Skip to main content

Neighbourhood crime and smoking: the role of objective and perceived crime measures



Smoking is a major public health problem worldwide. Research has shown that neighbourhood of residence is independently associated with the likelihood of individuals' smoking. However, a fine comprehension of which neighbourhood characteristics are involved and how remains limited. In this study we examine the relative contribution of objective (police-recorded) and subjective (resident-perceived) measures of neighbourhood crime on residents' smoking behaviours.


Data from 2,418 men and women participating in the 2007/8 sweep of the West of Scotland Twenty-07 Study were analyzed. Smoking status and perceived crime were collected through face-to-face interviews with participants. Police-recorded crime rates were obtained from the Scottish Neighbourhood Statistics website at the datazone scale. Adjusted odds ratios and 95% confidence intervals were estimated for the likelihood of current smoking using logistic regression models. Adjusted mean daily amount smoked and F statistics were calculated using general linear models. Analyses were conducted for all respondents and stratified by sex and age cohort.


Compared to individuals living in low crime areas, those residing in an area characterized by high police-recorded crime rates or those perceiving high crime in their neighbourhood were more likely to be current smokers, after controlling for individual characteristics. The association with smoking was somewhat stronger for police-recorded crime than for perceived crime. Associations were only slightly attenuated when adjusting for either the objective or subjective crime measures, suggesting that these indicators may exert an independent influence on the risk of smoking. Stronger effects were observed for women compared to men. Police-recorded crime rates were more strongly related to smoking status among older respondents than among the younger cohort, whereas the strongest effect for perceived crime was observed among younger participants.


Our findings highlight the relevance of paying attention to both objective and perceived neighbourhood crime measures when aiming to prevent smoking.

Peer Review reports


Smoking is a major public health challenge and the principal risk factor for cancer, cardiovascular diseases and tuberculosis. Together, these are responsible for 70% of all deaths that occur in adults over 30 years-old worldwide[1]. As such, smoking is the main modifiable cause of premature death [1] and exposures that influence smoking and could be targeted in preventive efforts should be identified.

It is increasingly recognized that where people live influences their health behaviours, including smoking [211] and there have been calls for a better understanding of the mechanisms involved in producing these "place effects" [10]. A small number of studies have found aspects of the local neighbourhood such as safety and crime to be related to smoking [8, 1117] even after controlling for residents' individual characteristics. Crime and safety measures represent socio-cultural features of neighbourhoods which may act as chronic stressors [18] and as such, may influence smoking through pathways such as stress or psychological well being.[10, 19] Most studies of crime-related exposures and smoking have investigated subjective measures of safety as perceived by respondents [8, 1215], or municipal authorities [17] while few have relied on more objective measures of crime such as rates recorded by police authorities [11, 16]. We borrow from Weden et al. (2008) the terms subjective measure which refers to an "individual-level assessment of a resident's neighbourhood" and objective measures which refer to "area-level indicators that can be characterized independent of a resident's perception" of his neighbourhood [[19] p.1257]. Since objective and subjective crime indicators may be tapping into different concepts [19, 20], it appears relevant to study both exposures in the same study. We found no study on smoking that had done so and thus sought to fill this gap found in the literature by studying how both objective (police-recorded) and subjective (resident-perceived) measures of crime influenced smoking, as well as whether, and if so how, each altered the other's association with the likelihood of smoking and with smoking intensity.

The study's objective was to investigate the association between smoking behaviour and neighbourhood crime measured in two ways: as recorded by police and as perceived by participants. Associations were investigated separately for men and women since some studies have found gender differences in the magnitude of associations between individual health and health behaviours and experience or perception of neighbourhood conditions [5, 8, 2124]. Analyses stratified by age group were also conducted since neighbourhood exposures may be perceived differently by younger versus older individuals and exert a differential effect on their health [23, 25]. For example, younger working-age cohorts may be more mobile than older groups. As such they may be less exposed to, and influenced by, their residential environment than less mobile groups [26, 27].


