Rule-enforcer or rule-breaker ? A quantitative study on the effect of organizational culture on police deviance in the Belgian local police

In spite of the paramount importance of police integrity in a democracy, police deviance continues to be a problem. One much-discussed factor for explaining police deviance, is police culture. However, quantitative empirical research on the relationship between police or organisational culture on the one hand and police deviance on the other is scarce. The aim of this study was to investigate the impact of (different dimensions of) organisational culture in the police on (different types of) police deviance. The central research question of this study was: "To what extent does organisational culture in the police have an impact on police deviance?".

To answer this research question, we first developed a theoretical framework that was based on the grid-group cultural theory and typology originally proposed by Mary Douglas (1970) as well as on earlier research by Maesschalck (2004b) and Wouters (2016). Regarding our independent variable ‘organisational culture’, we developed a framework consisting of 15 cultural dimensions that together specify eight positions in a two-dimensional diagram. This framework encompasses dimensions of traditional police culture (e.g. internal solidarity, law-enforcement orientation, results-orientation), but also dimensions from the broader literature on organisational culture (e.g. self-interest, rule-orientation). The same two-dimensional diagram with eight positions was also used for the conceptualisation of the dependent variable ‘police deviance’. This resulted in a new typology consisting of 16 types of police deviance (e.g. insufficient effort, property deviance, limitless obedience).

Grid-group cultural theory was also used to formulate hypotheses regarding the relationship between organisational culture and police deviance. Based on one of the central propositions of grid-group cultural theory – called the ‘socio-cultural viability proposition’ – we expect that each cultural dimension will lead to certain types of police deviance, unless its impact is compensated for by the strengths of the other cultural dimensions. An over-emphasis on hierarchy, for example, can lead to rigid rule following if the other cultural dimensions (e.g. individualism with its focus on autonomy) do not adequately compensate for this excess. This general hypothesis was then further elaborated in two hypotheses for two types of cultural dimensions: those with a narrow range and those with a wide range. For the ‘narrow’ cultural dimensions, we hypothesized – based on the literature – a strengthening impact on a limited number of deviance types. Particularly, we hypothesized (1) a strengthening impact on those types of deviance which are situated on the same position in the diagram and which are on the same level (i.e. internal – external) and (2) no or a preventive impact on all other types of deviance. The second hypothesis about the ‘wide’ cultural dimensions is largely similar, but hypothesizes a broader strengthening impact: not only on the types of deviance on the same position in the diagram, but also on those on the neighboring positions in the diagram. Based on these two hypotheses, we then formulated hypotheses for all (i.e. 15) cultural dimensions. In particular, the theory led to a specific hypothesis for every possible relationship between the 15 cultural dimensions on the one hand and the 16 types of police deviance on the other. This resulted in (15x16=) 240 specific hypotheses.

These hypotheses were subsequently tested by means of an online survey in 64 local police forces in four provinces in Flanders, Belgium (n = 3869) in the period March – June 2017. The online survey included two new measurement instruments: one for organisational culture (in the police) (the Leuven Organisational Culture Questionnaire (LOCQ)) and one for police deviance. The police deviance instrument consisted of two parts: a self-report measure (asking for respondents’ own behaviour) and an observer-report measure (asking for occurrence of the behaviour in the respondents’ work unit). Confirmatory factor analyses showed that the proposed factor structures (i.e. the 15-dimensional cultural framework and the 16-dimensional typology of police deviance) could also be found empirically. Consequently, reliable scales could be developed for the 15 cultural dimensions and the 16 types of deviance. Moreover, multidimensional scaling showed that most dimensions of organisational culture were empirically situated on the position where we expected them based on the grid-group cultural typology. For the dependent variable ‘police deviance’, this was, however, not the case.

For the actual test of the hypotheses about the impact of the 15 cultural dimensions on the 16 types of deviance, the 240 specific hypotheses were further specified in statistically testable hypotheses. Since the dependent variable was measured in two different ways (an observer-report and a self-report), the 240 hypotheses all had to be tested twice. This resulted in 480 statistically testable hypotheses. Moreover, each of those 480 hypotheses could potentially also be tested at four levels of analysis in the multilevel analysis. Specifically, we aimed at testing two combinations of levels: individual level + unit level and individual level + organisational level. In practice statistical conditions were not always satisfied for all combinations. Ultimately, this all resulted in 1080 statistically testable hypotheses.

