Transparency is in the eye of the beholder: the effects of identity and negative perceptions on ratings of transparency via surveys

Abstract

Surveys are a commonly used means of measuring transparency levels, but they are potentially vulnerable to perceptual biases. This study sought to examine perceptual differences by the respondents’ identities as general citizens or public employees, and the possible negative perceptions that one group may have of the other concerning responses to a survey-based measure of transparency. The survey was designed on the basis of existing literature, suggesting that transparency has up to six facets. Two samples were taken: from citizens who visited district offices to file civil applications during the survey period; and from public employees involved in processing these applications. A total of 472 surveys were used for analysis: 233 citizens and 239 public employees. The results indicated that the two groups had different understandings of transparency. Data from public employees produced a three-factor solution, which was labeled as Efficiency, Reliability, and Access. For citizens, a two-factor solution was a better fit, with the factors being described as Accessibility (a wider notion than Access) and Utility. The findings suggest that public employees adopt a somewhat technical view of transparency, whereas citizens have more practical concerns about it. Only citizens’ unfavorable perception of public employees had a negative influence on the level of transparency. This study contributes to the understanding of how public employees and citizens have qualitatively different perceptions of transparency.

Points for practitioners

To assess progress in governmental transparency, we must measure it, and surveys offer an accessible and potentially cost-effective approach. However, the survey responses of citizens and public employees show that they understand transparency in qualitatively different ways, with citizens’ perceptions of transparency also influenced by their perceptions of public employees. If governments are to increase public trust in policymaking and administration, they must focus on improving transparency as it is understood by the public rather than how it is understood by public servants.

Keywords

administration and democracy e-government public administration transparency trust

Introduction

Studies have highlighted the positive effects of transparency on democracy and political legitimacy, good governance, the elimination of corruption, trust, accountability, and national competitiveness (Bauhr and Grimes, 2012; Grimmelikhuijsen and Meijer, 2014; Heald, 2003; Hollyer et al., 2011; Park and Blenkinsopp, 2011; Rawlins, 2009; Vishwanath and Kaufmann, 2001). Many governments and agencies have shown a strong commitment to transparency and have taken steps to incorporate it into their policies (Coglianese, 2009; Otenyo and Lind, 2004: 288; Sternstein, 2011). In order to assess the progress being made in improving transparency, it is vital to be able to measure it. Yet, it has been found to be difficult to measure directly. When attempting the task, proxy measures have been used, which include access-to-information laws, e-government, official websites, a free press, data dissemination, feedback on public policies and practices, delays in information disclosures, and the number of citizen complaints about the quality of information (Bertot et al., 2010; Hollyer et al., 2011: 1194; Islam, 2006). However, surveys remain the most widely used method of gauging transparency (Da Cruz et al., 2015: 10; Rawlins, 2009). Any survey of transparency is inevitably a survey of perceived transparency, a ‘measure of opinion’ regarding ‘what [people] think of transparency in government’ (Sternstein, 2011: 25), with citizens’ evaluations being influenced by many factors, ‘in complex and changing ways’ (Wang and Gianakis, 1999: 550). Therefore, responses to survey-based measures of transparency may be prone to perceptual biases, such as self-serving interests and one’s negative perception of another, depending upon who completes the survey. Surveys asking about the perception of transparency in government agencies or public services risk obtaining data that are as much subjective as a rating of satisfaction with public service (Rawlins, 2009). Furthermore, public employees have all the information on the transparency levels of public services they provide, whereas citizens do not. While public employees are required to grasp the problems in information disclosure and to improve their ability to increase transparency, citizens’ perceptions of transparency may vary greatly from those of public employees. This is not necessarily a problem in cases where agencies are concerned with what citizens think about transparency, but it is a concern for researchers looking to use perception of transparency as a proxy for actual transparency.

Unlike the majority of earlier transparency studies tied to identifying properties of transparency and assessing the level of government transparency using an index, this study focuses on whether or how transparency measurements based on survey methods are affected by the respondents’ identities and negative perceptions. The administration of civil applications was chosen as a case to examine the differences of perceived transparency between two groups: citizens using the service and public employees working within it. This study contributes to the understanding of how survey-based measurement of transparency can be biased by the respondents’ identities, the negative perceptions one group may have of the other, and how public employees and citizens have qualitatively different perceptions of transparency.

Literature review

Transparency and its measurements: an overview

Over the last decade, there has been a rapidly growing consensus among researchers and practitioners about the need for greater transparency in government and public administrations (Bauhr and Grimes, 2012). The improvement of transparency as a policy initiative, particularly along with good governance, has been well noted to involve the public in government decision-making (Coglianese, 2009; Islam, 2006). Transparency, in a governmental context, can refer to: transparency in decision-making processes or of an organization in and of itself; transparency of public services, budgets, or policy content; or transparency of policy outcomes or effects (Grimmelikhuijsen and Welch, 2012; Rawlins, 2009; Relly and Sabharwal, 2009). The present study focused on transparency in public service processes, in this instance the processing of citizens' applications for permits, licenses, and registrations.