Study population

The Twenty-07 Study has been following people in three age cohorts (born around 1932, 1952 and 1972) sampled from the Central Clydeside Conurbation, West of Scotland, for 20 years [28]. In this paper, data from the fifth sweep of the study conducted in 2007/8 were used (n = 2,604 respondents, of whom 2,459 resided in the Greater Glasgow area at the time of interview and were considered for the present analyses). Respondents who participated at baseline have been shown to be representative of the general population of the sampled area [29]. Ethics approval was gained for each wave from the NHS and/or University of Glasgow Ethics Committees. All analyses were conducted in the summer 2010.

Individual-level variables

Individual-level measures were collected through face-to-face interviews conducted in participants' home by trained nurse interviewers. Information was collected on participants' socio-demographic characteristics (education, income, occupation, marital status), health conditions and behaviours including smoking, alcohol intake and physical activity, and everyday life including respondents' perceptions of neighbourhood conditions. Age was operationalized as a categorical variable corresponding to the study cohort (aged around 35, 55 and 75 when interviewed in 2007/8). Socio-economic status (SES) was operationalized with the head of household's social class defined using the 1980 Registrar General's classification: I (professional), II (intermediate), IIINM (skilled non-manual), IIIM (skilled manual), IV (partly skilled manual) and V (unskilled manual) [30]. Employment status indicated whether the participant was in employment or not at the time of interview. The unemployed included the retired, women on maternity leave and participants temporarily or permanently sick. The final sample for analysis is composed of 2,418 individuals (1,073 men and 1,345 women) for whom complete data for the variables of interest were available.

Perceived crime was used as an individual, resident-based subjective indicator of neighbourhood crime. It was measured by summing, for each participant, responses to the question: "Around where you live would you say that any of the following problems exist?" and included the items vandalism, burglaries, assaults and disturbances by children and youth. Problems were measured on a three-point scale using "not a problem" (score 1), "a minor problem" (score 2) and "a major problem" (score 3). A score was constructed by summing responses to each of the four items and scores were subsequently divided into quartiles, following their natural distribution in the sample. Higher scores indicated higher perceived crime.

Neighbourhood-level variables

Neighbourhoods were operationalized as datazones, the key small area statistical geography in Scotland. Scotland is divided into 6,505 datazones comprising on average 500 to 1,000 inhabitants. Datazones have been created using 2001 Census data so as to be relatively socially homogeneous. In general, they also respect physical boundaries and natural communities [31]. Participants in our sample were distributed across 1,256 datazones with an average of two individuals per zone (ranging from one to 12).

Crime data for the years 2007-2008 were extracted from the Scottish Neighbourhood Statistics database [32]. This publicly available database contains data from various sources including the Scottish Census, police records and population surveys. Crime rates per 10,000 population were available at the datazone scale for selected crimes of violence, domestic housebreaking, vandalism, drug offences and minor assault. Total area-level crime rates were attributed to participants based on their residential datazone and were categorized as low, medium and high rates based on tertiles created using the entire sample distribution.

Outcome variables

Participants were categorized as never, ex, or current smokers using the questions "Do you ever smoke tobacco now? (including pipe, cigars and roll ups)", and "If no, did you ever use to smoke any sort of tobacco?". Current smokers went on to report the number of cigarettes or cigars they smoked per day or per week for non-daily smokers, which we refer to as 'intensity'. Daily smoking intensity was analyzed as a continuous variable.

Statistical analyses

Clustering of smoking across datazones was found not to be statistically significant in our sample. First, the intra-class coefficient from a null multilevel model including only smoking and datazones was non-significant. Second, results from logistic regression models and generalized estimating equation models with robust standard errors (often used to analyze correlated data [33]) were compared. Odds ratios and standard errors were nearly identical between the two sets of results, again suggesting non-significant clustering. We thus present results from logistic regression and general linear analyses carried with SPSS v.16.0. Logistic regression models were used to investigate the likelihood of being a current smoker as compared to a non-smoker (combining never and ex-smokers), given exposure to different levels of crime. The lowest crime category was used as a reference in all analyses. Linear regression models were used to estimate the adjusted mean number of cigarettes smoked daily across crime categories among current smokers.

Models were built in a step-wise fashion, separately for exposure to objective and perceived crime. Crime measures were entered first, followed by individual socio-demographic variables (sex, age cohort, SES, employment status). Perceived crime was entered last in the crime rate model, and vice versa. Analyses stratified by sex and age cohort were conducted to explore modification by these variables.


