Public trust in the police in Taiwan: A test of instrumental and expressive models

Abstract

Although public assessments of the police have become the focal concern of a substantial amount of research efforts since the 1970s, a very small number of studies have analysed public opinions on the Taiwan police. Using survey data collected from three cities and two counties in 2010, this study expands the existing literature by assessing whether Taiwan residents’ perceptions are distinguishable in terms of procedural-based trust and outcome-based trust and whether both the instrumental and expressive models are predictive of Taiwanese trust in the police. Findings revealed that the Taiwanese tended to conflate procedural-based and outcome-based trust. Both the instrumental model (concerns about safety) and the expressive model (trust in neighbours and perceived quality of life) were significantly linked to Taiwanese trust in the police. Satisfaction with government performance and media influence were also predictive of police trustworthiness. Directions for future research and policy are discussed.

Keywords

expressive model instrumental model procedural-based trust Taiwan police trust in the police

Introduction

Public assessments of the police have become the focal concern of a substantial amount of research efforts since the 1970s. This line of inquiry is pivotal in establishing theoretical explanations of policing in general and public opinions on the police in particular. Citizen evaluations of the police also represent a critical issue for police administrators during the information technology era as locality-based crime prevention and fighting strategies and tactics become essential in performing a new model of policing that is mainly driven by data and technology (Rosenbaum, 2007). In this sense, understanding citizens’ attitudes towards the police represents key information for police managers to improve public participation in crime prevention programs and police–community relations.

The primary purpose of this study is to assess public trust in the police in Taiwan. Despite the importance of the topic, only a handful of studies have assessed factors that influence public attitudes towards the Taiwan police (Cao and Dai, 2006; Lai et al., 2010; Wu et al., 2012). Although a larger number of research reports have been published in Chinese, a general limitation of this line of studies is that they have tended to be descriptive in nature, elucidating the general patterns and trends of public opinions without much investigation into their causal mechanisms (see e.g. Huang, 2006; Yang et al., 2010). Indeed, past research has rarely tested the explanatory power of variables along theoretical frameworks, greatly reducing its potential contribution to theory construction and elaboration.

This study specifically addresses two concerns about the existing literature in public assessments of the Taiwanese police. First, recent research has acknowledged the complexity and multidimensional character of public attitudes towards the police in both Western and Chinese societies (e.g. Jackson et al., 2011; Stoutland, 2001; Sun et al., 2012). Trust in the police, for instance, can be delineated along two dimensions, procedural-based trust and outcome-based trust (Tyler, 2006; Tyler and Huo, 2002). Nevertheless, studies on public assessments of the Taiwan police have remained narrowly focused, with a single-item indicator reflecting trust (Wu et al., 2012) or confidence (Cao and Dai, 2006; Lai et al., 2010) in the police. This is problematic as single-indicator measures are prone to generating simplistic and even misleading assessments of social phenomena (Hudson and Kuhner, 2010). This study employs multiple items to measure trust in the police and examines whether these items can be distinguished along two conceptual aspects, procedural-based trust and outcome-based trust.

Second, recent studies, especially these conducted in the United Kingdom, have made conceptual and analytical distinctions between the instrumental model and expressive model of public trust in the police, with the former emphasizing the police’s responsibility for curbing crime and maintaining safety while the latter stresses their roles in upholding community values and promoting informal social control (Jackson and Bradford, 2010; Jackson and Sunshine, 2007). Such theoretical elaborations have yet to be explicitly tested using data from non-Western societies. By following the conceptualization and operationalization of these explanatory models of trust in the police in the Western literature, this study can expound to what extent theoretical frameworks and key concepts developed in highly industrialized, stable countries can be applied to transitional societies, such as Taiwan.

Using survey data collected from three cities and two counties in Taiwan, this study intends to assess public trust in the police along two research questions: (1) does the concept ‘trust in the police’ entail distinguishable multi-dimensions for Taiwan residents; and (2) are the instrumental and expressive models predictive of Taiwanese trust in the police? This study contributes to the existing literature of public trust in the police in a non-Western democracy by examining whether the multi-dimensional nature of trust in the police and the comparative influence of expressive and instrumental concerns recently established in Western societies can be applied to Taiwan.

Police–citizen relations in democratic Taiwan

Taiwan’s successful economic development and peaceful political transformation over the past three decades have been well-documented (e.g. Roy, 2003; Tien, 1988). The government promulgated a series of laws and statutes to regulate the post-martial law era (i.e. after 1987) of political and social order in the island. Along with the transformation towards democratic governance, the Taiwan police have inevitably and gradually shifted its primary role from the maintenance of the status quo and political control to the protection of individuals’ rights and crime control (Sun and Chu, 2006). Legal mandates have clearly become the police’s main sources of authorization. Changes in legislation and political environment have led to noticeable developments in police functions and organizations, including, among others, a refocus of police responsibilities on serving the public interest, a move towards professional leadership, and the establishment of more specialized units.

To establish a democratic police force, however, is not an easy task especially during rapid social and political changes. The Taiwan police faced great challenges in performing their duties particularly because of a continued increase in crime rates, the enhanced local control of police, and their vulnerability to media influence, all of which bear some implications for police–community relations. For instance, Taiwan’s total crime rate rose more than 40 per cent between 1980 and 1987. The rate doubled over the next decade, climbing from 41.9 per 10,000 residents in 1988 to 88.6 in 1998.¹ While crime rates have remained stable, and even shown a slight downward trend over the past decade (National Police Agency, 2012), this improvement has not been accompanied by a corresponding increase in public trust in the police.