For full transparency in providing public services, all information should be accessible to anyone, at any time, and in any place, so that citizens are informed of the whole public service or decision-making process that might affect their interests. However, the simple availability of information does not constitute full transparency. Useful information should be disclosed in a timely and convenient way, so that people can easily determine the expected benefits and risks. Raw data need to be processed to meet public interests, producing complete and substantial information. Based on these characteristics of information required for full transparency, scholars have developed a multidimensional approach to measuring transparency. Rawlins (2009: 84) suggested for organizational transparency that information should be ‘complete, relevant, verifiable, accurate, balanced, comparable, clear, timely, reliable, and accessible’. When discussing transparencies of public services, Vishwanath and Kaufmann (2001) identified five attributes: accessibility, comprehensiveness, relevance, quality, and reliability.

Most instruments developed to gauge transparency were for subjective assessments. For example, Bauhr and Grimes (2012) measured government transparency via a survey, asking public administration experts to respond to some pertinent questions concerning transparency. Caamaño-Alegre et al. (2013) used a Likert-type survey questionnaire composed of 15 items to measure budget transparency in 33 small municipalities. The International Institute for Management Development and the World Economic Forum have measured the transparency of government policy or policymaking by an expert or business leader survey on a regular basis in order to announce rankings of national competitiveness. However, there have also been studies that have employed objective indicators. For example, Hollyer, Rosendorff, and Vreeland (2014) suggested a transparency index based on information available on official local government websites. Esteller-Moré and Otero (2012) developed an index of fiscal transparency computed by identifying whether a municipality has provided the required budgetary information on the Internet. Da Cruz et al. (2015: 10) stated that the use of surveys in assessing transparency levels entails the problem of self-administered responses and, in turn, of inconsistencies with ‘the actual level of transparency’.

Identity, self-interest, and biased views

The results of transparency surveys may be influenced by respondents’ identities. Drawing on Erikson (1964), Vadera et al. (2009: 559) note that ‘identity is rooted in the very core of one's being', and stressed that identity affects a person's cognition, judgments, and behaviors. Identity has been identified as a typical source of self-serving bias in a survey, which is the tendency for people to interpret information in ways that serve their own interests. This occurs ‘where an individual’s preferences affect his beliefs in an optimistic direction, one favoring his own payoff’ (Kaplan and Ruffle, 1998: 243). Most people tend to have a bias in self-assessment, believing that they are ‘above average’ in their abilities and performance (Mezulis et al., 2004; Zábojník, 2004). This bias often appears as the result of the efforts to increase or protect one’s self-esteem (Felson, 1981). It operates in eliciting judgments of performance via a survey, and especially tends to be greater when not enough information is given, or when the criteria are unclear for assessing performance.

The relationship between the principal and the agent, which is used widely in the study of performance management in the public sector, is a special arrangement in which principals and agents differ in their interests (Heinrich and Marschke, 2010: 187). Agents are likely to perceive that they are trying fully to meet the principals’ expectations, while principals are seldom satisfied with the agents’ performances. Public employees are likely to report a higher perceived transparency for several reasons. First, when administering the system, they are aware of more channels and procedures to disclose information produced by the government, and thus may estimate a higher level of transparency than citizens. Second, their identity as public employees would be threatened by acknowledging that they worked for an institution that had poor transparency, consequently reporting higher levels of transparency than actually exist. Agents are likely to inflate their own performance, while the principals who pay for it might query or dispute this performance.

Negative perceptions

Negative perceptions are also likely to foster a biased report of transparency between citizens and government, which is often based on opposing needs or demands. Dissatisfaction with and mistrust of each other promote negative perceptions, which lead to a bias in an unfavorable direction. On the other hand, a favorable perception in a dyadic relationship may also have a significant impact on performance ratings (Lefkowitz, 2000: 69; Lefkowitz and Battista, 1995; Varma and Pichler, 2007; Varma et al., 2005). According to Marvel (2015: 2, 21), citizens have traditionally developed negative attitudes regarding government performance through ‘repeated exposure to anti-public sector messages’, and, as a result, people ‘automatically and unconsciously associate public sector organizations with inefficiency, inflexibility, and other pejoratives’. The negative views of such citizens may influence individuals to further downgrade performance in the public sector to lower than it actually is.

Hypotheses

Based on the previous literature reviews on the relationships between transparency, identity, and negative perception, along with the aims of this study, the following hypotheses were tested:

Hypothesis 1 (H1): A person’s identity (citizen versus public employee) will influence his or her reports of perceived transparency.

Hypothesis 2 (H2): Citizens will tend to perceive lower levels of transparency in public services than public employees.

Hypothesis 3 (H3): A person’s negative perception (citizen versus public employee) of the other will influence his or her reports of perceived transparency.

Hypothesis 4 (H4): Citizens with increased negative perceptions of public employees will perceive lower levels of transparency in public services.