Table 1 presents descriptive statistics for participants according to their smoking status. Smoking prevalence was higher among the younger cohorts than among participants in the oldest cohort. There was a statistically significant gradient according to SES and neighbourhood crime rate, with higher proportions of current smokers belonging to lower social classes and residing in neighbourhoods characterized by higher crime rates. A significantly higher proportion of participants perceiving high crime in their neighbourhood were current smokers compared to those perceiving low crime in their area, among both males and females. Objective and perceived crime measures were not strongly correlated with Pearson correlation coefficients ranging between 0.20 among women and 0.25 among men (data not shown).

Table 1 Characteristics of study participants by smoking status

Table 2 presents odds ratios and 95% confidence intervals for the association between neighbourhood crime measures and the likelihood of being a current smoker. Although interaction terms between crime measures and gender were statistically significant for perceived crime only (p value at the 0.05 significance level = 0.035, data not shown), we present results for all respondents and for men and women separately for both exposures, for reasons mentioned in the introduction. Adjusted Nagelkerke R2 are presented as indicators of model goodness-of-fit. After adjusting for individual characteristics (model 2), residents living in areas characterized by high or medium crime rates had a statistically significant higher likelihood of being current smokers, compared to residents of low crime areas. The effect was slightly stronger among women than men for residents of high crime areas (O.R (95% C.I) of 2.09 (1.48-2.96) and 1.75 (1.21-2.54) respectively). Adjusting for perceived crime (model 3) only reduced the association by about 1% in men and 5% in all respondents and in women alone and results remained statistically significant. Participants perceiving the highest level of crime had a higher likelihood of being a smoker (O.R (95% C.I.) of 1.43 (1.11-1.84)), compared to those perceiving the lowest level of crime in their area, after accounting for individual characteristics (model 2). After stratifying by sex, the association remained statistically significant in women only (O.R (95% C.I.) of 1.51 (1.07-2.13)). Adjusting for police-recorded crime rate reduced the effect of perceived crime (model 3). Odds ratios still remained above 1.0, ranging between 1.12 and 1.36, but results were below the significance level.

Table 2 Association between crime measures and the likelihood of being a current smoker, ORa (95% CI)

Results of analyses stratified by age cohort are shown in Table 3. After adjusting for individual characteristics, participants from all three age cohorts living in a neighbourhood characterized by high police-recorded crime had a higher likelihood of being a current smoker, compared to those residing in low crime areas. Adjusting for perceived crime reduced the effect of crime on smoking among participants aged 30-50 years-old and rendered it non-significant. Perceived crime did not considerably affect the magnitude and significance level of odds ratios for participants aged between 50 and 70 years or those older than 70 years (O.R (95% C.I.) of 2.16 (1.45-3.24) and 2.29 (1.20-4.36) respectively). For measures of perceived crime, the only statistically significant results were found for the youngest cohort (fully adjusted O.R (95% C.I.) of 1.63 (1.04-2.55)).

Table 3 Association between crime measures and the likelihood of being a current smoker, by age cohort, ORa (95% CI)

In terms of smoking intensity, Table 4 shows the F statistics and p values for significance of the difference between mean number of cigarettes smoked (unadjusted and adjusted) across exposure categories. Unadjusted models suggested that respondents living in areas with high crime rates or who perceived high crime in their neighbourhood smoked on average one to two and a half cigarettes per day more than those living in low crime areas. None of the associations between police-recorded or resident-perceived crime and smoking intensity were statistically significant at the 0.05 level, save for the unadjusted association between police-recorded crime rate and smoking intensity among all respondents. Analyses stratified by age cohort suggested that respondents aged 30-50 years-old and living in high police-recorded crime areas smoked on average more cigarettes than those residing in low or medium crime areas (statistically significant fully adjusted means of 12.24, 11.93 and 14.92 cigarettes/cigars per day for residents of low, medium and high crime areas respectively (data not shown)). Among the older two cohorts, none of the associations between police-recorded crime rate and smoking intensity reached statistical significance at the 0.05 level (data not shown). A high level of perceived crime was not associated with smoking intensity in models stratified by age cohort (data not shown).