These hypotheses were initially tested by means of multilevel Poisson regression analyses. Based on these analyses, we found that only 766 of the 1080 (or 71%) statistically testable hypotheses were empirically confirmed and that these mainly concerned hypotheses in which no effect or a preventive effect was expected. For the statistically testable hypotheses in which a strengthening impact was expected, we could only find empirical evidence in about 1 out of 6 cases. The hypothesis that the cultural dimensions at a certain position of the diagram inevitably lead to the associated types of police deviance (i.e. the types at the same and/or the neighboring positions in the diagram) could thus not be confirmed. The relationship between organisational culture and police deviance was also studied by means of latent profile analysis combined with Kruskal-Wallis tests. The different work units within the local police forces were first divided into five classes based on their scores on the cultural dimensions. The Kruskal-Wallis tests then showed that those classes differed in the incidence of police deviance. In particular, this incidence was lower in the more ‘egalitarian’ work units (i.e. work units with much solidarity, consensus and good relationships with citizens and little self-interest, distrust and injustice). Overall, the proposed theory is not empirically confirmed: cultural dimensions do not always lead to their expected excesses, the other dimensions do not systematically compensate for those excesses and more ‘balanced’ cultures do not lead to less deviance.

Despite this global rejection of the theory, the study did yield a number of important insights. First, we found that the individualistic and fatalistic cultural dimensions (e.g. self-interest, futility, distrust towards citizens) tend to strengthen police deviance, while the egalitarian cultural dimensions (e.g. internal consensus, organisational reputation) tend to prevent deviance. Second, we found that for some types of police deviance the cultural dimensions had an impact that was the opposite of what was expected. For example, some cultural dimensions that are typically perceived as ‘problematic’ (e.g. organisational injustice, internal distrust) had in fact a preventive impact on certain types of deviance. Likewise, some cultural dimensions that are usually considered ‘desirable’ (e.g. internal solidarity) had a strengthening impact on some types of deviance. Third, some externally oriented dimensions of organisational culture (e.g. distrust towards citizens) turn out to strengthen some internally oriented types of deviance (e.g. insufficient effort). Likewise, some internally oriented cultural dimensions (i.e. internal solidarity) had a strengthening impact on some externally oriented types of deviance (partiality and nepotism). Finally, we found that organisational culture particularly impacts less serious types of deviance (e.g. insufficient effort, disruptive and uncooperative behavior, rigid rule following). More serious types of deviance (e.g. corruption, property deviance, manipulation of information) are only impacted to a (very) limited extent.

The study also offers some useful conclusions for police practice. First, this study showed that some types of behaviour score relatively high. For example, about 70 respondents (+-2%) reported that they themselves used inappropriate and/or disproportionate force against a citizen in the last 12 months. About 100 respondents (+-3%) reported that they drew up documents or official reports that do not correspond to reality (in order to protect themselves or a colleague) in the last 12 months. In addition to these serious behaviours, a number of behaviours that are generally perceived to be less serious also scored relatively high. For example, nearly 40% of the respondents reported that colleagues within their work unit have at least occasionally (response categories "sometimes", "often" or "very often") reported falsely in as ill in the past 12 months. 545 respondents (16.3%) reported that they themselves have unlawfully consulted the police databases at least once in the last 12 months. More random checks and further awareness-raising campaigns regarding the right to privacy therefore seem desirable. These figures certainly deserve attention from police management. One response, particularly to the more serious types of deviance, could be to increase control and sanctions. That is unlikely to be sufficient, though. More supportive instruments such as training and individual coaching are also necessary.

Second, this research showed that the cultural dimension ‘futility’ was one of the strongest predictors for police deviance. This is in line with much earlier research that finds a feeling among police officers that their actions make little difference, for example because ‘criminals’ are insufficiently punished by ‘mild magistrates' or ‘inadequate legislation’ (Chan, 2007: 143; Loftus, 2009: 14; Loyens & Maesschalck, 2014: 150). Police leaders, of course, cannot do much about the (perceived) causes of that sense of futility, but they might reduce that feeling by the way in which they communicate with their staff. They can express appreciation for the good work that employees have done, even if, for example, the suspect is released a few hours later. The cultural dimension ‘internal positional authority’ also had a strengthening effect on many types of police deviance. As a police leader, one can possibly respond to this by more often involving employees in decision-making processes. A third important predictor for police deviance turned out to be, somewhat surprisingly, the cultural dimension ‘internal solidarity’. As a police leader, it is therefore important to be aware of the serious risks of a group solidarity that is too strong. Finally, the cultural dimension ‘rule orientation’ only had a very limited preventive effect on police deviance. Creating new rules therefore does not seem to be the obvious solution to avoid police deviance. The values-based approach to ethics management should also receive the necessary attention.

For more information on this project, see https://www.law.kuleuven.be/linc/english/research/researchonpoliceintegrity.html.