Another important development is the decentralized control of local police. Taiwan has a centralized police force with the National Police Agency (NPA) having administrative jurisdiction over all police departments throughout the country. The central control of police, however, has been weakened after the passage of the Local Autonomy Statute in 1999, which empowered the localities with the authority to appoint local police chiefs. Since then, elected city and county executives have chosen police chiefs from a list of candidates recommended by the NPA and have been responsible for evaluating their performance. Residents can freely express their concerns and opinions about their local police chiefs and departments directly or indirectly through political representatives.

Although the localization of the police has enhanced police accountability to local politics and communities, it has made the force much more vulnerable to political influences. For example, police operations against illegal businesses, such as gambling and prostitution, are frequently interfered with by local city and county councilmen who often have close ties with owners of illegal businesses (Chu and Sun, 2007), and are even underworld figures themselves (Chin, 2003). These councillors could cut the police budget if the police failed to follow their requests. Such interference hinders the Taiwan police’s crime-fighting ability and subsequently damages their images in the public eye.

The media has also surfaced as a major source of public scrutiny of the police. Unlike during the martial-law era when the media was owned and tightly controlled by the state, privately owned news agencies in Taiwan now face furious competition for market share. They thus tend to actively search for sensational news on the government and public figures to attract audiences. Negative, sometimes exaggerated, reports of police misconduct on television, in newspapers and on the Internet apparently serve such a purpose fairly well (Wu et al., 2012). A recent study found that being exposed to and believing negative news about the police lowered citizens’ satisfaction with the police in Taiwan (Huang, 2012).

Hofstede’s (2001) cultural dimension of ‘power distance’, referring to the extent to which less powerful people expect that power is distributed unevenly in society, may be used to explain how social relations between the police and public have evolved in Taiwan over the past several decades. In ‘high power distance societies’ the public expects legal authorities to exercise decisions with little input from citizens, whereas in ‘low power distance societies’ the public expects to have input into governmental decision making and the way that they are governed. Under the authoritarian rule of the nationalist party before 1987, Taiwan was a higher power distance country in which the police–community relations were largely autocratic and paternalistic in the public eye, with the police being superiors and citizens subordinates.

The democratic transformation has gradually switched such power relations to more equal and consultative interactions, with citizens participating in decision making in public policy and exercising greater control of the police (Sun and Chu, 2006). Accordingly, as Tyler (1990) asserts, procedural justice should represent a more valued concept and matter more than outcome justice in shaping the public’s expectations of legal authorities in a low power distance society.

Scholars posited that, compared to the 1980s and 1990s, police–community relations have become more stable during this past decade due chiefly to enhanced police accountability and professionalism and fewer police–citizen confrontations resulting from large-scale demonstrations (Wu et al., 2012). Public opinion surveys consistently indicated that Taiwan residents displayed favourable attitudes towards the police, with about two-thirds of respondents expressing general satisfaction with the force (Huang, 2006). Nevertheless, studies touching on specific areas of police behaviour and performance, such as trustworthiness, effectiveness and integrity, showed much lower ratings of public satisfaction. For example, based on a World Values Survey, a study found that only 38 per cent of Taiwan respondents expressed their confidence in the police in 2005 (Lai et al., 2010), a decline from 58 per cent in 1995 (Cao and Dai, 2006). Data from the Asian Barometer Survey also revealed that less than half of the Taiwan respondents trusted the police quite a lot or a great deal (Wu et al., 2012). In another study, only 40 per cent of the respondents reported that they were highly satisfied or satisfied with police performance in maintaining law and order (Yang et al., 2010).

Measuring trust in the police

A large number of concepts have been utilized to depict public opinions about the police. This study focuses on the measurement of trust in the police. Trust can be understood as a state of mind that goes beyond tangible experience. It is chiefly ‘the belief that a person occupying a specific role will perform that role in a manner consistent with the socially defined normative expectations associated with that role ... Trust deals with specific individuals occupying specific social roles’ (Hawdon, 2008: 186). Applying this to the profession of policing, public perceptions of police trustworthiness imply that the police would display right intentions towards citizens and are capable of carrying out various tasks expected by the public, and such belief subsequently shapes our expectations of future encounters with the police (Jackson et al., 2011).

Recent research has operationalized and tested distinctive types of trust in the police. For example, studies have distinguished between procedural-based trust and outcome-based trust (Hawdon, 2008; Hinds and Murphy, 2007; Tyler, 2006; Tyler and Huo, 2002). The former is posited to be cultivated primarily by people’s judgments about the fairness of the police decision-making process, while the latter is influenced chiefly by people’s judgements about fair distribution of police services across people and communities (Sunshine and Tyler, 2003).

Using data collected from Boston’s residents, Stoutland (2001) extended the construction of trust in the police into four dimensions. She found that the police were less trusted on the dimensions of shared priorities (‘Do the police share local residents’ priorities and concerns?’) and respectfulness (‘Are the police respectful, courteous, and fair in their interactions with residents?’), whereas they were generally trusted on the dimensions of dependability (‘Are the police dependable?’) and competency (‘Do the police have the knowledge and skills to effectively fulfil their tasks?’). In line with Tyler’s (1990) procedural justice model, the first two aspects (i.e. priorities and respectfulness) indicate procedural-based trust, while the latter two (dependability and competency) imply outcome-based trust (Hawdon et al., 2003). Similar dimensions of trust have been enunciated in the British literature. Jackson and Bradford (2010: 245), for instance, operationalized trust in the police through the three pivotal areas of effectiveness (e.g. ‘tackle drug dealing’), fairness (‘treat people with respect’) and community engagement (‘listen to the concerns of local residents’) (see also Hohl et al., 2010; Jackson and Bradford, 2009).

