Factors Influencing Public Trust in the Police in South Korea: Focus on Instrumental, Expressive, and Normative Models

Abundant studies examining public trust in the police have applied several theoretical models including instrumental, expressive, or normative models. However, few studies have attempted to simultaneously assess the empirical validity of these theoretical models of public trust in the police. In addition, there has been little research on public trust in police in East Asia; most of the empirical research on this topic has been explored in Western societies. To extend the knowledge of public trust in the police, the current study investigated to what extent factors drawn from three models influence public trust in the police using a sample of South Korean citizens. The results show that, consistent with prior research, police effectiveness, procedural justice, and social cohesion had significant, positive effects on public trust in the police. Police effectiveness was the most influential factor followed by procedural justice and social cohesion. Implications for practice and future research are discussed.


Introduction
Although advances in science have driven social system evolution, human activities still remain predominantly focused on policing because police activities are so intertwined with the public. Since public trust in the police underpins the entire system of police, a large body of research has emphasized the importance and the value of public trust in the police, and many police scholars have paid attention to this topic.
Prior studies have shown that building public trust in the police tends to enhance public cooperation with police in preventing crimes, arresting offenders, reducing fear of crime, and advancing neighborhood safety (Nix et al., 2015;Tyler, 1990). More importantly, public trust in the police is a crucial factor for improving police legitimacy (Hawdon et al., 2003). Specifically, overall effectiveness of police may be lessened when high levels of trust between police and the public are not warranted. If police officers perform their duties in a way contrary to citizens' expectations, police agencies are less likely to experience trust, support, cooperation, and voluntary public compliance. Conversely, if police agencies maintain a high level of public trust in the police, citizens are more likely to comply with the police which in turn, reduces levels of crime and disorder and improves the quality of life in the neighborhood.
Recent studies on public trust in the police have identified conceptual and empirical distinctions between instrumental and expressive models of public trust in the police (Bradford & Myhill, 2015;Jackson & Bradford, 2009;Jackson & Sunshine, 2007;Jang & Hwang, 2014;Sun et al., 2013). From the instrumental perspective, citizens evaluate the police in terms of their ability to prevent crimes, apprehend criminals, and maintain safety effectively. On the other hand, the expressive model posits that public trust in the police is determined by their expressive concerns related to neighborhood disorder and community cohesion. That is, higher levels of incivilities and lower level of social cohesion in neighborhoods tend to deteriorate public trust in the police because citizens feel that the police fail to ensure the moral 1068504S GOXXX10.1177/21582440211068504SAGE OpenLim and Kwak research-article20222021 1 DaeJeon University, South Korea 2 Chungnam National University, DaeJeon, South Korea structure and value of community (Jackson & Bradford, 2009;Jackson et al., 2012;Jackson & Sunshine, 2007).
In contrast, Tyler (2001Tyler ( , 2005 argued that citizens' perceptions of fair policing play a greater role than police performance in explaining citizens' trust in the police. More specifically, Tyler's studies indicate that quality of treatment drives public perceptions of trust in police. This finding has been discussed under the framework of procedural justice in normative models where judgments about the fairness of the police is the most important factor in such processes (Bradford, 2014). Furthermore, the level of normative compliance to the law emanate from judgments concerning the legitimacy in the intergroup context created by the police (Stott et al., 2012). In other words, public trust in the police is heavily influenced by the normative concerns that police officers treat citizens fairly, respect citizens' rights, interact citizens with respect, and listen to and care about citizens' concerns.
Many studies examining public trust in the police employed one or sometimes two of the three leading models (i.e., instrumental, expressive, or normative models) as theoretical frameworks for explaining the empirical relationships (Bradford & Myhill, 2015;Chambers et al., 2020;Jackson & Bradford, 2009;Jackson & Sunshine, 2007;Jang & Hwang, 2014;Park et al., 2021;Sun et al., 2013;Tyler, 2001Tyler, , 2005. However, relatively few studies have attempted to simultaneously test the empirical validity of these theoretical models of public trust in the police and to examine the relative effects of three models. The Korean National Police Agency (KNPA) has experienced a decline in public trust in the police and the levels of public trust in the KNPA remain lower than expected, irrespective of numerous KNPA attempts at police reform (Cao & Dai, 2006;Hwang et al., 2005). For instance, a comparative study confirmed that Americans were roughly twice as likely to report a higher level of trust in the police than South Koreans (Boateng et al., 2016, p. 299). Given the different context in which South Korea is placed, it may not be accurate or helpful to simply generalize to the South Korea context with findings from studies conducted in Western countries. In this regard, it is necessary to examine the factors that may influence public trust in the police in non-Western societies such as South Korea.
