondents, and (iv) fail to provide a generalised, as opposed to an

incumbent-driven, assessment of the political system.

The first concern relates to the notion that trust is a simple concept capable of direct

measurement. Single-item survey measures assume a one-to-one correspondence between

the indicator and the background concept. This assumption is sometimes made explicit

(although more often remains only implicit); for example, Selnes (1998), Metlay (1999) and

Lang and Hallman (2005) argue that trust is a unidimensional construct that, moreover, is –

as Selnes (1998) argues – directly accessible to survey respondents.4 Yet even if a concept is

found to be unidimensional, this does not entail that it can adequately be measured through

a single-item indicator, unless it can be shown that the indicator fully encapsulates the

meaning of the concept.5 Moreover, theoretical and empirical analyses of political trust

usually suggest that the concept is not unidimensional, but instead contains various

dimensions (relevant studies from the fields of political science, risk management,

organisational science and health studies include: Citrin and Muste, 1999; Grimmelikhuijsen

and Knies, 2015; Johnson, 1999; McEvily and Tortoriello, 2011; Earle, 2010; Ozawa and

Sripad, 2013). If trust does comprise various components or dimensions, it would seem

unlikely that any single-item indicator is able to fully capture the concept.6 Moreover, not

only are such indicators unlikely to fully map the concept of interest, but their reliability is

2 The International Social Science Programme is an exception in maintaining the NES measure of trust in

government with the referent to do what is right . 3 For example, the World Values Survey asks about people s confidence in organisations, while Eurobarometer asks about their trust in nine national organisations plus EU institutions. 4 Hetherington suggests that trust in government is a simple concept, about which almost

all people will express true attitudes . However, he provides no evidence to justify this claim. 5 A single-item indicator may be justified if this is held to provide a summary, or global , measure of a concept. Here, although a concept may be held to comprise various elements or components, a

single indicator is capable of capturing the essence of the concept that underlies each of these

elements (Fuchs and Diamantopolous, 2009: 200; Petrescu, 2013). Yet among the numerous studies

that rely on the conventional single-item indicator of trust, little evidence – or even argument – is

produced to justify the indicator on these grounds. 6 When it comes to using trust as an independent variable to explain various outcomes, multi-item

indicators of concepts such as social trust have been found to yield a stronger set of relationships than

single-item indicators (Zmerli et al, 2007: 46-50), presumably because the former provide a better

measure of the concept than the latter.

4

difficult to assess, while they are also prone to measurement error (Churchill, 1979; Zeller

and Carmines, 1980: ch3; deVellis, 1991; Spector, 1992: 4-5; Heath and Martin, 1997).

Many other concepts closely linked to political trust are deemed by scholars to be multi-

dimensional in nature and hence capable of measurement only through multiple indicators

rather than single ones. This list includes: trust in risk management (Poortinga and Pidgeon,

2003; Allum, 2007) and in the healthcare system (Egede and Ellis, 2008), confidence in the

legal system (Gibson et al, 2003; Hamm et al, 2011; Jackson et al, 2011) and the political

system (Keller et al, 2013), societal unease (Steenvoorden, 2015), interpersonal trust

(Yamagishi and Yamagishi, 1994), dispositional trust (Frazier et al, 2013), political efficacy

(Morrell, 2003), political cynicism (Cheng et al, 2012), democratic norms (McClosky and

Zaller, 1984), democratic support (Magalhaes, 2014) and system justification (Kay and Jost,

2003; Rutto et al, 2014). Indeed, early efforts to measure concepts like political cynicism

(Agger et al, 1961), political alienation (Finifter, 1970) and political trust itself (Craig et al,

1990) employed multi-item measures. The assumption behind these measures is that the

concepts are latent rather than being directly observable and, moreover, are multi-

dimensional, thus requiring multiple indicators to fully tap their various components. Yet

this insight has not been applied – in recent years at least – to the measurement of political

trust.

The second concern with the conventional measure of political trust is that survey questions

that explicitly reference potentially complex concepts like trust might induce different

understandings of the concept among respondents. Studies on interpersonal or social trust

have suggested that single-item indicators provide imperfect measures where the meaning

of the concept is imprecise or variant (Miller and Mitamura, 2003; Sturgis and Smith, 2010;

Bulloch, 2013; Crepaz et al, 2014). We know from qualitative research that complex terms

such as democracy are interpreted in different ways within and across populations

(Canache et al, 2001; Carnaghan, 2010; Schaffer, 2010). Yet one principle in the design of

survey questions is that these should be understood in comparable terms across the

population (Fowler, 2009: 91-94). Where survey questions are understood and interpreted

differently among respondents, we cannot rule out the possibility that any variations in the

distribution of responses – between groups or over time – might reflect differences in

comprehension and interpretation rather than in substance (Sturgis and Smith, 2010: 89).

Relatedly, the third concern with conventional measures is that survey items that squeeze a

range of potential evaluations into a single expressed opinion risk understating the level of

uncertainty and ambivalence in people s attitudes. Citizens may judge political actors and

institutions in both positive terms ( the government is competent ) and negative terms ( the

government is not concerned about people like me ). Qualitative research has suggested

that one reason people find it difficult to answer questions about general concepts such as

confidence or trust is because they feel unable to summarise competing – positive and

negative – feelings towards an object (Charlton et al, 2011).7 Studies of “merican citizens

7 Faced with summarising their opinions on a potentially complex issue like trust in a single general

survey measure, many people may avoid careful consideration of the factors or dimensions that

underlie the concept – particularly where this may induce ambivalence – and instead resort to a more

immediate reaction. This response is likely to reflect people s general feelings about the object under

5

attitudes towards their political institutions have identified widespread ambivalence:

between 50 to 70 per cent of Americans have been found to hold attitudes towards Congress,

the Supreme Court and the incumbent president that contain both positive and negative

feelings (McGraw and Bartels, 2005; also Gainous et al, 2008). Similarly, studies that have

examined public scepticism towards politicians – by including in their trust scale a mid-

point explicitly labelled as an absence of trust or distrust – have found high proportions of

citizens aligning themselves at this point (Mishler and Rose, 1997; Cook and Gronke, 2005;

also Gainous et al, 2008). Although single-item questions may be accompanied by response

scales that include such mid-points to capture the absence of clear views or the presence of

conflicting opinions, ambivalent feelings about political actors and institutions are better

gauged through multi-item measures that allow participants to respond positively on some

components of trust, but negatively on others.