This study examines whether the recently established multi-dimensional nature of trust in the police in the West (mainly USA and England and Wales) can also be applied to Taiwan. Specifically, it assesses whether Taiwan residents’ trust in the police can be separated into two notions: procedural-based trust and outcome-based trust.

Instrumental and expressive models of trust in the police

The existing research has conceptually disentangled and empirically assessed several theoretical frameworks of public assessments of the police. One line of inquiry contrasts the instrumental model with the expressive model (Jackson and Bradford 2009; Jackson and Sunshine, 2007; Jackson et al., 2009). The instrumental model contends that the police are instrumental in fulfilling their central role of preventing people being victimized and making people feel safe. If the police are perceived as not achieving such expectations, then the public may take this as an indication that crime cannot or will not be dealt with effectively, accordingly lowering their support for and trust in the police. This line of argument has some overlap with the accountability or performance model of public satisfaction with the police, which argues that the police should be held accountable for their responsibilities of combating crime and maintaining order and that their performance determines levels of public trust (Skogan, 2009; Van Craen, 2012; Zhao et al., 2012).

The expressive model posits that public assessments of police trustworthiness are influenced more by their expressive concerns about neighbourhood order and cohesion than by instrumental concerns about personal safety and crime (Jackson and Sunshine, 2007). The public expects the police to be not only crime fighters but also defenders of neighbourhood morals and values. Low community integration and high disorder and incivilities are likely to erode public trust in the police because residents feel that the police fail to reassure the moral structure of a community (Jackson and Sunshine, 2007; also see Tyler and Boeckmann, 1997). One can connect the expressive perspective to studies drawing upon the social disorganization tradition, which found residents of disadvantaged neighbourhoods having high levels of legal cynicism and dissatisfaction with the police (Reisig and Parks, 2000; Sampson and Jeglum-Bartusch, 1998; Wu et al., 2009). The expressive model also echoes the central arguments of social capital theory, which proposes that social networks that build around and extend beyond one’s immediate surroundings are positively related to confidence in governmental institutions (Brehm and Rahn 1997; Putnam, 2000).

Studies assessing variables derived from both models simultaneously have produced inconsistent results. In the United Kingdom, for instance, Jackson and his colleagues (Jackson and Bradford, 2009; Jackson and Sunshine, 2007) found that public confidence in the police was influenced more by concerns about neighbourhood stability and social cohesion than by worries about the risk of crime victimization. Using time series regression to analyse 84 months of the British Crime Survey data, the same dataset used by Jackson and his co-researchers, a recent study found contradictory results, with victimization rates significantly related to monthly levels of public confidence in the police, whereas public perceptions of crime, disorder and social cohesion were not predictive of confidence in the police (Sindall et al., 2012).

The existing literature on Taiwanese evaluations of the police has generated very little evidence concerning the instrumental and expressive models. Variables representing either model were not included in two of the studies (Cao and Dai, 2006; Lai et al., 2010). A third study found that sense of safety and victimization were uncorrelated with Taiwanese trust in the police, whereas stronger interpersonal trust led to greater trust in the police (Wu et al., 2012). Mixed results were found in recent studies on public trust in the police in mainland China. One study showed that higher levels of a sense of safety and social capital (i.e. trust in neighbours and neighbourhood resident committees) were accompanied by greater perceptions of police trustworthiness (Sun et al., 2012). Another study found that fear of crime did not but perceived quality of life did influence levels of trust in the police (Wu and Sun, 2009).

While findings from this thin line of inquiry are clearly inconclusive, we would speculate that instrumental concerns are more important than expressive concerns in influencing Chinese evaluations of the police for two reasons. First, compared to the US and some Western countries, Taiwan is a culturally and racially homogenous society. Chinese culture stresses the importance of moral and ethical governance and collective responsibility in societies. Successful economic development has greatly improved the standard of living and quality of life, with highly disadvantaged neighbourhoods, such as urban ghettos in the US, being uncommon in Taiwan. High cohesion and low disorder make expressive concerns about neighbourhood social and physical conditions less relevant to the evaluations of the police.

Second, while Taiwan’s crime rates have stabilized over the past decade and remain relatively low compared to those in Western countries, recent national victimization surveys have still revealed a steady high fear of crime, and that under-reporting of crime by victims and under-recording by police officers of certain crimes, such as fraud, drug crimes, domestic violence and Internet related offences, were pervasive (Sheu et al., 2011). Among victims who did not report their victimization to the police, 20 per cent of them stated that they chose not to do so because the police were incapable of resolving the incidents (Sheu et al., 2011). The high relevance of concerns about crime and safety may directly affect public trust in the police. The current study will compare and contrast variables representing both the instrumental and expressive models of trust in the police.

Methodology

Data collection and sample

Data used in this study were from a research project conducted in December, 2010. The project, entitled Police Images and Public Satisfaction with the Police, was carried out by a team of researchers from several universities. The team designed a 50-item survey instrument that captured an array of information related to public evaluations of the police as well as respondents’ demographic characteristics, experience with the police, attitudes towards the government and legal system, the media, and neighbourhood structural and organizational characteristics.

A professional poll company was contracted by the research team to conduct a telephone survey of residents in three metropolitans, Taipei, Taichung and Kaohsiung, and two counties, Yunlin and Taidong, in Taiwan. These cities were selected because they are the three largest metropolises located in northern, central and southern Taiwan. The two agricultural counties were included mainly for the purpose of comparison. Table 1 displays the demographic, socio-economic and crime statistics associated with each of these sample sites.

Table 1.

Demographic, socio-economic, and crime statistics of sample sites.