Methods

Research design and data collection

We selected the administration of civil applications as a particularly appropriate site for the study, being that it is a process in which systems must be rigorously transparent in order to prevent corruption in government (Cho and Choi, 2004). The research design – in which samples were taken from citizens who visited district offices to file civil applications during the survey period and from the public employees involved in processing these applications – allowed us to analyze differences in perceived transparency between citizens and public employees, and thus to assess the influence of the identity and negative perceptions of each group. Both parties were directly involved in the process so we could be confident that these perceptions were grounded in actual experience. Data were collected from three districts (Yeongdeungpo-Gu, Seodaemun-Gu, and Gangdong-Gu), which were randomly selected out of the 25 autonomous districts of the Seoul Metropolitan Government, South Korea. After gaining permission from the district offices, two investigators visited them to distribute and collect the surveys. Face-to-face surveys were administered to 200 individuals from each district for a total of 600 individuals: 300 citizens and 300 public employees. Participants were assured of confidentiality, and it was explained that the data would be used for academic purposes only. Data gathering was completed in July 2010. Of the 600 citizens and public employees that were approached to participate, 485 completed a survey, giving a response rate of 80.8%. Removing incomplete surveys resulted in a total of 472 surveys for analysis: 233 citizens and 239 public employees. Table 1 provides detailed demographic information on the sample.

Table 1.

Demographic characteristics of the sample.

	Public employees	Citizens	Total
Gender
Male	115 (48.1)	148 (63.5)	263 (55.7)
Female	124 (51.9)	85 (36.5)	209 (44.3)
Age (years)
<30	24 (10.0)	41 (17.6)	65 (13.8)
30–39	74 (31.0)	43 (18.5)	117 (24.8)
40–49	94 (39.3)	63 (27.0)	157 (33.3)
≥50	47 (19.7)	86 (36.9)	133 (28.1)
Level of education
High school	31 (13.0)	69 (29.6)	100 (21.2)
4-year university	192 (80.3)	154 (66.1)	346 (73.3)
Postgraduate	16 (6.7)	10 (4.3)	26 (5.5)
Total	239 (50.6)	233 (49.4)	472 (100.0)

Notes: Figures in parentheses are the percentages of respondents for each condition. N = 472.

There were some demographic differences between the two subsamples. For example, public employees were more likely to hold four-year university and postgraduate degrees. Nonetheless, the results were considered to have appropriately represented the demographic samples of both citizens and public employees.

Measures

Scholars have developed various indexes for measuring different kinds of transparencies, including transparencies in websites (Hollyer et al., 2014; Pina et al., 2007), budget or fiscal transparencies (Caamaño-Alegre et al., 2013; Heald, 2003), local government transparencies (Da Cruz et al., 2015), transparencies in service provision (Vishwanath and Kaufmann, 2001), and organizational or government transparencies (Bauhr and Grimes, 2012; Rawlins, 2009). Since the purpose of this study was to measure the perceived transparencies of the government’s provision of public services, a scale was developed that comprised 18 questionnaire items, drawing on Vishwanath and Kaufmann’s (2001: 42) conceptual definition of transparency in providing services. The 18 items comprised three for each of the six attributes of transparency: access, comprehensiveness, timeliness, relevance, quality, and reliability. Public employees and citizens were invited to respond to these items as they related to the transparency of the Gu Office administration of civil applications on a five-point scale, where 5 = strongly agree and 1 = strongly disagree (see Table 2).

Table 2.

Question items for transparency.

Attributes	Items
Access	t1	Few expenses are needed for citizens to get information.
	t2	Citizens can readily access necessary information anywhere.
	t3	The information is available when needed.
Comprehensiveness	t4	The Office provides information that is easy to understand.
	t5	Clear explanations are given.
	t6	Application instructions are easy to follow.
Timeliness	t7	Citizens can confirm the status of their application at any time.
	t8	The Office provides information when changes are made.
	t9	The information is provided in a timely fashion.
Relevance	t10	Civil application guides include all needed information.
	t11	The Office provides information that is essential for the applicants.
	t12	The disclosed information is relevant to citizens in helping them not to make mistakes in their applications.
Quality	t13	The disclosed information is complete.
	t14	The Office provides information that is accurate.
	t15	There are seldom flaws in the information.
Reliability	t16	The information is largely trustworthy.
	t17	The information is correct.
	t18	The Office provides information that is reliable.

First, an exploratory factor analysis for a total sample of 472 responses was completed in order to test the scale’s validity of transparency and to uncover the factor structure underlying that transparency. The results are shown in Table 3.

Table 3.

Results of factor analysis on the items for transparency.

