‘What discretion do you need?’ Factors influencing police decision-making in possession of cannabis offences

Introduction

At any one time significant numbers of people are in possession of drugs, particularly cannabis, with an estimated 2.6 million self-reported cannabis users in England and Wales in 2019/2020 (Office for National Statistics (ONS), 2020). In the year ending March 2021, 165,450 drug possession offences were recorded by the police, of which 133,805 (81%) were for cannabis possession (ONS, 2021). This suggests that only a very small proportion of cannabis users come to the attention of the police in any given year and that the risk of a person in possession of a controlled drug being apprehended is very small (Caulkins and Reuter, 2009; Nguyen and Reuter, 2012). Relatedly, with most drug offences pertaining to cannabis, the vast majority of police time dealing with drug offences is focused on low-volume, low-harm offending behaviour. While in recent years a number of national and regional authorities have experimented with the decriminalisation and legalisation of cannabis (see Bacon, 2022; Colson and Bergeron, 2017), in England and Wales, there remains a strong focus on the criminalisation of cannabis possession. Furthermore, research suggests that those most likely to be policed for cannabis possession are young, ethnically minoritised, urban males belonging to ‘suspect’ groups. Despite the importance of the issue, remarkably little research has been conducted on the policing of cannabis possession with the last detailed fieldwork dating back to 2004/2005 (May et al., 2007). This study aimed to explore the enforcement of cannabis possession offences in a Northern England force employing a mixed-methods approach to examine police decision-making in such cases, the findings from which are explored in this article.

Background

Three different sets of policing guidance on cannabis offences were issued in the early 2000s (Association of Chief Police Officers (ACPO), 2003, 2007, 2009), each in response to amended or new legislation (including reclassification ¹ ). The latest of these introduced the three-point escalating response to cannabis possession that was still national guidance at the time of writing (ACPO, 2009). This consists of a tiered, escalation process whereby, in most circumstances, the first offence should result in a cannabis warning; ² the second, a Penalty Notice for Disorder (PND); and the third in arrest ³ (or, where time is short and the offender’s identity not an issue, voluntary attendance at a police station (Lloyd et al., 2018a).

While in the past, national and force policy with regard to cannabis possession policing has been reasonably consistent, there now appears to be greater divergence in police force policies. Many forces have continued to apply the most recent of the three national guidance documents as outlined above (ACPO, 2009). However, since the National Police Chiefs’ Council (NPCC, 2017) report on out-of-court disposals, some forces have moved to a different, tiered approach whereby an offender is given a Community Resolution at first offence, a Conditional Caution at a second and (as with the earlier ACPO guidance) is arrested and charged at the third. Individual forces can decide whether to adopt this approach, risking inconsistency in policing. For example, Community Resolutions do not count as convictions but are recorded on local police information systems and are equivalent in this respect to a cannabis warning. However, they cannot be considered for offenders if they have committed a similar offence in the previous 12 months. Thus, if two offences are committed within a 12-month period, a Conditional Caution will be issued, resulting in a criminal record, rather than a PND, for a second offence. This may result in a more rapid escalation to a criminal record for cannabis users in force areas that have introduced variants of the NPCC system.

As common offences that often represent someone’s first contact with police as a potential suspect, experiences of being policed for cannabis possession can be very significant – especially as many suspects will see the offence as trivial and police action as unnecessary. May et al. (2002) emphasise the implications for police legitimacy particularly clearly:

. . . cannabis possession is one of the offences that are most likely to bring people into ‘adversarial’ contact with the police. The scope for erosion of police legitimacy is obvious. If the laws that the police most frequently enforce are regarded by the policed as unreasonable and unnecessary, it is unlikely that police powers will be regarded as legitimate. (p. 33)

Cannabis possession is also an offence that potentially involves a high level of police discretion, which can be defined as the freedom individual officers have to make a choice about whether or not they enforce the law (Aplin, 2021: 4). Huff (2021) argues that having this choice renders ‘police discretion a key decision point in the criminal justice process’ (p. 1). Lord Scarman in his report investigating the Brixton riots in London in 1981 also highlighted the centrality of individual officer discretion: ‘Neither politicians nor pressure-groups nor anyone else may tell the police what decisions to take or what methods to employ, whether to enforce the law or not, in a particular case . . .’. He went on to say that ‘the exercise of police judgement has to be as independent as the exercise of professional judgement by a doctor or a lawyer’ (Scarman, 1982: 104). As Scarman also pointed out, however, this independence has to be coupled with operation within the law and with the support of the community; otherwise, such discretionary decisions can ‘be the trigger for an unexpected and unprecedented disintegration of normal community order’ (Bronitt and Stenning, 2011: 319). Therefore, to be seen as legitimate, discretionary decision-making by the police, particularly ‘on the street’ where there is inevitably less scrutiny of those decisions, must be principled, accountable and not discriminatory (Bronitt and Stenning, 2011: 320).