Taipei City	Taichung City	Kaohsiung City	Yunlin County	Taidong County
Total population^a	2,619,000	1,082,000	1,530,000	718,000	231,000
Geographic area (square kilometres)^a	272	163	154	1,291	3,515
Population density (per square kilometres)^a	9,634	6,623	9,961	556	66
Percent minority population^a	0.51	0.68	0.76	1.09	34.63
Average household disposable income (US$)^a	24,085	17,362	18,025	12,636	13,689
Unemployment rate^a	5.8	5.9	5.9	5.8	5.8
Population mobility^b	−7.51	3.76	0.18	−0.55	4.39
Number of police stations^a	14	6	10	6	4
Number of police field stations^a	94	32	50	65	61
Number of police officers^a	7798	3342	4391	1,629	1,076
Citizen/police ratio^a	335.86	323.76	348.44	440.77	214.69
Total crime rate/100,000 habitants^a	1096.70	1406.97	2394.61	1089.94	644.26
Violent crime rate/1,000 habitants^ac	0.20	0.21	0.46	0.14	0.18
Property crime rate/1,000 habitants^ad	5.45	8.16	11.14	4.39	2.83

Notes: ^a2010 data; ^b2009 data; ^cviolent crimes including homicide, rape, robbery, forceful taking, kidnapping, intimidation and extortion, and aggravated assault; ^dproperty crimes including larceny, theft, motor vehicle theft and motorcycle theft.

A sample size of 2,250 in total was targeted, including 600 from Taipei, 550 from Kaohsiung, 400 from Taichung and Yunlin, and 300 from Taidong, which are proportional to their population sizes. A slightly higher number of respondents (2,289) were randomly selected and surveyed using the computer assisted telephone interviewing (CATI) system.² The actual success rate (i.e. the percentage of the number of completed surveys relative to the number of households phoned) ranged from 4.1 per cent in Taichung to 8.7 per cent in Yunlin. The average time taken to complete a successful survey was about 12 minutes. All respondents were at least 20 years old. They were informed about the principles of voluntariness and anonymity for their participation. Based on the respondents’ age, gender and residence area, survey data were weighted after each successful survey using raking ratio estimation to ensure that the difference between the sample and population is not significant. Cases with missing values were excluded from the analyses. The final sample for this study consisted of 1,716 respondents, comprising 472 from Taipei, 440 from Kaohsiung, 293 from Taichung, 282 from Yunlin, and 229 from Taidong. An analysis of the data indicated that there were virtually no differences between respondents included and excluded in the sample, thus the exclusion of missing data did not confound the results in any significant way.

Dependent variable

The dependent variable is trust in the police. Seven items tap into two types of trust in the police: procedural-based trust and outcome-based trust. These items are provided in Table 2. Response options for all these evaluative items are: disagree strongly (1), disagree (2), agree (3) and agree strongly (4). Factor scores derived from factor analysis were used as the dependent variable in multivariate regression (see discussion in the results section for the construction of this variable).

Table 2.

Construction of trust in police variables.

Variables	Survey items	Factor loading
Procedural-based trust	1. Do you think most police are fair in handling incidents?	.72
	2. Do you think most police officers are polite and friendly?	.77
	3. Do you think police have patience in assisting citizens?	.80
	4. Do you think officers give citizens opportunities to explain their situations?	.68
Outcome-based trust	1. Do you think police anti-crime performance is good?	.66
	2. Do you think police are efficient in handling public emergency and service calls?	.72
	3. Do you think police do a good job in traffic enforcement?	.60

Independent and control variables

The independent variables were constructed along the instrumental and expressive models of public assessments of the police. Two variables, concern about safety and victimization experience, belong to the former model. Respondents were asked whether they were worried about their own or family members’ personal safety in their neighbourhood. Responses ranged from not at all (1) to very much (4). Respondents were also asked whether they or any family members had been a victim of crime in the past twelve months (0 = no; 1 = yes). Three variables, trust in neighbours, quality of life and physical disorder, were indicators of the expressive model. The first two were constructed based on a single item asking respondents about their trust in neighbours (1 = not at all; 4 = very much) and satisfaction with quality of life in neighbourhoods (1 = highly dissatisfied; 4 = highly satisfied). Three items asked respondents whether there are empty houses, vacant lots and abandoned vehicles in their neighbourhoods, indicating neighbourhood physical disorder. A positive response to any of the three items was coded as 1 representing the presence of physical disorder in the neighbourhood.

Control variables are divided into four groups: background characteristics, satisfaction with government performance, media influence and locality.³ Background characteristics include gender, age, educational attainment, and family income. Gender was coded as a dummy variable with 1 representing male. Age is a four-category variable (1 = 20–39; 2 = 40–49; 3 = 50–59; 4 = 60 and above), educational attainment is a three-category variable (1 = middle school and below; 2 = high school; 3 = college and above), and monthly family income (in US dollars) is a five-category variable (1 = no income; 2 = below $650; 3 = $650–1,300; 4 = $1,301–2,000; 5 = above $2,000).

The second group includes three variables measuring respondents’ satisfaction with the performance of the central government, local government and judiciary. Responses to the first two variables ranged from highly dissatisfied (1) to highly satisfied (4). The last variable is an additive scale of two items indicating satisfaction with prosecution and sentencing.

Two variables related to the media were employed. Knowledge about police misconduct came from an item asking how often the respondents have heard or read news (from radio, TV, newspapers and the Internet) about police misconduct (such as brutality, abusive languages and corruption) (1 = never; 2 = rare, 3 = sometimes; 4 = often). A second measure reflects to what extent the respondents actually believe the negative news of the police reported by the media (1 = don’t believe; 2 = slightly don’t believe; 3 = believe; 4 = strongly believe). Finally, a series of dummy variables were created to represent the five localities in this study. Taipei City is the reference group in the regression analysis.

