Associations between neighborhood characteristics and dating violence: does spatial scale matter?

Paul Rodrigues; Martine Hébert; Mathieu Philibert

doi:10.1186/s12942-022-00306-3

Associations between neighborhood characteristics and dating violence: does spatial scale matter?

Rodrigues, Paul; Hébert, Martine; Philibert, Mathieu 2022-06-20 00:00:00 Background: Dating violence (DV ) is a public health problem that could have serious repercussions for the health and well-being of a large number of adolescents. Several neighborhood characteristics could influence these behav- iors, but knowledge on such influences is still limited. This study aims at (1) evaluating the associations between neighborhood characteristics and DV, and (2) assessing how spatial scale influences the estimations of the latter associations. Methods: The Québec Health Survey of High School Students (2016–2017) was used to describe DV. Neighborhoods were operationalized with polygon-based network buffers of varying sizes (ranging from 250 to 1000 m). Multiple data sources were used to describe neighborhood characteristics: crime rate, alcohol outlet density (on-premises and off-premises), walkability, greenness, green spaces density, and youth organizations density. Gendered-stratified logistic regressions were used for assessing the association between neighborhood characteristics and DV. Results: For boys, off-premises alcohol outlet density (500 m) is associated with an increase in perpetrating psy- chological DV. Crime rate (500 m) is positively associated with physical or sexual DV perpetration, and crime rate (250 m) is positively associated with physical or sexual DV victimization. Greenness (1000 m) has a protective effect on psychological DV victimization. For girls, walkability (500 m to 1000 m) is associated with a decrease in perpetrat- ing and experiencing psychological DV, and walkability (250 m) is negatively associated with physical or sexual DV victimization. Conclusions: Several neighborhood characteristics are likely to influence DV, and their effects depend on the form of DV, gender, and spatial scale. Public policies should develop neighborhood-level interventions by improving neigh- borhood living conditions. and 9% (ranging from < 1 to 54%) for sexual victimization Introduction [1]. Psychological violence was not assessed in this meta- Teen dating violence (DV), which can be described as analysis, but it appears to be the most common form of psychologically, physically, or sexually abusive behaviors DV. A systematic review suggests that the prevalence of from a dating partner, is a major public health problem. psychological violence ranges from 17 to 88% [2]. In addi- A recent meta-analysis estimated an overall prevalence tion to its high prevalence, DV is associated with negative of 20% (ranging from 1 to 61%) for physical victimization repercussions on the health and well-being of victims, such as anxiety, depression, post-traumatic stress symp- toms, as well as suicidal ideation [3–5]. *Correspondence: philibert.mathieu@uqam.ca Research on determinants of DV has mainly focused on Département de Sexologie, Université du Québec à Montréal, Succursale individual (e.g., antisocial or risky behaviors), family (e.g., Centre-Ville, Case postale 8888, Montréal, Québec H3C 3P8, Canada © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 2 of 21 exposure to family violence, lack of parental supervision), positive associations between perceived neighborhood and peer (e.g., affiliation with deviant peer) factors [6–8]. disorder and DV [21–24], while others reported no Empirical analyses of the association between neighbor- significant associations between crime rate, a dimen - hood characteristics and DV are scarce. Among studies sion of neighborhood disorder, and DV [25, 26]. The exploring such an association, most studies analyzed the association between alcohol outlet density and DV has effect of neighborhoods’ characteristics such as the soci - been observed among young adults [27–29], but studies odemographic composition, neighborhood disorder (i.e., have yet to assess this relationship among adolescents. visible social (e.g., crime) and physical (e.g., vandalism) However, alcohol outlet density could have a different signs of decay), collective efficacy and access to alcohol effect on this population as selling alcohol to minors is outlets, and led to inconsistent findings [9]. Yet, charac - prohibited. teristics of the physical environment are seldom analyzed Different mechanisms could explain the effect of neigh - in relation to DV despite reports of associations with borhood disorder and access to alcohol outlets on DV. adolescents’ behaviors, possibly reflecting socioenviron - Neighborhood disorder is often incorporated into social mental and psychobehavioral mechanisms. For example, disorganization theory, which describes it as a marker of neighborhood greenness, walkability, access to green the lack of order in the community and the inefficiency spaces, and access to community organizations could of social control [30, 31]. Disorder may refer to minor enhance social cohesion and reduce adolescents’ aggres- offences (e.g. graffiti, vandalism) which may lead to sion [10–12], which could positively affect DV [7–9]. more serious crimes and create fear and distrust among However, to our knowledge, the influence of these physi - residents [31]. Such conditions could encourage violent cal environment factors on DV has never been assessed. behaviors, including DV [9]. Exposure to neighborhood Furthermore, previous studies on the neighborhood violence, a component of disorder sometimes used to effects on DV have used census tracts to operational - measure the effect of neighborhood disorder, may also ize neighborhoods. There is little discussion of how to come to foster the normalization of violence and increase operationalize neighborhoods and the scale of analysis, frustration and anger in adolescents, which may influ - but these choices could potentially influence the estima - ence the adoption of such behaviors [9]. tion of the effect of neighborhood factors [13]. In addi - The density of alcohol outlets in neighborhoods is a tion, administrative units, such as census tracts, may not potential risk factor for many violent behaviors, includ- accurately reflect individual experience of space. Egocen - ing interpersonal violence. The presence of alcohol out - tric neighborhoods, defined as a buffer around a location, lets provides opportunities for consumption [32], which such as individuals’ homes, could better address these could result in risky drinking behaviors [19]. Alcohol limitations [14–16], but this approach is little used in abuse could, in turn, increase the risk of interpersonal research on DV. Against this backdrop, the current study violence victimization [8] and perpetration [7]. In socially aimed to analyze a range of neighborhood characteristics disorganized neighborhoods, alcohol outlet density may possibly related to DV, many of which have not yet been also enhance neighborhood disorder by encouraging the explored, and explore the effects of the spatial scale of gathering of people likely to adhere to norms and atti- analysis. tudes conducive to alcohol consumption and violence [19]. Context Neighborhood risk factors and DV Neighborhood protective factors and DV The relationship between neighborhood characteristics While neighborhood disorder and alcohol outlet density and DV has been explored through social disorganization could negatively impact adolescents’ behaviors, some theory focusing on neighborhood-level risk factors [9]. environmental and institutional resources could be asso- This theory posits that violence and criminality are more ciated with positive effects [33]. In particular, the level likely to occur in socially disorganized neighborhoods of greenness, the density of green spaces, the density of due to the community’s inability to collectively manage community organizations and walkability could prevent problems within the neighborhood [17]. Such context DV through several mechanisms. could also foster neighborhood disorder [17, 18] and Greenness and accessibility to green spaces could exacerbate negative consequences of alcohol outlets (e.g., affect some physiological and psychosocial processes, criminality, alcohol abuse) [19, 20], thereby influencing which in turn could influence DV. According to the bio - individuals’ violent behaviors. philia hypothesis, exposure to nature may improve men- Several empirical analyses assessed the effects of tal health by restoring cognitive functions and reducing neighborhood disorder and density of alcohol out- mental fatigue [34]. A higher level of greenness in lets on DV, but results are mixed. Some studies found neighborhoods may also be associated with a lower risk R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 3 of 21 of depressive symptoms [35, 36] and aggressivity [37] 26] and the density of alcohol outlets [27–29], but this in adolescents, both of which have been identified as choice remains little discussed, despite being known to potential risk factors of DV perpetration [7] as well as geographers to affect statistical analyses. victimization [8]. A study on intimate partner violence Two biases are related to neighborhoods’ operation- also found that rates of aggression were lower in neigh- alization: the Modifiable Areal Unit Problem (MAUP) borhoods characterized by a high level of greenness [38]. and the Uncertain Geographic Context Problem Access to green spaces, such as public parks, could pro- (UGCoP). The MAUP suggests that the shape and size mote participation in physical and recreational activities of spatial units could lead to variations in the estimated and encourage social interactions [39]. Physical activity effects of neighborhood characteristics [58, 59]. The may improve self-esteem and have a protective effect on UGCoP argues that the estimation of neighborhood the onset of depression and anxiety [40], all three deter- effects may be affected by the definition of neighbor - minants of DV [7, 8]. Social interactions may increase hoods and the use of inappropriate spatial units [60]. social cohesion in neighborhoods [41], which is likely to Both biases could limit the ability to observe associa- reduce the risk of DV [42, 43]. tions between neighborhood characteristics and a given Geographic access to community organizations (e.g., outcome. For example, census tracts were used in pre- social clubs, sports clubs) may promote adolescents’ par- vious studies on DV even though they may not be an ticipation in structured and supervised activities [10, 44]. adequate representation of neighborhood-level influ - Such resources provide opportunities to develop social ences. The MAUP and UGCoP could also partly explain ties between adolescents as well as