Explaining confidence in the police within transitional Hong Kong: the influence of postmaterial values

ABSTRACT The Hong Kong Police Force (HKPF) has a tradition for being among the most respected national police organisations globally. More recent political events in Hong Kong have however placed the HKPF in highly visible public order situations. A wide array of political factions have emerged as a consequence of the handover of control in Hong Kong, many of which have involved young people protesting the curtailment of democratic rights. In this paper, we utilise Wave 7 of the World Values Survey (n = 2075). We utilise Inglehart’s (1971. The silent revolution in Europe: intergenerational change in post-industrial societies. American political science review, 65 (4), 991–1017) concept of postmaterialism to examine how far young people’s more negative attitudes to the police can be underpinned by cohort-based (or generational) explanations. Our findings identify associations between postmaterialist values and lower confidence in the police. Those that favoured stricter immigration controls, experienced higher fear of crime, perceived drug selling in their area, and consumed certain types of media (notably television and new media) were also less likely to have confidence in the HKPF.


Introduction
Studies of public confidence in the police remain heavily focused on Anglophone contexts, especially from the USA and UK. More recently, scholars have raised questions regarding the generalisability of certain factors in accounting for public confidence in the police (Tankebe 2013, Cao and Graham 2019, Jackson and Bradford 2019, Liu 2019. Important cultural and historical differences in policing exist across different nations, variations in levels of crime, as well as significant political regime transitions affecting perceptions of State institutions, including the police.
The Hong Kong case is an important one given the legacy of colonialism and historically high levels of confidence in the police (Sinclair and Ng 1997). Hong Kong also has historically low crime levels (Broadhurst et al. 2008), despite possessing some of the largest wealth gaps in the developed world Zhang 2005, Chim and). More recently, aggregate level reports of public opinion in Hong Kong have shown a marked shift in distrust towards the police (Hong Kong Public Opinion Research Institute (HKPORI) 2020). This distrust has been identified in the wake of several high-profile conflicts between citizens and the police arising from public demonstrations as a result of concerns about the transition of power between Britain and China (Cheng 2016, Purbrick 2019, Adorjan et al. 2021, Ho 2021. These public demonstrations have also included a rise in the use of paramilitary-style policing to counter protesters (Ho 2020a, 2020b, Stott et al. 2021.
Part of the interpretation for the reduction in aggregate-level confidence in the Hong Kong police force involves age differencesgenerally youth have less favourable attitudes to the police compared to adults. Although this trend is well-known in studies of public attitudes towards the police (See Stewart et al. 2014), the Hong Kong case is more complex. Unlike many of their parents who grew up in Mainland China, most youth are born and raised exclusively in Hong Kong, thus forming a more sensitised political consciousness to values of democracy and 'independence' of Hong Kong (CCPOS 2020, Lee et al. 2020). This accords with a body of research within political science concerning the transition between materialist and postmaterialist values. Whilst materialist values (more likely associated with those born prior to 1970) are associated more with economic security, work and social stability, post-materialists (typically born post 1970s) prioritise political rights, values of democracy and freedom (See Ingelhart 1971, 1981, Wong and Wan 2009). These trends towards postmaterialist values have been explained by rising living standards in many Western nations, as well as remarkable periods of social stability contrasted with the early-mid twentieth century (Ingelhart 1971(Ingelhart , 1981. To date, examinations of age differences in understanding patterns of public trust and confidence in the police have yet to adopt cohort or generation-based explanations. The Hong Kong case is a fertile one with which to interrogate why youth appear to possess lower confidence in the Hong Kong police force (HKPF) compared to older generations. In this paper, we assess the characteristics of those that share lower levels of confidence in the HKPF (HKPOPI 2019). In examining this question, we utilise the recent Wave 7 of the World Values Survey (WVS) (2017-2020), where we restrict our analysis to Hong Kong respondents (n = 2075). The WVS is an important cross-national resource for assessing not only public confidence in the police, but also other key explanatory variables including experiences of crime, security and victimisation, media consumption practices, postmaterial values, and a range of key demographics. Testing the concept of postmaterialism (Ingelhart 1971(Ingelhart , 1981 we find that the effect of age and education are diminished, pointing to a cohort (or generation) of Hong Kongers that have more sceptical views of the police as a by-product of generational political values. Interestingly, we also show that lower confidence in the police was connected with restrictive views on immigration. This likely reflects anxieties about the increase in Mainland Chinese people entering Hong Kong, as well as wider contempt towards Mainland China which has emerged from recent political protests Chou 2018, Adorjan et al. 2021).