Therefore, the present study aims to shed light on what extent factors drawn from instrumental, expressive, and normative models influence public trust in the police in the context of South Korea by analyzing survey data from Daejeon Metropolitan City. Particularly, the current study attempts to examine the relative impact of the three models and to suggest policy implications to promote public trust in the police.

Public Trust in Police
Since trust can be defined as the belief that those who have authority to make a critical decision pursue acceptable outcomes, trust suggests that those in authority are reliable and valuable, and that they make decisions based on the right information and motivation (Jones, 2002). Trust is also associated with implicit or apparent expectations that others act in ways that are expected. In general, public trust can be described as "unquestioning belief in and reliance upon a group to which one belongs or a public institution established to protect citizens" (Cao, 2015: 242).
Trust is a particularly important concept for police agencies because citizens generally have limited knowledge of police practices and lack of expertise in evaluating their performance. The trust that a person has in police tends to be based on limited personal experience that conveys little information about police intentions and characteristics (Jackson et al., 2012(Jackson et al., , p.1054. Specifically, public trust in police can be defined as the belief that police officers have the right intentions for citizens and are competent to act in a particular way in certain circumstances (Hardin, 2002). That is, public trust in the police refers to the citizens' belief that police officers have appropriate motives and are competent in carrying out their duties as the public expects. Thus, public trust is related to public confidence that police officers would have the greatest interests in the community and would exercise their authority consistent with those interests (Tyler & Huo, 2002).
When the police fail to earn the public trust, their competency to maintain public order is likely to diminish, which may lead to disastrous consequences in neighborhoods. Since trust in the police is an important precursor to a perception of legitimacy, public trust in the police is an indispensable factor in policing and leads people to willingly obey the law and cooperate with the police (Bradford, 2014). While the police restrict citizens' rights and freedom due to the nature of their duties such as crime prevention and criminal investigation, citizens who trust the police tend to voluntarily accept such restrictions. Thus it is argued that public trust in the police is a prerequisite for successful police activities (Nix et al., 2015).
More specifically, the effects of public trust in the police are as follows. First, public trust in the police may promote citizens' compliance with the law (Jackson et al., 2012). It is assumed that citizens voluntarily abide by the law when they trust the police. Otherwise, citizen compliance with law can be obtained through threat or actual use of force. Police should rely on extensive and voluntary law-abiding activities so that police can concentrate their resources on specific cases and certain situations where compliance has not voluntarily been obtained (Sunshine & Tyler, 2003;Tyler & Huo, 2002). Second, public trust in the police may encourage citizens to cooperate with the police. When citizens proactively cooperate with the police, they are more likely to report crimes, provide information about criminals, and participate in crime prevention activities (Nix et al., 2015;Tyler, 1990). Third, public trust in the police may imbue the police with authority. The empowerment is associated with citizens' intentions to accept discretionary judgment by the police.
Police officers' discretion can be exercised because citizens have authorized the police to do so. Fourth, public trust in the police may foster citizen perception of police legitimacy. Trustworthy police are seen by citizens to be effective, fair, and have the shared value with, and strong commitment to the community (Tyler & Huo, 2002).

Instrumental Models
The instrumental model posits that public trust in the police is associated with how effectively the police do their job, especially in reducing crime, and fear of crime. The model suggests that the police are instrumental in playing a central role of crime prevention and making people feel safe (Sun et al., 2014: 127). Accordingly, if the police are perceived as failing to achieve citizens' expectation, citizens perceive this as an indication that police are ineffective at controlling crime. This may result in public distrust of the police. For instance, several studies suggested that public willingness to support and cooperate with the police increased when the police are perceived as reliable institutions that deliver a credible threat of sanctions for wrongdoing, along with a fairly distributed police service (Murphy & Cherney, 2011;Sunshine & Tyler, 2003).
The instrumental model has some overlap with the accountability model and the performance model of public satisfaction with the police (Skogan, 2009;Sun et al., 2014;Van Craen, 2012). According to the accountability model proposed by Skogan (2009), citizens expect the police to take full responsibility for crime prevention and reducing fear of crime in the neighborhood. Similarly, the performance theory suggests that public trust in the police is closely associated with their perceived police performance and effectiveness. That is, people tend to report a higher level of satisfaction with the police when police agencies perform their duties effectively (Espinal et al., 2006).