Scale/items		Factor loading
Scale/items		F1	F2
t1	Few expenses are needed for citizens to get information.	.248	.737
t2	Citizens can readily access necessary information anywhere.	.280	.881
t3	The information is available when needed.	.311	.834
t4	The Office provides information that is easy to understand.
t5	Clear explanations are given.
t6	Application instructions are easy to follow.
t7	Citizens can confirm the status of their application at any time.
t8	Office provides information when changes are made.
t9	The information is provided in a timely fashion.
t10	Civil application guides include all needed information.
t11	Office provides information that is essential for the applicants.	.666	.252
t12	The disclosed information is relevant to citizens in helping them not to make mistakes in their applications.	.715	.288
t13	The disclosed information is complete.	.738	.278
t14	Office provides information that is accurate.	.741	.279
t15	There are seldom flaws in the information.	.767	.198
t16	The information is largely trustworthy.	.762	.363
t17	The information is correct.	.796	.236
t18	Office provides information that is reliable.	.808	.236
Eigenvalues		6.190	1.150
Cumulative %		56.273	66.726
Reliability (Cronbach’s α)		.916	.833

Notes: Factor loadings > .60 are in boldface.

Data from the whole sample gave a two-factor solution having an eigenvalue of >1.0, where factor F1 was labeled as Information Quality and F2 as Access, while a total of seven items were eliminated because they failed to meet the minimum criterion of having a factor loading of .4 or above. The proportion of variance accounted for by these two factors was 66.73%. Cronbach’s α values for the two factors, commonly used as a measure of internal consistency, were .916 and .833, respectively. To identify whether there were systematic differences between citizens and public employees in how they responded to a survey-based measure of transparency, factor structures were examined using an exploratory factor analysis on each of the two subgroups. The results are shown in Table 4.

Table 4.

Results of factor analysis on the items for transparency.

For public employees (n = 239)				For citizens (n = 233)
Scale/items	Factor loading			Scale/items	Factor loading
Scale/items	F1	F2	F3	Scale/items	F1	F2
t1	.086	.277	.720	t1
t2	.253	.135	.868	t2	.624	.321
t3	.341	.138	.779	t3
t4				t4	.778	.196
t5				t5	.713	.333
t6				t6	.808	.286
t7				t7	.820	.192
t8	.689	.166	.341	t8	.730	.295
t9	.772	.315	.190	t9	.732	.354
t10	.690	.272	.300	t10
t11	.824	.200	.173	t11	.232	.776
t12	.783	.241	.091	t12	.295	.787
t13				t13	.298	.807
t14				t14	.286	.747
t15				t15
t16	.271	.771	.318	t16
t17	.247	.875	.153	t17
t18	.359	.849	.171	t18
Eigenvalues	5.771	1.216	1.112	Eigenvalues	6.008	1.215
Cumulative %	52.463	63.518	73.623	Cumulative %	54.620	65.665
Cronbach’s α	.878	.896	.800	Cronbach’s α	.903	.848

Notes: Factor loadings > .60 are in boldface.

The results of the factor analysis revealed that data from public employees alone produced a three-factor solution: F1 = Efficiency, F2 = Reliability, and F3 = Access. For citizens, a two-factor solution was a better fit, with the factors referred to as F1 = Accessibility (a wider notion than Access) and F2 = Utility. The factor structures of the two groups differed from each other and those of the whole sample. The factor analysis produced an apparently robust two-factor solution, but a three-factor solution emerged in the sample of public employees. Another two-factor solution (differing from that for the whole sample) emerged for the subsample of citizens. This result suggests that the two groups used different dimensions when perceiving the transparency of the same public service.

Previous literature on negative perceptions between two or more individuals or groups has suggested two primary dimensions: relationship and task (Solansky et al., 2014: 83; Szulanski et al., 2008: 467). When one individual or group has a negative perception of the other, they are prone to underestimate their counterpart in terms of relationship and task. To measure negative perception, 10 items were created. Each dimension of relationship and task was comprised of five items, drawing on previously developed conflict scales (Rahim, 1983). After instructing the public employees to indicate their perceptions of citizens who visit district offices for the public service of civil applications, we asked them to respond to 10 questions on a five-point scale (5 = strongly agree to 1 = strongly disagree). This was repeated with the citizens who visited to file their applications, asking about their perceptions of the public employees. The items are shown in Table 5.

Table 5.

Question items for negative perception.

	Items
Relationship	Public employees (or citizen applicants):
	c1	They tend to be not favorable to me.
	c2	They sometimes make me feel unpleasant.
	c3	They are not my supporters.
	c4	They have different views from mine.
	c5	They are in opposition to me.
Task	c6	They tend not to agree with my opinions about the work.
	c7	They use different standards.
	c8	They have different interests from mine.
	c9	They hardly cooperate with me.
	c10	They are in conflict with me.

A factor analysis was run to explore the underlying structures of the 10 items developed to measure negative perceptions. Contrary to our expectation that two clusters of items – relationship and task – would be identified, only one component was extracted, which explained 70.11% of the variance. The reliability α for the items of a factor (n of items = 10) was .952. The responses to the 10 items were averaged to form a variable of negative perception.

