![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
Personal contact with the police is mentioned as one of the key predictors of individuals’ opinions towards the police. But findings from earlier research contradict each other. Some researchers found an asymmetric relationship with a much stronger effect from unsatisfactory contact, compared with satisfactory contact, with regard to trust in the police and police legitimacy. Others found a more symmetrical relationship. In a way, these different findings can be due to different measures of trust and legitimacy. In the literature there is no consensus about the meaning and measurement of these concepts. The purpose of this study is therefore to test the (a)symmetrical relationship while taking into account criticisms about trust and legitimacy. More precisely, we consider trust in police procedural justice and trust in police effectiveness as two components of trust having an influence on police legitimacy in the form of moral alignment. Feeling an obligation to obey the police is considered as an outcome of moral alignment. We used path models in MPlus to do the analyses, which were conducted on data collected from the Social capital and Well-being In Neighbourhoods in Ghent (SWING) survey, Belgium. The results show a more symmetrical relationship between contact and both components of trust. Furthermore, trust in police procedural justice was found to be a stronger predictor for moral alignment than trust in police effectiveness. Moral alignment itself seemed to be a strong predictor for feeling an obligation to obey the police.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1. In his study, it was called confidence but after examining the measurement, it can be considered as trust as we use it in our study.
2. See FootnoteNote 1.
3. In the study of Bradford et al. (Citation2009a) a three-factor solution fitted the data better than a two-factor solution. That is why they made a distinction between perceptions of police effectiveness, procedural justice and community engagement as components of trust. Some researchers also add trust in distributive justice as a component. In the current study, this component was not taken into account. We believe that distributive justice is closely related to procedural justice. Handling all kinds of people in the same way can also be considered as fair handling. However, we are aware that procedural and distributive justice can be conceptually and practically distinguished.
4. For an extended version of this ongoing debate we refer to Van Damme et al. (Citation2013).
5. The focus in the current study is on perceptions of legitimacy. Thus, when we use the term legitimacy without an adjective, we refer to empirical or perceived legitimacy.
6. We do consider perceived legality as a dimension of perceived legitimacy, but we could not take this into account because the survey did not contain a question referring to perceived police legality.
7. In the ESS, for example, obligation to obey was measured by three items beginning with: ‘To what extent is it your duty to…’ Duty can be interpreted in different ways, which is why interviewers were given an instruction to clarify duty as meaning in the sense of citizens’ moral duty to the state, when people asked. In a way, this could counter Tankebe’s arguments, but we do not believe that the instruction given to the interviewers was strong enough.
8. We would like to thank the following researchers for providing these data: Hardyns, W., Vyncke, V., Pauwels, L. and Willems, S.
9. With neighbourhood we mean a statistical sector which comprises the smallest administrative unit of analysis at which demographic and socio-economic information is systematically gathered in Belgium and can be compared to the US census tract level (on average 1319 residents/neighbourhood).
10. In the current study we only used individual level variables.
11. The ESS is an academically driven social survey that maps the attitudes, values and opinions of the inhabitants of more than 30 European countries and it shows how they evolve. In round 5 (2010), special attention was paid to trust in the criminal justice system. Because of the high quality standard, most of the measures in this study are based on those from the ESS.
12. Cronbach’s alpha is a coefficient of internal consistency. All items have to measure the same concept. A value of less than 0.80 indicates little consistency between the items. In practice, values of 0.60 or higher are acceptable (Pauwels Citation2012).
13. A factor analysis is used as a critical test for the scale constructs. First, we need to be sure of the one-dimensional characteristic of the scale. Second, items with low factor loadings (<0.40) need to be eliminated because they do not contribute to the factor with measurement errors as a result (Pauwels Citation2012).
14. For these latent variables we implemented the Expectation Maximization imputation method for reducing the non-response as much as possible.
15. When adding contact as a dichotomous variable (Yes/No) into the path analysis, no significant effect was found. As we already stated in the literature review, this was not surprising. This finding does not mean that we can conclude that earlier contact does not influence people’s opinions of the police. It only confirms that future research has to take more advanced measures of contact into account.
16. Two new variables were created with the ‘visual binning’ procedure in SPSS, one based on the standard deviation of the procedural justice scale and the other on the standard deviation of the police effectiveness scale. The standard deviation measures the spread of the data by regarding how far data are situated from their mean. Notwithstanding this procedure leads to a loss of information, it offers the opportunity to interpret descriptive results in an understandable way (Pauwels Citation2012). The new variables give an indication of the number of respondents with the lowest, neutral and highest trust in police procedural justice and police effectiveness.
17. We do have to remark that a mean of 15.99 on trust in the procedural justice scale for the unsatisfied group is not that bad for a scale ranging from 5 to 25. In fact, trust in procedural justice seems to be rather high in our total sample.
18. We do have to remark that a mean of 8.99 on the trust in police effectiveness scale for the unsatisfied group is not that bad for a scale ranging from 3 to 15.
19. Before doing this analysis, we carried out some statistical controls. We controlled for the effect of gender, ethnicity and age in several block wise regression analyses. Only a very small effect from age was found on unsatisfactory contact, trust in police procedural justice, moral alignment and obligation to obey. But in line with Wikström (Citation2007), we do not consider this as problematic. Although, we cannot prove causation by means of the current study, we are interested in the mechanism that produces an outcome. In this regard, Wikström argues that when there is not a connecting process (mechanism) that brings about the effect, the ‘cause’ at best represents a symptom or ‘marker’ or is purely accidental. Gender and age, e.g., are such markers of real causative factors among predictors in studies. Wikström therefore states that if we can measure the real causative factors, there is no need to include attributes such as gender and age.
20. In the path analysis the neutral and unsatisfied groups are combined. There are several reasons why we have done this. First, because it is in line with earlier research. It makes it possible to compare findings. Second, we have done an analysis in which we took three dummy variables into account (neutral, satisfied and unsatisfied). Based on these results, we found that unsatisfactory contact has a little stronger impact on trust than satisfactory contact, but the results are not that stable because the number of respondents in the neutral and unsatisfied group is too low to offer statistically significant results. Third, post hoc analyses showed no significant differences between these two groups.
21. Because of the ongoing debate about the measurement of the concept of legitimacy, we also analysed a path model with moral alignment and obligation to obey as components of perceived legitimacy. The value of the RMSEA showed a worse model fit (RMSEA = 0.047; AIC = 13 474.723; BIC = 13 558.146).
22. See FootnoteNote 20.
23. Exploratory analysis confirms this assumption, although the results are not that stable because of the low number of respondents in both the neutral and unsatisfied group.
24. In the case of police behaviour, the number of complaints can give an indication. But then again, in most cases citizens do not report police misbehaviour.