Prior studies based on the instrumental perspective have examined two main determinants of public trust in the police: crimes (e.g., crime rates, victimization experience) and fear of crime. Many studies of public trust in the police focusing on crimes showed that police effectiveness played a critical role in shaping public trust in the police (Boateng, 2017;Cao et al., 2012;Sahapattana & Cobkit, 2016;Sampson & Bartusch, 1998;Sindall et al., 2012;Solakoglu, 2016;Tankebe, 2008). For example, Sampson and Bartusch (1998) reported a negative association between neighborhood homicide rates and the evaluation of the police after controlling for other neighborhood factors and individual factors. Tankebe (2008) also found the positive influence of perceived effectiveness on trustworthiness of the police in Ghana. It also confirmed that the effect of perceived effectiveness of the police would be stronger with higher level of perceived procedural fairness. In addition, confidence in the police is higher in countries with more government efficiency and is lower among residents of countries with higher homicide rates (Cao et al., 2012). Using a time series analysis of British Crime Survey data, Sindall et al. (2012) found that confidence in the police was not related to aggregate worry about crime but to perceptions of crime and the property crime rate. More recently, Boateng (2017) confirmed that police effectiveness such as controlling crime and providing service had a positive effect on trust in the police. With regard to relative effects of police effectiveness, Sahapattana and Cobkit (2016) found that public attitude toward crime suppression had the strongest effect on public confidence in the police followed by attitude toward crime prevention. In another study, Chambers et al. (2020) found that compared with the expressive model, the instrumental model was better in predicting procedure-based trust in the police. That is, expressive concerns are significantly associated with procedurally based trust, whereas instrumental concerns are not predictive of outcome-based trust.
A few studies have indicated that fear of crime had a significant effect on public trust in the police (Bradford & Myhill, 2015;Jackson & Bradford, 2009;Jackson & Sunshine, 2007;Kääriäinen, 2008). Specifically, Kääriäinen (2008) observed that the relationship between insecurity and trust in the police was inverse, suggesting that a feeling of insecurity in one's neighborhood lowers one's level of trust in the police. However, other studies found that fear of crime had no or limited effect on public trust in the police (Merry et al., 2012;Park et al., 2021). Specifically, Merry et al. (2012) found that fear of crime was not significantly related to public trust in the police.
In the context of South Korea, Hwang (2013) revealed that fear of crime had significant effects. Interestingly, the direction of the impact between Seoul citizens and Korean Americans in Detroit was inconsistent. In the case of Seoul citizens, the fear of crime had a positive effect on trust in the police while there was a negative relationship between fear of crime and public trust among Detroit Koreans (also see Jang, 2014). However, a recent study by Park et al. (2021) found that perceived fear of crime along with police effectiveness had a limited effect on public confidence in the police in South Korea.
Based on prior research, it can be hypothesized that police effectiveness and fear of crime significantly influence public trust in the police when individual characteristics such as age and gender are held constant.

Expressive Models
The main argument of the expressive model of public trust in the police is that citizens' views on trust in the police are shaped by expressive interests in community order and cohesion rather than by instrumental interests in crime and personal safety (Jackson & Sunshine, 2007;Sun et al., 2013). In other words, citizens expect the police to be not only crime fighters but guardians of local morality and values as well. Theoretically, this model is linked to social disorganization theory (Reisig & Parks, 2000;Sampson & Bartusch, 1998) and social capital theory (Brehm & Rahn, 1997). It is argued that high levels of disorder and low levels of community integration tend to undermine public trust in the police, as residents believe that the police fail to maintain the moral structure of the community. The expressive model also reflects the main premise of social capital theory in which a social network that forms around an environment is positively related to trust in government agencies (Brehm & Rahn, 1997). Particularly, since the police play an important role in local government, citizens' assessment of trust in the police can be heavily influenced by their social capital.
In addition to social capital, informal social control and social cohesion are closely associated with the formation of moral order and social order within a neighborhood. Prior studies have found that expressive concerns are more important than instrumental concerns (Cao et al., 1996;Jackson & Bradford, 2009;Sun et al., 2013). For example, Jackson and Bradford (2009) argued that people generally evaluate their local police with social cohesion and moral consensus (i.e., expressive concerns) rather than the risk of victimization (i.e., instrumental concerns). Cao et al. (1996) also indicated that contextual variables such as informal collective security and community disorder showed larger effects than crimerelated factors in terms of trust in the police (also see Chambers et al., 2020).