Results

To examine the magnitude and direction of the association between the variables, a Pearson’s correlation analysis was performed. Table 6 reports the means, standard deviations, and correlations among the study variables.

Table 6.

Descriptive statistics and correlations between the important variables.

	MEAN	SD	TR1	TR2	ID	NP	GN	AG	ED
TR1	3.66	.79	1.00
TR2	3.71	.65	.620***	1.00
ID	.51	.50	.358***	.328***	1.00
NP	2.76	.88	−.067	−.162***	.198***	1.00
GN	.56	.50	.000	.043	−.155***	−.052	1.00
AG	2.84	1.14	.029	−.053	−.136**	.092*	.099*	1.00
ED	1.84	.49	.015	.000	.193***	.031	−.007	−.215***	1.00

Notes: N = 472. Correlations with ID and GN are Spearman’s rho. The responses for gender were coded as 1 = male, 0 = female; those for age as 1 = <29, 2 = 30–39, 3 = 40–49, and 4 ≥ 50; and those for level of education as 1 = less than a high school diploma or equivalent, 2 = junior college degree, 3 = four-year university degree, and 4 = postgraduate degree. AG: age; ED: level of education; GN: gender; ID: Identity (public employees = 1, citizens = 0); NP: public employees’ negative perception of citizens or vice versa; TR1: Information Quality; TR2: Access. *p < .05; **p < .01; ***p < .001; two-tailed tests.

TR1 (Information Quality) and TR2 (Access) were very closely related to each other (r = .620, p < .001). ID was significantly associated with TR1 and TR2 (r = .358, p < .001; r = .328, p < .001), indicating that public employees had a higher tendency to positively rate the two types of transparencies than citizens. NP (negative perception) did not have a significant relationship with TR1 (r = −.067, p > .05), but its relationship with TR2 was significant and negative (r = −.162, p < .001). NP had a positive relationship with ID (r = .198, p < .001), indicating that public employees experience greater negative perceptions of citizens. Regarding demographic variables, some significant associations between ID and GN, AG, and ED indicate significant differences in the ratios of gender, and the distribution of age and education level, between citizens and public employees. For these reasons, demographic variables were controlled for, testing the hypothesis regarding negative perception and examining how they affect the perceptual differences between citizens and public employees.

Effects of respondent identity on transparency

As noted earlier, we found that public employees and citizens had qualitatively different understandings of the level of transparency. Responses from public employees indicated a three-factor solution (Efficiency, Reliability, and Access), while those from citizens produced a two-factor solution (Accessibility and Utility). These results showed a significant difference between citizens and public employees with regards to their perceptions of transparency. The results supported H1 (a person’s identity (citizen versus public employee) will influence his or her reports of perceived transparency). This study examined in more detail whether citizens, as compared to public employees, perceived lower levels of transparency in public services, cross-correlating the means of the two groups in terms of perceived transparency (see Table 7).

Table 7.

Differences in transparency perception between public employees and citizens.

	Dimensions of transparency	Mean (s.d.)
Public employees (n = 239)	Efficiency	3.78 (.63)
	Reliability	4.01 (.65)
	Access	3.93 (.64)
Citizens (n = 233)	Accessibility	3.22 (.75)
Citizens (n = 233)	Utility	3.46 (.71)

The means of the three dimensions (Efficiency, Reliability, and Access) for public employees was greater than those of the two dimensions (Accessibility and Utility) for citizens, and high enough to ensure significance. H2 (citizens will tend to perceive lower levels of transparency in public services than public employees) predicted that citizens have fewer positive perceptions regarding the transparency of public services than public employees do. This hypothesis was supported.

Effects of a negative perception on transparency

Before testing the effects of a negative perception on transparency, we also compared the means of negative perception between the two groups (citizens and public employees) to identify which group was more negatively perceived by the other. When first running Levene’s test for equality of variances, the two groups were found to be equal (F = .411, p = .522), so a t-test was run with an assumption of equality (see Table 8).

Table 8.

Differences of negative perception between public employees and citizens.

	Mean (s.d.)		Mean differences	t-value	Sig. (2-tailed)
	Public employees (n = 239)	Citizens (n = 233)	Mean differences	t-value	Sig. (2-tailed)
Negative perception	2.93 (.87)	2.58 (.86)	.35	4.379	.001

The results showed a significant difference between citizens and public employees in their negative perceptions of each other. The mean negative perception of public employees was 2.93, while that of citizens was 2.58, and the mean difference between the two groups was .35 (t = 4.379, p < .001). This suggests the view that public employees rate citizens more negatively than vice versa, at least with regard to the public service of civil applications. To analyze the effects of negative perception, we performed a linear regression analysis, controlling for demographic variables. Table 9 details the results of the regressions of negative perception on three types of transparency.

Table 9.

Effects of negative perception on transparency (public employees).