A number of studies have provided valuable insights into the ways that officers employ discretion in their policing of drug possession. For example, Warburton et al. (2005) found that while a third of the 150 patrol officers in their study reported always arresting someone in possession of cannabis, two-thirds did not. Demeanour of the suspect was found to be important: ‘If you stop someone and they are not aggressive or abusive and they don’t have a criminal record, I can’t see the point in giving them a criminal record for a bit of cannabis’ (p. 118). Others never arrested anyone: ‘I never nick anyone for cannabis, and never will, unless it’s a van load’ (p. 119). Suspects that did not attempt to conceal or deny the offence were more likely to be dealt with sympathetically. Similarly, May et al. (2002, 2007) asked officers if they had ever dealt informally with cannabis possession. Sixty-nine percent of officers in 2000 and 48% in 2004/2005 said that they had. In both studies, officers reported dealing with cases informally by getting rid of the cannabis, giving the offender an informal warning and not recording the offence. The researchers suggest that the reduction in informal disposals between 2000 and 2004/2005 may be because some cases previously dealt with informally were given formal ‘street warnings’ in the later years (a disposal introduced by the ACPO 2003 guidance and later renamed ‘cannabis warnings’). The main explanations for using formal disposals given by officers were beliefs that cannabis acts as a gateway drug (and so, early sanctions could deter escalating harm), that cannabis possession arrests lead to the detection of more serious offences and that the confusion brought about by the change of classification made these offences difficult to deal with informally.

Another important feature of drug offences is that possession rarely yields a direct victim or aggrieved party likely to push for police action, providing particular scope for police discretion (Bacon, 2016; Beckett et al., 2005; Houborg et al., 2016; Shiner, 2015; Warburton et al., 2005). Beyond this discretionary leeway, it is also important to emphasise that drug possession policing is to some degree ‘accidental’, in that research shows that officers do not generally seek out such offences proactively (Lloyd et al., 2018a). For example, May et al. (2007) found across their three study sites that cannabis possession most often came to light while investigating other offences, particularly when searching apprehended pedestrians or drivers. This noted, the initial justifications for stopping people often reflected vague suspicions relating to how people look, where they were and the time of day or night. As May et al. (2002) point out, the most likely explanation is that cannabis users have other, more visible, ‘suspicious’ attributes or behaviours that result in their increased contact with the police.

It is argued, therefore, that those who are apprehended for drug possession tend to be those in public spaces who seem suspicious to the police and this has meant that young men – and young Black men in particular – are disproportionately sanctioned. The much larger literature on the overrepresentation of young Black men among those subjected to stop and search cannot be wholly included here. However, the most recent of two studies that looked specifically at drug offences (Eastwood et al., 2013; Shiner et al., 2018) found that the age-adjusted rate of police drug stops per 1000 of the population varied from 18 for Black people to 2.5 for White. This gross disproportionality was even more marked in some individual forces. In Dorset Police, the population-adjusted rate among Black people was 26.5 times higher than among White.

In a study of drug policing in four Danish police districts, Houborg et al. (2016) explored the potential for drug enforcement to be used selectively to control marginalised young men. Reflecting the findings of May et al. (2007), they emphasised the importance of ‘suspicion formation’, whereby police officers draw on appearance, behaviour, time and location, and intelligence in deciding whom to approach on the street. Compared to a general population comparison group, suspects were more likely to be male, young, less educated, unemployed and have a criminal record. The researchers categorised 317 drug possession cases by the nature of the initial police contact and found, that in 27% of cases, the offence was a by-product of police work unrelated to drugs – the majority being initially suspected of a non-drug offence. The authors describe these cases as involving ‘reactive’ policing – that is, they had not been planned or sought out. In another 29% of what they referred to as ‘proactive’ cases, officers decided to approach suspected drug users on factors such as whether they were known to them or were behaving suspiciously. In interviews, officers talked about who they would approach proactively and described two types: ‘young guys, scruffy, unhealthy, slight of build’ (p. 336) and muscular men ‘with tattoos and gold chains’ (May et al., 2007) who could be involved in gangs. They also referred to particular kinds of cars – ‘old, big cars’ (May et al., 2007) – and particular kinds of drivers arousing suspicion. The authors conclude that the disproportionate criminalisation of unemployed young men for drug offences is, at least in part, associated with the fact that ‘these people get into circumstances where they attract the attention of the police’ (May et al., 2007: 338).

