The effects of e-government transparency and interactivity on trust in government

‘The effects of e-government Transparency and Interactivity on Trust in Government’

Author: Sebastian Walters Student No. 10488499

Graduate School of Communication University of Amsterdam

Abstract

This research attempts to probe the relationship between citizens’ e-government use and trust in government. An often contested relationship, stemming from the debatable influence that government websites can have on trust, the study proposed in isolating specific qualitative features of e-government – government Transparency and Interactivity – possibly more clearer and robust conclusions can emerge. To this end, a between- group experiment was designed with 140 participants viewing one of 4 government webpages designed with varying degrees of transparency and interactivity, and completing a post-test questionnaire that

gauged their trust in government after viewing the stimuli. Initial analyses revealed, that the notoriously nebulous concept of ‘trust in government’, actually measured two constructs – trust in the ‘performance’ of government, and trust in its ‘intentions’. Thus whilst the generally positive results of the experiment cannot speak directly to the influence of transparency and interactivity on overall trust in government, they do however show that government websites that maximize transparent and interactive features produce higher appreciations from citizens of governments intentions and performance. Providing some support to digital strategies that focus on transparency and interactivity. The results of the experiment appear to mesh well with more positive expectations of the influence of e-government.

The effects of E-government Transparency and Interactivity on Trust in Government Recent years have witnessed a great momentum in pushing government services online. Fuelled by the apparent cost savings; the ease of information dissemination to citizens and businesses; the prospects for citizen involvement and government transparency amongst other proposed benefit’s, governments worldwide are improving and developing their digital presence.

For instance The Open Government Agenda of the Obama Administration, the EU Malmö Declaration or the ‘Digital by Default’ strategy of the UK government all prioritize citizen participation and transparency in Government and politics (Lörincz et al. 2011), and they identify digital government, also known as ‘e-government’, the relatively new mode of citizen-to-government contact founded in information and communication technologies (ICTs), as the decisive medium in achieving these goals. Yet the expected potential of e-government is barely visible, hardly measured, and crucially its effects on citizen’s attitudes towards their governments has seldom been investigated.

Nevertheless, an emerging body of research focuses on the relationship between e-government and citizen trust in e-government. For many, including both academics and policy makers, e-government is seen as a potentially transformational medium – a mode of contact that could dramatically improve citizen perceptions of government service delivery and possibly reverse the long-running decline in citizen trust in government (Morgeson, Mithas & Van Amburg, 2011).

To date, however, the literature has left significant gaps in our understanding of the e-government – citizen trust relationship. For instance most of the extant empirical research on the impact of e-government on citizens’ trust in government remains at the macro-level and misses out on the complexities of the interaction between e-services and citizens’ trust in

government (Parent, Vandebeek & Gemino 2005; West, 2004, Tolbert and Mossberger, 2006; Moon, 2003). This research strives to address these complexities.

To this end, two proposed benefits of e-government: Transparency and Interactivity as outlined in the European Commission’s ‘eGovernment Benchmark Pilot on Open

Government and Transparency’ (Lörincz et al., 2011) are investigated for their effect on trust in government. Querying whether enhanced transparency and interactivity cues on

e-government websites can engender trust in e-government institutions. The following research question is asked:

RQ: What effects do e-government transparency and interactivity cues have on citizen trust in government?

The research’s’ scientific and social relevance is evident insomuch that it can be placed firmly in the debate regarding the potential democratic benefits of the internet. Further this research extends beyond mere academic curiosity – with implications for governments regarding e-government policy and how governments communicate with their citizenry online.

If the research finds a significant influence of e-government transparency and

interactivity on citizens trust judgements of government then the rationale for a digital agenda that focuses on transparency and citizen participation is strengthened. On the contrary if little evidence is found, then one can downplay e-government transparency and interactive

measures or policies. Further one could question the purported benefits of digital government more generally.

Theoretical Framework E-government: Optimists vs Pessimists

Not only is the literature on e-government nascent, there is sparse theory development (Couresy & Norris, 2008). However a burgeoning amount of studies have examined the relationship between citizens’ adoption and use of e-government and citizen trust and confidence in government (Bertot & Jaeger, 2006; Norris, 2001; Parent et al., 2005; Thomas & Streib, 2003; Tolbert & Mossberger, 2006; Welch et al. 2005; West 2004). In many of these studies, e-government is deemed to be a potentially transformational technological innovation, a manner of citizen government contact that could improve the services delivered to citizens, boost citizen satisfaction with government, and possibly even help combat the decline in citizen trust in government (Morgeson et al., 2011).

The basic argument is that as e-government becomes more innovative and takes advantage of new technologies, distrust will be countered by increasing perceived efficiency, quality of outcomes, and opportunities for policy inputs (Parent et al, 2005). This optimistic and familiar contention has considerable support from policy makers and academics and provides the rationale for many governments’ reforms of online provisions – this and the lure of cost savings (Kolsaker & Lee-Kelley, 2008).

Nevertheless, the theoretical premise of the e-government advocate argument has a genesis in a much broader debate that centres on the democratic benefits of the internet. Mostly concerned with political participation but still relevant is this research, the dispute is between scholars on the one hand who view emergent technologies as inherently

democratising with real benefits for political participation and efficacy; and those whom are less optimistic and expect little, if any benefits (Anstead & Chadwick, 2009; p58).

Transposed to the e-government – citizen trust in government relationship, the debate takes a similar guise. That is, the more optimistic accounts are countered by pessimistic

opponents that emphasize the perverse effects of new technologies. Insomuch, that the traditional ways of producing trust - direct contacts; contextualized interactions, are seen as superior to new mechanisms of trust (Grimmelikhuijsen & Meijer, 2012).

Such discord between scholars is reflected in the extant literature and importantly in their findings. Morgeson et al. (2011) note that the results and conclusions of a number of studies — at least those that formally test these e-government use and trust in government propositions — ‘have been mixed, with researchers expressing varying opinions on the ability of e-government to actually build citizen trust and confidence in government’ (p259).