Predictors	Dependent variables: Transparency
Predictors	Efficiency		Reliability		Access
PNP	−.072	(−.101)	−.101*	(−.137)	.014	(.019)
GN	−.054	(−.044)	.049	(.038)	−.061	(−.048)
AG	.090	(.130)	.134**	(.187)	.111*	(.158)
ED	.030	(.021)	−.051	(−.035)	−.017	(−.012)
Constant	3.713***		4.022***		3.656***
Adjusted R square	.009		.043		.007
F-value	1.512		3.691		1.446
Sig.	.200		.006		.219

Notes: N = 239. The figures in parentheses are standardized regression coefficients. See Table 4 for the three types of transparencies. AG: age; ED: level of education; GN: gender; PNP: public employees’ negative perception of citizens. *p < .05; **p < .01; ***p < .001; two-tailed tests.

When three types of transparencies were regressed on independent variables, only the Reliability model, which represents the reliability of government information provided to the public, was significant (F = 3.691, p = .006). The variance explained was .043, and PNP (public employees’ negative perception of citizens) was significant but negatively associated with perceived transparency (b = −.101, p < .05). This indicates that public employees tend to rate the level of transparency higher (reliability of information provided by themselves) when their negative perception of citizens is low. Table 10 presents the regression results of citizens’ negative perceptions of public employees on two types of transparencies.

Table 10.

Effects of negative perception on transparency (citizens).

Predictors	Dependent variables: Transparency
	Accessibility		Utility
CNP	−.299***	(−.339)	−.234***	(−.281)
GN	.124	(.079)	.083	(.056)
AG	.057	(.100)	−.048	(−.089)
ED	−.070	(−.049)	−.165	(−.122)
Constant	3.859***		4.442***
Adjusted R square	.110		.092
F-value	8.151		6.910
Sig.	.000		.000

Notes: N = 233. The figures in parentheses are standardized regression coefficients. See Table 4 for the two types of transparency. AG: age; CNP: citizens’ negative perception of public employees; ED: level of education; GN: gender. *p < .05; **p < .01; ***p < .001; two-tailed tests.

Unlike the results from the analysis of public employees, both the Accessibility and Utility models of data from citizens were significant. In the Accessibility model, the variance explained was .110 (F = 8.151, p < .001). The more negative the perception civil applicants had of public employees, the lower they rated the level of transparency regarding the accessibility to provided information (b = −.299, p < .001). The Utility model, which represents the usefulness of information, was also significant (F = 6.910, p < .001), with about 9% of the variance explained. Controlling for other variables, negative perception was a significant predictor of Utility (b = −.234, p < .001). In the two models, civil applicants’ unfavorable perception of public employees had negative effects on the perceived level of transparency. Compared with Table 9, the results show that civil applicants’ negative perceptions decreased their perceived level of transparency more than vice versa. H3 predicted that between principal and agent, a person’s negative perception (citizen versus public employee) of the other will influence his or her reports of perceived transparency. According to the regression results, H3 was partially supported for public employees, but fully supported for citizens. Hypothesis H4 predicted that citizens with more negative perceptions of public employees will perceive lower levels of transparency in the public service; this was supported. The results of this analysis showed that the effects of negative perceptions on transparency differed between the samples of public employees and citizens. In addition, the influence of citizens’ negative perception of public employees on perceived levels of transparency was greater than that of the public employees’ negative perception of citizens.

Discussion

The findings reveal that the two groups do, indeed, respond differently to a survey-based measure of governmental transparency. The sample of public employees used three dimensions to understand transparency – Efficiency, Reliability, and Access – while citizens described transparency in the two dimensions of Accessibility (a wider notion than Access) and Utility. In some cases, the items were loaded onto similar factors for both samples, but in other cases, it is clear that they had different or almost opposing meanings; for example, items 8 and 9 loaded on Efficiency for public employees but on Accessibility for citizens, while items 11 and 12 loaded on Efficiency for public employees but on Utility for citizens. Furthermore, public employees had more favorable views of transparency than citizens. In contrast, in responding to the same items about the administration of civil applications, citizens reported lower ratings of transparency when compared to public employees. Concerning negative perception, public employees rated citizens more negatively than vice versa. Citizens’ negative views of public employees produced a perception of lower transparency. Of the three types of transparency from the sample of public employees, only the Reliability model significantly presented the effect of negative perception on transparency, while both models from the sample of citizens were significant. This indicates that citizens’ negative perceptions of public employees had a consistently negative and much greater influence on the level of transparency than public employees’ negative perceptions of citizens. These results will be valuable and will offer new insights into the study of measuring and improving transparency. If participants (as citizens or public employees) interpret the items differently, then it becomes much more difficult to compare their responses. The findings also suggest that public employees adopt a somewhat technical view of transparency, whereas citizens have more practical concerns about it.

The survey method is claimed to offer greater validity and reliability than qualitative methods. A key element is its assumed consistency: all participants are asked to respond to the same items with the same choices of answer. It is believed that a survey method produces the exact attributes of a population if a sufficiently large number of people answer the survey. However, this study shows that when a survey method is used to measure levels of transparency, responses could be significantly refracted through respondents’ identity as general citizens or public employees, and through the negative perceptions between the two groups that, notably, work in a principal–agent relationship. To use a survey method as a valid method for data gathering, researchers should understand such perceptual biases stemming from subjectivity.