The fourth concern about the use of single-item indicators to measure trust goes back to the

debate – now four decades old – over whether these indicators tap a particular or

generalised evaluation of the political system. Even where single-item trust measures lack

an explicit reference to a particular government, there remain concerns that these indicators

tap people s feelings about incumbent political actors more than their feelings about the

operation of the political system. In the case of attitudes to the US Supreme Court, for

example, it has been shown that single-item measures of confidence tap people s feelings about the institution s performance rather than its broader standing or legitimacy (Gibson et

al, 2003). Moreover, in a study of US citizens attitudes towards Congress, it was found that measuring congressional approval in a single-item form ( Do you approve or disapprove of

the way the US Congress has been handling its job? ) attracted a stronger influence from

evaluations of the incumbent president than did multi-item measures of congressional

approval that tapped the various dimensions of legislative support (Mondak et al, 2007: 41).

Hence, single-item questions on attitudes towards a specific institution risk tapping

judgements that are contaminated by people s feelings towards the incumbent personnel.

Multi-item measures that direct respondents towards the specific aspects or dimensions of a

concept appear less prone to such incumbency contamination.

It is clear from the earlier discussion that the NES indicators provide a contentious, and

arguably flawed, way of measuring political trust. It is not clear, however, that the way in

which most other surveys assess trust provides a more robust gauge of the concept. Yet,

while the NES indicators have attracted numerous criticisms, fewer critical lenses have been

trained on the single-item indicators of trust that are routinely used on national and

international surveys. However, these single-item indicators arguably provide an imperfect

review rather than more considered evaluations of the object s features and properties. Singular trust

questions may thus encourage schematic responses, based on general feelings and impressions, rather than piecemeal responses, based on consideration of particular positive and negative

assessments (the distinction between schematic and piecemeal reasoning is made by Fiske [1986;

also Fiske and Neuberg, 1990]). Hence, general single-item survey measures of trust questions risk

picking up superficial or routine feelings, while more specific questions are more likely to attract

considered judgements (Citrin, 1977: 384). In addition, since contemporary social discourse

continually informs us that people in authority are not to be trusted, the very mention of the word

trust in the question might exert a negative priming effect, which would also serve to skew

responses.

6

way of measuring the concept of trust, since they potentially fail to capture its full and

multi-dimensional nature, establish no clear and invariant meaning across respondents,

restrict the ability to capture attitude ambivalence, and attract incumbent-based evaluations

rather than evaluations of the wider political system.8 These concerns are serious ones, and

behove researchers to think carefully about developing alternative, and more robust,

measurement instruments.9 Before I turn to this, however, I should at least consider the

potential virtues of the single-item measure of political trust.

The first of these advantages is that trust has often been asked about in consistent ways for

long periods of time and across different countries, facilitating longitudinal and comparative

analysis. The second advantage is that the trust question is simple and short. It takes up little

space within the confines of the typical opinion survey, thereby also allowing researchers to

ask respondents about their trust in many different political institutions (as noted earlier, up

to 18 institutions on the World Values Survey). A third advantage is that the wording of the

question ( How much do you trust [institution] in [country]? ) is neutral, since it does not

specify the grounds on which trust might be granted or withheld. Hence, there is no

question of conflating the measurement of the concept with its potential causes.

These advantages are neither groundless nor insignificant. Yet neither do they represent

strong arguments for limiting the way we measure political trust to the conventional

indicators. If, for reasons of replicability and comparability, researchers wish to retain

single-item indicators of trust, there is no reason why they cannot do so. Surveys could be

designed to accommodate both the traditional indicator of trust and any new indicators

designed to provide more robust measurement of the concept. When it comes to space

considerations, the merits of being able to measure trust across a range of institutions may

be more apparent than real. Analysts have consistently shown that expressions of trust in

different institutions reflect a consistent or common underlying attitude rather than discrete

attitudes (Abramson and Finifter, 1981; Feldman, 1983; Hooghe, 2011; Marien, 2011; Dekker,

2011; Rose, 2014).10 Given this, it is not fatal if single-item measures are replaced with multi-

8 In addition to these concerns, there are other potential flaws with the indicators typically used to

measure political trust. For example, response options that force survey respondents to assess

whether they trust an institution almost always , most of the time , only some of the time or almost never may provide a misleading way of categorising levels of trust. Numerical response

scales may be an improvement on categorical response options, but unless these explicitly distinguish

between trust and distrust in an institution, they potentially over-inflate the distribution of critical

attitudes among the population (Cook and Gronke, 2005; Gershtenson and Plane, 2011). However,

while these are important concerns, any inadequacies in the response options offered on survey items

are easily remedied; the more fundamental issue is surely to design more valid measures of the

concept. 9 I should acknowledge that many researchers who draw on existing survey measures of political

trust are well aware of their potential weaknesses, but rely on them nonetheless since they often

represent the only available source of data. By contrast, researchers measuring social trust are in the

fortunate position of being able to draw on behavioural measures of the concept alongside attitudinal

measures. 10 If there are different types of institutional trust, these reflect distinctions in judgements between

representative bodies (governments, parliaments, politicians, political parties and civil servants) and

implementation bodies (courts, the judiciary, the police and the army) (Rothstein and Stolle, 2008).

7

item measures, even if the latter consume more survey space and thus preclude asking

about trust in a variety of different institutions. The final advantage of the conventional

measure of political trust – its neutrality or absence of referent – is potentially more serious,

and is thus considered in more detail in the next section.

In short, while the traditional measures of trust have various weaknesses, their virtues are

relatively limited. This is not to say that the practice of measuring trust via single-item

questions should be discontinued. Instead, we should recognise that these indicators might

provide only a partial way of measuring the concept. Researchers concerned to understand

the relations between citizens and political authorities should think seriously about devising

alternative or complementary indicators that might do a better job of measuring political

trust.