However, on reviewing the literature, significant gaps emerge. That is, that most accounts concentrate on the macro-level surveys and tentatively account for individual citizens direct interaction with government websites and trust in government (for example Parent et al., 2005; West, 2004, Tolbert & Mossberger, 2006; Moon, 2003). They fall short in explaining or investigating what it is specifically about e-government that can possibly effect trust in government. With this in mind, my research will offer an important contribution by attempting to probe cues in e-government that can be directed at improving trust and subsequently taps citizen’s responses to these cues. I propose that in looking at more micro level causal mechanisms of trust perhaps clearer and more robust conclusions can emerge.

Thus, against this background I will attempt to isolate the specific features of e-government that are purported to be its main benefits and investigate their relationship with perceived trustworthiness of a government organization.

The two paramount features of modern digital strategies as discussed earlier are said to be Transparency and Interactivity, and crucially – theoretically at least – they have been linked to improving citizen trust in government (Welch & Hinnant, 2003, West, 2004). However before further exploring these theoretical propositions a better look at the concept of ‘trust in government’ is needed.

Defining and Measuring Trust in Government

What is ‘trust in government?’ and more importantly - why does it matter? Such simple questions, do not yield simple answers. Just defining ‘trust’ alone has been

problematic for many scholars across a spectrum of disciplines. Indeed, understanding why and how people trust has been a central focus of research for psychologists, sociologists, political scientists, economists and organizational scientists (Grimmelikhuijsen, Porumbescu, Hong & Im, 2013; p577). Unsurprisingly then, a myriad of definitions, concepts and

operationalization’s exist.

Nevertheless, despite there being little agreement in the literature on defining citizen trust in government, most writers agree that it is an important mechanism of public action and cooperation, and that it has been declining for years (Putnam, 1995; Norris, 2001; Welch & Hinnant; 2003). They discern that ‘trust’ in the faculties of government is crucial to

progressive public policy and economic development (Kong, 2012), thus without it governments are impotent and ineffective.

In this respect, beyond the question of voting and participation, trust is deemed important for the legitimacy and stability of the political system. ‘Trust in government encourages compliance with laws and regulations’ (Tolbert and Mossberger, 2006; p355). What is clear here is that a widely accepted thesis is that some form of trust is required for effective government. Less clear is precisely what trust in government is and how can it be measured?

According to Rousseau, Sitkin, Burt and Camerer (1998), all definitions of ‘trust’ are premised on the presence of some form of positive expectation regarding the intentions and behaviour of the object of trust. In the context of this study then this can be adapted to citizens’ positive expectation regarding the behaviour and intentions of their government.

Listhaug and Miller articulate the concept well and define trust in government as an evaluation of ‘whether or not political authorities in institutions are performing in accordance to normative expectations held by the public’ (Miller and Listhaug cited in Tolbert and Mossberger, 2006; p). Smith (2010) further narrows down the trust in government concept by identifying it as an ‘institutional trust’, as a relationship where a truster (citizen) places trust in the rules, roles and norms of an institution, independent of the people occupying those roles (Smith, 2010; p226). Thus trust is directed towards the institution of government not necessarily the bureaucrats that run the institution.

With these conceptualisations in mind, my research subscribes to the definition provided by Warren (2004):

‘Placing trust is an act that is based on the interpretation of information that provides the reason to trust in the institutions’ competence and motivations’ (Warren, 2004; p244)

This definition focuses this study’s interest in trust in several ways. Firstly it tells us that trust towards government to some extent is based on citizen’s assessments of government institutions’ behaviour or performance. Further it also conveys an assessment of a more ethical dimension, in terms of an institutions’ motivations or intentions; and most importantly it implies how it can be measured.

For instance along these lines, Grimmelikhuijsen et al. (2013) construct a

measurement of trust in government that combines three interrelated but separate elements that fit well with the definition above outlined by Warren (2004): Perceived competence’, ‘perceived benevolence’ and ‘perceived honesty’ of an institution (p.577). The competence element is linked to how effective an institution is perceived to be i.e. an evaluation of its performance. Benevolence is linked to the intention of government action or commitment. ‘It expresses some kind of interest by the one being trusted by others’ (Grimmelikhuijsen et al., 2013; p 577). Whilst honesty reflects its integrity and willingness to tell the truth

(Grimmelikhuijsen et al, 2013). Both benevolence and honesty reflect a more ethical facet of trust in government.

These measures thus speak to our definition as they combine competency and motivation evaluations. It is argued that when citizen’s form trust judgements, they do so with these perceptions in mind.

This multifaceted approach to measuring citizen’s trust in government will be used here as it potentially can grasp the complexities of the e-government – citizen trust in

government relationship. Deconstructing the perceived trustworthiness of government in such a way may provide the research with a greater understanding of the influence of

e-government on trust. (See the Method section for information on measures of these dimensions).

Transparency and Interactivity – Trust Mechanisms?

So far this paper has reviewed the proposed relationship between e-government use and citizen trust in government. I have also looked closer at the concept of trust in

government - what it is and how it can be measured.

Specific attention now turns to the purpose of this study and establishes its point of departure. It considers the two heralded mechanisms of e-government: ‘Transparency’ and ‘Interactivity’ and hypothesises their effects on citizens’ trust in government as outlined in the previous section.

Firstly an explanation of the interest in these two mechanisms is needed. A subsection of the literature on e-government, focuses on the normative models of e-government websites and the transformational trajectory of many governments’ online offerings. Darrell West (2004) identifies four stages of e-government transformation (1) the billboard stage; (2) the partial-service-delivery stage; (3) the portal stage, with fully executable and integrated

service delivery; and (4) interactive democracy with public outreach and accountability enhancing features (p17). Each stage is viewed as democratically better than the previous (West; 2004). The fourth model – the most optimal, is based around qualitative features that are transparent and interactive.

