Predicting police officer reported compliance and willingness to cooperate with internal affairs units: application of the relational model of authority

ABSTRACT

Internal accountability mechanisms, such as internal affair units, are an important aspect of misconduct and corruption prevention. Key to internal mechanisms is the fostering of compliant behaviour. Drawing on the Relational Model of Authority, this research investigated whether police officer perceptions of procedural justice and legitimacy, relevant to an internal affairs unit, influenced reported compliance and willingness to cooperate with the unit. An online survey was completed by 1195 police officers. Structural equation modelling revealed an indirect effect of perceived procedural justice on reported willingness to cooperate with the internal affairs unit via the subscales of legitimacy; perceived obligation to obey and perceived trust. The findings of this research provide important insight into theoretical understandings of compliance with authority in a policing context and will be of value to agencies seeking to improve compliance and cooperation with internal affair units.

KEYWORDS:

• Internal affairs
• compliance
• procedural justice
• legitimacy

Disclosure statement

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

Notes

1 Morality can also lead to normative compliance, however legitimacy is thought to be more useful as it can provide discretionary authority that individuals may not accept but will defer to, whereas personal morals can lead to both compliance or non-compliance depending on how the rule in question aligns with an individual’s moral code (Tyler 2006).

2 It is acknowledged that procedural justice is a concept that originates from the wider field of organisational justice, and that within this field there are disputes regarding its exact nature and how it relates to other organisational justice constructs (Colquitt et al. 2001, Bies 2005, Colquitt & Shaw 2005). This study will follow the conceptualisation of Tyler and associates as research supports the relevance of this approach to legitimacy within policing contexts (Tyler et al. 2007, Bradford et al. 2014, Tyler 2017, Van Craen & Skogan 2017).

3 Whilst a more intentional approach to sampling was preferred, the goal was to maximise responses to ensure an adequate sample size for structural equation modelling. Given the target population and topic of the project, it was anticipated that recruitment of a sample might pose a challenge and so convivence sampling was utilised.

4 This understanding of legitimacy has served as a ‘conceptual touchstone’ in the field, but some scholars have begun to criticise and recommend changes to the underlying theoretical basis (Bottoms & Tankebe 2012, Worden & McLean 2017, p.483). For example, Bottoms and Tankebe (2012) present a comprehensive theoretical discussion that suggests that neither trust nor the obligation to obey can be equated to true legitimacy and that these measures should be revised. Measurement issues have also been evident as conceptual differences have contributed to substantial variance in the measurement of legitimacy across the field of research (Reisig et al. 2007, Johnson et al. 2014). Many studies have reported difficulty when combining the two scales into a single construct, with some having found each subscale to have different effects on dependent variables (Reisig et al. 2007, Gau 2011, 2014, Johnson et al. 2014, Sun et al. 2017). Of note are instances of the obligation to obey having no significant effect on reported cooperation and compliance once separated from trust (Reisig et al. 2007, Reisig et al. 2014).

5 Specifically, Principal Axis Factoring (PAF) using direct oblimin rotation.

6 Model fit was χ² = 140.913, df = 5, p < 0.001, CFI = 0.970, RMSEA = 0.151, SRMR = 0.0325. Though the RMSEA was out of typical ranges, it is worth noting that the causes and meaning of contradictory indices are unclear (Lai & Green 2016), high but accurate RMSEA values (i.e. up to 0.2) are possible due to model conditions (i.e. high factor loadings, McNeish et al. 2016), and as similar issues have not been reported in past literature (Reisig et al. 2007, Gau, 2011, 2014), it is more likely results were based in measurement error (i.e. the wording of items) rather than theoretical explanations.

7 The structural equation model was specified with two existing error co-variances. Failure to connect appropriate mediator errors can result in misspecification (Preacher & Hayes 2008). Trust and Obey were co-varied given their theoretical mutual contribution to legitimacy and their consistent significant correlation in past literature and the CFA conducted above. Connecting the residual errors of compliance and cooperation was rationalised on the basis that the two behaviours are similar and are likely explained by similar factors outside of those specifically identified by the model (e.g., personality characteristics).

8 In comparison the second largest modification suggested was between COMP4 and Procedural Justice (MI = 0.57, EPC = 0.199).