Introduction

Police dispersed dozens of remaining protesters Friday [April 11] morning as lawmakers prepared to attend the Legislative Yuan’s first plenary meeting in more than three weeks. The protestors stayed in the square in front of the main Legislative Chamber overnight after most of the student-led occupation movement against a service trade agreement with Beijing ended late Thursday. After repeated warnings, the police began their dispersal around 7 a.m. taking some people away after handcuffing them. In the background, workers continued to work on restoring the appearance of the Nation’s Parliament, including handing back a plaque over the main entrance to the Legislature that was taken down on March 18, when the students stormed in the compound and began their 24-day occupation …. (The China Post 2014a, April 12).

The recent confrontation between the police and college students in Taiwan, Taipei, drew worldwide attention. The student protests were occasioned by the passage of the Cross-Strait Service Trade Agreement (CSSTA) by the then-ruling Kuomintang party [KMT, Chinese Nationalist Party] at the Legislature without a customary clause-by-clause review. The Sunflower Student Movement, driven by a coalition of college students, occupied the Legislative chamber on March 18 and, later, stormed the Executive Yuan of Taiwan on March 23. During this student-led protest movement period, two large-scale confrontations occurred on March 24 and April 11, respectively. More than 2000 college students stormed and occupied the Executive Yuan. In time, they were forcibly removed by approximately 1000 police officers equipped with riot shields, batons, and two water cannon trucks. The student/police conflict involved over 1000 additional protesters swarming various police precincts for several hours (The China Post 2014a, March 24). These protests ultimately ended with the police chief issuing a public apology (The China Post 2014a, April 12). Following Sunflower Student Movement, later social protests such as the anti-nuclear protest taking place on April 27 and the May 1 Labor Day protest have been organized, mobilized, and implemented by college students. Although scholars have opined that college students should be strongly encouraged to pay attention to and participate in public affairs-related public policy advocacy, only a few have given attention to the intimate relationship obtaining between the police and college students.

This study analyzes college students’ satisfaction with the police in Taiwan. It advances the existing literature on Taiwanese attitudes toward the police in three noteworthy ways. First, to the best of our knowledge, no extant studies systematically review the police-college students’ confrontations in Taiwan after the termination of the state of martial law in 1987. Second, although the past decade has witnessed the burgeoning of studies on Taiwanese perceptions of the police (Cao and Dai 2006; Lai et al. 2010; Sun et al. 2013a; Wu et al. 2012), college students’ perceptions of police have been largely overlooked within that spate of recent scholarship. By contrast, a growing body of research focuses on students’ perceptions of police in Mainland China. For example, Wu and Sun conducted a series of studies focused upon college students’ perceptions of police (Wu and Sun 2010), and contact experiences with the police (Sun and Wu 2010). Moreover, they also compared college student evaluations of police performance in China and the USA (Wu 2010a). Their studies explored this topic by applying a theoretical framework derived from research studies conducted in Western societies. As college student/police interactions are taking on more significance in Asian countries in recent years, and are likely to maintain or even gain in salience in the future, more research on regional differences in assessments of police roles and performance among college students is clearly called for on the part of scholars and researchers.

It is also the case that the authors of several recent studies on public attitudes toward the police in Asian countries have stressed the point that Chinese culture should be introduced and integrated into theoretical frameworks developed from research carried out in the West (Cao and Dai 2006; Lai et al. 2010; Wu and Sun 2010; Wu et al. 2012). While it is the case that Western societies focus heavily on individualism and personal liberty, the traditional Chinese culture extols collectivism and encourages people in many social settings to sacrifice their interests for the sake of promoting harmony within society. Strongly under the influence of collectivism, Chinese people are acculturated into the expression of benevolence when interacting with others; their lives are highly structured in a social hierarchy of clear social stratification (Shi 2001; Yin 2003). In reflection of this sentiment of high regard for benevolence, some public officials and many police family dependents echoed the view that Taiwan’s college students should hold benevolent and empathetic attitudes toward the police as they witnessed student protesters swarming police precinct (Chinatimes 2014, April 10). To date, however, only quite limited research (e.g., Wu et al. 2012) featuring Chinese traditional cultural norms and beliefs in the assessment of public sentiments toward the police has been published.

Drawing on survey data collected from a sample of 688 college students in five cities and one county prior to the 24-day Sunflower Student Movement, this study attempts to investigate empirically the level and the correlates of Taiwanese college students’ satisfaction with police performance. Three specific questions guide this research: (1) What social movements can be seen to be at play in police/college student confrontations in Taiwan over the course of the past three decades? (2) What is the overall level of satisfaction with police performance among contemporary Taiwanese college students? and (3) What are the factors, such as the Chinese virtue of benevolent sympathy, that account for variation in Taiwanese college students’ satisfaction with the police? The findings reported in this study can establish our baseline knowledge about Taiwanese college students’ evaluations of the police, and likewise provide the basis for formulating some timely implications for police research, training, and policing practice.

Literature Review

Police-College Students’ Confrontations in Taiwan

In 1987, the 26-year declaration of martial law was officially lifted and Taiwan moved into the company of transitional societies—i.e., counties moving from autocratic rule to democratic governance (Lai et al. 2010). Accordingly, along with the movement toward democratic government and competitive electoral politics, the Taiwanese police have inevitably (however gradually) shifted their primary role from the maintenance of the status quo and political control of the Kuomintang to the protection of individuals’ rights and proactive crime control (Gingerich et al. 2011; Sun and Chu 2006). Unfortunately, the cardinal task of establishing a democratic police force on the foundation of an authoritarian past during a period of rapid social and political change was a very difficult one to accomplish. Over the past three decades, while some significant broad-scale factors (e.g., the rise of critical citizens and the decline of public support for government) have affected public evaluations of the police since the 1990s (Norris 1999), several challenging social movements and open conflicts between protesters and police occurred forcing the police to have to adjust their roles and functions in some major ways. As a consequence of these forces, Taiwanese citizens have continuously held relatively low levels of confidence in their police (Cao and Dai 2006; Lai et al. 2010; Wu et al. 2012).