Most prior studies examining the effects of expressive concerns on public trust in the police indicated that perceived law and order, neighborhood disorder, and quality of life were significantly associated with public trust in the police (Boateng, 2017;Han et al., 2017;Nix et al., 2015;Sprott & Doob, 2009;Sun et al., 2013). Using survey data collected from Chinese citizens, Sun et al. (2013) found that expressive interests had a significant impact on public trust in the police. In particular, a sense of safety increased the odds of public trust in the police and citizens' trust in the police was predicted by trust in neighbors. Boateng (2017) also suggested that people are more negative about the police when they perceive high levels of community disorder (also see Nix et al., 2015;Sprott & Doob, 2009). Put differently, levels of public trust in the police were predicted by collective efficacy (Nix et al., 2015) and neighborhood cohesion (Han et al., 2017). On the other hand, Sindall et al. (2012) found that citizens' awareness of crime, disorder, and social cohesion did not have a significant effect on public trust in the police.
With regard to research in South Korea, Hwang (2013) reported that neighborhood cohesion had a positive effect on South Koreans' trust in the police. Jang (2014) also found that neighborhood cohesion and informal social control were positively associated with public confidence in the police whereas perceived disorder was negatively related to public confidence in the police. However, a more recent study revealed that perceived social cohesion had no effect on public confidence in the police (Park et al., 2021). Rather, the study found that individuals with a higher level of trust in formal control (i.e., criminal justice policies) retained greater confidence in the police.
Building upon these existing findings, social cohesion and informal social control are hypothesized to positively affect public trust in the police while neighborhood disorder is expected to exhibit a negative relationship.

Normative Models
Based on a procedural justice perspective, the normative model has suggested that the perceived fairness of police influences citizens' trust in the police (Gau, 2014;Reisig et al., 2007;Tankebe, 2013;Tyler, 2006). When people believe that the police fairly treat them, they are more likely to trust in the police (Murphy et al., 2014). According to the normative model, citizens decide whether to support the police based on procedural fairness rather than on instrumental concerns. That is, the level of normative compliance to the law emanate from judgments concerning the legitimacy in the intergroup context created by the police (Stott et al., 2012). Tyler (2005) presents two main elements of procedural justice: quality of decision making and quality of interpersonal treatment. Most studies have indicated that procedural justice is a significant predictor of public trust in the police (Jackson & Bradford, 2009, 2010Nix et al., 2015). More specifically, Tankebe (2008) found that those who perceived the police to behave fairly are more likely to report greater trust in the police. More recent studies also confirmed that evaluation of procedural justice had a positive effect on public trust in the police (Nix et al., 2015;Van Craen & Skogan, 2017). In the context of South Korea, Lim (2018) reported that perceptions of procedural justice had a strong effect on adolescents' trust in the police. The results of the structural equation models (SEM) analyses also showed that perceived police fairness was the primary determinant of confidence in the police in South Korea (Park et al., 2021). Based on these findings, it is hypothesized that procedural justice is positively related to public trust in the police.
While findings from this line of inquiry are somewhat inconclusive, we speculate that instrumental concerns are more important than expressive concerns or normative concerns in influencing South Koreans' trust in the police for the following reasons. First, compared to Western societies, South Korea is a culturally and racially homogenous society (Kim & Jeon, 2017;Shin, 2006). Thus, high levels of neighborhood cohesion and low levels of neighborhood disorder in South Korea may make expressive concerns less relevant to the evaluations of the police. Second, South Korean culture generally emphasizes the importance of moral and ethical behavior of citizens. Citizens are inclined to have a considerably obedient attitude toward police officers (Morash et al., 2008). Finally, as many crimes against women have recently occurred, it resulted in high levels of fear of crime among women in South Korea (Choi et al., 2020).

Data
The data for the current study were collected from a community survey in the Daejeon metropolitan city, which is 13th largest city in South Korea. Daejeon Metropolitan City is geographically located in the center of South Korea with the approximate population of 1.5 million and is subdivided into five districts ("Gu") including 53 neighborhoods 1 ("Dong"). To ensure representativeness of the sample, 20 survey staff members (mainly graduate students and undergraduate students) for this study were assigned 53 neighborhoods across the five districts and randomly chose 15 households from each neighborhood. The survey staff members visited each selected household and asked to fill out the questionnaire during June, 2018. 2 The total of 800 survey questionnaires was distributed and 783 were completed and returned (response rate = 97.88%). 3 Of 783 returned survey questionnaires, 11 incomplete questionnaires were excluded, resulting in 772 respondents for further analyses.