The context of policing in Hong Kong
Hong Kong is a former British Colony that is now a Special Administrative Region (SAR) of the People's Republic of China. Its handover in 1997 led to what is referred to as a 'one country and two systems' approach. A mini constitution of the SAR is underpinned by the Basic Law, which, in theory, protects rights such as freedoms of expression and assembly. This includes the People's Republic of China's promise to maintain the capitalist system after the sovereignty retrocession, leaving the prevailing British colonial system of governance in Hong Kong unchanged.
A general perception towards protest policing in the British colonies is that the policing authorities would liberally use force to control disturbances (Sinclair 2007). However, the strategy was never as overtly coercive as widely expected in Hong Kong. Since the 1970s, structural reforms have been introduced to professionalise the operations of the Hong Kong Police Force (HKPF), and also market it as a 'community partner' for crime fighting (Lau 2004, Lo and Cheuk 2004, Ho 2020a. Both the legitimacy of the colonial government and public trust in the HKPF increased in line with growing economic prosperity since the 1980s. In the first two decades of the Hong Kong government, public satisfaction with police performance has been consistently over 50% from 1997, reaching 80% in 2008(Hong Kong University Public Opinion Program 2019, Adorjan et al. 2021). Compared to events post-2014, there were few large scale and sustained public disorderly incidents testing the capacity of the police mobilisation strategy in Hong Kong. This included the rally in support of the student-led protests in Tiananmen Square in 1989, and the protest against opposition of Basic Law Article 23 legislation in 2003. Several large-scale protests emerged after the 2010s, including the Anti-National and Moral Education, Conservation of Historical Heritages, Anti-high Speed Railway construction. In each of the situations these did not create any significant public order disturbances and were largely peaceful (See Ho 2021).
Everyday policing in Hong Kong has been conducted without recourse to most of the extreme public order measures noted above. Citizens also regarded social stability and wealth as the foundation of safety and security (Lau and Kuan 1990). The low crime rate in Hong Kong, despite high social inequality and fears of restriction on freedom, likely contributed to public trust in the police (Broadhurst et al. 2013). The more recent shift away from such overwhelming public confidence in the HKPF was acknowledged in research examining youth attitudes to the police conducted between 2011 and 2013, prior to the larger episodes of protest in Hong Kong (Adorjan and Lee 2017). This study identified a mixed picture of support for the police, combined with negative sentiments, including those relating to more regular street policing as well as wider concerns about the police's allegiance to the government and violation of essential freedoms in Hong Kong. Many of these previous appeasements regarding the relative stability of Hong Kong society and trust in the police, have however, come under challenge following the transfer of power from Britain to China in 1997. Of particular note, is the change of Beijing's policy which has been accused of eroding the principle of Hong Kong's governance under the 'One Country Two Systems' constitutional framework (effectively granting Hong Kong relative independence, yet under ultimate control from China).
The period underpinning the transfer of power between Britain and Hong Kong has been characterised by a series of high-profile events, both in response to increasing movement of Chinese nationals into Hong Kong, together with the wider pro-democracy events established. From December 2014 to April 2015, the 'shopping protest' was launched in Mongkok (Hong Kong) involving 'antimainland Chinese' protesters against the presence of traders in Yuen Long, Sheung Shui and Tuen Mun (areas in Hong Kong bordering mainland China). The most well-known of public order incidents was the Umbrella Movement (UM) in 2014 (Cheng 2016, Purbrick 2019, Stott et al. 2021. The UM was so called because of demonstrators' use of their umbrellas to defend against the pepper spray that the police wielded as a crowd dispersal tactic. The movement was sparked by the order of the Beijing government to screen candidates standing for the 2017 Hong Kong elections. After 79 days of continuous but otherwise peaceful occupation, the UM was dispersed by police by the court injunction applied by a group of pro-government lawyers. During this time, the police took proactive action on 28 September 2014 to use tear gas to disperse the crowd. No further repressive action was launched to confront the protestor occupiers. The protest ended with failure and the chief organisers were arrested and sentenced to prison for 8-16 months (Ho 2020a).
While political protests are not new in Hong Kong, the current situation has involved an intensity, higher frequency, and longer duration of clashes between protesters and public authorities. This has led to questions about the level of trust that can be placed in the Hong Kong government, as well as the police (See Adorjan et al. 2021). Most recently, the Anti-Extradition Bill Movement (AEBM) resulted in large-scale public order incidents across Hong Kong. The AEBM was sparked by a Uturn on the original law passed prior to the handover of British rule of Hong Kong to China in 1997 which allowed criminals to be tried and imprisoned solely in the Hong Kong territory. In 2019, the Beijing Government passed the Fugitive Offenders and Mutual Legal Assistance in Criminal Matters Legislation (Amendment) Bill allowing the extradition of offenders from Hong Kong to be tried and imprisoned in Mainland China (for 37 different offences). The bill led to mass protests across Hong Kong, including major conflict with the police. The bill was withdrawn in October 2019.
Central to this paper, we assess the extent to which this turbulent context reframes how Hong Kong residents formulate confidence in the police as a response to alleged collusion with the interests of the Hong Kong government, and thus its support of the Beijing regime. Although our data allows us to examine the timing of the Umbrella Movement protests, the AEBM protests unfortunately occured after Wave 7 of the World Values Survey was collected. Next, we consider how standard explanations of confidence in the police can be amended to better fit this new civic context in Hong Kong.