This has been reflected in policy too. The European Commission published report ‘eGovernment Benchmark Pilot on Open Government and Transparency’ advises member states that digital government should be based on three priorities: Transparency, Participation and Collaboration (Lörincz et al. 2011). The priorities, participation and collaboration are equally understood as based on citizens interacting and engaging with government. Thus this research takes the probe of transparency and interactivity features as its point of departure in the study of e-government – citizen trust in government relationship. It asks the question of what influence do these specific features have on citizen trust?

One theory of epistemology of trust is that when people make trust judgments, they look for trustworthiness cues (signs) that manifest trustworthy properties (Smith, 2010; p227). The theory borrows partly from Elaboration Likelihood Model (ELM) developed by Petty and Cacioppo (1986) which contends that individuals have limited time and capacities for processing information. In order to make decisions, people heavily simplify the options and information available. In this context one can never fully know all the information to trust the competence and motives of an institution. A decision to trust a government

organization may, therefore, not always be conscious and/or rational. Rather judgements can be based on judgements of cues that could be linked to trust (Smith, 2010). In this sense could transparent and interactive features serve as trustworthiness cues?

Transparency

A large body of literature states mainly positive effects of government transparency on trust (Grimmelikhuijsen et al., 2013; Meijer, 2009; Norris, 2001, Welch et al, 2005). Defined in its most rudimentary, ‘transparency’ is understood as ‘lifting the veil of secrecy’ (Davis, 1998: p121) or ‘the ability to look clearly through the windows of an institution’ (Den Boer cited by Meijer, 2009). Cornelia Moser (2001) elaborates on these metaphors and

defines transparency as the opening up of ‘the working procedures, not immediately visible to those not directly involved in order to demonstrate the good working of an institution’ (p3).

From this, and for the purpose of this research, transparency is therefore understood as being a feature in the relationship between the ruler and the ruled, whereby the operations of the “government is sufficiently open to public view and simple enough in its essentials that citizens can readily understand how and what it is doing” (Dahl cited by Moser, 2001;p3)

Given these attributes, it has been suggested that that transparent government’s lead to better informed citizens, that are expected to have more trust in government (Meijer, 2009). As it is argued that ‘one cause for the lack of citizen trust in government is that citizens are not often enough provided with factual documentation about government processes and performance’ (Duckhong cited by Grimmelikhuijsen et al., 2013; p577). In this respect government transparency is perceived to help people become more familiar with government and brings them closer and creates understanding and possibly trust (Welch et al, 2005).

Heald identifies different points at which government determines the level of transparency: (1) transparency of decision-making processes, (2) transparency of policy content, and (3) transparency of policy outcomes or effects (Heald cited by Grimmelikhuijsen et al., 2013). Along these lines Welch et al. (2005) posit that when ‘government provides citizens with reliable information about administrative processes, decisions, activities, outputs, and outcomes’, there is a positive influence on trust in that institution (p379).

A feature of modern day government transparency is that it is mediated. Christopher Hood (2006; p20) stresses that modern forms of transparency differ from direct face-to-face transparency in traditional town meetings. ‘Transparency has been mediated through the mass media for some time but the media, especially the Internet and other computerized systems, play a key role in mediating government transparency in our times’ (Meijer, 2009; p259). In terms of the specific discussion on trust outlined above, it is expected that when a

government provides citizens with information online aimed at transparency, this will have a positive effect on evaluations of their performance and intentions and thus trust. This leads to the first hypothesis:

H1: A high level of transparency cues on e-government websites leads to more perceived trustworthiness of the government

Interactivity

Interactivity is seen as one of the biggest benefits to e-government delivery. It has been identified as a key determinant in citizens’ e-government satisfaction (Welch et al., 2005). In most normative models of e-government development, the integration of interactive features and two – way communication is key and deemed the most optimal and democratic (West, 2004). Whereby increased possibilities for citizen participation, feedback, possibilities to make comments and other sophisticated features are viewed to boost democratic

responsiveness and leadership accountability. In terms of trust then, the argument is that increased efforts by governments to establish reliable information-exchange mechanisms with its citizens can enhance citizen trust in government (Welch et al., 2005).

If one considers the prevalent definitions of ‘interactivity’ then some of the reasons why it has been identified as a possible trust mechanism can be illustrated. Firstly, interest in interactivity has grown since the proliferation of new media and the internet and a number of

different definitions abound (Kiousis, 2002). However in most general discussions it has been understood as the ability for message receivers to respond to message senders (Kiousis, 2002). Williams et al introduce a semblance of power to the discussion, and recognise interactivity as the degree to which participants in a communication process have control over and can exchange roles in their mutual discourse (Williams et al in Kiousis, 2002; p368). In a similar vein Ha and James (1998) argue that ‘interactivity is the extent to which the communicator and the audience respond to or are willing to facilitate each other’s communication needs’ (p462).

What these definitions convey is that interactivity has a conversational character, that it facilitates discussion and has a manner of reciprocity (Kiousis, 2002). In this context, i.e. in terms of government-citizen communication, they also convey crucially that interactivity is linked to engagement and public participation (Rafaeli & Sudweeks, 1997). This link is perhaps described best by Welch et al. (2005) - and such is used in this present study – as they describe interactivity as ‘the means of describing the willingness or ability of an agency to be responsive to citizens’ (p379)

This has certain interesting implications, as political theorists have previously linked civic engagement with trust (Uslaner, 2004). Eric Uslaner (2004) views the ultimate payoff of civic engagement ‘as the trust it engenders with our fellow citizens’, with the view that engagement can combat the policy alienation many citizens cite for their distrust in government (p225).