2. Towards alternative measures of political trust

The starting point for any measurement of political trust must be a clear and precise

definition of the concept. Many scholars argue that, when it comes to trust, conceptual

clarity and unity remain elusive (Bauer, 2014). However, the conceptual nature of trust has

at least been extensively explored; in comparison, its measurement has been neglected. In

order to pass over conceptual matters with a view to addressing measurement issues, I

simply take as a definition a widely-cited version from the organisational science literature,

in which trust consists of:

the willingness of a party to be vulnerable to the actions of another party based on the

expectation that the other will perform a particular action important to the trustor, irrespective of

the ability to monitor or control that other party (Mayer et al, 1995: 712)

This definition suggests that the trust of person A in another person or organisation B rests

on a judgement by A about how far B will act in a way consistent with their (A s) interests.

This judgement is likely to rest on such features as B s perceived competence, their concern

with A s interests, and their reliability (Mayer et al, 1995; McKnight and Cervany, 1996;

Johnson, 1999; Metlay, 1999; Citrin and Muste, 1999; Levi and Stoker, 2000). Identification of

these features suggest that trust is a multi-dimensional concept; that to trust rests on

judgements about a number of different considerations, rather than comprising a singular,

generalised evaluation. If trust is indeed multi-dimensional, measurement of the concept

would ideally proceed by systematic and exhaustive identification of the concept s components, followed by the design of empirical indicators to cover each enumerated

feature of the concept, and finally empirical testing of the indicators to establish their

validity and reliability (Hox, 1997). These methodological principles have been followed by

many researchers concerned to develop measures of trust in the field of organisational

studies (see McEvily and Tortoriello, 2011).

However, measuring trust in terms of its core dimensions risks blurring the distinction

between the elements of the concept and its causes or antecedents (Craig, 1993: 8-9, 24-25;

Owen and Dennis, 2001). For example, competence or ability might be taken as one of trust s central components, and thus included among the empirical measures of the concept. Yet

the reasons that A has to trust or distrust B might also reflect A s assessment of B s

8

competence. In other words, in measuring a concept through its core dimensions, we risk

conflating the concept s nature with its causes.11 The potential for measures of trust to

overlap with the concept s potential antecedents has recently been demonstrated by Hooghe

(2011). Drawing on data collected in Britain and using factor analytic techniques, he shows

that survey indicators designed to tap expressions of trust in fact load strongly onto the

same attitudinal dimensions as indicators designed to tap the determinants of trust,

suggesting that definitional and causal indicators are not easily separable.

The potential for conflating the nature and causes of a concept is not a serious problem if the

analyst s goal is merely to measure a concept (as is the case with, for example, the recently

devised citizen trust in government organizations scale Grimmelikhuijsen and Knies, 2015). Since there is no attempt to identify causal relations, measurement and explanation

are kept distinct. But where the analyst wishes not only to measure a concept but also to

account for its incidence, the risk arises of conflating measurement and explanation. The

overlap between the two becomes particularly troublesome where the measured

components of a concept also appear within the list of its potential determinants. To avoid

analysis that descends into tautology, the factors taken to be causal of a concept must be

kept separate to the factors treated as definitional.

One way of avoiding undue overlap between the definition and the causes of trust would be

to measure the concept in behavioural form, rather than in its more conventional attitudinal

guise. Here, trust would consist not in the beliefs that someone has, but in the way they

behave, in particular in their interactions with others. Trust as a behaviour is separate to the

particular reasons someone might hold for trusting or distrusting another person or

organisation. Moreover, some scholars have suggested that manifested (behavioural) trust

provides a more accurate measure of the concept than stated (attitudinal) trust, since

people s engagement with others reveals a trust that in attitudinal form is often disclaimed

O Neill, . Empirical studies of interpersonal trust corroborate this argument, by

showing that behavioural and attitudinal expressions of trust covary only weakly (Ermisch

et al, 2009). There are, then, some advantages in seeking to measure trust as a behaviour.

However, in practical terms, it is difficult to identify patterns of behaviour that could be

taken to proxy political trust. Among the ways that citizens interact with political

authorities, some (eg. electoral participation) have only a weak conceptual and empirical

relationship with political trust (Cox, 2003; Franklin, 2004: 145-46), while others (eg.

compliance with collective obligations) are better considered as the consequences of trust

rather than as its manifestation (Marien and Hooghe, 2011).

Aside from its behavioural form, trust also arises – and might be measured – in the form of a

decision. In this guise, trust represents an intention on the part of A to rely on, or submit to,

another agent, B. Trust as an intention thus sits somewhere between trust as a belief and

trust as a behaviour (for the distinction between trust as a behaviour, an intention or a belief,

see Dietz and Den Hartog, 2006; McEvily and Tortoriello, 2011). Manifested as an intention

to act, political trust maintains a conceptual distinction with its putative causes, while also

11 This conflation can be seen in the tendency among some studies of trust in the field of risk

management to treat factors such as competence and concern as definitional of trust (eg. Earle, 2010),

while other studies treat them as its determinants (eg. Frewer et al, 1996; Johnson, 2010).

9

being more amenable to measurement than in its behavioural guise. Within the

organisational science literature, trust is sometimes measured as an intention to act, in the

form of a willingness to submit to another actor. This notion of willing submission makes

sense in the context of the workplace, since a worker can often choose whether to make

herself vulnerable to the actions of a manager (for example by volunteering for an

assignment). However, voluntary submission to an agent is less intuitive in the political

realm where citizens typically have little choice over whether or not to be vulnerable to a

political elite.12 Yet even if political relations do not involve the notion of willing

vulnerability, trust within those relations might still be measured in terms of A s acceptance of B s right to make and enforce decisions. In the field of politics, then, the intention to trust might be measured through indicators that tap whether people accept the role of leaders in

making binding rules.

To measure political trust as a decision or intention to submit to the rule of politicians, one

could draw on the items designed to tap organisational trust developed by authors such as

Mayer and Davis (1999) and Gillespie (2012). Adapting these items so that they apply to

political actors rather than to organisational ones yields indicators (couched in the form of

agree-disagree statements) such as:

We should let politicians get on with things.