From these studies, it can be seen that police detections of cannabis possession often occur as an accidental discovery in the course of investigations based on other suspicions and other aims. While the recorded grounds for a stop and search in England and Wales is most frequently drugs (63% in 2019/2020 (Home Office, 2020)), and of these, most often for low-level possession offences (HMICFRS, 2017), the original reasons for approaching particular individuals and vehicles may be different. Those that carry drugs on their person in public places and are more generally over-policed – either because of their appearance/behaviour or their previous criminal record – tend to make up the very small percentage of users that are policed for their drug possession. It follows, therefore, that the majority of people who are otherwise law-abiding; who use cannabis in private settings; and are unlikely to attract police attention when in public, are very unlikely to appear in the offence statistics. The corollary of this is that the type of profile described by Houborg et al. (2016) becomes something of a self-fulfilling prophecy: those policed for drug possession are likely to be young, male and economically marginalised, at least in part because this is the group most likely to come to the attention of the police (e.g. Quinton et al., 2000). These findings also chime with more general discussions around marginalised groups being perceived as ‘police property’ (Lee, 1981: 53), whose social control and segregation the rest of society are content to leave to the police (Reiner, 2010). Consequently, those seen as ‘police property’ may be more criminalised, not because they are more criminal, but because police officers see them as more suspicious.

As noted above, a potential positive impact of drug possession enforcement is the discovery of more serious offences. Officers in Bear’s (2016) detailed ethnographic study of the policing of drug possession in the Metropolitan Police Service found that this was one of the reasons officers gave for targeting petty drug offences. Similarly, reflecting May et al. (2007, described earlier), officers in Bacon’s (2022) recent study involving interviews with 46 officers across 11 English forces spoke of the importance of possession policing in providing intelligence about dealers. However, in an analysis of 30,000 custody records, May et al. (2006) found that ‘less than one per cent of simple possession offences “opened the door” to the discovery of more serious offences’ (p. 32). They conclude that this challenges ‘the widespread view that cannabis arrests frequently lead to the detection of more serious offences’ (May et al., 2006: 32).

Finally, there may be other functions for cannabis policing from a police perspective. Bear’s work describes how policing drug possession provides a tangible outcome. Policing frequently does not result in clear impacts but ‘finding a bit of cannabis allowed the officer to hold up their “trophy” to their colleagues’ (Bear, 2016: 323). Thus, cannabis policing may also serve performative, symbolic or status-related functions for police officers that have little to do with the impact on users or wider society.

The research methodology

Despite the salience of the issue of cannabis policing in the United Kingdom, the last detailed research dates back to fieldwork conducted in 2004/2005 (May et al., 2007). This mixed-methods study aimed to explore patterns of enforcement of cannabis possession offences by a Northern English Police force which, while having a generally low crime rate, had a higher than the national average rate of recorded drug offences in 2016. ⁴ For the quantitative analysis, police partners provided anonymised data on 4597 drug possession offences and 6020 associated individuals apprehended for drug possession offences over a period of 3 years (2013–2016).

The bulk of quantitative analysis centred on descriptive analyses of coded data. For this, we explored recorded characteristics of offences and offenders and the extent to which sanctions escalated over repeat offences, and we drew on labour force statistics to explore the relationship between ward-level deprivation and cannabis sanctions. We then inductively coded the descriptive information on file for 50% of the 2015–2016 cannabis offences (496 cases). From this, we produced an offence typology based on the characteristics of the most frequently encountered scenarios described by sanctioning officers, and applied this typology deductively to the remaining cannabis sanctions across all 3 years.