Variables and Measurement
Dependent variable. Public trust in police can be defined as citizens' expectations that individual officers will perform their assigned duties properly. The items to measure public trust in police have been represented and discussed in other studies (see Cao, 2011;Jackson & Bradford, 2010;Skogan, 2009;Tankebe, 2009). To measure public trust in the police, the current study adopted Tankebe (2009)'s five survey items including (1) The police are reliable; (2) I am proud of the police; (3) I have confidence in the police; (4) The police are mainly honest; (5) The police always act according to the laws. The responses were coded into a 4-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree). The factor structure might differ between countries, so an exploratory factor analysis (EFA) was conducted for South Korean data to test construct validity of those five items. In particular, EFA based on Maximum likelihood (ML) extraction with an oblique rotation method was conducted since it produces a better simple latent structure and more interpretable results than a principal components analysis with a varimax rotation which has been traditionally used in criminal justice research (Fabrigar et al., 1999;Morash et al., 2008). The factor analysis yielded one factor solution that roughly captured 82% of variance (factor loadings > .87, χ 2 = 74.44, df = 5, p < .01). For the measures of public trust in the police, Bartlett scores (factor scores) were developed for each scale, and thus each scale had a mean score of 0 and a standard deviation of 1. Finally, the reliability coefficient (Cronbach's α) of public trust in the police was .95, which indicates an excellent level of internal consistency among the items.
Independent variables. The primary independent variables included police effectiveness, fear of crime, social cohesion, informal social control, neighborhood disorder, and procedural justice. Police effectiveness means the ability of the police to fulfill their central duties not only in combating crime but also in providing a visible presence, policing public events, and responding to emergency calls (Stanko & Bradford, 2009). Police effectiveness was measured using three questions as shown in Appendix A. The conceptual definition of police effectiveness for this study is that police effectively suppress crimes and quickly respond to citizens asking for help. The items used to measure police effectiveness were adopted from Tankebe (2009) and Sunshine and Tyler (2003). The factor analysis identified a one factor solution which captured about 81% of variance (factor loadings > .89; Cronbach's α = .88).
Fear of crime refers to general and specific fear of crime which indicate a negative emotional assessment of fear and anxiety about being a victim of crime (Tyler, 2005). The Tyler (2005) items to measure perceived fear of crime were adopted to reflect fear of crime in neighborhoods (see Appendix A). The reliability of the scale was 0.95 and factor loadings ranged from 0.77 to 0.90 (explained variance = 71.86%, factor loadings > .77; χ 2 = 365.56, df = 20, p < .01; Cronbach's α = .95).
Social cohesion has been recognized as complex concepts that include such characteristics as social interaction, personal empathy, and value agreements. Similarly, Buckner (1988, p. 744) identified three key elements of social cohesion. These elements are: (1) the residents' sense of the community, (2) the degree of interaction, and (3) the degree of attraction. Thus, a community with strong social cohesion generally is the community where local residents feel a strong sense of community, frequently participate in community meetings, and want to keep living in their community.
The conceptual definition of informal social control is that community members are willing to participate in crime prevention events and to maintain social order and norms in the community (Nix et al., 2015;Wolfe et al., 2016). Four items adopted from Nix et al. (2015) were used to represent levels of informal social control in the neighborhoods (see Appendix A). The construct validity and reliability of the four items were confirmed through EFA and Cronbach's alpha statistics (explained variance = 69.41%, factor loadings > .80; χ 2 = 321.15, df = 2, p < .01; Cronbach's α = .85).
Neighborhood disorder was measured using Tyler's (2005) scales (e.g., "the garbage was dumped randomly around and it is messy," "there are many cars or empty buildings left unattended around," etc). The factor analysis confirmed a high level of construct validity among six items (explained variance = 57.74%, factor loadings > .73; χ 2 = 278.07, df = 9, p < .01). The reliability of the scale was 0.85 which is relatively higher than an acceptable reliability coefficient (Cronbach's α = .70; see DeVellis, 2003).
Lastly, procedural justice was measured by combining five items asking various questions about procedural justice as shown in Appendix A. The factor and reliability analyses confirmed that the scale had a high level of construct validity and internal consistency among the five items (explained variance = 79.34%, factor loadings > .84; χ 2 = 29.75, df = 5, p < .01; Cronbach's α = .94).