Among those social movements challenging ruling governments, three significant events took place which were organized and initially mobilized by college students—namely, the Wild Lily Movement in 1990, followed by the Wild Strawberries Movement in 2008, and most recently, the Sunflower Movement in 2014. The first Taiwanese student movement to take place in modern Taiwan occurred in March of 1990; this was a 6-day demonstration involving more than 6000 students calling for democracy after the lifting of martial law in 1987. This movement, initiated by students from National Taiwan University, occupied Chiang Kai-Shek Memorial Hall square (later re-dedicated as Liberty Square in 2007). The initial protest grew quickly, ultimately leading to the participation of an estimated 22,000 demonstrators. This demonstration and subsequent movement sought three principal outcomes: (1) Direct election of Taiwan’s president and the nation’s vice president, (2) New popular elections for all representatives in the national assembly, and( 3) The abolition of the widely despised Temporary Provisions Effective During the Period of Communist Rebellion. During the period of the movement, the demonstrators stated their concerns and expressed their opinions in commendably rational discourse and in peaceful ways to the press and to governmental agencies. More than a thousand riot police surrounded the demonstrators and separated protestors and opposing citizens in the square. Although some minor conflicts arose between the riot police and students, this first student-led movement in modern-day Taiwan ended peacefully. This movement, accompanied by other political, economic, and international factors, marked a crucial turning point in Taiwan’s transition to pluralistic democracy, leading in due course to the first direct presidential election which was held in 1996 (Chao and Myers 1998).

Second, student-led sit-in protests were conducted in response to incidents of police violence, brutality, illegal search and seizure, and abuse of power during the visit of Chen Yunlin to Taiwan for “The Second Jiang-Chen Talk” on November 3 of 2008. The demonstrators occupied the square in front of the Executive Yuan and issued several demands, including these: (1) A public apology be made by President Ma Ying-Jeou and Premier Liu Chao-Shiuan for police misconduct, (2) The immediate resignations of National Police Agency Director-General Wang and National Security Bureau Director-General Tsai, and (3) An immediate review and amendment of the Assembly and Parade Law to ensure the right of peaceable assembly and free expression of political views (Taipei Times 2008a, November 11). Although dialogues were arranged between the secretariat of the Executive Yuan and representatives of the protesting students, no agreement was made in the end. A request was made by college student leaders for permission to engage in protest activity, but the police denied the request; riot police officers were deployed to forcibly remove students, professors, and other protestors from the Executive Yuan and move them to other sites such as the campus of National Taiwan University, Liberty Square, and several other secure locations on the afternoon of November 7. At 7:00 pm, the police cleared the square of the Executive Yuan and those protestors were moved to Liberty Square for the duration of their month-long protest movement. Premier Liu rejected the demand for a public apology, and President Ma issued a statement indicating that “while he believes there was room for improvement in Wang and Tsai’s handling of public assembly issues, he saw no reason for them to step down” (Taipei Times 2008b, November 13). This movement pushed the Justices of the Constitutional Court into action and they upheld in their No. 718 Interpretation that the citizens’ ability for urgent and occasional assembly and parade activities is protected by Article 11 of the Constitution. In this case, the protesters did not apply for permission from the police.

Third, the most recent Sunflower Student Movement occupied the nation’s Legislative chamber for 24 days, stormed the Executive Yuan on March 29, and subsequently swarmed some police precincts on April 11 as earlier noted. On April 6, Legislative Yuan Speaker Wang Jin-Pyng visited the occupying students and promised he would postpone review of the trade pact until legislation monitoring all cross-strait agreements has been passed (The China Post 2014b, April 7). Although these protest events were set to end peacefully and without bloodshed in a sign of the maturity and rationality of Taiwan’s democracy, the Minister of Justice Luo noted nonetheless that the students’ 24-day occupation of the Legislative chamber and the Executive Yuan were illegal and consequently must be investigated. She opined that the students knew full well that they were breaking the law and were trying to evade legal accountability (Taipei Times 2014, April 10). In terms of public assessment of the protest, the National Development Council [NDC] (2014) conducted a countywide poll and documented the fact that while 96 % of respondents were aware of the occupation of the Legislative chamber by students in protest of the CSSTA, only 46 % supported the students’ occupation as a means of expressing their dissatisfaction. Over two thirds of the respondents (69 %) opposed actions such as the storming of the Executive Yuan and clashing with the police on the evening of March 23; only 13 % favored the students’ actions. Approximately 60 % supported police’s forceful evacuation of protesters that night, while 23 % did not support forceful evacuation. Nearly 44 % believed the evacuation was a legitimate and lawful action to protect the Executive Yuan, but about a third (32 %) viewed it as a “bloody crackdown” action on the part of the police.

Overall, these three student-led movements contributed to the legacy of confrontation between police and Taiwanese college students. In these confrontations, several serious, violent confrontations did take place in each police/college student conflict episode; however, both the students and the police acted in highly rational and restrained ways in virtually all cases. The college students tried to put safeguards into place to maintain the order necessary to formulate and express their concerns and articulate their opinions. As for the police, leadership and their officers were inclined to exercise constraint in resort to violence, minimize any brutality that might occur, and use as little direct force as possible to defend or respond to the college student protesters. This restraint was in clear evidence, notwithstanding the fact that some of the protestors did not follow orders from student leaders and engaged in irrational behaviors and/or deviant misconduct (e.g., lay on the roads and intersections; throw stones, water bottles, and backpacks at police officers). Second, while some people are sympathetic to students and support their position in all matters, most citizens are inclined to believe that Taiwanese college students should have taken more lawful forms of action and used more appropriate means to address their dissatisfaction with governmental policy. Most Taiwanese today believe that Taiwan is a rather mature democratic society, and that multiple opinions can be expressed by lawful and legitimate channels without inappropriate impediments. Third, those student-led protests viewed as catalysts of social change in due course led to some dramatic political revolutions and contributed to Taiwan’s transition toward becoming a democratic nation. Finally, the development of relationships across the Taiwan Strait, specifically during the last two waves of college student-led movements, played a key role in those events. In both cases, the “China factor” constituted a core underlying issue, and this issue nearly invariably evokes emotional sentiments among Taiwanese college students.