Politicians usually take the right decisions without us keeping a close eye on them.

I am happy to let politicians take major policy decisions that affect me.

I would rather let politicians take big decisions that affect the country than leave things

to citizens.

We can rely on politicians to do the right thing.

The less influence politicians have over our lives the better. (responses reverse coded)

These items are, individually and collectively, designed to capture citizens willingness to submit to the jurisdiction of politicians, and hence to tap the intention to trust. Yet one may

wonder how robustly such indicators capture the specific concept of trust, as opposed to a

related concept such as acquiescence or legitimacy. Researchers are normally more

comfortable in assessing trust via beliefs or judgements on A s part that B manifests

particular features or qualities that induce trust (or distrust) in A. There is no ability to

identify the core elements of trust in measures that simply tap A s willingness to be vulnerable to B. Hence, in spite of the definition of trust identified earlier – a definition that

is widely cited across disciplines and which centres on the notion of willing vulnerability –

most studies within the field of organisational studies continue to measure trust in the form

of a belief rather than as an intention or behaviour (Dietz and Den Hartog, 2006).

If trust is measured as a belief, however, there needs to be a clear separation between its

components and its potential causes. This point is emphasised by Dietz and Den Hartog

(2006), who distinguish between the content of a trust belief and the source of that belief. The

content of a trust belief relates to A s judgement that B possesses the qualities that render

them worthy of trust. In other words, trust beliefs comprise evaluations of A s

12 The choice available to citizens concerns the composition of their political rulers (who represents

them), rather than on whether to have rulers at all (whether they should be represented).

10

trustworthiness which, in turn, are usually seen to have various components (such as ability,

benevolence and reliability). The source of those trust beliefs relate to the reasons A has for

believing that B manifests these qualities. The content and the sources of trust are thus

distinctive, and must be kept so. If the analyst s only concern is to measure trust, this can be

achieved by reference to the relevant qualities of B. Yet what if the analyst also wishes to

explain the feelings of trust or distrust that A has for B? According to Dietz and Den Hartog

(2006), the antecedents of that trust lie in three places: (a) the characteristics of A, notably

their propensity to trust; (b) the characteristics or past behaviour of B, notably the extent to

which these reveal trustworthy qualities; and (c) the context in which B operates, notably

whether they are faced with appropriate incentives and sanctions.

If the antecedents of trust are taken to reside in the nature of the trustor, A, this avoids any

conflation between the definition of trust and its causes, since A s propensity to trust is independent of the way trust is defined and measured. Yet while social or interpersonal

forms of trust are frequently explained by reference to an individual s disposition (Rotter,

1967; Yamagishi and Yamagishi, 1994; Uslaner, 2002), it is less likely that A s trust in a political agent or institution will depend on such personality features.13 Individual

characteristics do not appear a very promising source for understanding why citizens trust

or distrust political actors.14 Yet seeking explanations for trust in the features or past

behaviour of those actors is also fraught. Accounting for trust by reference to the actions of a

trustee risks conflating the definition and antecedents of the concept, rather than

distinguishing them. If the content of A s trust in B reflects a perception that B is competent,

we cannot then account for A s trust by reference to previous manifestations of B s

competence. To do so would intertwine the content and the sources of the trust belief.

The separation between the definition (or content) and cause (or source) of trust is upheld

more stringently if trust is explained by reference to Dietz and Den Hartog s (2006) third

feature, namely the context in which the trustee, B, operates. Here, trust or distrust is

stimulated by observing whether B faces the type of incentives and constraints likely to

induce acting in A s interests (for the role of the institutional environment in shaping the

development of trust, see also Bachmann, 2011). Empirical measures capturing these

incentives and constraints have been developed in the field of e-commerce by McKnight et

al (2002). These gauge whether people feel able to trust online vendors by virtue of the

structural conditions or features of the internet that determine the safety of online commerce. In the field of political relations, the relevant contextual features or structural

conditions might include the presence of appropriate regulations, behavioural norms,

performance targets and redress procedures. Thus, while trust might be measured by

reference to whether B is deemed to possess trustworthy qualities (competence, concern,

13 While studies have shown that levels of political trust do covary with certain personality traits

(Mondak and Halperin, 2008), it is generally implausible to suggest that any randomly selected

individual is likely to trust or distrust a political actor based solely on their personal characteristics. 14 Levels of political trust often fluctuate sharply, suggesting that (stable) personality factors do not

play a significant causal role. In addition, if personality factors were important, one would expect to

see major variations in levels of political trust between different demographic groups within the

population; most empirical analyses, however, find only limited variation in political trust between

such groups.

11

integrity etc), the explanation of that trust would refer to the conditions facing B and

whether these are likely to induce B to act competently, responsively and with integrity.15 In

the Appendix, I set out some examples of indicators that might be used to measure both

trust and its potential causes. The indicators of trust (the trustworthiness of B; column II) are

designed to align with the various dimensions or components that are usually held to define

the concept (column I). The indicators tapping the potential causes of trust (column III)

relate to the context in which B is located, and whether this context is judged likely to induce

B to act in a trustworthy manner.

By identifying the potential causes of trust in the context within which actors are located, we

maintain a distinction between the content of a trust belief and its sources. Yet this

distinction assumes that the average citizen is both aware of the context in which political

actors operate, and competent to assess whether or not these are likely to induce trustworthy

behaviour. Given the widespread evidence of citizen ignorance about the design and

operation of contemporary political systems (Delli Carpini and Keeter, 1996), this

assumption is questionable. Qualitative studies probing the basis of people s trust judgements suggest that expressions of trust often reflect generalised impressions of a political actor, rather than any detailed information about that figure, their goals or the

incentives they face (Carnaghan, 2011; Dekker, 2011). Thus, people may be capable of

expressing trust in a particular actor or institution, but potentially less equipped to identify

the grounds on which that assessment is based. Any survey measures designed to elicit

evaluations of trustworthiness must take account of this potential constraint.

3. Measuring political trust: trade-offs and recommendations

The preceding text highlights the various dilemmas facing any analyst wishing to measure

levels of political trust among a population. They can continue the conventional practice of

gauging political trust through the use of single-item survey questions that explicitly

reference the concept in their wording. However, there are real concerns over whether such

indicators accurately and precisely capture the background concept. Yet once we consider

alternative ways in which trust might be measured, we are quickly confronted by other

problems and trade-offs. Given these, what should be done?