In this respect, institutions that explicitly seek to engage citizens may improve trust in that institution. For instance Welch et al. (2005) coherently make this proposed association, noting that “agencies that include citizens in discussions and decision-making processes about policy-relevant topics, regardless of whether citizens take advantage of the

communicative exchange that builds process trust, and encourage interpersonal interaction that supports mutual trust between citizen and institution” (p379). From this, a second hypothesis can be established:

H2: A high level of interactivity on e-government websites leads to more perceived trustworthiness of government

From the above discussion on e-government transparency and interactivity, a third speculative hypothesis can be made. In line with the literature that sees transparency combined with interactivity on government websites as most optimal, one may see a

reinforcing effect occur when effecting citizens trust judgements. Therefore the following is suggested:

H3: A high level of transparency cues and interactivity cues will reinforce each other and lead to more perceived trustworthiness of government

Research Method and Experiment Design

To test the three hypotheses outlined (H1-H3), a causal model is proposed, in which exposure to enhanced transparency and interactivity cues could engender citizen trust in government. As such an existing e-government site - the United Kingdoms’ GOV.UK, was manipulated, with its homepage displaying varying degrees of government transparency and interactive features. Citizens trust judgements of government were then gauged after

exposure to one of four randomized homepages by means of questionnaire.

The procedure thus consisted of three elements: (1) an instruction of the experiment with some general questions, (2) presenting the stimuli, and (3) a post-test questionnaire.

The experiment was administered via an online self-selection survey and resulted in 171 respondents. Participants were required to be British citizens and of voting age (18 and

over) to be included in the research. This sampling criterion and a completion rate of 81% returned 140 participants retained for the final analysis. The sample make up was 53% female to 47% male, with a mean age of 35.

It should be noted that the sample is not representative for the population of the United Kingdom (ONS, 2012). That said, in this research I am not interested in statistical generalization to the whole population per se, but rather to generalizing the theoretical relation between e-government transparency and interactivity with trust in government.

Nevertheless by carrying out such an experiment, it is the view that an empirically founded and more refined view on causal relationships between e-government transparency, interactivity and trust in government can be provided instead of a mere correlation

(Grimmelikhuijsen et al., 2013).

What is more, through administering it online many participants can be recruited relatively easily. Whilst this might come at the cost of a loss of control over the actions of participants, which in turn might weaken the internal validity of the experiment; this can be partially obviated by making clear-cut instructions for the participants to follow

(Grimmelikhuijsen et al., 2013). Materials

4 GOV.UK homepages were constructed with differing prominence of government transparency and government interactivity features. Figure 1 outlines the experiment conditions.

Figure 1. Table of manipulations. Condition 1 Transparency (-) Interactivity (-) Condition 2 Transparency (+) Interactivity (-) Condition 3 Transparency (-) Interactivity (+) Condition 4 Transparency (+) Interactivity (+)

To guide the webpage development, the webpages were kept as close as possible to the current design and layout of GOV.UK, and this was consistent throughout all 4

conditions. For instance each page was headed with the same GOV.UK banner, the fonts used, colours and positioning of text were kept as consistent as possible.The randomization process resulted in condition 1 N = 36, condition 2 N = 34, condition 3 N= 36, condition 4 N = 34 . The screen shots used in the experiment can be viewed in the Appendix.

Condition 1: neither transparency nor interactivity is manipulated (-/-). This was the control condition and all information is displayed in line with the most basic stage – ‘the billboard stage’ – of e-government as outlined by West (2004). Whereby there is static mechanisms to display information reports and publications, and no explicit transparency or interactivity features (West, 2004).

Condition 2: increased transparency, no interactivity (+/-). Here government transparency features are made prominent but interactive features are not enhanced. Manipulations are based on the Dahl definition of transparency as government operations being ‘sufficiently open to public view and citizens having the ability to see ‘how and what it is doing’, (Dahl cited by Moser, 2001; p3) and the transparency indicators adapted from the European Commission publication ‘2011 e-government Benchmark Pilot on Open

Government and Transparency’ (Lörincz et al., 2011). Which includes indicators of performance, policy making process, organizational structure and citizen’s access to

information (Lörincz et al., 2011). To this end, cues include ‘Access to information’, ‘How government is run’, the ability to review policies, and ‘watch decisions being made’ with a link to video footage of the government operations.

Condition 3: Transparency features are decreased and interactive mechanisms are enhanced with explicit cues that point to two-way communication between government and citizen (-/+). Also in line with e-government benchmark pilot, interactive indicators were

developed that engage social media; allow for citizen and government collaboration; and the ability to give feedback (Lörincz et al., 2011). For instance there is an ‘e-petition’ feature and readers are encouraged to ‘influence policy’ as well as information on ‘how to engage with government’, directly aimed at citizen participation.

Condition 4: Both transparency and interactivity mechanisms are enhanced (+/+), in what is designed to replicate the most optimal e-government stage outlined by West 2004, and replicate the best practice. All transparency and interactivity from condition 2 and 3 respectively are transferred to condition 4. Figure 1 illustrates the manipulations.

Manipulation check

In order to see if the webpage stimuli indeed did display varying degrees of

transparency and interactivity, a manipulation check was required. They were checked by asking a convenience sample (n28) to score the webpages on their perceived interactivity and transparency. Each participant viewed one of the randomized screen shots and then rated it 1 to 10 (10 being highest) on its interactivity (described as its ability to facilitate two-way communication) and its transparency (described as features that display the institutions’ working processes).

The manipulations proved successful. A one-way ANOVA showed the transparency evaluations differed significantly across all four conditions, F (3, 24) = 9.53, η2 = .54, p<.001. Post hoc testing revealed that participants indeed found the enhanced transparency condition (condition 2) considerably more transparent (M = 8.33, SD = .82) than the participants of the control group - condition 1 (M=5.13, SD= 1.25) p<.001. Condition 2, was also scored more transparent than the high interactivity, condition 3 (M=5.00, SD = 1.01) p<.001. As expected there was not a significant transparency mean difference between condition 2 and condition

4, as condition 4 contains identical transparency cues. However, condition 4 was rated as being significantly more transparent (M=7.12, SD=1.80) than condition 1 and condition 3.

Interactivity mean scores also differed significantly across the experimental groups, F(3, 24) = 33.20, η2 = .81, p<.001. A post hoc test showed that the enhanced interactivity page (condition 3) significantly scored higher (M=9.00, SD=.89), than condition 1 (M= 3.38, SD=1.85), p<.001, and condition 2 (M= 4.50, SD =.55), p<.001. Again Condition 4 which also includes interactivity cues didn’t differ significantly from condition 3, but was rated higher (M=7.88, SD= .55) than condition 1, p<.001.