Table 3 reports the descriptive statistics of all variables used in the study. Possible multicollinearity problems were checked by examining the matrix of two-variable correlations among all independent variables. The highest correlation between two independent variables (income and educational attainment) is .42, which is acceptable. The small magnitude of the correlations is confirmed by assessing variance inflation factors (VIFs), all of which are well below the generally accepted limit of 10.

Table 3.

Descriptive statistics for variables.

Mean	SD	Min	Max
Dependent variable
Trust in police	0.00	1.00	−2.65	2.56
Independent and control variables
Instrumental model
Safety concerns	2.42	.85	1.00	4.00
Victimization	.12	.33	0.00	1.00
Expressive model
Trust in neighbours	3.18	.69	1.00	4.00
Quality of life	2.95	.83	1.00	4.00
Perceived physical disorder	.40	.49	0.00	1.00
Background characteristics
Male	.45	.50	0.00	1.00
Age	2.31	1.10	1.00	4.00
Education	2.31	.77	1.00	3.00
Income	2.84	1.35	1.00	5.00
Satisfaction with government performance
Central government	2.23	.81	1.00	4.00
Local government	2.54	.85	1.00	4.00
Prosecution and sentencing	4.03	1.39	2.00	8.00
Media influence
Know negative reports on police	3.02	.82	1.00	4.00
Believe media reports on police	3.17	.60	1.00	4.00
Locality
Taipei City	.28	.45	0.00	1.00
Taichung City	.17	.38	0.00	1.00
Kaohsiung City	.26	.44	0.00	1.00
Yunlin County	.16	.37	0.00	1.00
Taitung County	.13	.34	0.00	1.00

Results

We conducted factor analysis and a reliability test to answer the question about whether procedural-based trust and outcome-based trust are distinguishable among Taiwanese respondents. The principal component analysis of the seven items of trust in police indicated that only a single factor could be extracted, with an eigenvalue of 3.52, explaining 50 per cent of the variance and factor loadings ranging from .60 to .80 (see Table 2). The reliability test showed that an additive scale consisting of all seven items would have a high internal consistency with a Cronbach’s alpha of .83. Deleting any of the items would result in a lower Cronbach’s alpha, indicating a decreased internal consistency of the scale. Thus, the concept of trust in police is rather one-dimensional among the surveyed Taiwan residents.

A set of regression models was estimated to find whether the instrumental and expressive models are predictive of Taiwanese trust in the police. Table 4 displays the regressions results for all three models. In Model 1, the two variables representing the instrument model are significant predictors of Taiwanese trust in the police. Consistent with our expectations, greater concerns about safety and victimization experience are associated with lower levels of trust in the police. The two variables together explain 10 per cent of the variation in trust in the police.

Table 4.

Multiple regression summary for trust in police (n = 1,716).

Model 1		Model 2		Model 3
Variables	B	SE	B	SE	B	SE
Instrumental model
Safety concerns	−.29***	.03	−.20***	.03	−.14***	.03
Victimization	−.10***	.07	−.07**	.07	−.04	.06
Expressive model
Trust in neighbours	.13***	.03	.12***	.03
Quality of life	.20***	.03	.14***	.02
Physical disorder	−.03	.05	−.02	.04
Background characteristics
Male	.00	.04
Age	.05*	.02
Education	−.03	.03
Income	−.01	.02
Satisfaction with government performance
Central government	.13***	.03
Local government	.06**	.02
Prosecution and sentencing	.23***	.02
Media influence
Know negative reports on police	−.13***	.03
Believe media reports on police	−.08***	.03
Locality
Taichung City	−.03	.06
Kaohsiung City	−.03	.05
Yunlin County	−.02	.07
Taitung County	−.03	.07
R ²	.10		.17		.33
Incremental R²	−		.07*		.16*

p < .05; **p < .01; ***p < .001.

Concerns about both safety and victimization remain significant predictors in Model 2 when the expressive model variables enter the regression. Two out of the three expressive model variables exert a significant effect on trust in the police. As expected, higher levels of trust in neighbours and perceived quality of life lead to greater trust in the police. The third variable, perceived physical disorder in the neighbourhood is not predictive of trust in the police. Adding the expressive model variables into the analysis significantly improves the predictive power of the model by 70 per cent, raising the R² from .10 to .17.

Concerns about safety, trust in neighbours, and perceived quality of life continue to be significant predictors in Model 3, while the significant effect of victimization disappears when control variables are introduced into the equation, suggesting that the impact of victimization might be spurious. Among background characteristics, age is the only significant predictor, with older people having greater trust in the police than younger people. All three variables in the next group, satisfaction with government performance, exert a significant positive effect on trust in the police, with greater satisfaction with the central government, local government and judiciary linking to higher levels of trust. Media influence also matters. Both knowing negative reports about the police and believing in such reports significantly reduce citizen trust in the police. None of the locality variables is significantly connected to trust in the police. Adding control variables into the model increases the R² from .17 to .33, a nearly 100 per cent rise in the explanatory power.

Discussion

Although a substantial number of studies have examined factors influencing public trust in the police, there is no evidence showing that the multidimensional nature of trust in the police uncovered in the Western literature can be applied to a non-Western setting. The existing literature also lacks consistent findings regarding the influences of variables formulated along two popular frameworks, the instrumental and the expressive models, inviting more empirical research. We assessed these two issues by analysing survey data collected from Taiwan. Our data reveal that Taiwan residents tended to conflate procedural-based trust and outcome-based trust in their evaluations of the police. This result echoes Michelson and Read’s (2011) recent finding that Chinese citizens do not seem to distinguish between procedural-based justice and distributive-based justice as clearly as Westerners do.