Antecedents of confidence in the police
A plethora of research offers insights into the key predictors of lower confidence in the police, especially in relation to demographic and neighbourhood level factors. Men have been found to possess lower levels of confidence in the police compared to women (Cao et al. 2012, Han et al. 2017). Men's contact with the police is generally higher courtesy of greater involvement in crime compared to women. Younger people also tend to have lower levels of confidence in the police , although this is contested elsewhere (Kwak et al. 2012). There are questions surrounding why young people may have lower confidence in the police. This may involve a combination of life course transitions inducing a shaping of more cynical attitudes during adolescence (e.g. Sindall et al. 2017, Nivette et al. 2020, or because of increased interaction with the police (Dirikx et al. 2012, Ilan 2018.
There are mixed results concerning the role played by education in explaining confidence in the police. Some studies indicate that higher educated young people have lower levels of trust in the police (e.g. Kwak et al. 2012, Sun et al. 2012, Zhang et al. 2021, whilst others find the opposite (e.g. Lee et al. 2019). It is possible that higher levels of education may operate differently in explaining confidence in the police, depending on assumptions regarding the police's display of legitimacy. In Asian contexts including Hong Kong, student movements have commonly formed part of protests against authoritarian regimes and support of more democratic reform (See Weiss and Aspinall 2012). The role of students in organising protests has been extensive in Hong Kong, especially after the Umbrella movement (e.g. Lee et al. 2017, 2020, Purbrick 2019. It is therefore possible that as part of the alleged connection between the Hong Kong government and police, that younger, university educated respondents will display more critical views of the Hong Kong police. Living in areas of concentrated disadvantage can also reduce confidence in the police (Sampson andBartusch 1998, Wu et al. 2009), partly out of dissatisfaction with police service response, legal cynicism and crime levels. Sindall et al. (2012) found that although neighbourhood-level factors were not responsible for police confidence levels among British residents, perceptions of crime and disorder and rates of property crime predicted confidence in the police. Levels of public confidence in the police may also be underpinned by views surrounding the stability of neighbourhoods and community cohesion, rather than by crime and victimisation risks per se Bradford 2009). Perceived safety is also a predictor of confidence in the police Hou 2001, Sun et al. 2012), and may be impacted by factors associated with personal vulnerability (Jackson 2009), or by perceived risks of crime in areas of residence ).