The least disruptive step would be to tweak the existing measure of political trust. If

researchers wish to retain the single-item, direct measure of trust, they should at least

consider amending the response options, so that scales are anchored by explicit trust and distrust labels, while also containing an explicit mid-point to allow for the expression of

uncertainty or ambivalence (Mishler and Rose, 1997; Cook and Gronke, 2005). The explicit

anchoring of response options is likely to induce respondents to more closely consider their

judgements and, by distinguishing people who are genuinely distrusting from people who

are simply sceptical, to potentially generate a more discriminating measure of political trust.

15 A recent analysis of survey items fielded on a population sample in Britain showed that items

capturing trust in political actors are empirically distinct from items capturing judgements about

whether the political system is likely or not to identify and punish incidents of wrongdoing by public

officials (Rose, 2014: 34-44). In other words, the content of the trust belief remained separate from its

potential source.

12

Researchers might also wish to consider different formulations of the trust question, to make

clearer the area or domain over which trust is being extended (Levi and Stoker, 2000: 498-

500; Bauer, 2014). At present, many indicators of trust follow a generic, non-specific

formulation: How much do you trust [institution] in [country]? . This formulation fails to

provide any particular area over which trust is to be evaluated, and thus runs the risk both

of respondent incomprehension and response heterogeneity. This could be remedied by

specifying the domain of trust, which would clarify for respondents the grounds on which

the particular object is being evaluated.16

Yet while these additions would help scholars to distinguish more clearly between

conditions of trust, distrust and uncertainty or ambivalence, these are only minor steps in

the direction of more accurate conceptual measurement. If trust is seen to comprise more

than just a simple and singular assessment of a political agent, attention must be given to its

core nature and/or dimensions. Conceptually, the nature of trust may be held to reside in a

willingness to accept vulnerability to another s actions. In this case, trust might be measured

through the way a person behaves or intends to behave. Yet when it comes to the former,

one runs up against the difficulty in identifying relevant behavioural indicators. Unlike in

settings such as the workplace, the realm of politics provides few examples of citizen

behaviour that could meaningfully be taken to comprise trust. While social trust can be

measured in behavioural form – by observing the reciprocal exchanges between individuals

(eg. Glaeser et al, 2000) – many political exchanges lack an element of choice and so reveal

little or nothing about trust. Thus, for example, actions such as complying with the law or

visiting a doctor s surgery tell us little about an underlying relationship of trust if citizens in

modern democracies are obliged to follow legal rules or to consult a particular medical

practice.

More realistically, political trust could be seen as manifested as an intention to act in a

trusting manner, defined in the case of political relations as a willingness to rely or depend

on a political agent. Even though a citizen may, in practice, have little choice over their

reliance on the agent, treating trust as an intention enables us to gauge how far citizens

accept their dependence. If political trust is held to be accurately and fully captured by the

notion of willing dependence, the empirical task then consists of adding to the example

indicators listed above, and testing to see how far these items correlate with one another,

with a view to forming a single, multi-item scalar measure of (intentional) trust.17 In effect,

the task is to develop an intention-to-trust scale for the political realm analogous to those

developed for the workplace (Mayer and Davis, 1999; Gillespie, 2012)

16 While identifying the grounds or domains on which trust is being extended would improve

conceptual measurement, it would of course run up against the problem of potentially conflating the

nature and causes of trust. For example, it would be tautologous to measure people s trust in an actor

to tell the truth , while also assessing whether perceptions of honesty and openness explain such

trust. 17 There is a further task of exploring how far trust (as intention) is predicted by the kind of indicators

of trustworthiness listed in the Appendix (column II). But this is a task of explanation, not of

measurement.

13

If political trust is deemed to be measured most accurately through indicators that tap its

various components or dimensions, the only route is to treat trust as a belief rather than as

an action or an intention to act. If this route is followed, the empirical analyst must take care

to distinguish the indicators held to capture the meaning of trust from those held to capture

the reasons for that trust. If the indicators of trust tap the sense of the trustor (A) of how

likely the trustee (B) is to act in their interests, the indicators held to tap the causes of that

trust must measure a distinctive set of judgements, relating to the conditions in which B

operates and whether these conditions are likely to induce responsive behaviour or not. If

this route to measuring trust is followed, the empirical endeavour becomes somewhat

complicated. First, indicators tapping both the core elements of trust and the judgements

about the contextual conditions must be identified (indicative items are set out in the

Appendix, columns II and III). The relationship of the indicators of trust (column II) to the

concept s conceptual dimensions (column I) must then be established. Following this, the

putative antecedents of trust (column III) must be related to the indicators of trust (column

II). Ideally, we would find that research participants experienced little cognitive difficulty in

understanding and assessing these contextual conditions. We should also find that those

conditions relate most strongly to the indicators and dimensions of trust to which they are

conceptually linked (in other words, we should find horizontal linkages between the

dimensions, measures and conditions of trust listed in the Appendix).

4. Conclusion

In analysing political trust, most scholarly effort to date has been devoted to the twin issues

of its causes and its consequences. Given their relevance, it is understandable that these

issues have been pursued at the expense of the similarly important, though perhaps rather

less exciting, issue of how best to measure trust. As a result, political scientists have fallen

back, often without much if any critical comment, on well-worn methods of measurement.

This has usually involved couching survey measures in ways that explicitly reference the

concept and that are applied in a single-question format to particular political institutions.

The use of single-item indicators to tap trust is not unique to political scientists, but is also

prevalent in the fields of risk management and organisational studies (Chryssochoidis et al,

2009: 149-53; Earle, 2010: 553, 562; McEvily and Tortoriello, 2011: 33). However, empirical

researchers in these fields have at least paid greater attention to alternative ways in which

trust might be measured, and such exercises have important lessons for researchers dealing

with trust in the field of politics.