Control variables

Background variables known to affect trust in government are considered to be gender, age, education, political preference, and whether an individual works for the for the government and will be controlled for in this research (King, 1997; Norris, 2001; Putnam, 2000 cited in Grimmelikhuijsen et al. 2013). A citizen’s familiarity with e-government, and public sector websites in general may also effect the results as it is likely to influence their evaluations of the e-government transparency and interactivity cues (Tolbert and Mossberger, 2006).

Further, another individual characteristic that may affect the results is the level to which an individual has a trusting personality. This is based on the assumption detailed by Jackob (2012), that people who generally tend to trust their fellow humans also express high levels of trust in institutions. This ‘interpersonal trust’ is described as a general predisposition or ‘willingness to ascribe benign intentions to others’ (Jackob, 2012; p100). Thus

interpersonal trust is held constant in this analysis. Interpersonal trust was measured using a 3 item scale constructed by Jackob (2012).

The respondents had to indicate whether they agreed to the following statements on a 5 point scale 1= strongly disagree to 5=strongly agree: “In principle, one can trust ones fellow citizens”, “Normally I regard other people with suspicion”, and “One can only rely on a few people” (Jackob, 2012; p110). A principal component analysis (PCA) was conducted on the 3 items. All items loaded on to one component (eigenvalue 1.83) and the reliability of the scale was relatively good (Cronbach’s α =.68)

Similarly an important individual variable in this research and expected to be significant in the context of trust towards government is a citizen’s political efficacy. Generally it is understood as ‘the feeling that individual political action does have, or can have, an impact upon the political process’ (Craig, Niemi, and Silver, 1990; p291). Empirical evidence has established that a citizen’s ‘self-efficacy and political involvement predict trust in government’ (Parent et al. 2005; p 724). Scholars dissect political efficacy in to two: (1) internal efficacy, referring to beliefs about one's own competence to understand and to participate effectively in politics; and (2) external efficacy, referring to beliefs about the responsiveness of governmental authorities and institutions to citizen demands (Craig et al, 1990; p292)

This is interesting too as it taps a citizen’s prior attitude towards government and attitude towards the democratic process in general, which may influence their trust judgements and also affect how they evaluate transparency and interactivity cues. Thus internal and external efficacy are controlled for. Internal and external efficacy were measured using scales developed by Craig et al (1990) for the National Election Studies. Internal

efficacy included a battery of 6 statements measured with a 5 point scale ranging from 1=strongly disagree to 5= strongly agree. All items loaded strongly on to one component (eigenvalue=3.60), and all items correlate positively with this component. The internal efficacy scale also proved to be reliable (α = .87).

External efficacy included 5 statements and again respondents were asked to score whether they agreed or disagreed (1=strongly disagree to 5=strongly agree). A PCA revealed that all items form a single uni-dimensional scale: only one component had an eigenvalue above 1 (eigenvalue=2.80), the scale had good reliability (α = .83).

Dependent variables

The main dependent variable Trust in Government was measured by asking participants specifically about the perceived competence, perceived benevolence and

perceived honesty of central government after they had viewed a webpage. All 12 items were measured on a 7 point scale 1= strongly disagree to 7=strongly agree; and were derived from a trust scale developed and validated by Grimmelikhuijsen and Meijer (2012).

A PCA with orthogonal rotation (varimax), revealed however that the 12 items formed two scales, as 2 components were extracted in the analysis, accounting for 74.47% of the variance. The items that cluster on the same component suggest that component 1

(eigenvalue= 7.74), measured honesty and benevolence evaluations of government, with the item ‘central government is honest’, having the strongest association (factor loading is .84). Whilst component 2 (eigenvalue = 1.19), had strong loadings from the competence items, ‘central government is capable’ had the strongest association (factor loading = .87). Table 1 shows the factors after rotation.

Table 1. Rotated Component Matrix of Trust in Government variable Component

1 2

‘The central government is honest’ ‘The central government honours its commitments’

‘The central government approaches its citizens in a sincere way’

‘The central government is sincere’

‘The central government acts in the interests of

.84 .84 .82 .80 .76 .36 .38

its citizens’

‘If citizens need help, central government will do its best to help them’

‘The central government is genuinely interested in the well-being of its citizens’

‘The central government is capable’ ‘The central government is effective’ ‘The central government is skilful’ ‘The central government is professional’ ‘The central government carries out its duties well’ .72 .68 .36 .37 .51 .40 .48 .87 .81 .79 .73 .70

*Bold figures represent factor loadings that indicate an independent scale. Coefficients suppressed if <.30

This was to be expected as we’ve seen from the previous discussion that ‘trust in government’ as a concept is an amalgamation of performance or competency evaluations and evaluations of intentions or motivations of the institution (Warren, 2004). As such the two components were retained in the analysis and two separate but related scales were created. The first scale I called ‘intentions trust’, as it combined the honesty and benevolence items; the scale had a strong reliability (α=.94). The second scale was named ‘performance trust’ as it pooled all competence items, also had good reliability (α=.91). Survey items for variables used in the post-test questionnaire can be found in the appendix.

Analyses

Univariate analyses of covariance (ANCOVA) tests were carried out to test the effects of the websites on trustworthiness evaluations of government. An ANCOVA approach was considered to be suitable because of the categorical nature of the independent variable (webpage condition) and continuous nature of the two dependent variables,

PERFORMANCE trust and INTENTIONS trust. Further, the use of an ANCOVA implies the existence of covariates. Covariates used were the background variables of gender; age;

education level; political preference; interpersonal trust; internal efficacy; external efficacy; whether a citizen was employed by government and how often they visit government

websites. These were used to control for confounding influences of background variables that are known to have a potential effect on trust in government. A dummy variable of

transparency was coded 0 and 1 if the webpage displayed transparent features (condition 2, 4), and a dummy variable for interactivity was coded in the same way for interactive pages (condition 3 and 4), to further aid the analysis.

Results

The hypotheses H1-H3 proposed that increased levels of transparency (H1) and interactivity (H2) on government websites will lead to increased perceived trustworthiness of government and that the interactive and transparent features will reinforce each other and again lead to increased perceived trustworthiness of government (H3).

From the initial factor analyses, it was necessary to create two dependent variables as the proposed trust in government scale, noticeably measured two separate aspects of trust, which is the performance and intentions of government. As such the hypotheses are discussed here in relation to these particular aspects of trust in government.

An ANCOVA with ‘Performance’ trust as the dependent variable, revealed that external efficacy was the only covariate significantly related to trust in the performance of government F(1,127) = 14.39, p <.001. After controlling for the effect of external efficacy, there was a significant effect of website condition exposure on evaluations of the

performance government, F(3,127) = 5.83, p <.05, η2 = .12. These results should be approached with caution however as Levene’s test is significant F(3,136) = 3.13, p= .028, indicating that the group variances are not equal (the assumption of homogeneity of variance has been violated). Nevertheless the standard deviations of the four conditions (condition 1

SD = 1.47, condition 2 SD = 1.18, condition 3 SD = .98, condition 4 SD = .99) do not differ too much to be problematic and the ANCOVA can be judged as reliable.

Pairwise comparisons showed that condition 2 the enhanced transparency page (M = 4.34), condition 3 the enhanced interactivity page (M = 4.51) and the combined transparent and interactive, condition 4 (M = 4.54) resulted in significantly higher performance

evaluations of government than the control group, condition 1 (M = 3.60), p <.05. These results lend some support to H1 and H2 as the perceived performance aspect of trust in government is significantly higher after viewing one of the treatment conditions. Table 2 displays the adjusted means after controlling for covariate external efficacy.

Table 2 Performance Trust means after viewing webpages

Condition Mean Std. Error N

1 (control group) 2 (enhanced transparency)

3 (enhanced interactivity) 4 (transparency and interactivity)

3.60 4.34 4.51 4.54 .18 .19 .18 .19 36 34 34 36

For a more forensic look, the analysis was run again, this time with use of the dummy variables, transparency and interactivity. The analysis showed a significant main effect of transparency F(1,127) =4.26,p<.05, partial η2 = .03 which again in part supports H1. Also whether a webpage was interactive was significantly related to trusting the performance of government F(1,127) = 9.06, p<.05, partial η2 = .07 lending some support to H2. However there proved to be no significant interaction effect between transparency and interactivity F(1,127) = 3.32, p = .07, which suggests that transparency and interactive features do not reinforce each other when related to trusting the performance of government, partially rejecting H3. Nevertheless it was interesting to note that this only just exceeds the significance threshold of p<.05

A second ANCOVA was run with the ‘Intentions’ trust as the dependent variable. Of the potential covariates tested, External efficacy again proved to be significant when related to the intentions aspect of trust in government F(1,127) = 9.12, p<.005 and Age also had some effect F(1,127) = 8.02, p<.005. After controlling for the effects of both covariates there was still a significant effect of website condition on the evaluations of the intentions of government F(3,127) = 3.41, p<.05, partial η2 = .08. The mean comparisons showed that only condition 3 (enhanced interactivity) significantly lead to higher trust in the intentions of government (M = 4.26, SD = 1.28), p<.05 , when compared to the control condition (M= 3.41, SD = 1.26).

Again a more detailed look at what was occurring was called for and the analysis was run again with the transparency and interactivity variables. The ANCOVA showed that pages with interactive features had a significant effect on trusting the intentions of government F(1,127) =4.17, p<.05, partial η2 =.03, suggesting some support for hypothesis 2. However transparent features did not affect citizens trust in government significantly F(1,127) = 1.14, p=.29. Testing for an interaction effect of transparency x interactivity on trusting the

intentions of government did show evidence of some effect, F(1,127) = 4.38, p<.05 partial η2 =.03. Which could suggest that when it comes to evaluating trust in the intentions of

government, transparency features alone are not sufficient in having an effect , however when combined with interactivity they appear to increase trust in the intentions government. Table 3 illustrates these results.

Table 3

Tests between – Subject Effects. Dependent variable: Intentions Trust

Source Df F Partial η2 p Transparency Interactivity Transparency x Interactivity (interaction effect) External efficacy Age 1 1 1 1 1 1.14 4.17 4.38 9.11 8.02 .01 .03 .07 .06 .03 .289 .043 .003 .005 .038 Error 127

Discussion and Conclusions

This research sought to probe the relationship between e-government use and trust in government. A more micro level approach than existing research was attempted through an experiment that isolated Transparency and Interactivity cues on government websites as potential mechanisms in inducing trust in government. With a research question that queried what effects do e-government transparency and interactivity cues have on citizen trust in government? It was hypothesized that websites with enhanced transparency will result in citizens perceiving central government as more trustworthy (H1); websites with enhanced interactivity will also lead to higher trust perceptions of government (H2) and finally that transparency and interactivity features will interact, to lead to higher trust in government (H3).

What do the results of the experiment display? Do transparency and/or interactivity cues on government websites lead to increase trust in government? Can they be classed as

trust mechanisms as hypothesized? Perhaps providing definitive answers to these questions eludes this present research as the initial analysis revealed that the proposed uni-dimensional scale of ‘trust in government’ used, actually measured two separate but related concepts: trust in the ‘performance’ of government and trust in the ‘intentions’ of government. Nevertheless, deconstructing trust in government in such a manner proved insightful, as it was interesting to note what precise facet of trust was being affected by the stimuli.

For instance, a positive effect of transparency on performance trust was found. Means differed significantly from the control group in the terms that webpages with transparent features lead to citizen’s trusting the performance of government more. This effect was further confirmed by follow up analysis that showed that whether a page was transparent or not, significantly affected trust in government performance. Whilst these results, do not directly speak to the overall effect of transparency on trust in government (H1), they show that an important aspect of trust is positively affected by the experience of transparency cues on government websites. The findings seem to ratify some of the discussed literature’s assumptions, which expect positive outcomes from transparency policies (Grimmelikhuijsen et al., 2013; Meijer, 2009; Norris, 2001, Welch et al, 2005).

However, there appeared to be no significant increase in trusting the ‘intentions’ of government after exposure to transparency cues. Which shows that the overall optimism regarding e-government transparency should be nuanced. Nevertheless some logical explanations can be made from this outcome. The previous discussion on government

transparency and trust noted that a prominent feature of transparency is that it illuminates ‘the working procedures not immediately visible to those not directly involved in order to

demonstrate the good working of an institution’ (Moser, 2001; p3). This emphasis on revealing the ‘good working’ of an institution, quite naturally lends itself to evaluations of performance.

Interactive cues also seem to engender higher trust in the performance of government. The treatment pages with enhanced interactivity significantly led to higher means than that of the control condition. The more forensic analysis further revealed that whether a webpage had interactive features or not significantly had a positive effect on citizens judgements on the performance of government.

A similar effect occurred in evaluating the intentions of government. With results revealing that interactive features significantly affected trust in the intentions of government. Whilst transparency appeared to have no effect on increasing trust in the intentions of

government, it was evident that interactivity seemed to be important. Again some logical inferences can be reasoned from this. Government interactivity, here, and in other research has been conceptualised as ‘two-way communication’, features that explicitly aim to involve and engage citizens in the decision making process. This arguably lends to favourable judgements of the motives of government, thus the subsequent positive effect on the

judgements of the intentions of government is understandable. Though these findings, once more cannot directly speak to overall trust in government, they do lend some support to the second hypothesis, as citizen’s seem to react more favourably to government interactivity in regard to evaluating these two important aspects of trusting government.

Perhaps the most interesting and noteworthy findings came from the analyses that sought to find out whether transparency and interactivity would interact and induce higher trust judgements of government (H3). In terms of the two trust measures of performance and intentions, some apparently contradictory results emerged.

Indeed there appeared to be some significant interaction effect between transparency and interactivity in producing more favourable evaluations of the intentions of government. This interaction effect appears to be lost however when citizens are asked to judge the performance of government as the analysis returned non-significant results for an interaction

effect occurring on trust in the performance of government. As such it’s difficult to conclude whether transparency and interactivity do complement each other in inducing favourable trust judgements of government.

Nevertheless these results should not be discounted as it was intriguing to note that the interaction effect in regards to performance judgements only slightly exceeded the significance threshold (p=.07). Possibly with follow-up research, and a larger sample than present (n=140), we may see a significant and stronger effect occur. A suggestion made even more poignant given the trajectory of a number of digital strategies towards prioritizing both transparency and interactivity in their digital offerings (West, 2004).

Certainly if follow-up research was conducted, some important limitations of this study could be rectified. For instance, the study could of benefited with not only a larger but more representative sample of the UK population. This would of allowed for more robust conclusions in terms of generalizations of the effects of transparency and interactivity.

Further, a particular constraint of this research was that the experiment design heavily relied on the manipulated, static screen shots. In reality government webpages are full of hyperlinks that lead to innumerable other pages. Thus attempting to grasp citizen’s reactions to just one page may have affected the present results. This is perhaps noticeable in the means of the pages that significantly produced higher trust judgements, as they tended to centre around the ‘neither agree nor disagree’ rating. A solution may be to add a longitudinal aspect to this study in search of more definitive answers, as some scholars note that trust or distrust in government is an attitude garnered over time (Kolsaker and Lee-Kelley, 2008) . Which begs the need to study the effects of citizens’ repeat experiences with e-government with enhanced transparency and interactivity features.

In addition, the highly manipulated stimuli in the form of the GOV.UK screen shots, may have also compromised some external validity of the research. As government websites

display a wealth of information and services that in this experiment were necessarily altered. However a core ambition of my research was to isolate transparency and interactivity features and observe whether they can affect trust in government, so these cues were made more or less prominent accordingly with this goal in mind.

These caveats aside, there is some evidence here that transparency and interactivity overall did seem to affect trust judgements in a positive direction. This speaks to the argumentation that trust judgements can be formed from less involved thinking and can be influenced by certain qualitative cues, in-line with the Elaboration likelihood model outlined by Petty and Cacioppo (1986).

Noticeably, of the variables expected to have a confounding effects in the analyses, only external efficacy had a strong significant effect on both trust judgements. The results show that this construct, more so than the transparency and/or interactivity of a website, significantly influences and explains citizens judgements of the performance and intentions of government. As citizen’s whom believe government is receptive and responsive to their demands and participation in politics are more likely to have higher trust judgements (Parent et al, 2005). Government’s may thus chose to focus efforts on individuals with higher pre-existing external efficacious attitudes if they wish for their online efforts succeed. However they may be buoyed by the fact that there are still some significant positive influence of transparent and interactive features on citizen’s attitudes after controlling for the external efficacy variable.

In sum, what are the theoretical implications of these findings? Where do they fit in terms of the e-government optimist vs pessimist debate? Clearly they appear to mesh well with more positive e-government evaluations as there is evidence of some small but significant causative power of e-government transparency and interactivity on higher trust

judgements; though the expectations of e-government transparency and interactivity as a panacea for government distrust should muted.

Whilst the rationale of governments to focus on transparency and interactivity is not necessarily bolstered by evidence of mostly neutral responses, some weight can be placed behind the fact that webpages with transparency and interactivity produce more favourable trust judgements than webpages that function as ‘billboards’, displaying only service

information, reports and publications (West 2004).

At most, this research shows a need for a more forensic look at e-government features and their effects on citizen’s attitudes. At the least, it contributes to an under researched aspect of political communication and highlights the importance of a growing function of government operations.

References

Anstead, N., & Chadwick, A. (2009). Toward a comparative institutional approach. In A. Chadwick & P. N. Howard (Eds.), The Routledge handbook of Internet politics (pp.56–71). Abingdon, UK: Routledge. Retrieved from http://www.handbook-of-internetpolitics.com/pdfs/Nick_Anstead_Andrew_Chadwick_Parties_Election_Campa igning_and_Internet.pdf

Bertot, J. C., and P. T. Jaeger. (2006). User-centered e-government: Challenges and benefits for government web sites. Government Information Quarterly 23:163–68.

Coursey, D., & Norris, D. F. (2008). Models of e‐government: Are they correct? An

empirical assessment. Public Administration Review, 68(3), 523-536.

Craig, S. C., Niemi, R. G., & Silver, G. E. (1990). Political efficacy and trust: A report on the NES pilot study items. Political Behavior, 12(3), 289-314.

Davis, J. (1998). Access to and Transmission of Information: Position of the Media.

Grimmelikhuijsen, S. G., & Meijer, A. J. (2012). The Effects of Transparency on the Perceived Trustworthiness of a Government Organization: Evidence from an Online Experiment. Journal of Public Administration Research and Theory.

Grimmelikhuijsen, S.G, Porumbescu, G., Hong, B., & Im, T. (2013). The Effect of

Transparency on Trust in Government: A Cross‐National Comparative

Experiment. Public Administration Review.

Hardin, R (2013) Government without trust, Journal of Trust Research, 3(1), p32-52

Ha, L., & James, E. L. (1998). Interactivity reexamined: A baseline analysis of early business web sites. Journal of Broadcasting & Electronic Media, 42(4), 457-474.

Hood, C(2006) Transparency in historical perspective. In Hood, C and Heald, D, (eds.) Transparency: the key to better governance? Proceedings of the British Academy (135). Pp 3-23 Oxford University Press, Oxford, UK

Jackob, N (2012) The tendency to trust as individual predisposition–exploring the

associations between interpersonal trust, trust in the media and trust in institutions. Communications- European journal of Communication research. 37 (1), 99-120

Kiousis, S. (2002). Interactivity: a concept explication. New media & Society, 4(3), 355-383. Kolsaker, A., & LeKelley, L. (2008). Citizens' attitudes towards government and

e-governance: a UK study. International Journal of Public Sector Management, 21(7), 723-738.

Kong, D. T. (2012). Intercultural experience as an impediment of trust: Examining the impact of intercultural experience and social trust culture on institutional trust in

Lörincz, B., Tinholt, D., Van der Linden, N., Tol, T., Van Velzen, A., Schindler, R. . . . Cattaneo, G. (2011). 2011 eGovernment Benchmark Pilot on Open Government and Transparency

Moon, M. J. (2003, January). Can IT help government to restore public trust? Declining public trust and potential prospects of IT in the public sector. InSystem Sciences,

2003. Proceedings of the 36th Annual Hawaii International Conference on (pp. 8-pp).

IEEE.

Morgeson, F. V., VanAmburg, D., & Mithas, S. (2011). Misplaced trust? Exploring the structure of the e-government-citizen trust relationship. Journal of Public

Administration Research and Theory, 21(2), 257-283.

Moser, C. (2001). How open is ‘open as possible’? Three different approaches to transparency and openness in regulating access to EU documents. IHS Political

Science Series, 80.

Norris, P. (2001). Digital divide: Citizen engagement, information poverty and the Internet worldwide. New York: Cambridge University Press.

Office of National Statistics, 2011 Census: Population Estimates for the United Kingdom (2012) Retrieved from http://www.ons.gov.uk/ons/dcp171778_292378.pdf

Parent, M., Vandebeek, C. A., and Gemino, A.C.(2005). Building citizen trust through e-government. Government Information Quarterly 22, 720–36.

Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In Communication and Persuasion (pp. 1-24). Springer New York.

Putnam, R. D. (1995). Bowling alone: America's declining social capital.Journal of

democracy, 6(1), 65-78

Rafaeli, S., & Sudweeks, F. (1997). Networked interactivity. Journal of Computer‐Mediated

Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Not so different after all: A cross-discipline view of trust. Academy of management review, 23(3), 393-404. Smith, M. L. (2010). Building institutional trust through e-government trustworthiness

cues. Information Technology & People, 23(3), 222-246.

Tolbert, C. J., & Mossberger, K. (2006). The Effects of E‐Government on Trust and

Confidence in Government. Public Administration Review, 66(3), 354-369

Thomas, J. C., and G. Streib. (2003). The new face of government: Citizen-initiated contacts in the era of e-government. Journal of Public Administration Research and Theory 13:83–102

Uslaner, E. M. (2004). Trust, civic engagement, and the Internet. Political Communication, 21(2), 223-242.

Warren, M.E. (2004), “Trust in democratic institutions”, in Ankersmit, F. and te Velde, H. (Eds), Trust: Cement of Democracy, Peeters, Leuven

Welch, E. W., & Hinnant, C. C. (2003, January). Internet use, transparency, and interactivity effects on trust in government. In System Sciences, 2003. Proceedings of the 36th

Annual Hawaii International Conference on (pp. 7-pp). IEEE.

Welch, E. W., Hinnant, C. C., & Moon, M. J. (2005). Linking citizen satisfaction with e-government and trust in e-government. Journal of Public Administration Research and

Theory, 15(3), 371-391.

West, D. M. (2004). E‐Government and the Transformation of Service Delivery and Citizen

Appendix 1.

Webpage stimuli Condition 1

Appendix. 2

Post-test questionnaire

Items for “Perceived trustworthiness of government” Perceived competence

Based on this website (Condition: 1, 2, 3, 4) I think that….

1. central government is capable. 2. central government is effective. 3. central government is skillful. 4. central government is professional.

5. central government carries out its duty very well. Perceived benevolence

Based on this website (condition: 1,2,3,4) I think that…

1. If citizens need help, central government will do its best to help them.

2. That central government acts in the interest of citizens. 3. The central government is genuinely interested in the well-being of citizens.

4. The government approaches citizens in a sincere way. Perceived honesty

Based on this website (1,2,3,4) I think that,

1. The central government is sincere.

2. The central government honors its commitments. 3. The central government is honest.