Application to the Hong Kong context
With transitional territories such as Hong Kong, during which the police's recent high-profile role in restricting protest movements alongside deployment of paramilitary-style policing, it is prudent to consider how current social and political tensions inform peoples' orientations towards the police. The extent to which the police are perceived to be protecting social order or quashing a legitimate protest movement is a key consideration, and here we find a pivotal role played by the media. Indeed, 'news coverage is fundamental to a protest's viability' (Kilgo and Harlow 2019, p. 508). In relation to attitudes to the police, different types of media can play a role in shaping these views. Greater television watching has been associated with higher confidence in the police (Callanan andRosenberger 2011, Graziano and, although use of online news and social media reduces confidence in the police (Intravia et al 2018, Gauthier and Graziano 2018). 1 It could be that these differences reflect the type of reporting, with arguably more positive representations of the police and protests in television news than in online news or social media (See Lee et al. 2017).
The wider politics of Hong Kong complicate how we understand the role of certain values in shaping attitudes to the police. One such example is anti-immigration attitudes which likely plays an indirect role in shaping attitudes to the police. This may be as a result of persons holding more supportive attitudes to authority also being more likely to hold anti-immigration views. Possessing anti-immigration views may correspond with right wing authoritarian attitudes as recognised in recent political campaigns in Europe (de Zavala et al. 2017). Anti-immigration attitudes may also be conditioned by the context of frustrations with the State's performance (Knigge 1998, Rooduijn et al. 2016, providing opportunities for protest parties to emerge and thus further fuelling anti-immigration views. Recent research by Dzordzormenyoh and Perkins (2021) finds that fear of immigrants increased support for police stops of immigrant groups within the U.S. In the context of Hong Kong, we recognise that the politics of immigration are however quite different to those witnessed in North America and Western Europewithin the latter regions more often framed through the othering of minority groups, as opposed to a situation in which a minority territory like Hong Kong fears its larger, more politically powerful neighbours. Despite the residence of Westerners in Hong Kong from the early stages of the British colonial rule during the mid-late 1800s, the more recent politics of immigration, especially since the 1990s, has been about the intensification of anti-Mainland (China) sentiments. As Lee and Chou (2018) argue, anti-immigration sentiment towards persons arriving in Hong Kong from Mainland China has been more prevalent among those Hong Kong residents with a 'pro-Hong Konger identity'.
As outlined previously, measures of age and education have widely been used as demographic variables in testing attitudes to the police. However, it is possible that there is more nuance to these factors, in particular, recognising the possibility of cohort or generational differences underpinning demographic differences. According to Inglehart (1971) materialist societies tend to involve a greater emphasis on prioritising economic values, and favouring the stability and security of society via law and order. Following periods of comparable affluence in Western nations including the UK and USA, postmaterialist values have grown. Compared to materialism, postmaterialism has a greater concern with issues such as rights, personal improvement, democracy, freedom of expression and environmentalism (Inglehart 1981, Inglehart andAbramson 1994). Inglehart (1981) argues that these differences between materialist and postmaterialist values are generation, rather than age specific. Subsequent research has demonstrated the extent to which a postmaterialist generation varies across different national conditions and political culture (Veugelers 2000, Grasso et al. 2019, as well as inducing more authoritarian views among youth compared to older generations (Jin and Zhou 2021). Empirical tests of the postmaterialist concept have mixed results in the context of Hong Kong society (e.g. Wan 2009, Lee 2018). Ingelhart's (1981) original conception focused on postmaterialist values emerging out of those born during the 1970s. Studies of Hong Kong have identified that postmaterialism was observable slightly later for those born during the 1980s (Lee 2018). Part of the explanation for why a stronger postmaterialist orientation might be likely among younger people born during the 1980s and later is that by this time in history, Hong Kong had already established considerable economic affluence and stability. In addition, the majority of younger people were born in Hong Kong, compared to their parents that were more likely to have been born in Mainland China (Sing 2010, Wong and Tang 2017, Ku 2019, Tang and Cheng 2021. This context may be responsible for establishing a stronger Hong Kong identity, especially in cultivating pro-democracy views (See Wan 2009, Lee 2018). Therefore, although we consider age and higher education important in assessing attitudes to the police, it may well be much broader social and political values shared by the postmaterial generation which are responsible for explaining more distrust towards the police, especially when located in the current politics of democracy in Hong Kong.