The preceding discussion makes clear that there is no single ideal way of measuring a

complex concept like political trust. Instead, there are various ways in which trust might be

identified and assessed. None of these is either wholly satisfactory or wholly suitable for the

different tasks that empirical researchers might have. The main message of this review is to

urge scholars to think more carefully about how they measure political trust, and to relate

their choices to their particular research purpose. Rather than uncritically replicating

existing ways of measuring trust, researchers should consider if alternative indicators might

yield stronger measures of the concept. Fairly minor adjustments to existing measures can

be made by introducing new response options to survey items. Yet if existing survey

indicators are felt to provide an inadequate way of measuring trust, amending their wording

or format is unlikely to significantly strengthen their capability. Instead, analysts might need

14

to think about more radical changes to the way they measure trust. This could involve

conceptualising and measuring trust in behavioural or intentional forms. Alternatively, it

could involve a continued focus on trust as a belief, but measured through indicators

designed to fully capture the concept and its core components. These latter steps would

clearly involve much additional research effort. They would also upset the link with the way

in which political trust has conventionally been measured, going back five decades. Yet

what is sacrificed in continuity might be more than gained in more accurate and effective

measurement of a concept that most agree is central to the relations between citizens and

political actors.

Appendix A: Indicative measures of trust and of conditions leading to trustworthiness

I: DIMENSION OF TRUST II: MEASURES OF TRUST III: CONDITIONS INDUCING TRUSTWORTHINESS

Competence

Competence Generally speaking, politicians are competent In politics, only the most able people get to the top

Political parties ensure that only capable people get selected as

election candidates

Inefficiency Governments waste a lot of public money There is not enough scrutiny over how ministers spend public

funds

Expertise Politicians generally know what they are doing Once elected, politicians are given a good training

Concern; benevolence

Concern Politicians usually try to help their constituents Only politicians who can show they are responding to people s Politicians tend to look after their own interests rather

than trying to help others

concerns win elections

Understanding In the main, politicians don t really understand the problems facing ordinary people

The average politician is very like the average citizen

Integrity

Honesty; truthfulness In the main, politicians tell the truth If politicians don t tell the truth, they are likely to get caught out

Openness Government ministers tell us as little about what they

get up to as they can

There are clear rules about what politicians must disclose

Governments distort the facts to make their policies look

good

There is not enough factual information available to citizens to

allow them to judge government decisions

In the main, when things go wrong politicians admit

their mistakes

Politicians who admit to mistakes tend to get punished by voters

Reliability

Promise-keeping Politicians are happy to promise things at an election but

forget their promises afterwards

If politicians go back on promises made to voters, they get

punished at elections

Politicians try to keep their promises

16

Consistency Politicians change their minds all the time Politicians must respond to voters; if voters change their minds, so

must politicians.

Fairness

Equity Governments treat each group within society equally Governments that favour some social groups over others will not

last in office for long

“bramson, Paul R and “da W Finifter On the Meaning of Political Trust New Evidence from Items Introduced in , American Journal of Political Science, 25:2,

297-307

“gger, Robert E, Marshall N Goldstein and Stanley “ Pearl Political Cynicism Measurement and Meaning , Journal of Politics, 23:3, 477-506.

“llum, Nick Empirical Test of Competing Theories of Hazard-Related Trust: Case of

GM Food , Risk Analysis, 27:4, 935-46

”achmann, Reinhard “t the Crossroads Future Directions in Trust Research , Journal

of Trust Research, 1:2, 203-13.

Bauer, Paul C (2014) Conceptualizing and Measuring Trust and Trustworthiness’, Working Paper

61, IPSA Committee on Concepts and Methods

”ulloch, Sarah L Seeking Construct Validity in Interpersonal Trust Research: A

Proposal on Linking Theory and Survey Measures , Social Indicators Research, 113:3,

1289-1310

Canache, Damarys, Jeffery J Mondak and Mitchell “ Seligson Meaning and Measurement in Cross-National Research on Satisfaction with Democracy , Public

Opinion Quarterly, 65:4, 506-28.

Carnaghan, Ellen The Difficulty of Measuring Support for Democracy in a Changing Society: Evidence from Russia , Democratization, 18:3, 682-706.

Charlton, Michelle, Sarah Morton and IPSOS Mori (2011) Exploring Public Confidence in the

Police and Local Councils in Tackling Crime and Anti-social Behaviour, Research Report

50, London: Home Office

Cheng, Helen et al The Measurement and Evaluation of Social “ttitudes in Two ”ritish Cohort Studies , Social Indicators Research, 107:2, 351-71.

Chryssochoidis, George, “nna Strada and “thanasios Krystallis Public Trust in Institutions and Information Sources Regarding Risk Management and

Communication Towards Integrating Extant Knowledge , Journal of Risk Research,

12:2, 137-185.

Churchill, Gilbert “ Jr “ Paradigm for Developing ”etter Measures of Marketing Constructs , Journal of Marketing Research, 16:1, 64-73.

Citrin, Jack Comment The Political Relevance of Trust in Government , American

Political Science Review, 68:3, 973-88.

Citrin, Jack Political “lienation as a Social Indicator , Social Indicators Research, 4:1,

381-419.

Citrin, Jack and Christopher Muste Trust in Government , in John P Robinson et al, eds, Measures of Political Attitudes, Volume 2, San Diego: Academic Press

Cook, Timothy E and Paul Gronke The Skeptical “merican Revisiting the Meanings of Trust in Government and Confidence in Institutions , Journal of Politics, 67:3, 784-

803

Cox, Michaelene When Trust Matters Explaining Differences in Voter Turnout , Journal of Common Market Studies, 41:4, 757-70

Craig, Stephen C (1993) The Malevolent Leaders: Popular Discontent in America, Boulder, CO:

Westview Press

Craig, Stephen C et al Political Efficacy and Trust “ Report on the NES Political Survey Items , Political Behavior, 12:3, 289-314

Crepaz, Marcus et al Trust Matters The Impact of Ingroup and Outgroup Trust on Nativism and Civicness , Social Science Quarterly, 95:4, 938-59.

18

Dalton, Russell J (2004) Democratic Challenges, Democratic Choices: The Erosion of Political

Support in Advanced Industrial Democracies, Oxford University Press

Dekker, Paul Political Trust What Do we Measure? Should Politicians and the Public

Trust Us? , paper for 6th ECPR General Conference, Reykjavik, 25th-27th August.