Western scholars have acknowledged that in some countries procedural factors may play a less important role than outcome factors in shaping public perceptions of legal authorities (Tyler, 1990). For instance, Tankebe (2009) found that public cooperation with the Ghana police is influenced by utilitarian concerns, such as perceptions of police effectiveness in fighting crime, rather than by procedural fairness. Indeed, concerns about procedural justice are likely to be more important in the US than in non-Western societies, given America’s strong focus on the protection of individual rights and due process (Tonry, 2007).

While the democratization of policing has greatly enhanced the rule of law and the protection of human rights in Taiwan, our findings indicate that such positive development does not necessarily translate into a stronger awareness of the difference between procedural-based trust and outcome-based trust among local residents. Why do conceptually sound and empirically supported multi-aspects of trust in the police in the West not apply to the Taiwanese? One might speculate that the legal culture in Taiwan that stresses substantive justice over procedural fairness (Lo, 2007) may contribute to a lower level of sensitivity to the issue of procedural justice and a blurred distinction between the two types of trust. In addition, our data indicate that less than 40 per cent of the respondents had had direct contacts with the police over the preceding year. It is possible that limited direct encounters with the police make the detection of the differences between trust developed based on procedural fairness and intervention results less likely. These potential explanations should be further investigated in future research.

Our results also suggest that variables derived from both the instrumental and expressive models play a roughly equal role in predicting public trust in the police.

The standardized regression coefficients (B = .14) associated with concerns about safety (a variable representing the instrumental model) and perceived quality of life (a variable representing the expressive model), indicate that they have an identical strength in predicting trust in the police. A second variable from the expressive variable, trust in neighbours, has a slightly lower strength of influence compared to perceived quality of life B = .12). Not surprisingly, those who had greater concerns about safety tended to report lower levels of trust in the police, while those who expressed higher levels of trust in neighbours and perceived quality of life were more likely to view the police as trustworthy. It appears that Taiwan residents have both personal safety and neighbourhood social cohesion and order in mind when making judgements about the police.

This study confirms the importance of concerns about safety in explaining trust in the police found in both Chinese (Sun et al., 2012) and Western societies (e.g. Sprott and Doob, 2009; Van Craen, 2012). It also underlines the pivotal role of neighbourhood social capital and quality of life in evaluating police trustworthiness as suggested in past research (e.g. Sun et al., 2012; Van Craen, 2012; Wu et al., 2012). Conversely, our findings are incongruent with results from recent UK studies showing the primacy of the expressive model in accounting for confidence in the police (Jackson and Bradford 2009; Jackson and Sunshine, 2007; Jackson et al., 2009), and do not support the case for a weakened role of the instrumental model (see Cao et al., 1996). A weak link between two other commonly investigated variables, victimization and perceived disorder, and trust in the police also contradicts findings from previous studies (Cao et al., 1996; Jackson et al., 2009; Skogan, 1996). Nevertheless, our findings and the comparisons should be taken with caution because the operationalization of instrumental and expressive variables is not entirely identical between our study and past research. The complex and inconclusive nature of these findings calls for more research attention to clarify their possible connections.

Two additional groups of variables stand out as important predictors of public trust in the police. Three variables measuring the performance of the central government, local government and judiciary respectively were positively related to public trust in the police, with satisfaction with the judiciary being the strongest predictor in the regression model. The high relevance of this group of variables does not come as a surprise. As the police are an integral part of the political system, it is reasonable to expect that public evaluations of the trustworthiness of police are also influenced by their satisfaction with other governmental agencies. Our findings echo the institutional performance perspective of public trust in the police, which posits that citizens’ perceptions of the police are shaped by their evaluations of police performance in particular and government performance in general (e.g. Larsen and Blair, 2009; Tyler, 1990). Indeed, citizens’ evaluations of the performance of government and non-police political institutions are not stand-alone sentiments but highly intertwined with their assessments of the police (Wu et al., 2012). One may thus also speculate that satisfaction with or trust in local police could be indicative of one’s trust or confidence in government. Given that only a small number of studies have analysed this association (see Albrecht and Green, 1977; Ivković, 2008; Wu et al., 2012), this line of variables should continue to be included in future assessments of public trust in the police.

A second group of predictors affecting trust in the police is media influence. Both being exposed to negative reports about the police and actually believing in such reports significantly reduced perceived trustworthiness of the police. Despite being commonly viewed as one of the most influential factors, the media’s role in shaping public opinions on the police has rarely been empirically investigated in Taiwan (see Huang, 2012 for a notable exception). This study provides some preliminary evidence supporting the media’s influence in Taiwan. Our findings are also in tune with results from an established line of research showing the impact of the media on public attitudes towards the police in Western societies (e.g. Calanan and Rosenberger, 2011; Chermak et al., 2006;Dowler, 2002; Kaminski and Jefferis, 1998; Surette, 2007; Wu et al., 2011). Media influence clearly needs to be considered in future research of public opinions on the police.

Future research should be conducted along two additional avenues. First, more research should analyse the conceptual relationships between measures of public attitudes towards the police. Recent studies have assessed the conceptual and analytical distinctions between trust and confidence (Hohl et al., 2010; Jackson and Bradford, 2010), trust and legitimacy (Hawdon, 2008), legitimacy and satisfaction (Hinds and Murphy, 2007; Tankebe, 2010), and global satisfaction and specific satisfaction (Wu et al., 2011). Albeit valuable, this vein of investigation is still thin and vital relationships between key concepts, such as satisfaction with and trust in the police, remain under-explored. More research utilizing elaborated, multiple-item scales should be initiated to examine the potential distinctions between concepts in the same level and across different levels.