Study details
We develop the following hypotheses regarding predictors of lower confidence in the police in Hong Kong. Our analysis is designed to illuminate (a) the main demographic differences underpinning lower confidence in the police, (b) the extent to which perceptions of crime affect attitudes, (c) whether anti-immigration sentiment and type of media consumed reduce confidence in the police, (d) the extent to which groups exhibiting lower confidence can be understood as sharing certain political values associated with postmaterialism.
Starting with general demographic characteristics underpinning lower confidence in the police, we expect men to be more critical of the police than females (H1), and we further hypothesise that the same holds for younger (H2) and more educated (H3) residents. This follows research which has found at the aggregate level, more critical views of the police among the young (Center for Communication and Public Opinion Survey, CUHK 2020), with many of the 'pro-democracy' protests also popularised in higher education establishments in Hong Kong (Sing 2010, Ku 2019, Tang and Cheng 2021. Although Hong Kong has some of the lowest crime rates in any major developed city, following past research, we test whether perceptions of crime at a neighbourhood level and fear of crime can influence attitudes to the police (H4). The direction of these effects can vary, from more critical views being exhibited via legal cynicism and social disadvantage in areas where crime is high (Sampson andBartusch 1998, McCarthy et al. 2020), or otherwise to increase support for the police precisely because of higher perceptions of crime (Sindall et al. 2012). In Hong Kong, patterns of social disadvantage are not as spatially marked as in the USA or other similar locations, therefore we are uncertain on the influence of crime in shaping lower confidence towards the police, although fear of crime is more likely to be associated with these outcomes.
Research evidence linking anti-immigration attitudes to the police is rare, despite indirect suggestions that it can influence authoritarian attitudes or connect to more critical views of the State (e.g. Rooduijn et al. 2016). We expect that anti-immigration views in Hong Kong will be bound up with more critical views towards the police (H5), noting that the postcolonial trajectory in Hong Kong (See Adorjan et al. 2021) and successive protests are strongly associated with condemnation of closer Beijing government involvement (Lee and Chou 2018). We also hypothesise that lower confidence in the police is associated with consumption of different media, with online news and social media highlighted in particular Graziano 2018, Intravia et al. 2018). We expect to observe this pattern in our models (H6). Finally, noting our earlier age hypothesis (H2), we investigate whether demographic differences can be trumped by the possession of postmaterialist values. This reflects more the influence of generations and political cohorts, which in Hong Kong are important features to examine. We expect that those that adopt more postmaterialist values will have lower confidence in the police, as well as eroding the earlier influence of age and educational characteristics (H7).

Analytical strategy
To test the list of hypotheses, we develop and estimate a series of multivariate models. In doing so, we utilise the recently released Wave 7 of the World Values Survey (WVS). The WVS is a nationally representative sample of 79 counties with over 127,000 respondents. The survey was established in 1981, and is repeated approximately every five years. All major continents are represented in the survey which is supported by a number of collaborative partners. 2 A range of questions are covered in the WVS, including moral and political values, experiences and perceptions of crime, as well as confidence in public authorities, which makes it particularly suitable for the current study. The WVS is based on representative sampling, either as full probability or a combination with stratified sampling techniques.
The Hong Kong wave of the WVS was conducted across 18 districts between July and November 2018 (October was the peak period of data collection). The majority of surveys were conducted via Paper and Pencil Interviewing (57.4%) compared to Computer-Assisted Web Interviewing (42.6%). For further methodological details of the WVS, see Haerpfer et al. (2020), World Values Survey (2020).