Table 4 Mean number of cigarettes/cigars smoked per daya


In this study, residents living in areas characterized by high and medium police-recorded crime rates were more likely to be current smokers than residents of low crime areas. Similarly, individuals perceiving high crime in their neighbourhood were more likely to be smokers than those perceiving low crime levels. These associations remained statistically significant even after adjusting for individual characteristics. Odds ratios were faintly reduced after adjusting for either the objective or subjective crime indicator, suggesting that these might exert an independent effect on the likelihood of smoking. However, the perceived crime-smoking association was slightly more attenuated on adjustment for objective crime than the reverse, suggesting that part of the association between perceived crime and smoking is attributable to actual police-recorded crime rates in the neighbourhood. The association between both crime measures and smoking was more pronounced among women than men. A number of studies have previously found that area characteristics may be more strongly associated with women's smoking [11] or health in general [24]. This could be due to women being more aware of, or sensitive to, what happens in their neighbourhood [5]. Alternatively, women might be more exposed to their local area than men due to their spending more time there, on a daily basis [24]. In our study, high police-recorded crime rates were more strongly associated with being a current smoker among older respondents than among the younger cohort. A differential effect of neighbourhood exposure on the health of various age groups has been documented in other studies [34, 35]. It could be attributable to older individuals being more exposed to actual crime rates in their residential neighbourhood given that they are often less mobile than the younger, working-age groups [26]. Older participants may also have had a longer lifetime exposure to their neighbourhood due to accumulating a longer residence time. Unfortunately, we did not have any information on the time participants actually spent in their neighbourhood on a daily or weekly basis, and we only had limited information relative to length of residence. Thus, we could not adjust for these variables in the models. We found the strongest effect for highest perceived crime to be among the younger participants. This could be due to younger groups' perceptions being more influenced by other factors, for example the media. Daily smoking intensity was slightly higher among residents of high crime areas or perceiving high crime in their neighbourhood in all groups except for the older age cohort. However, none of these results reached statistical significance save for the association between high police-recorded crime rate and smoking intensity among the youngest cohort.

This study is one of the few to have investigated the association between objective (police-recorded) and subjective (resident-perceived) crime measures and smoking. Our results concord with two studies which had reported an association between police-recorded crime and smoking [11, 16]. The current findings of an association between high perceived crime and a higher likelihood of being a smoker add to the equivocal results found in the literature. Indeed, they concord with results from two studies having investigated perceived safety aggregated at the area level [17] or as an individual-based indicator [8] while two other reports had not found perceived safety to be associated to smoking [12, 14].

To our knowledge, this study is the first to have examined the relative contribution of objective and subjective crime measures to the likelihood of smoking. Both indicators were associated with smoking status, however, when adjusting the crime rate-smoking association for perceived crime, or vice versa, estimates were slightly modified towards the null. This finding does not lend support to strong confounding, especially by the subjective crime measure of the association between objective crime and smoking. This is understandable in light of the low correlation coefficients between objective and perceived crime measures. This weak correlation mirrors that found in other studies [20, 36] and suggests that these indicators may exert an independent influence on the risk of smoking. Crime rate and perceived crime are both chronic environmental stressors, defined as "insidious, with a slow and imperceptible onset and an open ended recurring character" [[37] p.2605], experienced at the neighbourhood level. They may be associated with stress and depression [37, 38] and therefore influence smoking through stress-related mechanisms. However, objective crime rates may measure structural features not necessarily perceived by residents. Conversely, residents may not all be aware of, or influenced by, crime incidents in their local area defined with administrative boundaries [19, 20]. As well, perceptions may be influenced by various factors other than actual crime activity such as residential tenure [25]. Thus, each crime indicator may measure neighbourhood features not encompassed by the other and capture neighbourhood characteristics which influence smoking through different mechanisms. For example, subjective measures of crime might be more proximal causes of stress and psychosocial disorder influencing health [19] and smoking, while objective crime rates may reflect other health-influencing neighbourhood features such as deprivation or lack of resources and amenities which might not be perceived negatively by residents [19] but which have also been found to be related to smoking.