We also drew on routinely available datasets to explore associations between ward-level measures of deprivation (taken from ONS ward profiles for 2011) and rates of cannabis sanction. Rates of sanction were calculated as sanctions per year per ward resident, ⁵ and we had data both on the location of sanctions (where cannabis was found) and sanctioned individuals’ home wards. Ward-level unemployment rates and claimant count rates were used as proxy indicators of deprivation. Analyses were simple tests of association, using Pearson’s r. To control for night-time economy (and town centre) effects, we routinely repeated analyses while excluding town centre wards, which routinely evidenced both high levels of unemployment/claimant count, and distinctive (very high) patterns of sanctions. Analyses excluding weekends (and weekend evenings) were also conducted. We did not identify any instances where excluding wards, days or periods of time from analysis had any notable impact on analyses, suggesting a strong and persistent relationship between our chosen indicators and rates of cannabis sanction.

The qualitative component of this study was informed by this statistical analysis, comprising semi-structured interviews with 37 officers in a variety of rural and urban police stations across the force area to explore in depth their decision-making processes in the policing of cannabis. Officers were asked specifically about what factors might affect their decision-making in the most frequently encountered scenarios drawn from the statistical analysis. While counties with very small ethnically minoritised populations have been found to have grossly disproportionate stop and search rates (Shiner et al., 2018), we could find no evidence at all of disproportionality within our dataset – within both rural and urban wards and across ethnic groups, the proportions of those sanctioned reflected local populations. Consequently, this study presented an opportunity to examine other factors that might also influence police decision-making in cannabis possession offences.

Thirty of the interviewed officers were constables and seven were sergeants. ⁶ All officers were in response roles, primarily responding to emergency calls. Their length of service ranged from 1 month to 29 years with an average length of service of 8.6 years. Available officers on duty at the time of fieldwork visits were interviewed in rooms in police stations where conversation could not be overheard. Interviews lasted between 23 and 68 minutes, were digitally recorded and transcribed in full. NVivo 9 was used to code and analyse all transcripts using a thematic approach (Braun and Clark, 2013). Common themes and sub-themes were independently identified by the three researchers with a final coding scheme agreed upon following iterative discussions. Key themes included awareness and understanding of national and force guidance, use of the escalation process, perceived benefits and impact of policing cannabis offences, discretion and informal actions, views on classification of cannabis, public attitudes and expectations of policing drugs, reasons for searching and arresting and policing young people’s drug use.

This article provides a summary of the statistical findings from the quantitative data before moving to a detailed discussion of qualitative findings exploring the reasons behind officers’ decision-making in cannabis possession offences. ⁷

Ethics

Ethical approval was obtained from the Department of Social Policy and Social Work’s ethics committee at the University of York.

Findings

Decision-making in frequently encountered scenarios

As discussed above, statistical analysis of police data provided a typology of the most frequently encountered scenarios in which officers pursued cannabis possession. We consequently wrote these up in five typical vignettes, and, in interview, asked officers how they would approach these situations (Table 1), which were:

Stopping or approaching a car then smelling cannabis coming from the vehicle and deciding to search the person(s) and/or car for cannabis (CAR ¹⁰ ).

Seeing cannabis while attending a house and deciding whether to take action for possession of the drug (HOUSE).

While on patrol receiving a report from a member of the public that they can smell cannabis coming from a group of young people in a public place and deciding whether to detain and search the group (YOUNG PEOPLE).

Finding cannabis on a person when searching them in relation to another offence in police custody, and deciding whether to take action for possession of the drug (OTHER OFFENCE).

OPEN IN VIEWER

Table 1.

Officers’ searching decisions in most common scenarios.

Scenario
Decision to search	CAR	HOUSE	YOUNG PEOPLE	OTHER OFFENCE
Yes would search	25	22	5	33
No would not search	3	5	10	0
Depends on circumstances	6	9	20	1
Other	0	0	1	0
Missing	3	1	1	3
Total	n = 37	n = 37	n = 37	n = 37

All interviewed officers were asked how they would respond in these situations and Figure 1 summarises interviewees’ preferred course of action in these most frequently encountered scenarios.

OPEN IN VIEWER

Figure 1.

Officers’ searching decisions in most common scenarios.

The majority of officers said that they would search a car and its occupants if they suspected cannabis possession. The most frequent reasons given were the risks associated with drug-driving.