To avoid severe multicollinearity (MC) problems among independent and control variables, the variance inflation factor (VIF) and tolerance statistics were calculated. Results confirmed that there were no severe MC problems among the variables (VIF < 1.80; tolerance > 0.56). Thus, the current study simultaneously included all available variables in the regression models.

Analytic Strategy
Ordinary least squares (OLS) regression analysis was used to estimate the effects of the independent variables on public trust in the police while controlling for demographic factors. First, police effectiveness, fear of crime, social cohesion, informal social control, neighborhood disorder, and procedural justice were entered to examine their relationships with public trust in the police. Second, four control variables were entered to control for any possible intervening effects of demographic characteristics variables (gender, age, home ownership, residence period) of individual respondents. Table 1 illustrates the descriptive statistics for dependent and independent variables used in this analysis. Of the total 772 respondents, about 50% were male and most citizens own their houses (70%). More than half of the respondents were younger than 49 years old (cumulative percent = 59.4%). In terms of residency period, 429 respondents (55.1%) have lived less than 15 years in the current neighborhood. Table 2 includes results from a correlation analysis among independent and dependent variables. As shown in the Table  2, while there was no correlation between fear of crime and public trust in the police, the other six independent variables were statistically associated with the dependent variable at the .05 level. More specifically, police effectiveness, social cohesion, informal social control, and procedural justice were positively correlated with public trust in the police. That is, as police effectiveness, social cohesion, informal social control, and procedural justice increased, the public trust in police increased. However, there was a negative relationship between neighborhood disorder and public trust in police (r = −0.09). The magnitudes of the correlation coefficients were relatively small or weak for most of the significant variables except police effectiveness (r = 0.73) and procedural justice (r = 0.64) which is considered high. The results from the OLS regression analyses were presented in Table 3. In the first model presented in the righthand column of the table, the statistically significant four predictors were police effectiveness, social cohesion, neighborhood disorder, and procedural justice. The four predictors explained approximately 59% of the variance in public trust in the police. After adding the control variables in the second model, three of the same predictors remained statistically significant. The additional variance explained by control variables was 1.0%, and the increase was statistically significant at the 0.05 level (R 2 change = .01, F (4, 762) = 2.90, p < .05).

Findings
However, neighborhood disorder was no longer a significant predictor of public trust in the police. To compare the relative strength of the effects, β weights (standardized regression coefficients) were also calculated. Based on these coefficients, the strongest predictor for public trust in the police was police effectiveness (β = .54) followed by procedural justice (β = .27) and then social cohesion (β = .12). These findings indicated that the respondents who perceived that their police officers effectively responded to calls for service and applied the rules equally to citizens had a higher level of public trust in police. Lastly, the citizens who actively participate in community activities reported a higher level of public trust in police. Regarding the control variables, residence period was negatively associated with public trust in police (β = −.08). That is, respondents who lived longer in the neighborhood had lower levels of public trust in the police.  Note. b = b coefficients; SE = standard errors; β = beta weights (standardized coefficients). a Bartlett factor scores. b Number of observations after listwise deletion of incomplete cases. *p < .05. **p < .01.

Discussions
To extend the knowledge of public trust in the police, the current study investigated to what extent factors drawn from instrumental, expressive, and normative models influence public trust in the police and examined the relative impact of the three models. The results revealed that police effectiveness, social cohesion, and procedural justice had significant effects on public trust in the police. In particular, among these significant predictors of public trust in the police, police effectiveness was the strongest predictor for public trust in South Korean police followed by procedural justice and social cohesion. On the other hand, fear of crime, neighborhood disorder, and informal social control were not significantly associated with public trust in the police. These findings indicated that instrumental models, expressive models, and normative models are somewhat applicable to the context of South Korea. First, this study found that police effectiveness had the strongest effect on public trust in the police. Consistent with our expectation and prior studies (Boateng, 2017;Karakus, 2017;Tankebe, 2008;Van Craen, 2012), the relative effect of police effectiveness on public trust in the police was found to be twice as stronger as procedural justice. That is, citizens with a higher perceived police effectiveness are twice as likely to report favorable attitudes and trust toward the police. This finding was consistent with performance theory explaining that citizens are more likely to be satisfied with the police when the police perform their duties effectively (also see Espinal et al., 2006). According to the instrumental theoretical framework, citizens' judgments of police effectiveness depend on the capability of the police to fight crime, to reduce fear of crime, and to enhance safety and security in a neighborhood (Jackson & Bradford, 2009;Sun et al., 2013). Thus, to improve levels of public trust in the police, the police need to more effectively implement their basic tasks of preventing and suppressing the crime.
Second, Sampson and Bartusch (1998) argued that macrolevel conditions affected public attitudes toward police. In other words, residents in underdeveloped areas are more likely to be cynical about the police. These findings expand macro-level studies and specify potential mechanisms in which neighborhood conditions may influence evaluation of formal social control agencies such as police agencies (Park et al., 2021).
Consistent with previous research (Han et al., 2017;Hwang, 2013;Jang, 2014;Karakus, 2017;Nix et al., 2015;Sun et al., 2014), the results also confirmed that social cohesion had a positive impact on public trust in the police. That is, citizens perceived their police as the representatives of Daejeon city to maintain social cohesion and moral standard. Accordingly, the police should help the community to voluntarily participate in community activities and strengthen social ties among residents. Moreover, as the current study found that social capital (e.g., trust in the neighborhood) may be accompanied by higher level of public trust in the police, police should actively engage in community programs for strengthening social cohesion and increasing informal social control through mutual interactions with local residents. By doing so, trust relations among neighbors can be formed and community members are willing to work together to prevent crimes and to maintain community order, resulting in higher levels of public trust in the police.
Third, this study provides empirical evidence that public perceptions of procedural justice are related to public trust in the police. More specifically, consistent with prior research in Western societies (e.g., the U.S., U.K., & Australia), this finding could expand the current knowledge of public trust in the police to non-Western societies, including South Korea. Specifically, the current study found that public perceptions of procedural justice play a central role in the normative evaluation process of public trust in the police (also see Karakus, 2017;Nix et al., 2015;Tankebe, 2008;Tyler & Huo, 2002). In addition, the perceived legitimacy among people of the way they were treated had a significant impact on the internal dynamics as well as patterns of collective action among people (Stott et al., 2012). In other words, individuals who recognize that the police are treating citizens fairly are more likely to trust police officers.
To facilitate public trust in the police, police officers should be procedurally fair and treat citizens with a sense of respect. Police officers should bear in mind that the quality of interaction improve public trust in the police, but the effects of negative experiences may be greater than those of positive experiences (Skogan, 2006). In particular, to ensure that the police are seen as being procedurally fair, the police should provide the residents with the opportunity to explain how decisions are made and allow citizens to file complaints, along with courtesy and respect.
Based on a procedural justice perspective, the police can create motive-based trust and shared collective membership. Police also promote collaboration with citizens to maintain community values as a civil guardian. As a result, fighting against crime would be more efficient, more cost-effective, and more ethical by treating citizens fairly with dignity and respect. When the police are effective, procedurally fair, and concerned with local interests, this would not only make the police more responsible, but would also enhance moral relationships between citizens and the police, and promote active citizen engagement in community safety as well (Jackson & Bradford, 2010: 248). In sum, the results of this study suggest that the more effective and procedurally fair the police are, and the stronger social cohesion in neighborhoods is, the higher the levels of public trust in the police will be.
Based on the results, we suggest the following policy and practical implications to improve public trust in the KNPA. First, as people perceived higher level of police effectiveness, they reported higher level of trust in the police. Thus, police agencies should make an effort to improve police effectiveness and public perceptions of police effectiveness (Skogan, 2009). This does not imply that police effectiveness itself is a sufficient condition for public trust in the police. Rather, police effectiveness can be considered as a necessary condition for maintaining public trust in the police. Police performance is directly linked to the ability to respond to crimes, prevent crimes, and apprehend offenders. Effectiveness of crime control is also one of the basic factors in citizens' evaluation of police. That is, if a police agency is perceived as carrying out its duties in an effective manner, the police agency is more likely to be trusted by the public.
Second, external procedural fairness refers to procedural fairness in the relationship between the police and local residents. Procedural justice appears to play an important role in an individual's normative assessment of public trust in the police. Thus, KNPA needs to employ various organizational strategies to ensure that police officers actually apply the principle of procedural justice to every person they served. Supervision and discipline should be critical parts of such organizational strategies, and their effect may depend on monitoring thoroughly the delivery of procedurally fair services. In particular, training for line officers is another tool to provide a guideline for performing police officers' daily duties, and such training regarding procedural justice can affect police officers' perceptions toward their counterparts in the long run. Some studies provided evidence that training for procedural justice may positively influence police officers' view Wheller et al., 2013). For instance, in their quasi-experimental study of the short-term effects of training and the assessment of its long-term results in Chicago, Skogan et al. (2015) found that training makes Chicago police officers more supportive of the principles of procedural fairness.
According to the final report of President's Task Force on 21st Century Policing (2015), law enforcement agencies should adopt procedural justice as a guiding principle for internal and external policies and practices to guide their interactions with the communities they serve. However, the report also addressed difficulties and multi-dimensional issues for promoting procedural justice within police organizations (see MacQueen & Bradford, 2017). In particular, the task force examined the connection between fair supervision (i.e., internal procedural justice) and fair police activities (i.e., external procedural justice) and suggested that the perception of internal procedural justice stimulates police officers to implement external procedural justice. That is, internal procedural justice encourages police officers to develop external procedural justice (Van Craen, 2016, p. 275).
Police officers' perception of internal procedural fairness may affect their trust in citizens as well as supervisory officers. Specifically, fair treatment from supervisors can contribute to the formation of belief among police officers that most people can be trusted, which in turn facilitates implementation of external procedural justice through best practices (Van Craen & Skogan, 2017, p. 6). Therefore, police agencies should strive to enhance internal procedural fairness to improve public trust in the police.
Third, social cohesion may have an objective and subjective influence on public trust in the police. Residents proactively participating in various forms of social activities can enhance social capital by improving information flow, mutual communication, and support. These social capital activities thus increase public trust in the police. Accordingly, to strengthen social cohesion, police should continue to develop and invest in community programs to ensure that residents are well acquainted with each other. When residents have regular meetings to share what is happening in the neighborhood and try to help each other, residents will gain more trust among themselves.
People who distrust each other are less likely to trust public institutions. Rothstein and Stolle (2008), for example, argued that citizens' generalized trust (usually in others) was related to justice of law enforcement agencies. They also insisted that police officers act as important signals to citizens regarding moral standards in society. In other words, by acting fairly, police officers stimulate citizens to act fairly and encourage citizens to expect others to act in similar ways. Such actions result in generalized trust.

Conclusion
This study shows that police effectiveness, procedural justice, and social cohesion were found to be significant predictors influencing public trust in the police which indicates instrumental models, expressive models, and normative models are somewhat applicable to the context of South Korea. Although the current study makes significant contribution to the current knowledge of public trust in the police, there are several limitations. First, the findings of the current study are limited in their generalizability due to the smallsize sample collected only in Daejeon Metropolitan City, South Korea. In addition, using data gathered in a nonexperimental setting with a cross-sectional nature, this study was not able to control and account for temporal changes in public trust in the police. Future studies need to utilize representative samples using longitudinal data. In addition, the current study did not include the factors such as democracy or police corruption. Future research should incorporate the perception of democratic maturity and the perception of police corruption as explanatory factors of public trust in the police (see Hsieh & Boateng, 2015). Moreover, future studies should employ multilevel analyses to examine cross-level interaction effects of neighborhood and individual factors on public trust in the police. Finally, while current research is only concerned with the factors affecting public trust in the police, future researchers need to explore the potential consequences of public trust in the police. That is, future studies should investigate whether the levels of cooperation and compliance with police will be further improved when citizens trust the police. In addition, it is necessary to examine not only the levels of public trust in the police, but also the levels of police officers' trust in citizens because the attitude of trust is mutually influential.
Higher levels of public trust in the police are critical not only for fostering sound police-community relationships, but also for enhancing community safety. To win the hearts and minds of Korean people, KNPA should continue to reform efforts in the organization and place the highest priority on enhancing police effectiveness, procedural justice, and social cohesion among citizens. Essentially, public cooperation with police cannot be divorced from trust in the police (Sunshine & Tyler, 2003;Tyler, 2005 2. Neighbors will help girls in any way when they see them being bullied by bullies.

0.85
3. If neighbors organize a crime prevention team among themselves, they will support it.
0.80 4. If crimes occur frequently, neighbors will try to solve problems in some way. 0.84 Procedural justice 1. The police understand and apply the law accurately.
1-4 a 0.84 .94 2. The police make decisions based on facts, not on their own personal biases. 0.91 3. The police try to collect facts about the situation before deciding on an action.
0.89 4. The police give an honest explanation for their actions. 0.91 5. The police apply the rules equally to the people they deal with. 0.90 Note. The above scales were created using an Exploratory Factor Analysis based on Maximum Likelihood Extraction with direct oblimin rotation. a 1 = Strongly Disagree, 2 = Disagree, 3 = Agree, and 4 = Strongly Agree.