This study presents a number of advantages and limitations. Important advantages include the large sample size, which allowed for exploring effect modification by sex and age cohort, and complete data for neighbourhood-level crime rates. Limitations are that smoking status was self-reported and determined using a question combining cigarettes and cigars, which could result in error in the measure of our outcome. However, prevalence data on cigarette versus cigar smoking indicate that cigar smoking is low in this population [39] so this may have a relatively minor influence on our results. In addition, former and never smokers were combined in the reference category for logistic models, and this may have led to exposure misclassification, diluting the effect of crime measures on smoking. Another limitation resides in a potential discrepancy between timing of exposure to crime and timing of its effect on smoking. Since no information was found concerning the lag time between exposure to crime and its effect on smoking, crime rates for the years 2007-2008 (i.e. when respondents were interviewed) were used. However, high correlations (>0.90) were found between 2007 and 2008 crime rates and those from earlier years (2001-2004) suggesting that crime rates are considerably stable over time. Therefore, exposure measured in 2007-2008 may be representative of an exposure having taken place earlier. An additional limitation concerns neighbourhood definitions. Although datazones are considerably small areas, they might not correspond to the neighbourhood respondents actually use and are exposed to, leading to exposure misclassification. In future research, more precise measures of exposure should be aimed for, for example by measuring exposure to crime within an area defined by participants, by taking into account the time they spend in their neighbourhood or by considering their exposure to other places where they spend time [40]. Finally, the data we used were cross-sectional, hampering the identification of a causal association between crime and smoking.


In conclusion, high levels of police-recorded and resident-perceived crime measures were found to be associated with a greater likelihood of being a current smoker, over and above participants' individual characteristics. Attention should be paid to neighbourhood-level exposures, both objective and perceived by residents, to reduce smoking prevalence. Policy makers should thus focus on creating residential neighbourhood conditions which are less prone to crime, but which also make residents feel safe independently of actual crime rates in their area.


  1. Jha P: Avoidable global cancer deaths and total deaths from smoking. Nat Rev Cancer. 2009, 9: 655-664. 10.1038/nrc2703.

    Article  CAS  PubMed  Google Scholar 

  2. Ahern J, Galea S, Hubbard A, Syme SL: Neighborhood smoking norms modify the relation between collective efficacy and smoking behavior. Drug and Alcohol Dependence. 2009, 100: 138-145. 10.1016/j.drugalcdep.2008.09.012.

    Article  PubMed  Google Scholar 

  3. Chaix B, Guilbert P, Chauvin P: A multilevel analysis of tobacco use and tobacco consumption levels in France: are there any combination risk groups?. Eur J Public Health. 2004, 14: 186-190. 10.1093/eurpub/14.2.186.

    Article  PubMed  Google Scholar 

  4. Diez Roux AV, Merkin SS, Hannan P, Jacobs DR, Kiefe CI: Area characteristics, individual-level socioeconomic indicators, and smoking in young adults: the coronary artery disease risk development in young adults study. Am J Epidemiol. 2003, 157: 315-326. 10.1093/aje/kwf207.

    Article  CAS  PubMed  Google Scholar 

  5. Ellaway A, Macintyre S: Are perceived neighbourhood problems associated with the likelihood of smoking?. J Epidemiol Community Health. 2009, 63: 78-80.

    Article  CAS  PubMed  Google Scholar 

  6. Echeverría S, Diez-Roux AV, Shea S, Borrell LN, Jackson S: Associations of neighborhood problems and neighborhood social cohesion with mental health and health behaviors: The Multi-Ethnic Study of Atherosclerosis. Health Place. 2008, 14: 853-865. 10.1016/j.healthplace.2008.01.004.

    Article  PubMed  Google Scholar 

  7. Frohlich KL, Potvin L, Gauvin L, Chabot P: Youth smoking initiation: disentangling context from composition. Health Place. 2002, 8: 155-166. 10.1016/S1353-8292(02)00003-5.

    Article  PubMed  Google Scholar 

  8. Miles R: Neighborhood disorder and smoking: findings of a European urban survey. Soc Sci Med. 2006, 63: 2464-2475. 10.1016/j.socscimed.2006.06.011.

    Article  PubMed  Google Scholar 

  9. Pearce J, Hiscock R, Moon G, Barnett R: The neighbourhood effects of geographical access to tobacco retailers on individual smoking behaviour. J Epidemiol Community Health. 2009, 63: 69-77.