If there’s a smell of cannabis, and someone’s driving . . . my main concern would be there, that the driver’s behind the wheel. (Officer3)

Or the opportunity to search for evidence of other offences:

Having smelled the cannabis . . . I’d be looking at searching the [four] people and the car. Not only for cannabis, but for other offences. (Officer18)

Officers’ decisions around cannabis possession when attending a house were more nuanced, with around a third of officers saying that any action would depend upon circumstances. Decisions were often based around the context, the seriousness of what else was being dealt with at the time, the prominence of any visible drugs or drug use and whether they had sufficient grounds to take action. The welfare of vulnerable residents was also a significant factor – even if they were the person suspected of drug use. Some interviewees seemed broadly inclined towards restraint:

So an example of [deciding not to investigate / pursue . . .]. [T]he officers went round to a job. The suicidal female put a ligature round her neck, and the officer were going to go for her for a concern for safety. Whilst the officer was there he found a small amount of cannabis. And I did tell him not to pursue that . . . I said to him, the welfare and the safeguarding of that female to make sure that she gets better takes priority over that drug. I am not interested at this moment in time, about that small amount of cannabis . . . not in the public interest. (Officer9)

Linked to concerns around welfare, several officers also highlighted that they might be more lenient if the person they were dealing with when attending a house was a victim of crime.

When it’s a victim of a serious offence, and there’s not cannabis on a desk in front of me, but there’s a lot of evidence of use, and maybe a faint smell. . .Am I really serving a policing purpose by criminalising a victim of serious crime . . . I’m not particularly bothered about the cannabis, I’m more bothered about this very serious offence. (Officer33)

Nevertheless, it was felt by some officers that the cannabis would have to be dealt with at a later date:

Obviously the victim side of it takes priority, but that means that the cannabis still needs dealing with. (Officer35)

Other pragmatic considerations, such as proving whose cannabis it was in a multiply occupied house, also came into play.

If you went to a house and you saw the cannabis laid about if there’s five or six people in the house, I mean, how are you going to prove whose it is? (Officer10)

Searching groups of young people in public space was a complex situation, and only 5 officers said that they would definitely search the group, 10 officers said they would not search them and 20 officers said their decision would depend again on the circumstances. The size of the group and the resources available often dictated a decision:

How do I stop and detain ten people, when the tenth will have the cannabis on them, and they are just going to drop it anyway. Complete waste of time. (Officer8)

Two other officers felt that any group was likely to have dispersed by the time they got there, or as soon as they spotted them arriving:

But nine times out of ten if you approach kids like that they would run off anyhow. They’re normally quicker than me. (Officer23)

A fifth of officers felt that reports from the public were not enough to justify a search, and would require direct evidence of drug use to take the case any further.

I think I would personally want a little bit more than just somebody’s say so. So I’d be looking for the smell . . . or whatever to just give me that little bit more suspicion myself. (Officer14)

A number of officers mentioned that dealing with groups of youths in this scenario would be more likely to fall to Police Community Support Officers (PCSOs), at least initially; who, in their view, were more able to use a degree of discretion:

The PCSOs are probably the ones who would get involved more because they’ll be the ones who the kids trust. (Officer10)

Almost all interviewees said they would always take action if they found cannabis on a person they were holding on suspicion of another offence; particularly if they were going to be arrested anyway. However, even following arrest the option of giving a cannabis warning or drug referral was still open to officers.

Well, if they’ve been arrested for something, it’s very easy to tag a charge of possession of a drug . . . It might be that we give them a cannabis warning for the drugs but they can go to court for the assault, the burglary, whatever. (Officer28)

These findings again chime with Houborg et al.’s (2016) in that those who regularly come into contact with the police, either as a part of a ‘suspect population’ or because they are considered vulnerable in some way, are most likely to be dealt with for cannabis possession – even if they are victims of crime. In addition, the policing of groups of young people, in particular, appeared to be largely driven by pragmatic decision-making in circumstances of limited resources (i.e. the near-impossibility of one or two officers successfully detaining and searching a large crowd of young people in open, public space).

Findings from interviews

As well as asking for views on the most frequently encountered scenarios discussed above, interviews also explored officers’ opinions on cannabis policing more broadly.

Discussion

The findings from this study indicate that there is very little purposeful, proactive policing of cannabis possession, at least by response officers in the studied police force area. Instead such offences were usually discovered either in the investigation of other possible offences; because an individual is already suspected by or known to the police; or because they are judged by officers due to their behaviour or their location to be a suitable target for police attention.

Once cannabis possession offences came to light, most officers in this study felt that they had no choice other than to take action and enforce the law. Indeed, many felt that, due to what they perceived as the harms associated with cannabis, they were right to do so. In addition, officers appeared to feel accountable to both their managers and the public to take formal action. As a result, informal action for these officers was a practice they would never countenance. The introduction of a range of pre-court disposals, particularly cannabis warnings, was in most officers’ opinions sufficient to allow them enough low-level options for dealing effectively with straightforward cannabis possession. However, contrary to College of Policing guidance, searches were still being made on the basis of smell alone and a third of officers thought that this was sufficient reason to search. In addition, the strong smell of modern forms of cannabis made it difficult for officers to ignore the use of this drug in public spaces or if the smell was emanating from vehicles.

Police officers’ decision-making was also driven by a variety of contextual factors. Some of these factors were pragmatic, such as the practicality of searching large groups of young people, or identifying who owns cannabis discovered in a house. Other decisions were based on concerns about more serious offending such a drug-driving or dealing, or the welfare of individuals, particularly victims of crime.

Limitations

A clear limitation of this study is its focus on one force. In addition, our interview sample was limited to response officers and their sergeants, who may well have a more ‘reactive’ style of policing than say, neighbourhood police officers or PCSOs, who might work differently, particularly with young people, and/or detectives who would have more of an interest in intelligence gathering targeting drug dealers.

Given the considerable variation of practice evident even within this one force, there is clearly merit in a future study that would attempt to capture variations in cannabis policing around the country. Alongside the variation between those forces continuing to use cannabis warnings, there are also a number of forces experimenting with new approaches to policing drug possession around the United Kingdom.

Limitations on ward-level quantitative analyses should also be noted. As we relied on ward profiles, the only data available were from 2011. It is possible that some wards had changed significantly over the subsequent 2 years, before our sanction data began. It should also be noted that our proxies for deprivation are just that – they represent our best attempts at securing directly relevant data that can offer some insights into the characteristics of wards. Alternatives, such as the multiple deprivation index, offer a more direct measure of deprivation, but could not be reconciled (geographically) with the ward-level police data we had. Our analyses should consequently be approached with some caution – but the strong and persistent relationship between ward-level claimant counts and unemployment, in particular, do suggest that a persistent underlying relationship between policing and deprivation exists.

Finally, this study did not include the perspective of those policed for their cannabis use. Such a perspective would be invaluable in order to explore the consequences of that policing for them, which in turn might be effective in developing police drug policy and practice that aims to prevent more harm than it causes.

Conclusion

These findings chime with those from previous research (May et al., 2002, 2007; Warburton et al., 2005) in terms of the reasons given by officers for taking formal action in cannabis possession cases, albeit now officers are making those decisions within different policy frameworks that appear to shift them away from informal action. Officers in this study also expressed being concerned about greater scrutiny of and accountability for their on-street policing decisions both from the public and from their managers. Nevertheless, there was evidence in this study, of discretionary practices in officers’ decisions about who to approach, search and arrest. These decisions appeared sometimes to be based on reactive ‘suspicion formation’ (Houborg et al., 2016) about certain ‘suspect’ groups, individual officers’ personal views of the dangers of cannabis; or the hope that detection of more serious offending might be revealed. However, what is particularly concerning is that suspects with previous drug convictions or more simply just known to the police – ‘police property’ – were more likely to be approached and searched and more quickly escalated to arrest and prosecution, leaving the vast majority of cannabis users with little to worry about in terms of risking a criminal record for drug possession. These processes also played out the geographical level, in terms of the significantly greater levels of cannabis sanctions among people living in wards with high levels of welfare claimants – even though these people were most often picked up by the police outside their residential wards. These findings raise fundamental questions about quality and justice in the policing of cannabis possession.

As Bronitt and Stenning (2011) argue, it is possible that the risks of misuse associated with police discretion can be reduced by ‘structuring its exercise through administrative guidelines’ (p. 325). There was some evidence of the impact of such guidelines in this study – although any systematic, consistent or linear application of the escalation process embedded in the guidance was far less obvious. In addition, some officers continued to use smell as grounds for search against current College of Policing professional practice guidance.

At the time of writing, around a third of police forces in England and Wales, including most recently the Metropolitan police (The Guardian, 4 January 2022), have implemented or are considering implementing drug diversionary schemes (see Bacon, 2022). Through such schemes, those caught in possession of a drug, most often cannabis, are referred to drug education, harm reduction or treatment, as an alternative to arrest, prosecution or formally caution (Transform Drug Policy Foundation, n.d.). While this could be seen as a positive way forward for those brought into these schemes, it also further increases variation in police practice across the country. This begs the question as to whether such variety in policy and practice is desirable given the consequences of a drug possession conviction and given the targeting of already criminalised and marginalised groups for police attention.