Models that Shape Public Satisfaction with Police

Generally, the public attitudes toward the police (PATP) area of study refers to the subjective assessment of key police behaviors and outcomes such as public safety services provided, effectiveness and competence in carrying out work, law enforcement practices and policies, and officer and agency performance (Reisig and Parks 2000; Ren et al. 2005; Schuck et al. 2008). This area of study has been a focal concern among criminologists and criminal justice researchers since the 1970s, especially in democratic societies. This is the case because the police in democratic societies cannot thrive and achieve public safety provision goals without the confidence and trust from the public (Cao et al. 1998). With the support from the public, the police can both refine their crime control strategies and achieve a good measure of crime prevention (Sun et al. 2012). It is the case that police agencies endeavor to be viewed as legitimate in the communities in which they serve and to draw strength from that reserve of respect and trust. Lack of public trust, and subsequently citizen compliance and cooperation, can lead to serious performance problems and/or dire political consequences (Wu et al. 2012).

Our review of the PATP literature revealed the presence of three rather distinctive models commonly used to explain the sources of public satisfaction with the police—namely, models featuring a focus on personal experience and attitudes, neighborhood contexts, and demographic characteristics (Brown and Benedict 2002; Lai and Zhao 2010; Ren et al. 2005; Weitzer and Tuch 2006; Wu and Sun 2010). The Chinese cultural element, benevolent sympathy, can be classified into the personal experience and attitudes model.

Personal Experience and Attitudes

A substantial amount of contemporary research has suggested that the effects of police contact should be considered as being one of the major predictors of citizen assessments of police (Brandl et al. 1994; Ren et al. 2005; Rosenbaum et al. 2005; Scaglion & Condon 1980). For example, Schafer et al. (2003) argued that public evaluations of the police and police services are more likely to be derived from their personal experience with law enforcement than arise from their own demographic background characteristics. Of course, there are several types of police-citizen contacts and some are associated with favorable outcomes while others are not. In this regard, Decker (1981) has suggested that the most important distinction in police/citizen contact entails the voluntary/resident-initiated contact vs. the involuntary/police-initiated contact. Specifically, voluntary contact tends to produce more positive results, whereas involuntary interactions tend to occur in more adversarial situations. Subsequent research has generally supported Decker’s claims in this area. For example, Rosenbaum et al. (2005) found that people who initiated police interaction (i.e., calling for assistance) were more satisfied than persons who were targets of police-initiated encounters such as traffic stops.

Victimization is often found predictive of citizen perceptions of the police, in large part because citizens often hold police directly accountable for crime control and public safety in their neighborhoods (Skogan 2009; Wu et al. 2012; Zhao et al. 2014). By-and-large, the extant literature has shown rather conclusively that victimization tends to give rise to unfavorable attitudes toward the police (Lai and Zhao 2010; O’Connor 2008; Payne & Gainey 2007; Wu and Sun 2010; Wu et al. 2009). In line with this consensus among scholars, Cao et al. (1996) found that victimization had a greater impact on citizens’ confidence in the police than any of their demographic characteristics. As Ren et al. (2005) have noted, those who are victims of crime tend to believe that the police failed to protect them. It should be noted in this regard, however, that some scholars have reported null effects from crime victimization (Chu & Song 2008; Wu et al. 2012). Likewise, in one of their recent studies Sun et al. (2013a) found that there was no significant association between crime victimization and public trust in police.

A factor that deserves much more attention than it has received is that of cultural values and their role in citizen assessments of police performance in Chinese societies (Cao and Dai 2006; Lai et al. 2010; Wu et al. 2012). The core of traditional Chinese culture is Confucianism, a belief system that emphasizes personal reservation, collectivism, and ubiquitous obedience in an established social hierarchy. Consequently, respect for customs, adherence to norms, and observance of rules associated with one’s social status and position in family and society is a bottom line to build/maintain harmony with others (He and Marshall 1997; Wu and Sun 2009). Chinese have long been taught that they should be devoted to their parents, deferential to elders, and respectful toward teachers and governmental agencies. It follows that properly acculturated Chinese should be loyal to their superiors and, to the extent necessary and possible, they should sacrifice their own interests if social harmony is best served by doing so.

It is also the case that Chinese people being reared in their families have been taught to forgive transgressions, and to be benevolent and empathetic toward other people. In Chinese, this concept is known as “Ren” (benevolence). Benevolence is a major principle that is intended to frame communications and guide interactions between people (Hampton and Xiao 2009). Chinese people are nearly universally familiar with some idioms directly related to empathetic spirits, idioms to which they were exposed when they were in primary schools. Idioms such as “What you do not wish upon yourself extend not to others (己所不欲、勿施於人),” “The feeling of commiseration implying the principle of benevolence belongs to all men (惻隱之心,仁也,人人皆有),” and “Holding a benevolent and good-hearted attitude toward those in need (人飢己飢、人溺己溺).” The core element of this belief is to realize one’s own humanity through goodness or love (Bockover 2007).

This virtue of Ren is similar to the concept of “sympathy” developed in Western cultures. Sympathy can be considered a special kind of empathy, viz. empathy coupled with a benevolent attitude toward other people. Stated more precisely, sympathy not only includes empathizing but also entails having a positive regard or a non-fleeting concern for the other person; the concept of “stepping into another’s shoes” captures this thought well (Chismar 1988). In Chinese societies, “benevolent sympathy” is an appropriate term for capturing the respectful and considerate associations one might have with non-relatives, specifically with official public servants such as school teachers, nurses, or even police officers. It reflects the values of Chinese tradition in which people are obligated to be charitable in their remarks and ascriptions of competence of other individuals, not because it is cost-effective or Machiavellian in some way, but rather because these “others” have paid their dues and are due respect as a consequence. The idiom “Even if he didn’t have prominent achievements, he still made some contributions (沒有功勞,也有苦勞)” is a good example. Adversely, “Hard working to sweat is still to be criticized (工作到流汗,被嫌到流涎)” refers to people holding a non-benevolent sympathy view toward others who do not grant the fact that there are often things beyond the control of otherwise deserving people.

Although the benevolence concept has been developed well in Chinese societies, and the values related to it have been integrated into the education of each Chinese and Taiwanese student for a long time, the concept is still not well understood nor often incorporated into quantitative research. Relatedly, Posick et al. (2014) found a similar dynamic at play when integrating “empathyFootnote 1” into the study of criminological discourse. In their exploratory study using data collected from 4882 in-home interviewers of youth nested in 132 middle and high schools across the USA, they found that empathy emerges as an important predictor of perceptions of police effectiveness. In light of the above-mentioned findings, we seek here to empirically examine relationships between benevolent sympathy and attitudes toward police services among college students in Taiwan. We further hypothesized that benevolent sympathy would be highly related to positive attitudes toward the police in Taiwan.