    Article  CAS  PubMed  Google Scholar 

  10. Pearce J, Barnett R, Moon G: Sociospatial inequalities in health-related behaviours: pathways linking place and smoking. Prog Human Geography

  11. Virtanen M, Kivimäki M, Kouvonen A, Elovainio M, Linna A, Oksanen T, Vahtera J: Average household income, crime, and smoking behaviour in a local area: the finnish 10-town study. Soc Sci Med. 2007, 64: 1904-1913. 10.1016/j.socscimed.2007.01.005.

    Article  PubMed  Google Scholar 

  12. Ganz M: The relationship between external threats and smoking in central Harlem. Am J Public Health. 2000, 90: 367-371.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Johnson S, Solomon B, Shields W, McDonald E, McKenzie L, Gielen A: Neighborhood violence and its association with mothers' health: assessing the relative importance of perceived safety and exposure to violence. J Urban Health. 2009, 86: 538-550. 10.1007/s11524-009-9345-8.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Parkes A, Kearns A: The multi-dimensional neighbourhood and health: a cross-sectional analysis of the Scottish Household Survey, 2001. Health Place. 2006, 12: 1-18. 10.1016/j.healthplace.2004.03.004.

    Article  PubMed  Google Scholar 

  15. Patterson JM, Eberly LE, Ding Y, Hargreaves M: Associations of smoking prevalence with individual and area level social cohesion. J Epidemiol Community Health. 2004, 58: 692-697. 10.1136/jech.2003.009167.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Tseng M, Yeatts K, Millikan R, Newman B: Area-level characteristics and smoking in women. Am J Public Health. 2001, 91: 1847-1850. 10.2105/AJPH.91.11.1847.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. van Lenthe FJ, Mackenbach JP: Neighbourhood and individual socioeconomic inequalities in smoking: the role of physical neighbourhood stressors. J Epidemiol Community Health. 2006, 60: 699-705. 10.1136/jech.2005.043851.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Nielsen L, Curtis T, Kristensen TS, Rod Nielsen N: What characterizes persons with high levels of perceived stress in Denmark? A national representative study. Scand J Public Health. 2008, 36: 369-379. 10.1177/1403494807088456.

    Article  PubMed  Google Scholar 

  19. Weden MM, Carpiano RM, Robert SA: Subjective and objective neighborhood characteristics and adult health. Soc Sci Med. 2008, 66: 1256-1270. 10.1016/j.socscimed.2007.11.041.

    Article  PubMed  Google Scholar 

  20. Oh AY, Zenk SN, Wilbur J, Block R, McDevitt J, Wang E: Effects of perceived and objective neighborhood crime on walking frequency among midlife African American women in a home-based walking intervention. J Phys Act Health. 2010, 7: 432-441.

    PubMed  Google Scholar 

  21. Ross CE: Walking, exercising, and smoking: does neighborhood matter?. Soc Sci Med. 2000, 51: 265-274. 10.1016/S0277-9536(99)00451-7.

    Article  CAS  PubMed  Google Scholar 

  22. Ellaway A, Macintyre S: Geographies of women's health. International Studies of Women and Place. Edited by: Dyck I, Lewis N, McLafferty S. 2001, London, New York: Routledge, 265-281.

    Google Scholar 

  23. Ellaway A, Macintyre S, Kearns A: Perceptions of place and health in socially contrasting neighbourhoods. Urban Studies. 2001, 38: 2299-2316. 10.1080/00420980120087171.

    Article  Google Scholar 

  24. Stafford M, Cummins S, Macintyre S, Ellaway A, Marmot M: Gender differences in the associations between health and neighbourhood environment. Soc Sci Med. 2005, 60: 1681-1692. 10.1016/j.socscimed.2004.08.028.

    Article  CAS  PubMed  Google Scholar 

  25. Roman CG, Chalfin A: Fear of walking outdoors: a multilevel ecologic analysis of crime and disorder. Am J Preventive Med. 2008, 34: 306-312. 10.1016/j.amepre.2008.01.017.

    Article  Google Scholar 

  26. Yen IH, Michael YL, Perdue L: Neighborhood environment in studies of health of older adults: a systematic review. Am J Preventive Med. 2009, 37: 455-463. 10.1016/j.amepre.2009.06.022.