Delli Carpini, Michael X and Scott Keeter (1996) What Americans Know about Politics and Why

It Matters, New Haven: Yale University Press

DeVellis, Robert F (1991) Scale Development: Theory and Applications, Newbury Park, CA: Sage

Dietz, Graham and Deanne N Den Hartog Measuring Trust within Organizations , Personnel Review, 35:5, 557-588.

Earle, Timothy C Trust in Risk Management “ Model-Based Review of Empirical

Research , Risk Analysis, 30:4, 541-574.

Easton, David “ Reassessment of the Concept of Political Support , British Journal of

Political Science, 5:4, 435-57.

Egede, Leonard E and Charles Ellis Development and Testing of the Multidimensional Trust in Healthcare Systems Scale , Journal of General Internal

Medicine, 23:6, 808-15.

Ermisch, John et al Measuring People s Trust , Journal of the Royal Statistical Society A,

172:4, 749-69

Feldman, Stanley The Measure and Meaning of Trust in Government , Political

Methodology, 9:3, 341-54.

Finifter, “da Dimensions of Political “lienation , American Political Science Review,

64:2, 389-410.

Fiske, Susan T Schema-Based versus Piecemeal Politics: A Patchwork Quilt, but not a

”lanket, of Evidence , in Richard R Lau and David O Sears, Political Cognition,

Hillsdale, NJ: Lawrence Erlbaum

Fiske, Susan T and Steven L Neuberg “ Continuum of Impression Formation, from

Category-Based to Individuating Processes: Influences of Information and

Motivation on “ttention and Interpretation , in Mark P Zanna, ed, Advances in

Experimental Social Psychology, Vol 23, San Diego, CA: Academic Press

Fowler, Floyd J Jr (2009) Survey Research Methods, 4th edition, Thousand Oaks: Sage

Publications

Franklin, Mark N (2004) Voter Turnout and the Dynamics of Electoral Competition in Established

Democracies Since 1945, Cambridge: Cambridge University Press

Frazier, M Lance, Paul D Johnson and Stav Fainshmidt Development and Validation of a Propensity to Trust Scale , Journal of Trust Research, 3:2, 76-97.

Frewer, L.J. et al What Determines Trust in Information “bout Food-Related Risks?

Underlying Psychological Constructs , Risk Analysis, 16:4, 473-86

Fuchs, Christoph and “damantios Diamantopoulos Using Single-Item Measures for

Construct Measurement in Management Research: Conceptual Issues and

“pplication Guidelines , Die Betriebswirtschaft, 69:2, 195-210

Gainous, Jason et al Social Welfare “ttitudes and “mbivalence about the Role of Government , Politics and Policy, 36:6, 972-1004.

Gershtenson, Joseph and Dennis Plane “n “lternative Measure of Political Trust Reconciling Theory and Practice , in John H “ldrich and Kathleen M McGraw, eds, Improving Public Opinion Surveys: Interdisciplinary Innovation and the American National

Election Studies, Princeton, NJ: Princeton University Press

19

Gibson, James L et al Measuring “ttitudes toward the United States Supreme Court , American Journal of Political Science, 47:2, 354-67.

Gillespie, Nicole Measuring Trust in Organizational Contexts “ Review of Survey-

”ased Measures , in Fergus Lyon et al, eds, Handbook of Research Methods on Trust,

Cheltenham: Edward Elgar (pp175-188).

Glaeser, Edward L et al Measuring Trust , Quarterly Journal of Economics, 115:3, 811-

46.

Grimmelikhuijsen, Stephan and Eva Knies (2015 Validating a Scale for Citizen Trust in Government Organizations , International Review of Administrative Sciences,

forthcoming

Hamm, Joseph et al Exploring Separable Components of Institutional Confidence , Behavioral Sciences and the Law, 29:1, 95-115

Heath, “nthony and Jean Martin Why are There so Few Formal Measuring

Instruments in Social and Political Research? , in Lars Lyberg et al, eds, Survey

Measurement and Process Quality, London: John Wiley

Hetherington, Marc J (2005) Why Trust Matters: Declining Political Trust and the Demise of

American Liberalism, Princeton, NJ: Princeton University Press

Hill, David ” “ttitude Generalization and the Measurement of Trust in “merican Leadership , Political Behavior, 3:3, 257-270.

Hooghe, Marc Why There is ”asically Only One Form of Political Trust , British

Journal of Politics and International Relations, 13:2, 269-75.

Hox, Joop J From Theoretical Concept to Survey Question , in Lars Lyberg et al, eds, Survey Measurement and Process Quality, London: John Wiley

Jackson, Jonathan et al (20 Developing European Indicators of Trust in Justice , European

Journal of Criminology, 8:4, 267-85.

Johnson, ”randen ” Exploring Dimensionality in the Origins of Hazard-related

Trust , Journal of Risk Research, 2:4, 325-54

Johnson, Branden B (2 Getting Out of the Swamp Towards Understanding Sources of

Local Officials Trust in Wetlands Management , in Michael Siegrist et al, eds, Trust

in Risk Management: Uncertainty and Scepticism in the Public Mind, London: Earthscan

Kay, “aron C and John T Jost Complementary Justice Effects of Poor but Happy and Poor but Honest Stereotype Exemplars on System Justification and Implicit “ctivation of the Justice Motive , Journal of Personality and Social Psychology, 85:5, 823-

837.

Keller, Carmen et al The General Confidence Scale Coping with Environmental Uncertainty and Threat , Journal of Applied Social Psychology, 41:9, 2200-29

Lang, John T and William K Hallman (2005) Who Does the Public Trust? The Case of GM Food in the United States , Risk Analysis, 25:5, 1241-1252.

Levi, Margaret and Laura Stoker Political Trust and Trustworthiness , Annual Review

of Political Science, 3: 475-507

Marien, Sofie and Marc Hooghe Does Political Trust Matter? “n Empirical

Investigation into the Relationship between Political Trust and Support for Law

Compliance , European Journal of Political Research, 50:2, 267-91

Mayer, Roger C et al “n Integrative Model of Organizational Trust , Academy of

Management Review, 20:3, 709-34.

20

Mayer, Roger C and James H Davis The Effect of the Performance “ppraisal System on Trust for Management: A Field Quasi-Experiment , Journal of Applied Psychology,

84:1, 123-36.

McCloskey, Herbert and John Zaller (1984) The American Ethos: Public Attitudes towards

Capitalism and Democracy, Harvard University Press

McEvily, ”ill and Marco Tortoriello Measuring Trust in Organisational Research Review and Recommendations , Journal of Trust Research, 1:1, 23-63.

McGraw, Kathleen M and ”randon ”artels “mbivalence toward “merican Political Institutions Sources and Consequences , in Stephen C. Craig and Michael D. Martinez, eds, Ambivalence and the Structure of Political Opinion, Palgrave Macmillan

McKnight, D Harrison et al Developing and Validating Trust Measures for e-

Commerce “n Integrative Typology , Information Systems Research, 13:3, 334-359.

McKnight, D Harrison and Norman L Cervany The Meanings of Trust , Management

Information Systems Research Center Working Paper 4, Minneapolis, MN: University of

Minnesota

Magalhães, Pedro C Government Effectiveness and Support for Democracy , European Journal of Political Research, 53:1, 77-97.

Marien, Sophie Measuring Political Trust “cross Time and Space , in Sonja Zmerli and Marc Hooghe, eds, Political Trust: Why Context Matters, Colchester, Essex: ECPR

Press

Metlay, Daniel Institutional Trust and Confidence “ Journey into a Conceptual

Quagmire , in George Cvetkovich and Ragnar E Loftstedt, eds, Social Trust and the

Management of Risk, London: Earthscan

Miller, “lan S and Tomoko Mitaruma “re Surveys on Trust Trustworthy? , Social

Psychology Quarterly, 66:1, 62-70.

Mishler, William and Richard Rose Trust, Distrust and Skepticism Popular Evaluations of Civil and Political Institutions in Post-Communist Societies , Journal of

Politics, 59:2, 418-51.

Mondak, Jeffrey J et al Does Familiarity Breed Contempt? The Impact of Information

on Mass “ttitudes toward Congress , American Journal of Political Science, 51:1. 34-48.

Mondak, Jeffrey J and Karen D Halperin “ Framework for the Study of Personality and Political ”ehaviour , British Journal of Political Science, 38:2, 335-62.

Morrell, Michael E Survey and Experimental Evidence for a Reliable and Valid Measure of Internal Political Efficacy , Public Opinion Quarterly, 67:4, 589-602

Muller, Edward N (1979) Aggressive Political Participation, Princeton, NJ: Princeton University

Press

Nannestad, Peter What Have We Learned “bout Generalised Trust, If “nything? Annual Review of Political Science, 11, 413-36.

O Neill, Onora A Question of Trust, Cambridge University Press

Owen, Diana and Jack Dennis Trust in Federal Government The Phenomenon and its “ntecedents , in John Hibbing and Elizabeth Theiss-Morse, eds, What is it About

Government that Americans Dislike?, New York: Cambridge University Press

Ozawa, Sachiko and Pooja Sripad How do you Measure Trust in the Health System? “ Systematic Review of the Literature , Social Science and Medicine, 91: 10-14.

Petrescu, Maria Marketing Research using Single-Item Indicators in Structural

Equation Models , Journal of Marketing Analytics, 1:2, 99-117.

21

Poortinga, Wouter and Nick Pidgeon Exploring the Dimensionality of Trust in Risk Regulation , Risk Analysis, 23:5, 961-72

Poznyak, Dmitriy et al (2014) Trust in American Government: Longitudinal Measurement

Equivalence in the ANES, 1964– , Social Indicators Research, 118:2, 741-58.

Rose, Jonathan (2014) The Public Understanding of Political Integrity: The Case for Probity

Perceptions, Basingstoke: Palgrave

Rotter, Julian ” “ New Scale for the Measurement of Interpersonal Trust , Journal of

Personality, 35:4, 561-65.

Rutto, Filippo, Silvia Russo and Christina Mosso Development and Validation of a Democratic System Justification Scale , Social Indicators Research, 118:2

Schaffer, Frederick Charles (2010) Thin Descriptions: The Limits of Survey Research on the

Meaning of Democracy , Political Concepts Working Paper 45, International Political

Science Association Committee on Concepts and Methods

Scholz, John T and Mark Lubell Trust and Taxpaying Testing the Heuristic “pproach to Collective “ction , American Journal of Political Science, 42:2, 398-417.

Selnes, Fred “ntecedents and Consequences of Trust and Satisfaction in ”uyer-Seller

Relationships , European Journal of Marketing, 32:3-4, 305-322

Spector, Paul E (1992) Summated Rating Scale Construction: An Introduction, Quantitative

Applications in the Social Sciences 82, Newbury Park, CA: Sage

Steenvoorden, Eefje H “ General Discontent Disentangled A Conceptual and

Empirical Framework for Societal Unease , Social Indicators Research, 124:1, 85-110.

Stokes, Donald Popular Evaluations of Government “n Empirical “ssessment , in Harlen Cleveland and Harold Lasswell, eds, Ethics and Bigness: Scientific, Academic,

Religious, Political, and Military, New York: Harper

Sturgis, Patrick and Patten Smith Assessing the Validity of Generalized Trust

Questions What Kind of Trust are we Measuring? International Journal of Public

Opinion Research, 22:1, 74-92.

Uslaner, Eric M (2002) The Moral Foundations of Trust, New York: Cambridge University

Press

Yamagishi, Toshio and Midori Yamagishi Trust and Commitment in the United States and Japan , Motivation and Emotion, 8:2, 129-166.

Zeller, Richard A and Edward G Carmines (1980) Measurement in the Social Sciences: The Link

between Theory and Data, New York: Cambridge University Press

Zmerli, Sonja, Kenneth Newton and Jose Ramon Montero Trust in People, Confidence in Political Institutions and Satisfaction with Democracy , in Jan Van Deth, ed, Citizenship and Involvement in European Democracies, Routledge