Neighborhood Contexts

This model broadened the scope of the analysis of public satisfaction with police by incorporating the characteristics of the neighborhood in which police operate and students attend college into the analysis. This also introduced new dynamics into the relationship between the public and the police (Ren et al. 2005). Previous studies generally demonstrate that individual neighborhood contexts have a significant impact on residents’ judgment of police (Reisig and Parks 2000; Sampson and Jeglum-Bartusch 1998; Weitzer and Tuch 2006). Overall, the common indicators used to tap into neighborhood environment are such things as fear of crime, level of social integration, sense of collective efficacy among residents, and degree of neighborhood disorder (Cao et al. 1996; Grank & Giacomazzi 2007; Huebner et al. 2004; Lai and Zhao 2010; Reisig & Giacomazzi 1998; Reisig and Parks 2000; Ren et al. 2005).

As for fear of crime, most studies have found that it was negatively related to public satisfaction with police (Cao et al. 1996; Reisig & Giacomazzi 1998; Weitzer and Tuch 2006). It could be argued that when the quality of life in a community is improved and the fear of crime abates, an increase in public confidence in local police is to be expected (Zhao Scheider and Thurman 2002; Zhao et al. 2014). Based on comparable college student samples from both China and the USA, Wu and Sun (2010) found that fear of crime had a significant negative effect on assessments of police demeanor, integrity, and effectiveness in both national settings.

Through a review of literature, the relationship between social integration, collective efficacy, and public satisfaction with the police has also been explored. Similar to past measures of social ties, social integration was assumed to exist when citizens knew their neighbors, spoke to them regularly, and felt that their neighborhood was a “real home” (Gibson et al. 2002; Morenoff Sampson and Raudenbush 2001). In other words, social integration is the initial step in which neighbors get to know one another in a positive way and thereby create mutual trust and respect that leads to the willingness of helping each other in times of need. At the same time, collective efficacy was originally developed by Sampson et al. (1997) to capture the notion that “willingness and intentions to intervene on behalf of the neighborhood would be enhanced under conditions of mutual trust and cohesion” (p. 921); this concept has been used in subsequent studies by numerous other scholars (e.g., Gibson et al. 2002; McGarrell Giacomazzi & Thurman 1997). For example, Cao et al. (1996) found that collective efficacy (also referred to as collective security) was a robust predictor of citizen confidence in the police (see also Huebner et al. 2004). Ren et al. (2005) reasoned that it was likely that confidence in one’s neighbors bred confidence in the police. By extension, it is likely that those who felt abandoned or alienated by their own neighbors might also feel abandoned by or alienated from the police.

Perceptions of neighborhood disorder tended to lessen citizens’ confidence in the police (Cao et al. 1996; Reisig and Parks 2000; Ren et al. 2005). It was plausible that neighborhood disorder, both in their physical and social aspects (e.g., noisy neighbors, loitering by rowdy teenagers, graffiti, deteriorating property), sent a message that law enforcement had lost control over the neighborhoods (Skogan 2009). If this is indeed the case, then neighborhood disorder might serve as a daily cue that the police could not be trusted to provide needed public safety protection.

Demographic Characteristics

Demographic characteristics, such as gender, age, major, household income, year in college, and locality, are the most commonly studied factors relating to confidence in the police (Wu and Sun 2010). Gender has been found by most previous studies as only weakly related to citizen perceptions of the police (e.g., Sampson and Jeglum-Bartusch 1998; Wu et al. 2012). Some research suggests that gender is a significant determinant of citizen confidence in the police (Cao et al. 1996; Correia et al. 1996; Lai & Zhao 2010; Lai et al. 2010; Reisig and Parks 2000). For example, females tend to hold more favorable attitudes toward the police than males. Cao et al. argued that it was reasonable to expect that females thought more positively about the police, inasmuch as women tended to have less antagonistic contacts with the police than did men. Other scholars have reported that there was not a significant relationship between gender and satisfaction with police in the populations they studied (Correia et al. 1996; Ren et al. 2005; Sampson and Jeglum-Bartusch 1998; Sun et al. 2013b; Wu et al. 2009, 2012).

Age was found to have predictive power in most analyses wherein it has been incorporated. In general, research has indicated a positive relationship between age and public satisfaction with the police (Cao et al. 1998; Correia et al. 1996; Lai and Zhao 2010; Ren et al. 2005; Reisig and Parks 2000; Sun et al. 2013a; Wu et al. 2009). Cao et al. (1998) have argued that aging was generally a process that promotes conservatism and more thorough integration into the institutional order. Younger individuals often perceive police as attempting to restrict their independence and ability to explore the environment, and as a result they tend to evaluate the police less favorably than their elders. More recent studies, however, have, in a number of cases, reported that there was no significant association between age and public satisfaction with police (Schafer et al. 2003; Wu and Sun 2010; Wu et al. 2012).

While the effect of educational attainment has been considered in some previous studies (e.g., Lai and Zhao 2010; Lai et al. 2010; Reisig and Parks 2000; Ren et al. 2005), less is known about the effect of the type of education on assessments of the police performance than is the case with other demographic factors (Brown and Benedict 2002). Specifically, different college majors may influence the association between education and perceptions of the police (Wu and Sun 2010). For example, Brown and Benedict (2002) hypothesized that students who majored in liberal arts and social sciences would hold less favorable attitudes toward the police than those majoring in a business-related or technical area of study, partially due to their varied ideologies (liberalism versus conservatism) resulting from different academic training. For example, based on a study of Chinese college students, Wu and Sun (2010) reported that those majoring in liberal arts and the social sciences hold a more critical view of police effectiveness than those who majored in natural sciences, engineering, and business.

There has been a good deal of research supporting the presence of a link between social class and perceptions of police, with people with lower socioeconomic status (SES) more likely holding an unfavorable attitude toward the police than the more affluent (e.g., Lai et al. 2010; Sampson and Jeglum-Bartusch 1998). For example, Wu et al. (2012) found that Chinese citizens self-rating themselves in a higher SES status reported a higher level of trust in the police than citizens rating themselves in a lower status. In contrast, a few studies have found that in some areas, people with higher income actually viewed the police less positively than those with lower income (e.g., Murphy & Worrall 1999; Weitzer and Tuch 2006). Worthy of note, most existing literature has shown either a weak link or no relationship between income and citizen evaluation of police performance (e.g., Brandl et al. 1994; Cao et al. 1996; Ren et al. 2005; Schafer et al. 2003; Sims Hooper & Peterson 2002; Webb & Marshall 1995).

Finally, the effects of residential area have been incorporated into some studies. An urban/rural division is found to be especially important, with urban residence displaying lower levels of trust in police than rural residence (Weitzer and Tuch 2006; Wu et al. 2011). In recent studies conducted in Chinese societies, researchers have found that urban dwellers reported higher levels of trust in police than their rural counterparts (Wu and Sun 2010; Wu et al. 2012). Researchers Wu et al. (2012) argued that rural Chinese are residing in less economically developed circumstances and in areas lagging behind in the modernization and democratization process; likewise, rural police are generally less professionalized, and rural residents tend to have closer interpersonal relationships and stronger adherence to traditional values. In contrast, using data collected from a telephone survey with 2289 respondents nested in five counties and several cities in Taiwan, a recent study showed that residents in a rural county expressed higher levels of satisfaction with the police than their urban counterparts (Huang et al. 2013).

Methods

Data Collection and Sample

Survey data were collected from a convenience sample of undergraduate students attending public and private universities during February and March of 2014; special attention was given to documenting the locations and demographics of the students. Six professors from six universities in five cities and one county gave permission to survey the students in their respective classes during regularly scheduled class meetings. Surveys were distributed to and collected from students through these professors, with the assistance of their research assistants. Students were encouraged to participate in this survey, and likewise encouraged to ask questions if they had any concerns. Students were assured that their participation would be completely voluntary and anonymous, and that non-participation would bring no negative consequence to their grade in the classes. Students agreeing to participate in the survey were informed that they were free to discontinue if they felt that they were unable to complete the questionnaire as they worked their way through it. Students who declined to participate were thanked for their consideration of participation and asked to wait quietly while their fellow classmates completed the questionnaire.

Participants were instructed not to write their names anywhere on the questionnaires, but were told to give their email address to the instructor/teaching assistant on a separate piece of paper if they wished to be informed of the results of the survey. All completed surveys were mailed to the researcher’s university and locked in secured office cabinets. The research team members were the only persons allowed access to the completed questionnaires. A total of 742 students participated in the survey; cases with missing data were dropped from the analysis, resulting in a final sample of 688 respondents. There were virtually no differences between respondents included in and excluded from the sample vis-à-vis background and demographic characteristics, and the exclusion process did not confound results in any significant way.

Dependent Variable

The dependent variable in this analysis is satisfaction with police services experienced in the respondents’ neighborhoods where they attended college. Nine separate items featured respondents’ assessment of the following services provided by the local police in their immediate neighborhood: (1) traffic law enforcement, (2) narcotics and drug enforcement, (3) responding to gang issues, (4) responding to illegal gambling, (5) responding to prostitution, (6) crime prevention efforts on juvenile delinquency, (7) responding to domestic violence, (8) interaction with citizens, and (9) clearance of crime investigations. Respondents were asked to rate police performance on a Likert-type attitude scale ranging from 1 (very unsatisfied) to 5 (very satisfied). The satisfaction with police services index of the nine items was calculated as the sum of scores divided by 9. The Cronbach’s alpha for this scale is 0.90, indicating a high level of congruence among the nine items.

Explanatory Variables

Three groups of independent variables were examined in this study: (1) personal experience and attitudes, (2) neighborhood context, and (3) demographic characteristics. Locality effect was controlled for in the analyses presented here.

Personal Experience and Attitudes

The model consisting of contact experience, victimization experience, and benevolent sympathy captured the respondents’ experiences with crime and police. First of all, three dichotomous variables were created to measure the contact experience with the police; these variables were created by categorizing responses to two distinct questions: (1) “During the past 12 months, have you had any contact with a police officer?” and (2) “Was this contact initiated by the police?” Following Sprott and Doob (2009). these two questions were converted into three dichotomous variables—namely, no contact (none), voluntary contact (i.e., citizen-initiated contact), and involuntary contact (i.e., police-initiated contact) with the police. Involuntary contact experience was treated as a reference group in the analysis. Second, victimization experience was a dichotomous variable, asking respondents if he/she had been a victim of crime at/nearby home or on campus over the past 12 months. Two items summed up and recoded as a dummy variable, with 0 = no victim experience (none) and 1 = victim of at least one time. Finally, a scale of four items was developed to measure benevolent sympathy. On a scale ranging from 1 (strongly disagree) to 5 (strongly agree), respondents were asked if they agree or disagree with the following statements: a police officer’s job is: (1) somewhat difficult, (2) pretty heavy, (3) very stressful, or (4) both arduous and thankless. The benevolent sympathy index of the four items was calculated as the sum of scores divided by 4. Using varimax rotation, it can be documented that all four items load on a single factor with an eigenvalue of 2.89 and Cronbach’s alpha = 0.87.

Neighborhood Context

The second group, respondents’ neighborhood context, includes four variables. First, fear of crime was measured by four items: (1) I am worried that my home would be burglarized when not there, (2) I am worried that I and my family would be the subject of a home invasion, (3) I am afraid of being out alone in my neighborhood at night, and (4) I am afraid of being out alone in my neighborhood during the day. Summed responses on a Likert-type scale ranged from 1 (not safe at all) to 5 (very safe). The fear of crime index of the four items was calculated as the sum of scores divided by 4. The Cronbach’s alpha for the fear of crime index was 0.84, with an eigenvalue of 2.72.

Next, a total of six questions were included to measure social integration and collective efficacy within respondents’ neighborhoods. Two items were designed to measure the key concept of social integration: (1) How many neighbors do you know by name (including nickname)? (2) How often would you say that you converse with your neighbors? Responses were recorded on a Likert-type scale ranging from 1 to 5. The social integration index of the two items was calculated as the sum of scores divided by 2. The next four items used to measure collective efficacy were originally developed by Sampson et al. (1997, p. 921) to assess the likelihood that “willingness and intentions to intervene on behalf of the neighborhood would be enhanced under conditions of mutual trust and cohesions.” These survey items have been used in many other studies since their first use by Sampson et al. (e.g., Gibson et al. 2002; Zhao et al. 2010). These items are as follows: (1) how likely are adults in your neighborhood to take responsibility for notifying the police about illegal activity occurring in this neighborhood; (2) if there is a suspicious person hanging around your block, someone is likely to call the police; (3) when you do a favor for a neighbor, can you generally trust the neighbor to return the favor; (4) if you were in need of help with your car stuck in front of your residence, how much faith do you have that your neighbors would come to your assistance. Responses for each item on a Likert-type scale ranged from 1 (never) to 5 (always). The collective efficacy index of the four items was calculated as the sum of scores divided by 4.

The extant literature suggests that these items form a single factor, reflecting one’s integration in the neighborhood and willingness to intervene in the case of need (Gibson et al. 2002; Taylor 2002; Zhao et al. 2010). However, the analysis done with the sample of Taiwanese college students identified two distinctive and separate factors (please refer to Table 1). The first two items were loaded on a first factor accounting for 18.75 % of the variance, and represented the perceptions of respondents’ social integration (eigenvalue = 1.13, Cronbach’s alpha = 0.77). The second factor extracted explained 42.73 % of variance, and the four items seemed to reflect the respondents’ collective efficacy (eigenvalue = 2.56, Cronbach’s alpha = 0.66).

Table 1 Results of factor analysis on SI/CE
Full size table

Finally, respondents were asked to assess the level of neighborhood disorder obtaining in their area of residence, and included the following items: (1) noisy activities (e.g., binge drinking); (2) groups of teenagers hanging out; (3) promiscuity and sex trade-related activities; (4) fighting/violent activities; (5) entertainment activities (e.g., arcade games, cybercafé, and pool halls, etc.); (6) vehicles parking inappropriately; (7) drunk/homeless people; and (8) graffiti/vandalism. For each type of neighborhood disorder, respondents were asked to rate the problem on a scale ranging from 1 (not serious at all) to 5 (serious problem). The neighborhood disorder index of the eight items was calculated as the sum of scores divided by 8. The factor loadings analysis using a varimax rotation revealed that all eight items loaded on a single factor with an eigenvalue of 4.64 and Cronbach’s alpha = 0.89

Demographic Characteristics

Demographic variables included gender, age, academic major, household income, and year in college. Gender is expressed as a dichotomized variable (0 = male and 1 = female). Age is measured in years. Academic major was coded as a dummy variable, with 1 representing disciplines related to the liberal arts, social sciences, and law and 0 representing disciplines in natural science, engineering, and business. Household income per month consisted of a choice among seven categories ranging from 1 (less than 20,000 NTD, equal to US$660) to 7 (120,000 NTD, equal to US$4000) and above. Year in college was coded as an ordinal variable ranging from 1 (freshman) to 4 (senior). Finally, a series of dummy variables were created to represent the five localities Footnote 2 in this study. Taipei City was treated as the reference group, whereas New Taipei, Tainan, and Kaohsiung Cities and Chiayi County were regressed with other explanatory variables in the final equation.

Possible multicollinearity problems were first checked for by the use of a correlation matrix containing all explanatory variables. The variance inflation factors (VIFs) computed by regressing each independent variable on other variables in the model was also employed as a good indicator of any problems in this regard (Judge et al. 1988; Tabachnick and Fidell 1996). A correlation check indicated that the two variables of age and year in college were highly correlated (r = 0.67), a covariance which is at an acceptable level. In addition, by entering both variables in the same model, multicollinearity examinations resulted which indicated a VIF of 2.0, suggesting that the index was lower than the tolerance statistic value of 4. Since VIF scores were below 4, it can be concluded that multicollinearity is not a problem of any significance in this study.

Results

Descriptive Statistics

The descriptive statistics for this study are reported in Table 2. The mean of the dependent variable, satisfaction with police performance, was 3.25, suggesting that overall residents tended to be satisfied with police services in their immediate neighborhood. In terms of the personal experience and attitudes variables, 53.6 % of respondents did not have contact with the police in the previous year, 22.5 % had voluntary contact, and 23.8 % experienced involuntary contact. Victimization experience showed that about 16.0 % of the respondents were a victim of crime at home or on campus during the 12 months prior to this survey. The mean rating of benevolent sympathy was 3.64, suggesting that most respondents believe that the job of the police is not an easy one.

Table 2 Descriptive statistics for all variables (N = 688)
Full size table

Regarding neighborhood context variables, respondents reported relatively high levels of social integration (M = 2.62, SD = 1.13) and collective efficacy (M = 3.20, SD = 0.79). Not surprisingly, they expressed relatively low levels of fear of crime (M = 2.44, SD =0.82) and relatively little concern for neighborhood disorder (M = 2.30 SD = 0.74).

Approximately 56 % of survey respondents were females and 44 % were males. The mean age of the respondents was 19.9 years, and about 58.3 % of the college student respondents majored in liberal arts, social science, and law majors whereas 41.7 % of the students were majoring in natural science, business, and engineering disciplines. Slightly over half of the respondents (50.2 %) reported that their total household income per month was less than 60,000 NTD (around US$2000). A large proportion of the students were freshmen (48.3 %), followed by juniors (18.9 %), sophomores (17.9 %), seniors (13.5 %), and postgraduates (1.5 %). Finally, 339 students (49.2 %) were from the country’s northern cities (i.e., Taipei and New Taipei cities), and the remaining 349 students (50.8 %) from the southern county and cities origins (i.e., Chiayi county and Tainan and Kaohsiung cities).

Multivariate Analysis

In the multivariate analysis, the variables derived from the personal experience and attitudes, neighborhood context, and demographic characteristics models were regressed on the scale of satisfaction with police. The results are presented in Table 3. First, among personal experience and attitudes model indicators, two variables—voluntary contact experience and benevolent sympathy—were significant predictors of satisfaction with police performance. While those who had voluntary contact experiences with police reported lower levels of satisfaction with police performance in their neighborhoods, respondents who had higher degrees of benevolent sympathy held favorable attitudes toward police work. Second, in terms of the neighborhood context model indicators, three out of four variables were significantly associated with the dependent variable. Those who perceived higher levels of fear of crime and greater neighborhood disorder in their neighborhoods reported lower levels of satisfaction with police. Moreover, collective efficacy significantly increased the levels of satisfaction with police work in his/her neighborhoods. Five variables derived from the demographic characteristics model produced insignificant association with satisfaction with police performance. With respect to locality, those who resided in New Taipei City and Chiayi County reported significantly lower levels of satisfaction with police performance compared to their counterparts who resided in Taipei City.Footnote 3 Among those significant variables, benevolent sympathy is the most robust predictor, followed by neighborhood disorder, New Taipei City, and collective efficacy. Overall, the explanatory variables were able to explain 24.3 % of the variance in the model.

Table 3 Multiple regression summary (N = 688)
Full size table

Discussion and Conclusion

Discussion

During the late 1980s, Taiwan entered a period of ongoing democratic development after the lifting of 26 years of martial law in 1987. Voices from diverse levels of political, racial, gender, age, and social life organized around their issues and mobilized for the exercise of influence in the political process. Social movements of various types arose, pushing Taiwanese society toward a pluralistic democratic society in the process. Democratic reforms and political liberalization followed apace (Gingerich et al. 2011). As the most visible aspect of government, the social role and function of police gradually (but steadily) moved from what Cao and Dai (2006, p. 72) describe as a “high policing” model—a model of policing that is repressive and seeks to eliminate “potential threats in a systematic attempt to preserve the distribution of power in a given society”—to a “low policing” model that champions the rule of law, individual liberties, and systematic attention to public preferences (see also, Brodeur 1983).

To be certain, the three waves of Taiwanese college student movements set within the context of other political, economic, and international influences contributed importantly to the democracy-promoting transformations of the police in the country. Somewhat surprisingly, the topic of college student attitudes toward the police has been largely overlooked in Taiwan. To fill the gap, using data collected in February and March of 2014 gathered from six universities prior to the Sunflower Student Movement, 688 college student evaluations of local police services were reviewed for insight into this process of societal and police democratization. The Chinese culture value of benevolent sympathy and the explanatory variables derived from Western societies were included in the analyses.

First, the Chinese cultural value, benevolent sympathy, is a significant and robust predictor of satisfaction with police services among college students. Compared to the Western societies that emphasize individualism, traditionally, the Chinese people view collectivism as priority. Chinese are taught to observe customs, follow rules, and be guided by norms. For the sake of harmony, Chinese should be guided by the attitudes of tolerance, forgiveness, and compassion toward others. In a likely manner, Chinese should be respectful of authority and follow the orders of duly authorized figures of social order such as teachers and the police. If possible, Chinese should avoid actions that conflict with authorities such as police.

Values associated with democracy and individualism derived from Western societies are overwhelmingly exerting an influence on Taiwanese, and over the past several decades, the core of Chinese cultural values have gradually succumbed to the societal changes accompanying democratization. One area where this cultural change is perhaps most dramatic is in that of respect for police; such respect is not in ample supply in contemporary Taiwanese society.

Taiwanese police are often referred to as “citizens’ baby sitters,” suggesting that police should play more of a role of taking care of citizens than of fighting criminals. It follows that the police should be both respectful and court public favor. Somewhat ironically, when Taiwan moved successfully from authoritarianism to democracy after the end of decades of martial law, the police ended up receiving lower marks on satisfaction and confidence from the public (Lai et al. 2010; Cao and Dai 2006; Wu et al. 2012). It is likely that the public is not fully aware of the dramatic improvements made in police professionalism and training and service provision. Fortunately, this study found that young people with benevolent sympathy cultural values are indeed satisfied with the police services being provided in the communities where they attend college. It appears that some young people still harbor the Chinese core values and are “stepping into police’s shoes” to experience the police as a key institution in Taiwan society. It also implies that some Taiwanese sufficiently understand that policing is difficult, demanding, and an important job which is not fully appreciated; such persons tend to express unconditional support for the police. Taiwanese police administrators and officers should be gratified.

In addition, consistent with findings reported in Western societies, both contact experience and neighborhood contexts significantly impact satisfaction with police services (Cao et al. 1996; Huebner et al. 2004; Reisig & Giacomazzi 1998; Reisig and Parks 2000; Ren et al. 2005; Rosenbaum et al. 2005; Weitzer and Tuch 2006). Unexpectedly, those Taiwanese college students who had voluntary contact experiences with police expressed lower levels of satisfaction with police services. A majority of research examining the impact of citizen-initiated interactions found a positive public rating effect on police performance (Rosenbaum et al. 2005). We speculate that Taiwanese college students are not satisfied with police services when they call for help from police. This may be because, compared to other people, college students have a substantial academic background and consequently are more “demanding” and “critical” in their expectations. At the same time, those young people who perceived relatively high levels of fear of crime and neighborhood disorders reported distinctly unfavorable attitudes toward the police. As Skogan (2009, p. 301) noted, citizens always “hold police accountable for local crime, disorder, and fear.” Hence, when residents’ perceptions of disorder and fear are high, the police will always receive relatively low levels of ratings from citizens. By contrast, those college students who perceived higher levels of collective efficacy in their neighborhood also expressed greater police satisfaction (Huebner et al. 2004).

Finally, while the locality indicator is treated as a control variable, results from ordinary least squares (OLS) were consistent with previous studies indicating that citizens who lived in rural areas tend to report lower levels of satisfaction with police work than their counterparts residing in a downtown area (e.g., Cao et al. 1998; Frank et al., 1996; Reisig and Parks 2000). As Wu and Sun (2010) noted, police in a downtown area tend to have higher levels of formal education and a more extensive training background than those in a rural area. Sun et al. (2013b). argued that lacking the adequate resources and equipment, appropriate training and political/administrative scrutiny, the rural police have failed to make expected progress in effectively dealing with crime and social disorder; nor have adequate accountability arrangements been developed to deal with a rapidly changing environment, and all these factors impact detrimentally on villagers’ views of the police. Applying these argument to this study, Chiayi County is an agricultural area and the participants’ university is located in a village where most citizens plant and harvest pineapples. It is not surprising to see relatively negative evaluations of police performance in that area. Moreover, although New Taipei City is a metropolitan municipality in the north of Taiwan, it covers both rural and urban areas. In other words, the quality of community life varies across the districts of New Taipei city, and expectedly, perception of the quality of police performance likely varies accordingly across the city’s broad jurisdiction. In this regard, the respondents’ university is located in a traditional village where most citizens are elders living in older houses and modest apartments there. Most students are not from the New Taipei City area and would have a higher level of sensitivity to differential treatment by the local police when compared to that of the police in Taipei City from whence they are most likely to come. Scholars have also posited that rural villagers’ dissatisfaction with quality of life led to lower ratings of local officials including police officers (Sun et al. 2012). More empirical data are needed to either support or refute these speculations, but they are likely good speculations based on the personal experience of the authors.

Policy Implications

The findings of the present study carry some implications for policy makers. First of all, the results suggest that cultivation of benevolent sympathy might have a positive influence on Taiwanese college students’ attitudes toward the police and other criminal justice agencies. Teaching about these Chinese traditional values might help college students cultivate positive attitudes not only toward police, but also toward other people and moderate the adoption of individualistic values via the mass media and global communication channels. Educators at the college level may wish to review the curriculum of their academic programs and include some attention to the purposive cultivation of good virtues in at least some of their courses. Nevertheless, it is never too late to seek effective ways to cultivate such venerable Chinese virtues. As Confucius has instructed, “What you do not wish upon yourself extend not to others.” This virtually universal “golden rule” common to most cultural systems and religions should be the basic principle in treating one’s fellow human beings and, accordingly, the police when they are attempting to do their difficult job in good faith. If education policy makers had appreciated the importance of benevolent sympathy earlier, perhaps the troublesome police-college student confrontations over the course of the past few decades could have been reduced. We believe that cultivating good traditional Chinese virtues, including particularly benevolent sympathy, among college students can serve as a starting point to improve the students’ perceptions of the police.

Second, the OLS results indicated that voluntary contacts with police significantly erode rather than build satisfaction with police performance among Taiwanese college students. Given these findings, greater efforts should be made to generate more positive experiences with the police for college students. For example, the police administrators should arrange targeted training programs designed to improve communication skills and to promote professional knowledge regarding the handling of confrontations with college students. At the same time, police supervisors should monitor frontline officers’ performance closely during any type of encounters with college students and set up clear policies and guidelines to ensure that polite and professional conduct is displayed and procedural justice is rendered in handing college students’ requests for assistance as well as traffic-related accidents (Sun and Wu 2010).

Third, efforts should be made to generate more positive experiences for the college students such as enhancing neighborhood-specific service courses for sworn police officers, especially those in rural areas. Police managers should encourage frontline officers to engage actively in non-enforcement related encounters, such as initiating causal conversations with citizens and offering assistance to community residents. Also, Taiwanese police administrators should introduce some evidence-based useful programs and practices developed to foster citizens’ understating of the police such as citizen police academy, citizen ride-along programs, school resource officers, and volunteers for police assistance, programs which have been widely used in the USA. These activities can be designed to specifically target college students and can be actively promoted on police departments’ websites and in public media (Wu and Sun 2010).

Finally, the police agencies should pay more attention to their behaviors and services in neighborhoods where disorder crimes and higher levels of fear exist; disordered neighborhoods and crime hot spots deserve particular police attention. As for solving neighborhood disorder problems, problem-oriented policing encourages officers to work with communities in reducing disorder and improving neighborhood conditions. For example, assisting residents and college students they served to organize neighborhood watches and/or citizen patrols by providing equipment such as batons, whistles, vests, pepper sprays, and patrolling vehicles, etc. (Lai 2013). In terms of fighting crime hot spots, deploying more patrolling officers on streets at night, installing CCTV cameras for monitoring hot spots areas, and taking proactive steps to reduce crimes regularly are good practical activities to improve better police-community relations. At the same time, local police agencies should make an effort to encourage residents on an ongoing basis to engage in collaborative partnerships to both shape positive attitudes toward the police and promote democratic values. Those neighbors who perceived higher levels of safety and collective efficacy also would tend to derive satisfaction from contributing to public safety and police performance, leading in turn to higher levels of satisfaction with and trust in the police (Huebner et al. 2004; Ren et al. 2005).

Limitations

Certain limitations of this study must be acknowledged. First, the findings from this study have limited generalizability because the data were collected by a non-random sampling method. While the gender, age, and academic major of those 688 respondents have been examined to compare to the population of all college students in Taiwan, results indicated that gender is significantly over represented based on results of chi-square statistics,Footnote 4 suggesting the opinions of female students are over represented. In addition, the universities selected from five cities and one county reflect substantial diversity (e.g., urban versus rural, public versus private, and high-ranking versus middle ranking), but they do not fully reflect the diverse nature of Taiwan’s colleges and universities.

Second, this study made use of the Chinese cultural value of benevolent sympathy to assess attitudes toward the police among college students. Although the measure developed for the study produced a positive association with police performance, it may not cover the components of the actual “benevolence” value. For example, in Hampton and Xiao (2009). benevolence is treated as a basic principle to be used to structure communications and frame interactions between people. Benevolence emphasizes harmonious interpersonal relationships and the component factors should include tolerance of others, seeking to work in harmony with others, the ongoing display of humility, loyalty to superiors, and kindness (understood as forgiveness and compassion). In this study, only one major sub-component, namely benevolent sympathy, has been explored. Consequently, the results reported here must be interpreted with caution.

Third, the possibility does exist that some students may be referencing the performance of the police in their home area rather than those in the area where they are attending school, despite instructions to the contrary. Future studies done on attitudes toward police among college students should control for the respondents’ length of residing on the college or university campus in question.

Finally, while the theoretical models and correlates are able to explain a good deal of citizens’ attitudes toward the police (more than 25 % of variance in the West (e.g., Cao et al. 1996; Lai and Zhao 2010; Ren et al. 2005), only 16 % of the variance in PATP is explained among Taiwanese students if the benevolent sympathy indicator is not included in the analysis. This finding suggests strongly that distinctive cultural elements should have been considered and introduced when applying models and factors developed in the West to non-Western nations and regions. In the meantime, the author also acknowledges that OLS regression produced biased results with regard to neighborhood effects because they cannot effectively control for endogenous membership problems and omitted variables bias. Instrumental variables and fixed effects models are sometimes recommended in place of OLS regression. Future investigations should keep these several limitations in mind.