Many critical questions have been raised regarding a survey conducted to determine the level of transparency of the White House, Congress, and other government agencies. According to them, the survey may be a measure of what people think of transparency in government – essentially a measure of opinion (Sternstein, 2011). Although it has inherent flaws, the survey method is important because both types of actual and perceived transparency are needed. Da Cruz et al. (2015: 20) stated that the policies for sustainable transparency practices can be developed based on citizen-centered or various stakeholders’ perspectives. Rawlins (2009: 73) maintained that ‘if the pragmatic value of transparency is to increase trust, then transparency needs to be measured from the perspective of the stakeholders’. In cases where agencies are concerned with what citizens think about transparency, a survey will be a useful tool. If not, an alternative approach for measuring transparency would be to employ objective measures.

This study also suggests how governments will be able to improve the level of transparency. Some scholars contend that society requires an optimal level of transparency (Cornand, 2008; Heald, 2003). The results of this study suggest that the optimal level in perceived transparency will differ depending on who evaluates it. Public employees may think that transparency has reached an optimal level already, while citizens may think it is still lower than the level they desire. This implies that the optimal level of transparency in public services can be reached by a different approach according to dimensions of transparency. Our results suggest that where the main objective is to increase levels of transparency assessed by citizens’ views, it would make sense to focus resources on activities that might improve accessibility and utility – the two dimensions that citizens used in perceiving transparency. The effects of negative perception that public employees have of citizens were partially significant when evaluating their own performance, whereas the negative perceptions that citizens have of public employees had a higher negative influence on transparency for the same government service. In this case, creating a positive feeling toward public services will help. One fundamental approach would be to develop public service motivation, in which the motivation of public employees – identified by their commitment to the public interest, compassion, and willingness for self-sacrifice – contributes to the sharing of their knowledge with others in the interests of serving the public (Chen and Hsieh, 2015). Another (more direct) way is to create a favorable feeling by promoting citizen involvement in the Internet and mobile space, which is widely known to be an effective means of accessing government information (Bertot et al., 2010; Grimmelikhuijsen and Welch, 2012; Pina et al., 2007).

Conclusions

Recently, many governments have sought to increase the volume of information they release to the public (Bertot et al., 2010; Pina et al., 2007). The next step will be to assess their progress by measuring transparency. However, few studies have addressed how to measure transparency and its limitations. This study sheds light on the limitations of survey-based measurements of transparency, showing that evaluations of transparency by citizens or public employees may differ if the citizen or public employee is affected by an identity and/or negative perception. Wang and Gianakis (1999) opined that performance measures are often invalid, in that public officials do not assess the activities or the results of governmental service simply. This applies when citizens perceive transparency levels. Although agreeing that an identity is likely to affect responses, Kaplan and Ruffle (1998) contended that there is a lack of evidence to support the bias assumption since alternative interpretations are possible. According to the authors, even though public employees perceived higher transparency, this might be due to the effect of their motives to increase the degree of their esteem or acts of bettering their performance in situations in which their work may be neglected (called the self-enhancement effect) but not of the effect of bias. Accordingly, contextual factors need to be examined in future studies. The data used for the analysis in this study were collected from citizens who had actual experience in perceiving transparency from visiting an office of civil applications. One limitation of the study is that the findings may not be generalizable to people who did not have such encounters due to the effects of various contextual factors on them; these factors govern the situations in which our survey was conducted.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Huengsik Park is Professor of Public Administration at Chung-Ang University in Seoul, South Korea. His research examines ethical issues in public administration, including whistle-blowing, ethical cultures, and transparency.

John Blenkinsopp is Professor of Organizational Behaviour & Human Resource Management at Hull University Business School, UK. He is interested in ethicality in public service organizations, with a particular focus on whistle-blowing and employee voice.

References

Bauhr M and Grimes M (2012) What is government transparency? New measures and relevance for quality of government. The Quality of Government Institute, Working Paper Series 2012:16. Available at: http://www.qog.pol.gu.se/digitalAssets/1418/1418047_f2012_16_bauhr_grimes.pdf.

Bertot

Jaeger

Grimes

(2010) Using ICTs to create a culture of transparency? E-government and social media as openness and anti-corruption tools for societies. Government Information Quarterly 27(3): 264–271.

Caamaño-Alegre

Lago-Peñas

Reyes-Santias

(2013) Budget transparency in local governments: An empirical analysis. Local Government Studies 39(2): 182–207.

Chen CA and Hsieh CW (2015) Knowledge sharing motivation in the public sector: The role of public service motivation. International Review of Administrative Sciences, advance online publication, DOI: 10.1177/0020852314558032.

Cho

Choi

(2004) E-government to combat corruption: The case of Seoul Metropolitan Government. International Journal of Public Administration 27(10): 719–735.

Coglianese

(2009) The transparency president? The Obama administration and open government. Governance 22(4): 529–544.

Cornand

(2008) Optimal degree of public information dissemination. The Economic Journal 118(528): 718–742.

Da Cruz

Tavares

Marques

(2015) Measuring local government transparency. Public Management Review 17: 3–28.

Erikson

(1964) Insight and Responsibility, New York: Norton.

10.

Esteller-Moré

Otero

(2012) Fiscal transparency: (Why) Does your local government respond? Public Management Review 14(8): 1153–1173.

11.

Felson

(1981) Ambiguity and bias in the self-concept. Social Psychology Quarterly 44(1): 64–69.

12.

Grimmelikhuijsen

Meijer

(2014) Effects of transparency on the perceived trustworthiness of a government organization: Evidence from an online experiment. Journal of Public Administration Research and Theory 24(1): 137–157.

13.

Grimmelikhuijsen

Welch

(2012) Developing and testing a theoretical framework for computer-mediated transparency of local governments. Public Administration Review 72(4): 562–571.

14.

Heald

(2003) Fiscal transparency: Concepts, measurement and UK practice. Public Administration 81(4): 723–759.

15.

Heinrich

Marschke

(2010) Incentives and their dynamics in public sector performance management systems. Journal of Policy Analysis and Management 29(1): 183–208.

16.

Hollyer

Rosendorff

Vreeland

(2011) Democracy and transparency. Journal of Politics 73(4): 1191–1205.

17.

Hollyer

Rosendorff

Vreeland

(2014) Measuring transparency. Political Analysis 22(4): 413–434.

18.

Islam

(2006) Does more transparency go along with better governance? Economics & Politics 18(2): 121–167.

19.

Kaplan

Ruffle

(1998) Self-serving bias. Journal of Economic Perspectives 12(2): 243–244.

20.

Lefkowitz

(2000) The role of interpersonal affective regard in supervisory performance ratings: A literature review and proposed causal model. Journal of Occupational and Organizational Psychology 73(1): 67–86.

21.

Lefkowitz

Battista

(1995) Potential sources of bias in supervisor ratings used for test validation. Journal of Business and Psychology 9(3): 389–414.

22.

Marvel JD (2015) Unconscious bias in citizens’ evaluations of public sector performance. Journal of Public Administration Research and Theory. DOI: 10.1093/jopart/muu053.

23.

Mezulis

Abramson

Hyde

(2004) Is there a universal positivity bias in attributions? A meta-analytic review of individual, developmental, and cultural differences in the self-serving attributional bias. Psychological Bulletin 130(5): 711–747.

24.

Otenyo

Lind

(2004) Faces and phases of transparency reform in local government. International Journal of Public Administration 27(5): 287–307.

25.

Park

Blenkinsopp

(2011) The role of transparency and trust in the relationship between corruption and citizen satisfaction. International Review of Administrative Sciences 77(2): 254–274.

26.

Pina

Torres

Royo

(2007) Are ICTs improving transparency and accountability in the EU regional and local government? An empirical study. Public Administration 85(2): 449–472.

27.

Rahim

(1983) Measurement of organizational conflict. Journal of General Psychology 109(2): 189–199.

28.

Rawlins

(2009) Give the emperor a mirror: Toward developing a stakeholder measurement of organizational transparency. Journal of Public Relations Research 21(1): 71–99.

29.

Relly

Sabharwal

(2009) Perceptions of transparency of government policymaking: A cross-national study. Government Information Quarterly 26(1): 148–157.

30.

Solansky

Singh

Huang

(2014) Individual perceptions of task conflict and relationship conflict. Negotiation and Conflict Management Research 7(2): 83–98.

31.

Sternstein

(2011) Tracing transparency. Government Executive 43(1): 24–29.

32.

Szulanski

Jehn

Greer

(2008) The effects of conflict types, dimensions, and emergent states on group outcomes. Group Decision and Negotiation 17(6): 465–495.

33.

Vadera

Aguilera

Caza

(2009) Making sense of whistle-blowing’s antecedents: Learning from research on identity and ethics programs. Business Ethics Quarterly 19(4): 553–586.

34.

Varma

Pichler

(2007) Interpersonal affect: Does it really bias performance appraisals? Journal of Labor Research 28(2): 397–412.

35.

Varma

Pichler

Srinivas

(2005) The role of interpersonal affect in performance appraisal: Evidence from two samples – US and India. International Journal of Human Resource Management 16(11): 2029–2044.

36.

Vishwanath

Kaufmann

(2001) Toward transparency: New approaches and their application to financial markets. World Bank Research Observer 16(1): 41–57.

37.

Wang

Gianakis

(1999) Public officials’ attitudes toward subjective performance measures. Public Productivity & Management Review 22(4): 537–553.

38.

Zábojník

(2004) A model of rational bias in self-assessments. Economic Theory 23(2): 259–282.