Second, future research should improve on the operationalization of key theoretical constructs explaining public trust in the police. For example, the instrumental model can incorporate actual crime rates and perceived crime problems into the analysis (Jackson and Bradford, 2009; Sindall et al., 2012). Similarly, more neighbourhood-based measures of the expressive model, such as collective efficacy among local residents and moral alignment between the police and the public, can be valuable. Hierarchical linear modelling approaches with multi-level data would be required in these studies.

Our findings carry some clear messages to the Taiwan police and government. Given that Taiwan residents show no particular preference for outcome-based or procedure-based police performance, Taiwan police should act as both crime fighters who are capable of curbing crime and reducing fear of crime and moral guardians who are willing to uphold community social cohesion and value structures. A service-centred policing strategy that integrates crime prevention activities into local government services would allow the police to capture the core values of both role orientations. Police administrators have to ensure that street-level enforcement and activities do not focus overwhelmingly on the instrumental roles with little care given to the expressive roles, which is highly possible under a performance evaluation system based largely on crime-related statistics, such as arrests and citations.

If public evaluations of the police are intricately intertwined with satisfaction with the central and local governments and non-police criminal justice agencies, then a better image for the police could be achieved only when the public is satisfied with the performance of relevant agencies. The responsiveness of government to the public’s needs, the integrity and legitimacy of government officials, a fair and equitable legal system, and the overall economic conditions of the nation are some critical indicators of government performance highly relevant to trust in the police (Wu et al., 2012). Public trust in the police will not change overnight, but establishing an open society with effective, efficient and fair central and local governance is no doubt a move in the right direction.

Footnotes

Notes

References

Albrecht

Green

(1977) Attitudes toward the police and the larger attitude complex: Implications for police–community relationships. Criminology 15: 67–86.

Brehm

Rahn

(1997) Individual-level evidence for the causes and consequences of social capital. American Journal of Political Science 41: 999–1023.

Calanan

Rosenberger

(2011) Media and public perceptions of the police: Examining the impact of race and personal experience. Policing and Society 21: 167–189.

Cao

Dai

(2006) Confidence in the police: Where does Taiwan rank in the world?. Asian Journal of Criminology 1: 71–84.

Cao

Frank

Cullen

(1996) Race, community context and confidence in the police. American Journal of Police 15: 3–22.

Chermak

McGarrell

Gruenewald

(2006) Media coverage of police misconduct and attitudes toward police. Policing: An International Journal of Police Strategies & Management 29: 261–281.

Chin

(2003) Heijin: Organized Crime, Business, and Politics in Taiwan, Armonk, NY: M.E. Sharpe.

Chu

Sun

(2007) A comparison of Taiwanese and American police supervisors’ attitudes. Police Quarterly 10: 63–86.

Dowler

(2002) Media influence on citizen attitudes toward police effectiveness. Policing & Society 12: 227–238.

10.

Hawdon

(2008) Legitimacy, trust, social capital and policing styles: A theoretical statement. Police Quarterly 11: 182–201.

11.

Hawdon

Ryan

Griffen

(2003) Policing tactics and perceptions of police legitimacy. Police Quarterly 6: 469–491.

12.

Hinds

Murphy

(2007) Public satisfaction with police: Using procedural justice to improve police legitimacy. Australia and New Zealand Journal of Criminology 40: 27–42.

13.

Hofstede

(2001) Culture’s Consequences: Comparing Values, Behaviours, Institutions, and Organizations Across Nations, Thousand Oaks, CA: Sage.

14.

Hohl

Bradford

Stanko

(2010) Influencing trust and confidence in the London Metropolitan police: Results from an experiment testing the effect of leaflet-drops on public opinion. British Journal of Criminology 50: 491–513.

15.

Huang

(2006) 2005 summary of public satisfaction with police study. Academic Journal of Central Police University 43: 1–32.

16.

Huang Y (2012) The media, political party orientation, and attitudes toward the police. Unpublished masters thesis, National Taipei University, Taipei, Taiwan.

17.

Hudson

Kuhner

(2010) Beyond the dependent variable problem: The methodological challenges of capturing the productive and protective dimensions of social policy. Social Policy and Society 9: 167–179.

18.

Ivković

(2008) A comparative study of public support for the police. International Criminal Justice Review 18: 406–434.

19.

Jackson

Bradford

(2009) Crime, policing, and social order: On the expressive nature of public confidence in policing. British Journal of Sociology 60: 493–521.

20.

Jackson

Bradford

(2010) What is trust and confidence in the police. Policing 4: 241–248.

21.

Jackson

Sunshine

(2007) Public confidence in policing: A neo-Durkheimian perspective. British Journal of Criminology 47: 214–233.

22.

Jackson

Bradford

Hohl

Farrall

(2009) Does the fear of crime erode public confidence in policing?. Policing: A Journal of Policy and Practice 3: 100–111.

23.

Jackson

Bradford

Hough

Kuha

Stares

Widdop

Fitzgerald

Yordanova

Galev

(2011) Developing European indicators of trust in justice. European Journal of Criminology 8: 267–285.

24.

Kaminski

Jefferis

(1998) The effect of a violent televised arrest on public perceptions of the police. Policing: An International Journal of Police Strategies & Management 21: 683–706.

25.

Lai

Cao

Zhao

(2010) The impact of political entity on confidence in legal authorities: A comparison between China and Taiwan. Journal of Criminal Justice 38: 934–841.

26.

Larsen

Blair

(2009) The importance of police performance as a determinant of satisfaction with police. American Journal of Economics and Business Administration 1: 1–10.

27.

(2007) The Legal Culture and System of Taiwan, Alphen aan den Rijn, Netherlands: Kluwer Law International.

28.

Michelson

Read

(2011) Public attitudes toward official justice in Beijing and rural China. In: Woo

Gallagher

Goldman

(eds) Chinese Justice: Civil Dispute Resolution in Contemporary China, Cambridge, MA: Harvard University Press, pp. 169–203.

29.

National Police Agency (2012) Offenses Handled by Police Organizations. Available at: http://www.dgbas.gov.tw/ct.asp?xItem=15440&CtNode=4634&mp=1 (accessed 27 December 2012).

30.

Putnam

(2000) Bowling Alone: The Collapse and Revival of American Community, New York, NY: Simon and Schuster.

31.

Reisig

Parks

(2000) Experience, quality of life, and neighbourhood context: A hierarchical analysis of satisfaction with police. Justice Quarterly 17: 607–630.

32.

Rosenbaum

(2007) Police innovation post 1980: Assessing effectiveness and equity concerns in the information technology era. IPV Review 1: 11–44.

33.

Roy

(2003) Taiwan: A Political History, Ithaca, NY: Cornell University Press.

34.

Sampson

Jeglum-Bartusch

(1998) Legal cynicism and (sub cultural?) tolerance of deviance: The neighbourhood context of racial differences. Law and Society Review 32: 777–804.

35.

Sheu

Chen

Tsai

Mon

Huang

(2011) Crime Victim Survey in Taiwan, Year 2010, Taipei, Taiwan: National Police Agency.

36.

Sindall

Sturgis

Jennings

(2012) Public confidence in the police. British Journal of Criminology 52: 744–764.

37.

Skogan

(1996) The community’s role in community policing. National Institute of Justice Journal 231: 31–34.

38.

Skogan

(2009) Concern about crime and confidence in police: Reassurance or accountability?. Police Quarterly 12: 301–308.

39.

Sprott

Doob

(2009) The effect of urban neighbourhood disorder on evaluations of the police and courts. Crime and Delinquency 55: 339–362.

40.

Stoutland

(2001) The multiple dimensions of trust in resident/police relations in Boston. Journal of Research in Crime and Delinquency 38: 226–256.

41.

Sun

Chu

(2006) Attitudinal differences between Taiwanese and American police officers. Policing: An International Journal of Police Strategies and Management 29: 190–210.

42.

Sun

(2012) Social capital, political participation, and public trust in police in urban China. Australia and New Zealand Journal of Criminology 45: 87–105.

43.

Sunshine

Tyler

(2003) The role of procedural justice and legitimacy in shaping public support for policing. Law and Society Review 37: 513–547.

44.

Surette

(2007) Media, Crime, and Criminal Justice: Images, Realities and Policies, 3rd edn. New York, NY: Wadsworth.

45.

Tankebe

(2009) Public cooperation with the police in Ghana: Does procedural fairness matter?. Criminology 47: 1265–1293.

46.

Tankebe

(2010) Public confidence in the police: Testing the effects of public experiences of police corruption in Ghana. British Journal of Criminology 50: 296–319.

47.

Tien

(1988) Social change and political development in Taiwan. In: Feldman

Kau

Kim

(eds) Taiwan in a Time of Transition, New York, NY: Paragon House, pp. 1–37.

48.

Tonry

(2007) Determinants of penal policies. In: Tonry

(ed.) Crime and Justice: A Review of Research. Vol. 36, no. 1, Crime, Punishment, and Politics in a Comparative Perspective, Chicago, IL: University of Chicago Press, pp. 1–48.

49.

Tyler

(1990) Why People Obey and Law, New Haven, CT: Yale University Press.

50.

Tyler

(2006) Psychological perspectives of legitimacy and legitimation. Annual Review of Psychology 57: 375–400.

51.

Tyler

Boeckmann

(1997) Three strikes and you are out, but why? The psychology of public support for punishing rule breakers. Law and Society Review 31: 237–265.

52.

Tyler

Huo

(2002) Trust in the Law: Encouraging Public Cooperation with the Police and Courts, New York, NY: Russell Sage Foundation.

53.

Van Craen M (2012) Explaining majority and minority trust in the police. Justice Quarterly, Online First. Available at: https://www-tandfonline-com.web.bisu.edu.cn/doi/full/10.1080/07418825.2011.649295 (accessed 27 December 2012).

54.

Sun

(2009) Citizen trust in police: The case of China. Police Quarterly 12: 170–191.

55.

Poteyeva

Sun

(2012) Public trust in police: A comparison between China and Taiwan. International Journal of Comparative and Applied Criminal Justice 36: 189–210.

56.

Sun

Smith

(2011) Race, immigration and policing: Chinese immigrants’ satisfaction with police. Justice Quarterly 28: 745–774.

57.

Sun

Triplett

(2009) Race, class or neighbourhood context: Which matters more in measuring satisfaction with police?. Justice Quarterly 26: 125–156.

58.

Yang

Cheng

Lou

(2010) The Investigation of Victimization and Public Satisfaction with Government Efforts in Maintaining Law and Order for the First Half of 2010, Chiayi County, Taiwan: Crime Research Centre, National Chung Cheng University.

59.

Zhao J, Tsai C, Ren L and Lai Y (2012) Public satisfaction with police control disorder crime: Does the public hold police accountable? Justice Quarterly, Online First. Available at: https://www-tandfonline-com.web.bisu.edu.cn/doi/full/10.1080/07418825.2012.667140 (accessed 27 December 2012).