    Article  Google Scholar 

  27. Rainham D, McDowell I, Krewski D, Sawade M: Conceptualizing the healthscape: contributions of time geography, location technologies and spatial ecology to place and health research. Soc Sci Med. 2010, 70: 668-676. 10.1016/j.socscimed.2009.10.035.

    Article  PubMed  Google Scholar 

  28. Benzeval M, Der G, Ellaway A, Hunt K, Sweeting H, West P, Macintyre S: Cohort profile: west of Scotland Twenty-07 study: health in the community. Int J Epidemiol. 2009, 38: 1215-1223. 10.1093/ije/dyn213.

    Article  PubMed  Google Scholar 

  29. Der G: A comparison of the West of Scotland Twenty -07 study sample and the 1991 census SARs. A Comparison of the West of Scotland Twenty-07 Study Sample and the 1991 Census SARs. 1998, Glasgow: MRC Medical Sociology Unit

    Google Scholar 

  30. Office for Population Census and Surveys: Classification of Occupations 1980. 1980, London: General Household Survey. HMSO

    Google Scholar 

  31. Scottish Neighbourhood Statistics Data Zones Background Information. []

  32. Scottish Neighbourhood Statistics. []

  33. Hanley JA, Negassa A, Edwardes MDD, Forrester JE: Statistical analysis of correlated data using generalized estimating equations: An orientation. Am J Epidemiol. 2003, 157: 364-375. 10.1093/aje/kwf215.

    Article  PubMed  Google Scholar 

  34. Ecob R, Macintyre S: Small area variations in health related behaviours; do these depend on the behaviour itself, its measurement, or on personal characteristics?. Health Place. 2000, 6: 261-274. 10.1016/S1353-8292(00)00008-3.

    Article  CAS  PubMed  Google Scholar 

  35. Datta GD, Subramanian SV, Colditz GA, Kawachi I, Palmer JR, Rosenberg L: Individual, neighborhood, and state-level predictors of smoking among US Black women: a multilevel analysis. Soc Sci Med. 2006, 63: 1034-1044. 10.1016/j.socscimed.2006.03.010.

    Article  PubMed  Google Scholar 

  36. McGinn AP, Evenson KR, Herring AH, Huston SL, Rodriguez DA: The association of perceived and objectively measured crime with physical activity: a cross-sectional analysis. J Phys Act Health. 2008, 5: 117-131.

    PubMed  Google Scholar 

  37. Matheson FI, Moineddin R, Dunn JR, Creatore MI, Gozdyra P, Glazier RH: Urban neighborhoods, chronic stress, gender and depression. Soc Sci Med. 2006, 63: 2604-2616. 10.1016/j.socscimed.2006.07.001.

    Article  PubMed  Google Scholar 

  38. Curry A, Latkin C, Davey-Rothwell M: Pathways to depression: The impact of neighborhood violent crime on inner-city residents in Baltimore, Maryland, USA. Soc Sci Med. 2008, 67: 23-30. 10.1016/j.socscimed.2008.03.007.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Robinson S, Bugler C: General lifestyle survey 2008: smoking and drinking among adults, 2008. 2010, Newport: Office for National Statistics

    Google Scholar 

  40. Kwan MP: From place-based to people-based exposure measures. Soc Sci Med. 2009, 69: 1311-1313. 10.1016/j.socscimed.2009.07.013.

    Article  PubMed  Google Scholar 

Pre-publication history

Download references


This work was undertaken while MS was on a 2 month internship at the MRC SPHSU working with AE. MS was supported by a Doctoral Research Award from the Canadian Institutes of Health Research (CIHR) and by an internship award from the Département de médecine sociale et préventive of the Université de Montréal, Canada. AE is employed by the UK Medical Research Council and leads the Neighbourhoods and Health Programme (wbs U.1300.00.011.). We are grateful to Geoff Der for statistical advice and to Laura Macdonald for assistance in data preparation.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Anne Ellaway.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

MS conducted the statistical analysis and drafted the manuscript. Both authors participated in the interpretation of the data and contributed to subsequent drafts. Both authors read and approved the final manuscript.

Rights and permissions

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and Permissions

About this article

Cite this article

Shareck, M., Ellaway, A. Neighbourhood crime and smoking: the role of objective and perceived crime measures. BMC Public Health 11, 930 (2011).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: