‘Seeing disorder’ in an English town

Abstract

Perceptions of low-level social and physical disorder loom large in criminological theory and research. When disorder seems high, trust in authorities is eroded, concerns about crime precipitated, and a general sense of unease develops. In this article, we use fine-grained survey data from a medium-sized town in the north of England to consider why some people experience their environment as disorderly while others do not. People are more likely (than others living in the same locality) to identify disorder as a problem when (a) they feel let down or abandoned by local and national authorities; (b) they are in an economically precarious situation; (c) they have been recent victims of crime and (d) when they are dissatisfied with the place they live. These findings illuminate the social and structural factors that underpin perceptions of disorder and, consequently, wider concerns about crime, institutions and social change.

Introduction

Perceptions and experiences of low-level ‘social disorder’ loom large in criminological thinking and research. Ever since publication of Kelling and Wilson’s (1982) Broken Windows, concern has consistently focused on youths hanging around, begging, homeless people, drug-taking and street drinking, and on graffiti, vandalism, litter and dog mess. Such incivilities are, among other things, treated as signs of neighbourhood decline, and as triggers for more serious criminal behaviour (Braga et al., 2015; Mackenzie et al., 2010; O’Brien and Ciomek, 2023). One important mechanism for such a process is thought to be the negative effect of disorder on social cohesion and collective efficacy – on people’s understandings of others in their community and of their collective ability to regulate it (Weisburd et al., 2024). Social disorder has also been found to be an important predictor of fear of crime (Farrall et al., 2009) and of trust in the authorities, particularly the police (Brown and Reed Benedict, 2002; Jackson et al., 2012). Activities in people’s immediate physical and social environments which are coded as disorderly constitute important signals of the presence of more serious criminality, social breakdown, the decline and loss of social control, and the failure of authorities.

Unsurprisingly, a significant body of research has therefore focussed on why people ‘see’ disorder in their communities or neighbourhoods. At stake in this literature are concrete signs of particular behaviours (littering, loitering, drug use), and also, arguably more importantly, whether these things are seen or identified as problems by local residents. Broadly, it is the social meaning that people attach to these signs or forms (Girling et al., 2000; Innes, 2004) that is important in terms of the outcomes noted above, and studies have considered how, inter alia, individual characteristics (Hipp, 2010), ideological stances (Jackson et al., 2018) and racial stereotyping (Sampson and Raudenbush, 2004) shape perceptions of disorder.

In this article, we add to this literature by, first, exploring why people living in a medium sized town in the United Kingdom, Macclesfield in Cheshire, come to see their immediate physical and social environment as disorderly. This is a rather different type of place to the large urban areas, often in the United States, that dominate many prior studies. We have discussed elsewhere the types of things people in this place find disorderly and the meaning they attribute to disorder (see Girling et al., 2025; Loader et al., 2025) – our focus here is on why some are more inclined to see disorder than others. Second, we foreground locality and place in our analysis. Taking advantage of a survey that sampled across the entire town, we can tie people not only to the neighbourhoods in which they live, but also, by definition, locate them all in the same place, that is, Macclesfield. Third, we broaden the range of potential predictors beyond crime and immediate local concerns to consider the wider social economic and political climates that shape people’s lives and outlooks.

Seeing disorder

Studies concerned with why people may or may not ‘see disorder’ (Sampson and Raudenbush, 2004) in their local areas tend to start from one of two different but not incompatible positions. On the first account, there is a ‘real’ level of disorder in a particular area, indicated by ‘objectively observable aspects of disorder such as garbage, broken bottles, litter, graffiti, abandoned cars, and drug paraphernalia’ (Sampson and Raudenbush, 2004: 321). Yet, people exposed to these cues are more or less likely to notice them, and/or code them as a problem, depending on their personal or social characteristics. Women, for example, may be more likely to see disorder because they perceive a greater threat from crime, and are thus more attuned to the character of their immediate environment (which can provide information on the level of threat) (Hipp, 2010). Here, variation in individual perception can be classified as a form of ‘error’, differential, more or less idiosyncratic, apprehensions of some underlying physical or social reality.

On the second account, regardless of whether it is possible to construe disorder as an empirical reality, some people are more inclined to code – and report – behaviours and situations as disorderly. Research has considered, for example, how experiences of crime shape perceptions of disorder (Mackenzie et al., 2010). Those with authoritarian attitudes may be more likely to see ‘teenagers hanging around’ as a problem (Jackson et al., 2018) because they are motivated to see the behaviour of young people as non-normative or transgressive. Those more sensitive or attuned to the presence of racial or ethnic minorities – for reasons of bias, prejudice or outright racism – may associate the presence of minorities with disorder, ‘read’ such presence as disorderly, and therefore report more disorder in areas with larger minority populations (Hinkle et al., 2023; Sampson and Raudenbush, 2004; Wickes et al., 2013). This account does not ultimately rest on an assumption that there is a ‘real’ level of disorder: here, disorder is about perception, all the way down.

Criminological interest has thus, usually, been focused on how and why differentially-situated individuals see disorder, and with the social meanings they attribute to its particular forms. This, in turn, is linked to associated political and policy discourses that have typically centred on defending neighbourhoods from undesirable ‘others’, for example via zero-tolerance or hot-spots strategies that prioritise police crackdowns on disorder. Disorder is thus ‘lifted’ from the range of troubles that may affect the liveability of a street, neighbourhood, town or city. It is given its own separate – and often prioritised – attention, and classified, broadly, as crime. And it is firmly positioned as an objective feature of local areas and properly the focus of sometimes aggressive intervention from state and other actors (Harcourt, 2005). The result is that the question of safer neighbourhoods gets cut off from the question of better neighbourhoods.

In this article, we are concerned less with the ontological status of disorder than with reconnecting perceptions of disorder with everyday understandings of place, and with the ways these link to wider social and political concerns. Previous studies have tended to assume that variation in perceptions of disorder stem from individual characteristics, and the way different people respond to the products of socio-structural processes (the presence or absence of crime, or of individuals or groups perceived as risky or threatening). Individuals who live in the same area, that is, perceive different levels of disorder because they have individual traits, characteristics and experiences that predispose them to do so. This remains our primary interest in this article – we ask, that is, why people who live in the same area, and who are thus exposed to a similar physical environment, may have differential propensities to identify visual and other cues in that area as problems of disorder. Yet, with the exception of the racial politics of the United States and elsewhere, often missing from prior accounts has been consideration of how people think about and assess the social and political forces that shape their physical (and social) environments. We seek to add such assessments to consideration of ‘seeing disorder’, with a particular focus on broader, non-criminological questions about people’s ability to live well, or at least bearably, within a particular place.

Clearly, people experience disorder somewhere, and in criminology this is generally conceived in terms of locality, neighbourhood and/or community. The experience of living in a place may thus affect people’s propensities to see disorder there, and a few extant studies have addressed this issue. Wallace et al. (2015), for example, found in Seattle that neighbourhood ‘behavioural attachment’ (knowing neighbours, watching neighbour’s property, etc.) was (inconsistently) associated with identifying disorder in their neighbourhood. In particular, watching neighbour’s homes was associated with a higher chance of reporting (that is, noting) a ‘disorder cue’ such as litter or trash (Wallace et al., 2015: 256). This study raises an important issue in many studies in this area, however. Given cross-sectional data, it is impossible to say whether it is that those who engage in informal social control (by watching out for neighbours) are more likely to identify disorder, or that those who see more disorder are more likely to engage in informal social control.

In this study we, too, rely on cross-sectional data. However, we conceptualise people’s relationship with place somewhat differently, as relating to satisfaction with life in the place one lives – whether it is a ‘good’ or a ‘bad’ place to be. Jackson et al. (2018) take a motivated cognition approach to perceptions of disorder. Assuming that psychological goals and motivations shape how people experience the world, interpret information and reach judgements (Jost et al., 2019), they consider how ‘instrumental’ motivations to understand and manage personal risk, and ‘relational’ motivations concerned with community cohesion and the importance of shared moral values, may influence perceptions of disorder. In a similar manner, we hypothesise that overall judgements of what it is like to live in a place – whether it meets one’s needs, provides an adequate standard of living, and is generally a good place to be – will predict perceptions of disorder. Due not least to the need to avoid cognitive dissonance, those who are satisfied with where they live will be motivated to see is as relatively free of disorder, while those who are dissatisfied will be motivated to see it has more prone to disorder.

Perceptions of disorder gain here an expressive quality. ‘Seeing’ disorder, and identifying it as a problem, is a way of saying something about place. As Whitehead et al. (2003: 4–5; in Millie, 2008: 382) note, ‘virtually any activity can be anti-social depending on a range of background factors, such as the context in which it occurs, the location, people’s tolerance levels and expectations about the quality of life in the area’ (emphasis added). Disorder can thus be used as a metaphor (Mackenzie et al., 2010) for other types of harm, wrong, or wider social breakdown or malaise: identifying one’s neighbourhood as disorderly is a way of saying something about it. Our first research hypothesis is thus the following:

• H1: Those who are more satisfied with Macclesfield as a place to live will be less likely to see disorder as a problem in their neighbourhood.

While people see disorder in a particular place, and may use this as a way of talking about that place, both disorder and their understanding of it will be shaped by the wider socio-economic context. For the current study, this context includes post-2010 austerity, Brexit, the aftermath of the Covid-19 epidemic and the cost-of-living crisis crystallized by the economic shock resulting from Russia’s invasion of Ukraine. This series of events have had profound impacts on the national and local economies of the United Kingdom as well as, of course, on ideas and hopes about the future and the trajectory of change at local and national levels. How and why might this wider context shape perceptions of disorder? We start from the premise set by a series of studies that have linked perceptions of disorder to trust in the police (e.g. Brown and Reed Benedict, 2002; Girling et al., 2000; Jackson and Bradford, 2009). These, in turn, link perceptions of disorder to underlying concerns about social change, and to questions of social cohesion and fragmentation. In particular, when people experience society as becoming less cohesive, they reach for the metaphor of disorder (usually positioned as a result of the behaviour of denigrated others – Mackenzie et al., 2010: 10) to help explain their concerns. Disorder stems from what might be termed social failure – a collective inability to maintain order – and trust in the police, as supposed guardians of society, suffers when people experience their environment as disorderly (Jackson and Bradford, 2009).

Such ‘social breakdown’ is clearly underpinned by economic forces. Failure to generate and maintain order and cohesion in local areas stems from issues of chronic under-resourcing, austerity budgets, and the decline of local government (see inter alia Morenoff et al., 2001; Sampson et al., 1997; Sampson and Wilson, 1995). Poverty (and crime) at the area level is linked to perceptions as well as incidents of disorder (Skogan, 1990; Steenbeek and Hipp, 2011; Wickes et al., 2013), and it may also be that the experience of economic stress primes people to see more disorder in their immediate physical environment. Disorder becomes, again, a metaphor, a way of saying something about the experience of living in a place – an experience which is inevitably shaped by economic, as well as social, capital. Our second hypothesis is, therefore, that

• H2: People in a more economically precarious position will be more likely to identify disorder as a problem in their neighbourhood.

Moreover, just as disorder is used as a metaphor for social and economic stress and decline, it may also be used as a way to think and talk about politics. To the extent that people blame the political process for the economic stress they are experiencing, or for perceived social and economic ills in a wider sense, this, too, might cause them to reach for the metaphor of disorder, and to be more likely to code events in their social and physical environments as disorderly (even as others, who live in the same place but are less exercised by political failures, do not). It may also be the case that disillusionment, disengagement and an associated loss of political efficacy undermines people’s sense that they have some control – through their elected representatives – over their physical and social environment. This could further motivate or predispose them to construe that environment as disorderly. We therefore hypothesise that disillusionment with and disengagement from politics will be associated with people’s propensities to see disorder, at least in part because seeing disorder is a way to name and attribute blame. ‘They’ have left things to decay and decline. Our third hypothesis has two parts:

• H3A: People who are more politically disillusioned will be more likely to identify disorder as a problem in their neighbourhood.

• H3B: Those who are more politically disengaged will be more likely to identify disorder as a problem in their neighbourhood.

Naturally, a wider sense of disillusion that spreads beyond politics (narrowly defined) might also be important. Research linking perceptions of disorder to trust in the police suggests that disorder indicates instrumental and symbolic failures on the part of police to adequately maintain civility and order. A sense of abandonment looms large in such accounts – that police and others are not present, not engaged, and thus do not care (Girling et al., 2000; Jackson and Bradford, 2009). Feeling police are absent, or at least less present than previously, may increase sensitivity to signs of disorder that, in turn, signals a lack of care, decay or danger. Simultaneously, noting – or perhaps ‘naming’ – disorder is a way of talking about that absence and lack of care. Our fourth hypothesis is, therefore, that

• H4: People who feel the police are absent from their community will be more likely to identify disorder as a problem.

Finally, it is plausible to suggest that economic distress and perceptions of political failure and abandonment affect the way people conceive of the place they live – their satisfaction with it – and through this their perceptions of disorder. Conceptions of place may thus mediate any association between the former and the latter, an idea that echoes the argument that people use understandings of place as, among other things, a linguistic device for saying things about the state of their lives, and the wider social and political world. Our fifth hypothesis is the following:

• H5: Satisfaction with place will mediate any associations between economic precarity, political disillusionment, police visibility and perceptions of disorder.

Naturally, the discussion above does not provide a full list of the potential perceptual and experiential predictors of ‘seeing disorder’. Perhaps most importantly, the ways people use space – their routine activities – might be important (Wallace et al., 2015). Those with children may be more attuned to the presence of disorder because they are more engaged with local facilities, the presence of young people, and so on (Hipp, 2010). Those who spend more time in their local neighbourhood – because they walk rather than use the car, for example – may be more likely to pick up cues that others miss. Women and older people may use spaces differently, and have different concerns, compared with men and younger people. While not a routine activity (for most), crime victimisation has also been linked to perceptions of disorder, and for similar reasons: recent victims of crime may be more attuned to signals of threat and danger (Mackenzie et al., 2010; Mellgren et al., 2010; Roccato et al., 2011). We therefore include measures representing victimisation and routine activities – and associated demographics – as control variables in our analysis.

Summary

In this article, we assume that perceptions of disorder are shaped by a range of factors associated with the ways people use, read and judge their immediate physical and social environment – the place that they live. Economic precarity, and perceptions of social, political and institutional abandonment, seem likely to shift conceptions of place and, at least in part through this, propensities to see disorder. Absent from our models are variables almost certain to correlate with perceptions of disorder but which have been positioned by other studies as outcomes of such perceptions, rather than predictors of them. These include trust in the police, concerns about crime, and perceptions of social cohesion/collective efficacy. In reality, all such concerns are likely to be mutually constituted with and by perceptions of disorder. But given the apparent centrality of perceived disorder in the formation of trust, fear of crime, and so on, our aim here is to consider what is it apart from these factors that attunes people to cues of disorder in their immediate physical and social environments.

Data and methods

Macclesfield is a mid-sized town in Cheshire, England, with a population of around 53,000 people. It is in some senses typical of many towns in its region and across the country. The remnants of old manufacturing industries sit alongside newer industries, while a significant number of commuters travel to nearby conurbations to work. There are pockets of significant deprivation, but it is generally considered a relatively affluent place (although less so than other nearby towns). The population is largely white.

As part of a larger mixed methods study of security, crime and disorder in the town, the opinion survey company ORS was commissioned to conduct two face-to-face surveys of Macclesfield residents in the summers of 2021 (n = 427) and 2022 (n = 502); we draw here on data from the 2022 wave. Addresses were sampled randomly from across the town. Pre-alert letters were sent to the preselected addresses, with interviewers subsequently making up to three visits to each address to secure an interview. Interviewees were selected at random from the people living at each address. The response rate was relatively low, at 20 percent. Within the 2022 sample: 49% were female; 23% aged under 35 and 30% aged 65 and over; 90% were of a white British ethnicity; 50% were in work, 7% unemployed, and 33% retired (the remainder included students and those not looking for work).

Crucially for current purposes the highly localised nature of the survey means that in the 2022 wave we have respondents in 117 of the town’s 180 Output Areas (OAs). OAs are census-based small area units; in Macclesfield the average population of an OA is around 300 people. Although they vary in size, most comprise only a few streets. It is reasonable to assume that all those living in an OA are exposed to a very similar neighbourhood environment. Yet, all respondents were clearly also from one particular place, Macclesfield. These two features of the survey frame the analysis presented below.

Dependent variable

Our dependent variable, perceptions of disorder, is a scale derived from nine items probing respondents’ views on ‘how big a problem’ a range of behaviours and issues are in their local area. These were drawn largely from the Crime Survey of England and Wales, and include teenagers hanging around on the streets, rubbish or litter lying around, drug dealing and speeding cars. To these we added some behaviours identified in the wider project as particularly relevant locally, most notably the way people park their cars. We used Confirmatory Factor Analysis (CFA) in the statistical package Mplus 7.2 to derive and validate this scale, as well as the others described below. ¹ See the Appendix Table for more details.

It is important to consider the implications of treating perceptions of disorder as a latent variable in this way. In essence, the CFA model proposes that people have an unobservable, latent, trait that can be characterised as their propensity to identify problems of disorder in their local area. This trait is measured by the observed indicators collected via the survey, that is, the items outlined above, but these items do not exhaust the types of situations and behaviours that could be seen as disorderly. It is the latent trait that ‘causes’ responses to the survey items, not the other way round. In this sense ‘perceptions of disorder’ refers, precisely, to the tendency to ‘see disorder’ without specifying what it is, exactly, that is seen as disorderly (although we might suppose that these will be the types of things identified by the observed indicators – it seems unlikely that anyone would code a well-tended and colourful road-side flower patch as disorderly, for example).

Independent variables

Place satisfaction is measured by a scale constructed from five survey items probing respondent’s sense that Macclesfield was a good place to live that provided them with facilities and opportunities (e.g. ‘I have access to shopping facilities’). For full item wordings see the Appendix Table. Alongside this, we added two dichotomous variables representing neighbourhood connections, which have been shown to be important for perceptions of disorder in previous studies: whether the respondent was born in Macclesfield (1 = yes) and whether they know their neighbours (1 = they knew the names of all their nearest neighbours).

Economic precarity was measured by summing two items: ‘Which of the descriptions on this card comes closest to how you feel about your household’s income nowadays?’ (four response categories ranging from ‘Living comfortably on present income’ to ‘Finding it very difficult on present income’); and ‘If for some reason you were in serious financial difficulties and had to borrow money to make ends meet, from the bank or family, how difficult or easy would that be?’ (five response categories, ranging from ‘Very easy’ to ‘Very difficult’). Higher scores in this measure represent greater economic precarity.

Political disillusionment and engagement were measured by two scales using items from the Political Efficacy Short Scale (Groskurth et al., 2021). Political disillusionment was measured by four items including ‘National politicians strive to keep in close touch with the people’ (reversed). Political efficacy – which we take to be an indicator of political engagement – was measured by two items, including ‘I am good at understanding and assessing important political issues’ – again, see the Appendix Table for full item wordings. Both scales were coded such that high equals more.

Perceptions of police presence were measured by two items indicating, first, how often respondents saw police in their area. This was generated by taking the mean of two items that asked about seeing police on foot (with six response categories ranging from ‘more than once a day’ to ‘never’) and in cars (with similar response categories). This variable was coded such that higher scores indicate seeing police more often. The second item represented whether respondents thought police were visible enough in their communities. This was generated from two items that followed the visibility questions, which asked whether the level of visibility was ‘enough’. In relation to police on foot, 69% said the current level of visibility was not enough; the equivalent figure for cars was 49%. We created a binary indicator scored 1 if a respondent said not enough in both cases (47% did so).

A number of different variables represented routine activities and related issues. Most took the form of binary indicators: whether children aged under 18 lived in the respondent’s household (1 = yes); whether they were a homeowner or not (1 = yes); and whether they had been a victim of crime in the past 12 months (1 = yes). To these we added gender (1 = female) and age (1 = aged over 65). Two additional measures represented how people moved about the town – how often they used the car and walked on foot to make journeys within Macclesfield. Both were entered as continuous variables with four levels; daily, at least once a week, at least once a month, rarely or never.

Contextual variables

To take some account of the ‘objective’ characteristics of the neighbourhoods we use the following measures from the 2021 Census and the 2019 Index of Multiple Deprivation. With the exception of crime, these were measured at the OA level. The measures are: household deprivation (the proportion of households deprived in at least two dimensions, as defined by the 2021 Census); the proportion of residents aged under 19 (Census); population density (Census); and whether the OA was in a high crime area (2019 IMD, defined as being within a Lower Super Output Area in the top two deciles for crime).

Analytical approach

To model perceptions of disorder we use linear random effects models, with the level 2 variable set to OA. By partitioning the variance in perceptions of disorder between area and individual, we look ‘within’ OAs to assess variation in perceptions among residents who live in the same small local areas who, we assume, are exposed to very similar physical environments.

Results

Table 1 shows results from the main analysis. We first estimated Model 0, not shown in the table, a variance components model with no predictors. This model had an intra-class correlation coefficient (ICC) of .17, suggesting around one-sixth of the variation in perceptions of disorder was explained at the OA level – by the characteristics of the area, not the people living there. Model 1 in Table 1 shows results from a model with only area-level predictors. Here, we find that perceptions of disorder tended to be higher in higher crime areas, in more densely populated areas, and in areas with more young people. Conditioning on these variables, the ICC reduces to .14.

Table 1.

Linear random effects models predicting perceptions of disorder.

	Model 1			Model 2			Model 3
	b	se(b)	B	b	se(b)	B	b	se(b)	B
Contextual variables
High Crime area	0.29*	0.13	0.12	0.25*	0.11	0.11	0.23*	0.10	0.10
Household deprivation	0.01	0.00	0.08	0.00	0.00	−0.01	0.00	0.00	0.01
Population aged under 19	0.02*	0.01	0.13	0.01	0.01	0.08	0.01	0.01	0.04
Population density	0.00**	0.00	0.18	0.00*	0.00	0.13	0.00*	0.00	0.11
Individual-level variables
Women (ref: men)				0.06	0.06	0.04	0.09	0.06	0.06
Older (ref: under 65)				−0.26***	0.07	−0.17	−0.24***	0.07	-0.16
Children under 18 at home (ref: none)				0.08	0.07	0.05	0.07	0.07	0.04
Home owner (ref: other tenure)				−0.04	0.07	−0.03	0.00	0.07	0.00
Walking journeys				0.00	0.03	0.00	−0.01	0.03	−0.01
Car journeys				0.01	0.03	0.01	−0.02	0.03	−0.03
Victim of crime (ref: no)				0.28*	0.11	0.10	0.25*	0.11	0.09
Political disillusionment				0.13**	0.04	0.13	0.03	0.04	0.03
Political efficacy				0.07+	0.04	0.08	0.13**	0.04	0.15
Economic precarity				0.10**	0.04	0.13	0.05	0.04	0.07
See police				0.15***	0.03	0.24	0.15***	0.03	0.24
Police not visible enough (ref: visible enough)				0.25***	0.06	0.17	0.24***	0.06	0.17
Place satisfaction							−0.38***	0.06	−0.29
Born in Macclesfield (ref: not)							−0.04	0.06	−0.03
Knows all neighbours (ref: does not)							−0.15*	0.06	−0.10
Constant	−0.75***	0.22		−1.20***	0.24		−0.84***	0.23
ICC	0.14			0.09			0.07
n	498			484			484

+

p < .1; *p < .05; **p < .01; ***p < .001.

Model 2 adds most of the individual-level predictors. Of the contextual variables only crime and population density retain significance at the 5% level in this model. The ICC also drops further, to .09, indicating that some of the area-level variation identified in Models 0 and 1 is actually explained by compositional effects, that is, the fact that different types of people live in different types of area. Considering the individual-level variables, we find, first, economic precarity, political disillusionment and political efficacy (although p = .07 in the latter case) were associated with perceptions of disorder. People in more economically precarious positions were more likely to see disorder in their local area, as were those who felt disillusioned by politics. But, contrary to expectations, those who felt more politically efficacious were perhaps also more likely to see disorder. Second, perceptions of police visibility are associated with perceptions of disorder. Both those who felt they saw the police more often and those who felt they did not see police enough were more likely to indicate that disorder was a problem in their neighbourhood. We return to these apparently contradictory results below. Third, among the (broadly defined) routine activities variables only age and victimisation were significant: recent victims of crime were more likely to see disorder, older people were less likely.

Model 3 in Table 1 adds conceptions of place. Notably, place satisfaction has a large and strongly significant association with perceptions of disorder. Those who were satisfied with Macclesfield as a place to live were substantially less likely to perceive disorder in their neighbourhoods. Respondents who said they knew all their neighbours were also less likely to see disorder, but whether someone was born in the town or not appeared to make no difference. Once conceptions of place were added to the model there are substantive changes in relation to other variables. Notably, the coefficients for economic insecurity and political disillusionment shrink in size and lose significance (p > .1 in both cases), while the coefficient of political efficacy grows in size, and the p value shrinks (p = .001). This suggests that conceptions of place, primarily place satisfaction, ² may mediate the association between these variables and perceptions of disorder, a point we pick up below. By contrast, the other variables associated with perceptions of disorder (age, victimisation and police visibility) are barely changed in Model 3.

The mediating role of place satisfaction

To further explore the potential mediating role of place satisfaction, we step out of the multilevel context and estimate a path model using the variables of interest from Model 3 in Table 1: perceptions of disorder, place satisfaction, economic precarity, political disillusionment and political efficacy, and police visibility. Results from this model are shown in Figure 1, and support the idea that place satisfaction plays the mediating role suggested.

Figure 1.

Path model exploring the mediating role of place satisfaction.

We find, first, that place satisfaction was associated with economic precarity, political disillusionment and political efficacy, and in predictable directions. Those in more precarious positions and who were more disillusioned with politics were less likely to be satisfied with the place they live; those who felt more politically efficacious were more likely to be satisfied. By contrast, though, police visibility was not associated with place satisfaction (but retained its direct association with perceptions of disorder). While in this model direct statistical effects of economic precarity as well as political efficacy on perceptions of disorder persist even conditioning on place satisfaction, ³ we also find significant indirect effects (IE), via place satisfaction, from economic precarity (IE = .07; p < .0005), political disillusionment (IE = .11; p < .0005) and political efficacy (IE = −.06; p < .0005) (note that the total effect of political efficacy was .07; p = .11, indicating that the positive direct association with perceptions of disorder was effectively cancelled out by the negative indirect effect). There is good evidence, then, that place satisfaction channels some of the statistical effect of all three variables towards perceptions of disorder.

Discussion

To return to our research hypotheses, we find strong support for the idea that those who were more satisfied with Macclesfield as a place to live were less likely to see disorder as a problem in their neighbourhood (H1). Similarly, people in a more economically precarious position, and those who were politically disillusioned, were also more likely to see disorder as a problem (H2 and H3A supported). Our findings in relation to H3B are more complex, but, overall, it seems that the more politically engaged were more likely to see disorder in their neighbourhood, the reverse of the hypothesised relationship. Similarly, while feeling that police were not visible enough was associated with a greater probability of seeing disorder, so too was greater reported police visibility (H4 partially supported). Finally, satisfaction with place mediated the associations between economic precarity and political disillusionment and perceptions of disorder, but not those of political engagement and police visibility (H5 partially supported).

Turning to the other variables in our models, we find only limited evidence of an association between routine activities and perceptions of disorder. Of the broad set of indicators we included under this banner, only victimisation was consistently positively associated with seeing disorder, and one might imagine this is less to do with routine activities per se than a sensitivity to cues of disorder that is heightened by recent victimisation. Perhaps surprisingly, older people were less likely to see disorder than younger people.

Taken together, and remembering that our analysis looked within small local areas to explore the views of residents likely to be exposed to very similar environmental conditions, these results lend support to the idea that ‘seeing disorder’ is in important ways an expressive function of wider social, political and economic experiences. In a proximate sense, identifying and talking about disorder may be a way of expressing dissatisfaction with place; it may also be that such dissatisfaction makes one more sensitive to cues that can be labelled disorderly. In an arguably more distal sense, economic troubles and political disillusionment may have similar effects, triggering or exacerbating a sense of dissatisfaction with place that finds expression in the identification of disorderly conduct within it. Our findings underline the idea that people experience disorder (or do not) in a particular place, and their understanding of that place and how it came to be the way that it is are likely to be important factors in their judgements that it is, or is not, ‘disorderly’.

Some of the findings here need further reflection. We found that political efficacy was associated with seeing more disorder, not less. While one might expect that a sense of political disempowerment would trigger a similar feeling as that raised by disillusionment, that does not seem to be the case. One possibility here is that those who feel politically engaged and efficacious in a place like Macclesfield will also tend to be the type of people who particularly care about it. All else equal, they may therefore be more sensitive to cues of disorder – perhaps most obviously littering and associated issues – that indicate that the place is heading in the wrong direction, decaying, or is under threat in some way. We found some support for this interpretation in our qualitative work (Girling et al., in press).

We also found associations between perceptions of policing and perceptions of disorder that were potentially contradictory. Those who felt they saw police more often were more likely to see disorder, but so also were those who felt they did not see police enough. Recalling again that we are looking inside small local areas – where one might imagine that the actual level of policing is fairly constant – two different things may be going on here. On the one hand, seeing ‘more police’ and ‘more disorder’ might both be indications of – indeed may co-constitute – an underlying sense of threat or danger. Or perhaps seeing police in itself indicates danger, increasing sensitivity to cues of disorder. On the other hand, feeling that one does not see police enough may trigger an expressive response to local conditions, motivating the identification of disorder in order to provide a reason for why this is a problem. All that said, it may be that the police visibility variables are tapping into processes occurring within OAs, where there may be specific locales (‘micro-places’) where there is both more disorder and more police presence; presence that, in turn, might be construed as inadequate as the disorder is (still) occurring. More research is needed to unpick these associations, but they indicate a complex relationship between police visibility – which has often been viewed as a sign of order, or at least the attempt to assert it – and perceptions of disorder.

Limitations

This study has, naturally, a number of limitations. First, it relies on cross-sectional data, and cannot therefore tap into or even approximate causal processes. Second, the set of explanatory variables is inevitably limited – and the sample size itself on the small side, limiting what can be attempted with it. Third, while it is arguably representative of a broad swathe of ‘middle England’, Macclesfield is just one particular place. It is easy to imagine that people living there think differently about some of the issues we address above than others living elsewhere, particularly in larger, more diverse urban areas (and indeed more rural locales). Future research could usefully address all these issues, and consider whether the associations we have identified and explore can be found in other types of places, and working with other types of data.

Conclusion

In line with previous studies, in this article we have shown that perceptions of disorder are linked to experiences of place and in turn to wider social, economic and political processes. Yet, we have also extended prior research by suggesting that those who feel – and indeed are – economically and politically marginalised, ‘let down’ and dissatisfied by and/or in the place they live are significantly more likely to ‘see disorder’ there. On this basis, it is plausible to suggest that economic processes drive not just the occurrence of disorder (as litter goes uncollected and vandalism unrepaired, for example) but also people’s propensity to see it (as austerity impoverishes people and local economies). To the extent that people use disorder as a metaphor to say something about place, change and institutional context, we can begin to see how individual perceptions of space and place are reflexively shaped by socio-structural processes operating at both local and supra-local levels.

To return to where we started this article, it seems to us that this provides a route into better understanding fear of crime, trust in police, and the other variables known to be closely associated with perceptions of disorder. On the account offered above, perceptions of disorder are not ‘free-floating’, idiosyncratic perceptions detached from concrete social and physical processes. But neither are they determined by those processes. Rather, there is a complex interplay between the objective nature of a place, the social and other forces that make it the place that it is and the ways people experience it which, together, shape perceptions of disorder and therefore, perhaps, the other variables consistently linked with perceived disorder. We can only understand ‘disorder’, ‘fear’ and ‘trust’, that is, if we understand both where people live and how they think about and make sense of their lives.