Data and measures
For this paper, we restrict the analysis to data from Hong Kong (n = 2075). Hong Kong is just above the global average (2.35) in terms of confidence in the police included in the European Values Survey and World Values surveys, with a mean score of 2.24 out of a maximum of 4 (i.e. 4 denotes lowest confidence). This ranks slightly lower than neighbouring nations including Taiwan (Table 1).
In the most recent wave of the WVS collected in 2018, overall support dropped to 67%. This is a marked reduction given the short timeframe between the previous wave of the WVS in 2014a period which was also associated with major police disruptions of groups within the context of the 'Umbrella movement'. Historically, differences in levels of support across age bands have been minimal, although with slightly higher levels of support among those over the age of 50. The overall drop in confidence has been observed across all age bands, but most evident among younger populations.

Dependent variables
The dependent variable of confidence in the police is measured with the following survey question: 'how much confidence do you have in the police'; 1 = a great deal, 2 = quite a lot, 3 = not very much, 4 = none at all. This was recoded into a binary scale as 1 = higher confidence in the police, 2 = lower confidence in the police.

Independent variables
Crime variables: These consisted of perceptions of violence in neighbourhood, perceptions of drug selling in neighbourhood, perceptions of robbery in neighbourhoodall recoded into 0 = quite frequently/very frequently, 1 = not at all/not frequent. Fear of crime coded as 0 = feel not at all/not very secure in neighbourhood, 1 = feel secure/very secure in neighbourhood.
Information sources: This included measures of consumption of newspaper news, television news, radio news, news from talking to friends. These were coded as: 1 = Never, 2 = Less than monthly, 3 = Monthly, 4 = Weekly, 5 = Daily. The creation of the Newmedia variable involved combining mobile, email, internet and social media news. The scale ranged from 4 (no use of new media) to 20 (very high use of new media). Immigration: This was coded into three dummy variables using the same reference category but with different contrast categories. Dummy 1: open borders (0) vs let people in as long as there are jobs (1).

Dummy 2: open borders (0) vs place strict limits (1), Dummy 3: open borders (0) vs prohibit all (1).
Post-Materialism: We used the derived 12-item postmaterialism scale. This is measured by asking respondents to choose the two most important issues from three batteries of questions. This is a widely used measure of postmaterialism across several studies (e.g. Inglehart and Abramson 1999, Wong and Wan 2009, Henn et al. 2018. The three batteries are: Battery 1 1. A high level of economic growth 2. Making sure this country has strong defense forces 3. Seeing that people have more say about how things are done at their jobs and in their communities 4. Trying to make our cities and countryside more beautiful Key demographic variables were also included in the models consisting of age (Age, 1 = 15-34 years, 2 = 35-54 years, 3 = 55 years and over), education (4-point categorical scale, ranging from to 0 = no education/primary-level educated, 1 = secondary-level, 2 = post-secondary, 4 = tertiary/ university educated), sex (0 = male, 1 = female).

Model design
In addition to the demographic variables of age, sex, and highest education level which test for hypotheses 1 through to 3 respectively, we also include four measures of crime related variables, that is, fear of crime, and three perception of crime variables (robbery, drugs, and violence) to test hypothesis 4; a variable that captures attitudes towards immigration to test hypothesis 5; a series of variables that capture media usage across various types of information channels to test hypothesis 6; and finally a measure of postmaterialism to test hypothesis 7.
Given the dependent variable is a binary outcome, we fit a series of binomial logistic regression models estimated in R (version 4.0.5) using the glm function. In order to provide a clearer assessment of the relative impact of each type of explanation for confidence in the police, we run five separate regression models. In the first stage, we examine demographic factors comprising age, sex, education (model 1), before assessing whether experiences of crime influence confidence in the police (model 2). In models 3 and 4 we add attitudes towards immigration, as well as sources of media consumption. Finally, in model 5, we include post-materialism as a predictor.

Results
Consistent with H1, model 1 (Table 2) shows that age (notably younger age groups), education (higher levels) and sex (females) are significant predictors of low confidence in the police. When incorporating variables on perceptions of crime, we see that the youngest age band (15-34) remains significant, whilst the older age groups do not, along with sex. Education at higher levels also continues to predict lower confidence in the police. In support of H2, Model 2 also indicates that higher perceived fear of crime and drugs in the local area are associated with lower confidence in the police.   Table 3 presents model 3 and 4. These incorporate the addition to attitudes to immigration and information sources. H3 is supported in the analysis. When we add immigration attitudes into the model, respondents that support greater restrictions on immigration were significantly associated with lower confidence in the police. H4 received partial support/ Model 4 shows that consumption of TV, newspapers and new media were associated with lower confidence in the police. This differed to our prediction that online media would prove to be significant. Most of the earlier variables, with the exception of age remained significant in the model.
Our final test outlined in Table 4 (model 5) examines the role of postmaterialist attitudes in predicting lower confidence in the police. In support of H5, we can see that the inclusion of the postmaterialism variable renders the age and education variables insignificant, whilst conversely sex (female) became significant again. All immigration, experiences of crime, and information source variables remained significant.

Discussion
The case of public attitudes towards the Hong Kong police force is an interesting area considering a rapidly changing context of Hong Kong society. Successive large-scale protests have occurred, voicing concerns about the violation of rights and freedoms of Hong Kong citizens as they encounter the transition from post-colonial rule from Britain to the People's Republic of China. Earlier periods of the transition between Britain and China after 1997 were 'not characterised by full-blown ideological competition or firm-handed interventions from Beijing' (Adorjan et al. 2021, p. 655). Rather, more recent events have intensified the discord between citizens (largely youth), the police and the Hong Kong government at largethe latter accused of forming close alliances with Beijing. When located in the context of these political transitions and protest events, the role of the Hong Kong police appears to have been brought closer into these complex situations of order maintenance policing, with some commentators accusing the police of exceeding powers deemed reasonable to preserve stability in these events (Stott et al. 2021, Ho 2021. Recent aggregate-level reductions in the levels of confidence in Hong Kong (See HKPORI 2020) also provide a wider lens with which to interrogate further the factors which operate beneath these figures in terms of explaining those least likely to report confidence in the Hong Kong police. Given the speed with which the political climate has changed in Hong Kong, including a rapid growth of protests against the Hong Kong government, our paper provides important insight into factors shaping lower confidence in the Hong Kong police. Our analysis began by examining common demographic characteristics of Hong Kong residents that exhibit lower confidence in the police. Previous studies have identified lower confidence among young people, including those with higher educational levels (e.g. Kwak et al. 2012, Sun et al. 2012, Zhang et al. 2021. Although higher educated young people were more likely to exhibit lower confidence levels in the Hong Kong police, these variables were no longer significant when adding our final variable of postmaterialism into the models. Following previous research examining the concept of postmaterialism (e.g. Inglehart 1981, Inglehart and Abramson 1994, Wong and Wan 2009, Lee 2018, we find that those respondents with a greater postmaterial orientation had lower levels of confidence in the police. Although the concept of postmaterialism has, to our best knowledge, not been applied to attitudinal studies of policing, we believe that it provides a richer explanation than standard measures of age and education. Research from Hong Kong Wan 2009, Lee 2018) indicates a growing number of people possess values of postmaterialism, which connect strongly with values of democracy. These values have, of course, resulted in considerable tension and escalation of protest in Hong Kong, including conflict with the police. It is therefore logical that we find postmaterialism strongly predicted lower confidence in the police. Our measure of confidence in the police is limited to a single question. Testing multiple items of confidence and legitimacy with the concept of postmaterialism would be useful scholarly additions, especially in nations experiencing rapid political and economic changes. In a more nuanced measurement of confidence or legitimacy, it is quite possible that postmaterialist values will accord with support for certain types of policing, especially those which have a strong democratic and procedurally just orientation (See Manning 2010).
Reports of the recent protest situation in Hong Kong arising from claims about the greater interventionist stance of the Beijing government have played a key role in initiating a strong Hong Kong identity, including views against the entrance of Mainland Chinese into Hong Kong territory (Lee and Chou 2018). We find that those respondents reporting a stronger anti-immigration stance were less likely to support the police. This runs counter to research which finds that more hostile attitudes towards immigrants are generally associated with higher levels of support for the police, although derived from U.S-based samples (e.g. Dzordzormenyoh and Perkins 2021). Located in Hong Kong, this association between pro-immigration views and lower confidence in the police likely stems from an indirect link between respondents that were more critical of both the Beijing government, and Mainland Chinese community at large (although we cannot say through our measures that antiimmigration views were exclusively directed at Mainland Chinese specifically, yet they make up the largest ethnic group entering Hong Kong, See So 2017). The causes of anti-Mainland Chinese sentiment in Hong Kong are heterogeneous and pre-date the recently intensified public protests in Hong Kong (Law andLee 2006, Lowe andTsang 2017). Yet, more recently, a strong 'Hong Konger' identity has developed into a political symbol in Hong Kong, precisely during the same period as social conflict in Hong Kong has intensified (See HKPORI 2019). In summary, that anti-immigration views are associated with lower confidence in the police suggests an indirect critical attitude formation, likely towards the police, Hong Kong government and its perceived collusion with the Beijing government.
In Hong Kong, the use of online media has been closely connected to more critical news reporting of Hong Kong politics, and used by protesters to fuel 'pro-democracy' messages, (See Lee et al. 2017). We find that more frequent consumption of new media (as well as more conventional newsprint and TV news) was significantly related to lower confidence in the police. These findings run counter to research which finds that higher consumption of TV news increases confidence in the police (Callanan and Rosenberger 2011, Graziano and Gauthier 2018), although we show support for research indicating that online news and social media is associated with lower confidence in the police (Gauthier and , Intravia et al. 2018. The results also demonstrate the significance of factors such as higher fear of crime and perceptions of crime and disorder (drug use) in predicting lower confidence in the police. The more common empirical mechanism put forward is that fear of crime and confidence in the police may be underpinned by perceptions of communities, such as levels of disorder which may communicate a lack of social order ).
The timing of our data in Autumn 2018, coincided with a background of the 'Occupy Central' protests through the second half of 2014 (encompassing the Umbrella Movement), as well as the 'Mong Kok Riots' in February 2016. In the latter, disturbances were organised by the 'radical local advocates' who emerged after the 'failure' of 'Umbrella Movement' in 2014. Furthermore, the decline in confidence in the Hong Kong Police which we report through the WVS is likely to be an underestimate of the full scale of discontent with policing. Occuring after the WVS, regular polls collected in Hong Kong show approximately two thirds of respondents were satisfied with the police, which dropped sharply to less than 40% in August 2019 at the height of the Anti-Extradition protests (HKPOPI 2020, Adorjan et al. 2021). More than 80% of citizens have also questioned whether police actions were an indiscriminate, necessary and proportionate exercise of violence and believe that independent investigations should be carried out in accordance with the legal procedures stipulated in the laws of Hong Kong (Center for Communication and Public Opinion Survey, The Chinese University of Hong Kong 2020). The extreme lockdown experienced in Hong Kong and its policy of 'disease eradication' has further intensified criticisms of the police in its enforcement of Covid-based laws (Chan et al. 2021).
In practical terms, the police's claims of maintaining social order in Hong Kong will inadvertently mean that these actions will draw criticism, even if defenders of the police will claim their actions are reasonable within the law and not consistent with the goal of suppressing political dissent. Noteworthy are efforts to legislate against protest in Hong Kong, which has given the police powers to prevent any kind of political protest (Amnesty International 2020). These political contexts raise the concern that the Hong Kong police's historical reign of consensual policing and high tradition of public trust becomes a near impossible feat to maintain, within conditions where they are increasingly exposed to the goals of 'order maintenance' tasked by the SAR government at large. Notes 1. All of these four studies assessing the effects of media on policing used age as a control variable. 2. See https://www.worldvaluessurvey.org/WVSContents.jsp

Disclosure statement
No potential conflict of interest was reported by the author(s).