between residents why the effect of certain factors, such as crime [25, 26], of the neighborhood (e.g., their parents) [10]. Commu- has not yet been observed. nity organizations are also safe places that could reduce In effect, administrative units, such as census tracts, exposure to the negative aspects of neighborhoods, such may not adequately represent individuals’ actual expo- as disorder [10, 45] and foster participation in supervised sure to their neighborhood. These units have artificial activities. Thus, adolescents attending these facilities boundaries, which implicitly assume that individuals’ could have a lower risk of risky behaviors, such as sub- environments are limited to the corresponding spatial stance use [10, 46] and delinquency [46–48], which are units, regardless of their real location inside the spatial determinants of DV [7, 8]. unit. However, individuals living closely to each other Walkability refers to the attributes of urban design (e.g., but in two different spatial units could have more simi - road network, land use) promoting walking [49]. Walk- lar exposures to neighborhood factors than two indi- ability is likely to promote social interactions, which viduals living further away but in the same spatial unit. could contribute to a better sense of belonging [11, 50] Neighborhood-level processes, such as social interac- and collective efficacy (i.e., the ability of the community tions or access to resources, may depend on geographical to act collectively to regulate deviant behaviors) [11]. proximity and can be observed across these boundaries These concepts are central to social disorganization the - [14, 16, 61]. In consequence, the accuracy of measures of ory and could influence DV [9]. To our knowledge, the exposure to neighborhood factors could vary depending effect of walkability on DV has not been described. How - on the location of individuals within a spatial unit. These ever, empirical studies suggest that increased walkability measures could be less accurate for people living fur- is associated with a decrease in homicide [51] and violent ther away from the centroid of the spatial unit [62–64]. behaviors [52] among adolescents. When geographic access to resources is assessed, using Finally, it is essential to note that the associations large spatial units such as census tracts could also lead to between neighborhood characteristics and DV may be aggregation error [62, 63]. In response to such a caveat, modified by gender. Most studies suggest that neighbor - egocentric neighborhoods were proposed as operational hood effects are stronger for boys than for girls [53, 54]. forms of neighborhoods which allow for integrating prox- Boys may have less parental supervision than girls [55], imity. They provide a person-centered approach and refer leading to greater exposure to their neighborhood [56, to the set of places located within a given distance from 57]. the individuals’ homes [15, 16, 65]. Egocentric neighbor- hoods are defined specifically for each individual and allow for overlapping neighborhoods, suggesting that Defining neighborhoods some individuals share common exposures to neighbor- In addition to the lack of knowledge about neighbor- hood factors depending on their proximity. An egocen- hood effects on DV, previous studies have not explored tric definition of neighborhoods could more adequately the impact of neighborhoods’ operationalization and reflect the real use of spaces by residents and their per - spatial scale of the estimated effects. Most studies on ception of the residential context [14, 16, 61]. DV used census tracts to measure the level of crime [25, Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 4 of 21 Objectives Québec, Canada) and is composed of 15 municipalities The current study had two main objectives: (1) to analyze (Fig. 1) [66]. About 85% of the island’s population lives in the association between several contextual characteris- the city of Montréal, the largest city in Québec and the tics (crime, density of off-premises alcohol outlets, den - second largest in Canada. sity of bars, density of community organizations, level of greenness, density of green spaces and walkability) and Dependent variables DV (victimization and perpetration) and (2) to assess Four variables describing different forms of DV were scale effects in studying these relationships. used. Two variables referred to victimization (e.g., “He/ she (…) me”), and two referred to perpetration (e.g., “I Data and method (…) him/her). Participants Psychological DV was assessed by items derived from The current study used data from the Québec Health two questions. One item referred to verbal violence Survey of High School Students (QHSHSS) 2016–17, (“I criticized him/her viciously about his/her physical a cross-sectional survey of secondary-school Québec appearance; I insulted him/her in front of people; I put youth (Grades 7 to 11). A three-stage stratified cluster him/her down.” / “He/she viciously criticized my physi- sampling was used to recruit participants in this survey. cal appearance; he/she insulted me in front of people; he/ Schools were randomly selected for each grade level and she put me down.”), while the other related to controlling each health region separately. Classes were randomly behaviors (“I controlled his/her outings, email conversa- selected from the selected schools. The QHSHSS pro - tions or cell phone; I prevented him/her from seeing his/ vides a representative sample for the province of Québec her friends.” / “He/she controlled my outings, my email and each health region. Only participants whose postal conversations or cell phone; he/she prevented me from code of residence was located in the island of Montréal seeing my friends”). The response scale ranged from 0 were included in this study (n = 2,687). The island of (never during the past 12 months) to 3 (three times or Montréal had a population of 1,942,040 inhabitants in more during the past 12 months). A dichotomous meas- 2016 (about 23.8% of the population of the province of ure was obtained by distinguishing participants who Fig. 1 Study area-island of Montréal R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 5 of 21 had experienced one of these events at least once (one theft, and robbery. First, the geolocation of crime events or more) from those who reported having never experi- allowed their number to be calculated for each egocen- enced these situations. tric neighborhood (buffer zone). Secondly, the popula - To assess physical violence, four items from the Con- tion size of each egocentric neighborhood was estimated ict T fl actics Scale [67] (e.g., “I slapped him/her”/ “He/She by summing the population of dissemination blocks (DB) slapped me”) were used. Sexual violence was assessed by included in the neighborhood. DBs are the smallest spa- two items referring to experiences of sexual activity with- tial units for which data on population size are dissemi- out consent (e.g., “I forced him/her to have sexual con- nated by Statistics Canada. If only part of the DBs was tact or sexual intercourse with me when he/she didn’t included in a buffer, only the proportion of its population want to” / “He/she forced me to have sexual contact or corresponding to the proportion of area included in the sexual intercourse when I didn’t want to.”). Due to the neighborhood was considered for summing the popula- small number of participants who reported sexual vio- tion. The crime rate is the number of crimes divided by lence perpetration, physical and sexual DV were merged. twice the population size (the number of crimes was esti- Dichotomous measures were derived by differentiating mated for two years) multiplied by 100. participants who had experienced physical or sexual vio- lence at least once from those who had never experienced Greenness and access to green space these situations. The Normalized Difference Vegetation Index (NDVI) is commonly used in epidemiological studies and provides Neighborhood‑level variables an objective and accurate measure of overall greenness Egocentric neighborhoods [69]. This index was first produced at a 30 × 30 m spatial Egocentric neighborhood refers to the area within a spe- resolution from Landsat 8 satellite data (United States cific radius around the individual’s home and is often Geological Survey, 2015–2016). The greenness of the operationalized using buffer zones [15, 16, 65]. In this egocentric neighborhoods was estimated using the aver- study, the participants’ place of residence was estimated age NDVI in the area covered by a buffer. from the centroid of the postal code area. Postal codes To assess access to green spaces, a map of public parks are managed by Canada Post for mail delivery and have and green spaces was first obtained for the island of an average of 14.5 dwellings in Montréal. Their cen - Montréal by cross-referencing information provided by troids provide a good approximation of the exact address the municipalities. The municipalities of the City of Mon - of participants [68]. Egocentric neighborhoods were tréal, Dollard-des-Ormeaux, Pointe-Claire, Kirkland, and operationalized from polygon-based network buffers East Montréal provided the location of parks and green around the participants’ place of residence, a method spaces as open access data. A map of parks and green of buffering providing an accurate representation of the spaces was created for the remaining cities using land spatial area used by individuals [65]. In this study, the use data at the parcel-level from the property assessment government of Québec’s official road network database roll provided by the Montréal Metropolitan Community (Addresses Québec) and the Network Analyst extension (2016) and information available on the municipalities’ of ArcGIS Pro 2.7.0 were used to generate all routes from websites. The parcels on the property assessment roll participants’ place of residence to a specific distance. were manually identified as green space based on the Polygon-based network buffers were then computed documents and maps available on the websites. A total of by connecting the endpoints of these routes, result- 1389 green spaces were identified for the island of Mon - ing in irregular polygons (Fig. 2). For each participant, tréal. Access to green spaces was estimated by the num- four buffer zones were created using different distances: ber of green spaces that intersected each buffer. 250 m, 500 m, 750 m, and 1000 m. Figure 2 provides an example of polygon-based network buffers for two postal Access to alcohol outlets codes. Alcohol outlets were located using data on private out- lets licensed by the Régie des alcools, des courses et des Criminality jeux (RACJ, 2016) and outlets administered by the Société The level of crime was measured using data from Ser - des alcools du Québec (SAQ, 2016). A total of 1432 bars vice de police de la Ville de Montréal (SPVM) for the and 1842 off-premises alcohol outlets (e.g., convenience years 2016 and 2017. This database provides the date stores, grocery shops and SAQ outlets) were identified and the location of crime events at the nearest inter- from these data. Using the same method as for green section. Events pertaining to one of the following six spaces, density measures have been used to estimate categories were retained: offences resulting in death, access to bars and access to off-premises alcohol outlets intrusion, mischief, theft in/on a vehicle, motor vehicle separately. Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 6 of 21 Fig. 2 Egocentric neighborhoods Access to community organizations (2016) was used to assess the land use mix. An entropy The Directory of Community Organisations from 211, index [70] was elaborated to describe the level of het- an information and referral public service, was used to erogeneity in land use considering four categories of identify and locate community organizations for young land use: residential, commercial, services and cul- people (e.g., youth centers, YMCA, sports associations). tural, recreational and leisure. Secondly, data from the Lists of community organizations in the different munici - property assessment roll were used to calculate the net palities of the island of Montréal, which are available on residential density. This variable refers to the number the corresponding websites, supplemented these data. of dwellings per hectare of the residential area included A total of 423 community organizations offering ser - in each buffer. Thirdly, data from Addresses Québec vices for young people were identified. To assess access (2016) were used to calculate the density of intersec- to community organizations, density measures were used tions with three or more segments of the road network by calculating the number of community organizations (excluding highways). The walkability measure used within each buffer. in this study is based on several studies reporting on the development of this index [71, 72]. It was obtained Walkability by the sum of the z-scores of all variables described For each egocentric neighborhood, walkability was above: z-score(land use mix) + z-score(residential den- measured using three variables: land use mix, residen- sity) + 1.5(z-score(intersection density)). In its complete tial density, and intersection density. Firstly, data from form, the walkability index also requires the retail floor the property assessment roll (parcel-level land use) area ratio [71], which could not be obtained as is the provided by the Montréal Metropolitan Community case for many studies. The weight of 1.5 for intersection R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 7 of 21 density (replacing a weight of 2 in the complete index) corresponding buffer. Median income was used as it is was proposed by Sundquist et al. [72] as a way to com- less sensitive to extreme values than average income and pensate for the lack of data on retail floor area. This walk - may therefore better capture the spatial distribution of ability index has been validated and is a good predictor income. of walking behaviors [72]. A study also found a strong The percentage of single-parent dwellings was used to correlation between the three-components index and the assess single parenthood. The number of dwellings and four-components index, and both were predictors of util- the number of single-parent dwellings were estimated for itarian walking [73], confirming the validity of using the each egocentric neighborhood by summing the frequen- three-components walkability index as a surrogate for the cies of the two variables at the DB-level. four-components index. Residential instability was measured using the percent- age of residents living in the egocentric neighborhood for Covariates five years or less. For each egocentric neighborhood, the Individual‑level covariates number of residents living there for less than five years Several individual-level covariates were used: gender (girl and the population size were estimated by summing the or boy), high school grade level (Grades 7 or 8, 9, 10, and frequencies of the DBs considered as part of the corre- 11), the highest level of parental education (high school sponding buffer. or less, college or professional training, university), fam- Finally, to assess ethnocultural diversity, a language ily structure (two parents, blended family, or shared cus- diversity index was developed. Languages spoken at tody, living with one parent or other family structure), home were divided into 16 classes based on the World and parental country of birth (two parents born in Can- Value Survey classification [74]. Frequencies of each class ada, at least one parent born outside Canada). were first obtained for each egocentric neighborhood by summing the frequencies of the corresponding DBs. Neighborhood sociodemographic characteristics covariates Shannon entropy index [70] was then calculated for each Five potentially confounding neighborhood sociodemo- egocentric neighborhood to estimate the level of hetero- graphic characteristics were measured using data from geneity/homogeneity of languages spoken at home. the 2016 Canadian census. Population density corresponded to the number of Statistical analysis inhabitants per hectare and was estimated from the pop- Data at the neighborhood level were matched with data ulation size and the area of the buffers. The population from the QHSHSS at the individual level using partici- size of each egocentric neighborhood was obtained by pants’ postal codes. Only adolescents who reported being summing the population size of the DBs (i.e., the smallest involved in a romantic relationship in the last 12 months spatial units for which population size is available) within and had no missing values for the DV measures were the corresponding buffer zone weighted by the propor - included (37% of participants with a postal code in Mon- tion of the area of the included DBs (a weight of 1 was tréal). Among them, 121 had at least one missing value used for DBs completely located in the buffer). for covariates and were excluded (12% of participants Although population size data are available at the DB- who reported a romantic relationship). The final sample level, no sociodemographic data are disseminated by consisted of 879 adolescents (Fig. 3). Statistics Canada at this scale. For this reason, the four The associations between neighborhood character - remaining sociodemographic variables were derived istics and DV were estimated with logistic regressions from variables at the dissemination area (DA) level, the using the SURVEYLOGISTIC procedure of SAS Enter- smallest spatial units for which sociodemographic data prise Guide 8.5 [75]. All models considered the sam- are available. As DBs are embedded into DAs (see Fig. 2), pling design and used bootstrap weights. They were the sociodemographic measures were assigned to the carried out separately for girls and boys to account for corresponding DBs, assuming a homogeneous distribu- gender differences. A Directed Acyclic Graph (DAG) tion within each DA. Similar to the method used to esti- was developed to identify confounding variables (Addi- mate population density, each egocentric neighborhood tional file 1). A DAG is a graphical representation of was composed of the DBs included in the corresponding causal assumptions regarding a set of variables and can buffer, weighted by the proportion of the area of included be used as a tool to identify confounders [76, 77]. This DBs. approach can also limit the risk of overadjustment bias Socioeconomic status (SES) was measured by [77]. This bias occurs when a model controls for vari - median income. For each egocentric neighborhood, ables that are not confounders. Assessing neighbor- the median income was estimated by calculating the hood effects could be subject to overadjustment bias population-weighted median of the DBs included in the due to the complex relationships and high correlations Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 8 of 21 Fig. 3 Flowchart for sample selection between neighborhood factors. The DAG used in the of crime were modelled controlling for all the variables current study thus provided a parsimonious approach described above, except for the density of community and allowed the identification of confounding factors organizations. tailored to each neighborhood factor specifically. Based At each stage and for each dependent variable, four mod- on this DAG, not all models required the same set of els estimated the association with a given neighborhood covariates. Associations between neighborhood char- characteristic at different scales, i.e. using different buffer acteristics and DV variables were estimated using dif- sizes (250 m, 500 m, 750 m, and 1000 m). For each of these ferent models in three stages. (1) Associations between models, covariates were modelled at the scale considered DV and walkability were estimated using models the most appropriate in the previous stage. Choosing the including individual-level variables and neighborhood most appropriate scale for modelling a given neighborhood sociodemographic characteristics as covariates. (2) characteristics was based on the comparison of the mod- Associations between DV and the level of greenness, els’ fit using the Akaike Information Criterion (AIC). The the density of green spaces, the density of community selected scale was the one with the lowest AIC. It should be organizations, the density of off-premises alcohol out - noted that the scale for neighborhood sociodemographic lets, and the density of bars were estimated by con- characteristics used as covariates in all models was simi- trolling for individual-level variables, neighborhood larly identified in preliminary analyses. In these analyses, sociodemographic characteristics as well as walkability. the effect of neighborhoods’ sociodemographic charac - (3) Associations between DV and neighborhoods’ level teristics on the DV measures were estimated separately in R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 9 of 21 models adjusted by individual-level variables only (results Our results suggest an effect of the density of off-prem - available upon request). For example, for perpetration ises alcohol outlets and crime rate on DV. There was a of psychological DV among boys, preliminary analyses positive association between the density of off-premises showed that the most appropriate scale for neighborhood alcohol outlets within a 500 m radius and psychological sociodemographic characteristics (lowest AIC) were DV perpetration (ß = 0.08; SE = 0.04; p = 0.049). Crime 1000 m for median income, 250 m for percentage of single- rate was positively associated with physical/sexual DV parent dwellings, 500 m for residential instability, 250 m for perpetration with radii of 500 m (ß = 0.97; SE = 0.28; ethnocultural diversity, and 1000 m for population density. p < 0.001) and 750 m (ß = 0.84; SE = 0.37; p = 0.024). The These variables were used in Stage 1 to estimate the effect AIC of the model using a 500 m buffer was consider - of walkability. In this stage, models showed that walkabil- ably smaller than that of the model using a 750 m radius ity within 250 m had the best fit (lowest AIC). Variables (ΔAIC = 3.931). There was also an association between selected in preliminary analyses and Stage 1 were then used crime rate measured with a 250 m radius buffer and to estimate the effects of all variables identified in Stage 2. physical/sexual DV victimization (ß = 0.18; SE = 0.09; In Stage 3, to assess the effect of crime rate, the scales used p = 0.050). No significant associations were identified for neighborhood characteristics covariates were based on between the density of bars and DV. results in all previous stages. This method provided parsi - The analysis of possible protective factors revealed only monious estimations of neighborhood factors for each out- a negative association between NDVI within a 1000 m come. Multicollinearity was assessed for all models using radius and psychological DV victimization (ß = − 4.24; Variance Inflation Factors (VIF) and the data showed no SE = 2.09; p = 0.044). multicollinearity problem (VIF < 4). In line with the second objective of this study, the influ - Models of associations between neighborhood ence of the spatial scale of analysis was evaluated by com- characteristics on DV among girls paring the AIC of models estimating associations between Results from logistic regression assessing the associations the same neighborhood-level factors and a given outcome between neighborhood characteristics and DV among across buffers (250 m, 500 m, 750 m, and 1000 m). As a rule girls are shown in Table 3. of thumb, an AIC difference greater than two units sug - Regarding risk factors, results revealed negative asso- gest that the model with the lowest AIC is the most predic- ciations between density of bars within a buffer of 250 m tive [78]. Scale effects were also investigated by calculating (ß = − 0.29; SE = 0.15; p = 0.045) and 1000 m (ß = − 0.03; Spearman’s rho correlation coefficients between all neigh - SE = 0.01; p = 0.049) and psychological DV victimization. borhood-level variables across buffers. Weak correlations The difference in AIC between the two models was small suggested differences in measurements of neighborhood- (ΔAIC = 0.204), suggesting that both spatial scales are level variables across buffers, while strong correlations comparable. No significant effects were observed for the suggested minor differences. Spearman’s correlations were density of off-premises alcohol outlets and crime rate. used instead of Pearson’s correlations because some vari- In analyzing possible protective factors, several asso- ables were skewed. ciations were found between walkability and DV. Walk- ability within a radius of 500 m (ß = − 0.19; SE = 0.07; Results p = 0.007), 750 m (ß = − 0.20; SE = 0.07; p = 0.007), and Sample description 1000 m (ß = − 0.19; SE = 0.07; p = 0.013) were negatively Most participants reported living in a two-parent family associated with psychological DV perpetration. The (64.74%) with at least one parent who has obtained a uni- model using a 500 m radius buffer had the smallest AIC versity degree (68.56%) (Table 1). Psychological violence (AIC = 496.119), but the difference with the other mod - was the most prevalent form of DV: 20.25% of adolescents els was negligible (ΔAIC < 1.464). There were associa - reported perpetration, while 28.78% reported experiences tions between walkability in a buffer of 500 m (ß = − 0.13; of victimization. Perpetration and victimization of physi- SE = 0.06; p = 0.028), 750 m (ß = − 0.17; SE = 0.06; cal/sexual DV were observed respectively by 14.10% and p = 0.008), and 1000 m (ß = − 0.17; SE = 0.07; p = 0.012) 19.94% of participants. For all forms, girls were more likely with psychological DV victimization. The model using a to report an experience of DV. 750 m radius had the lowest AIC (AIC = 580.321) but was comparable with the other two models (ΔAIC < 1.800). Models of associations between neighborhood Finally, a decrease of walkability within a radius of 250 m characteristics on DV among boys (ß = − 0.15; SE = 0.07; p = 0.029), and 500 m (ß = − 0.15; Table 2 summarizes the results from logistic regressions SE = 0.07; p = 0.049) were associated with lower physical/ analyzing the relationship between neighborhood char- sexual DV victimization. The difference between AIC is acteristics and DV among boys. small (ΔAIC = 1.233), suggesting that the two models are Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 10 of 21 Table 1 Descriptive statistics All (n = 879) Girls (n = 452) Boys (n = 427) Individual‑level variables % % % Psychological DV perpetration At least once 20.25 23.76 16.68 Never 79.75 76.24 83.32 Physical or sexual DV perpetration At least once 14.10 20.41 7.67 Never 85.90 79.59 92.33 Psychological DV victimization At least once 28.78 34.13 23.33 Never 71.22 65.87 76.67 Physical or sexual DV victimization At least once 19.94 22.83 17.01 Never 80.06 77.17 82.99 Grade level Grade 7 or 8 29.09 25.09 33.15 Grade 9 18.32 18.06 18.58 Grade 10 23.71 23.93 23.50 Grade 11 28.88 32.92 24.77 Parental country of birth Two parents born in Canada 42.21 37.88 46.61 At least one parent born outside Canada 57.79 62.12 53.39 Family structure Two parents 64.74 61.96 67.57 Blended family or shared custody 16.91 19.13 14.66 Living with one parent or other family structure 18.35 18.92 17.78 Highest level of parental education High school or less 13.06 16.55 9.50 College or professional training 18.38 20.20 16.54 University 68.56 63.25 73.96 Neighborhood‑level variables Mean (SD) Mean (SD) Mean (SD) 250 m Sociodemographic characteristics Median income 58,754.92 (25,196.54) 57,754.87 (22,883.30) 59,772.37 (27,339.35) Single parenthood 20.32 (7.63) 20.72 (7.40) 19.92 (7.84) Residential instability 38.94 (13.13) 39.39 (12.84) 38.48 (13.42) Ethnocultural diversity 1.28 (0.39) 1.30 (0.39) 1.27 (0.39) Population density 9.89 (6.07) 9.90 (5.88) 9.88 (6.25) Risk factors Density of off-premises alcohol outlets 0.75 (1.18) 0.74 (1.12) 0.76 (1.23) Density of bars 0.31 (1.00) 0.35 (1.03) 0.28 (0.97) Crime rate 1.54 (2.18) 1.64 (2.76) 1.44 (1.34) Protective factors Walkability − 0.47 (2.61) − 0.37 (2.57) − 0.58 (2.65) NDVI 0.50 (0.11) 0.49 (0.11) 0.50 (0.11) Density of green spaces 0.87 (0.95) 0.89 (0.92) 0.86 (0.99) Density of community organizations 0.39 (0.65) 0.37 (0.64) 0.40 (0.67) 500 m Sociodemographic characteristics Median income 57,714.61 (22,067.14) 56,744.45 (20,220.35) 58,701.65 (23,782.52) R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 11 of 21 Table 1 (continued) Neighborhood‑level variables Mean (SD) Mean (SD) Mean (SD) Single parenthood 20.33 (6.56) 20.74 (6.43) 19.91 (6.67) Residential instability 39.37 (11.88) 39.67 (11.50) 39.06 (12.26) Ethnocultural diversity 1.29 (0.37) 1.30 (0.38) 1.27 (0.37) Population density 8.61 (4.42) 8.65 (4.41) 8.58 (4.44) Risk factors Density of off-premises alcohol outlets 3.26 (3.91) 3.24 (3.77) 3.27 (4.04) Density of bars 1.50 (3.63) 1.51 (3.04) 1.49 (4.14) Crime rate 1.32 (3.16) 1.24 (1.01) 1.41 (4.37) Protective factors Walkability − 0.53 (2.66) − 0.44 (2.62) − 0.61 (2.69) NDVI 0.50 (0.10) 0.49 (0.10) 0.50 (0.10) Density of green spaces 2.86 (2.35) 2.9 (2.32) 2.82 (2.38) Density of community organizations 1.10 (1.24) 1.06 (1.15) 1.15 (1.31) 750 m Sociodemographic characteristics Median income 56,871.79 (19,941.78) 56,039.25 (18,387.81) 57,718.81 (21,395.60) Single parenthood 20.52 (5.86) 20.87 (5.73) 20.16 (5.97) Residential instability 39.68 (11.01) 39.97 (10.5) 39.40 (11.52) Ethnocultural diversity 1.29 (0.36) 1.31 (0.36) 1.28 (0.36) Population density 7.99 (3.80) 8.09 (3.79) 7.89 (3.81) Risk factors Density of off-premises alcohol outlets 7.32 (7.64) 7.29 (7.56) 7.36 (7.72) Density of bars 3.78 (7.98) 3.83 (7.55) 3.72 (8.41) Crime rate 1.18 (0.64) 1.18 (0.63) 1.19 (0.65) Protective factors Walkability − 0.55 (2.68) − 0.46 (2.63) − 0.65 (2.74) NDVI 0.49 (0.10) 0.49 (0.10) 0.50 (0.10) Density of green spaces 5.75 (4.18) 5.84 (4.29) 5.66 (4.07) Density of community organizations 2.08 (1.89) 2.00 (1.81) 2.16 (1.97) 1000 m Sociodemographic characteristics Median income 56,237.1 (17,931.19) 55,576.33 (16,738.62) 56,909.36 (19,064.11) Single parenthood 20.65 (5.40) 20.95 (5.34) 20.34 (5.45) Residential instability 40.08 (10.44) 40.39 (9.93) 39.76 (10.94) Ethnocultural diversity 1.30 (0.35) 1.31 (0.35) 1.29 (0.34) Population density 7.51 (3.40) 7.60 (3.35) 7.42 (3.45) Risk factors Density of off-premises alcohol outlets 12.69 (12.60) 12.73 (12.72) 12.65 (12.49) Density of bars 6.95 (13.98) 7.05 (13.57) 6.85 (14.40) Crime rate 1.18 (0.58) 1.18 (0.58) 1.18 (0.59) Protective factors Walkability − 0.54 (2.69) − 0.44 (2.63) − 0.64 (2.75) NDVI 0.49 (0.09) 0.49 (0.09) 0.49 (0.09) Density of green spaces 9.42 (6.38) 9.53 (6.60) 9.31 (6.15) Density of community organizations 3.42 (2.75) 3.29 (2.72) 3.54 (2.78) comparable. Significant positive associations were also p = 0.035) and 1000 m (ß = 4.25; SE = 1.82; p = 0.021) observed for NDVI and density of community organi- were related to greater physical/sexual DV victimiza- zations. NDVI in a buffer of 750 m (ß = 3.86; SE = 1.81; tion. Model using a 1000 m buffer had the lowest AIC Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 12 of 21 Table 2 Associations between neighborhood characteristics and DV among boys 250 m 500 m 750 m 1000 m ß (SE) AIC ß (SE) AIC ß (SE) AIC ß (SE) AIC Psychological DV perpetration Risk factors Density of off-premises alcohol outlets 0.15 (0.14) 363.351 0.08 (0.04)* 362.064 0.05 (0.03) 362.621 0.02 (0.02) 363.195 Density of bars − 0.18 (0.17) 363.249 − 0.03 (0.03) 363.620 − 0.02 (0.01) 363.449 − 0.01 (0.01) 363.380 Crime rate − 0.08 (0.13) 362.791 − 0.21 (0.34) 362.638 − 0.56 (0.35) 361.415 − 0.33 (0.39) 362.628 Protective factors Walkability 0.09 (0.08) 362.323 0.10 (0.09) 362.486 0.09 (0.09) 362.629 0.10 (0.09) 362.365 NDVI − 1.92 (2.29) 363.504 − 1.00 (2.37) 364.134 − 0.89 (2.53) 364.187 − 0.45 (2.54) 364.289 Density of green spaces − 0.28 (0.21) 361.705 − 0.07 (0.08) 363.399 − 0.06 (0.05) 362.343 − 0.04 (0.03) 362.239 Density of community organizations 0.24 (0.23) 362.165 − 0.02 (0.12) 364.297 − 0.07 (0.08) 363.611 − 0.03 (0.06) 364.085 Physical or sexual DV perpetration Risk factors Density of off-premises alcohol outlets 0.01 (0.17) 242.997 0.04 (0.06) 242.771 0.00 (0.03) 243.000 0.01 (0.03) 242.950 Density of bars − 0.32 (0.26) 241.920 − 0.03 (0.06) 242.897 − 0.03 (0.02)† 242.115 − 0.02 (0.01) 242.270 Crime rate 0.14 (0.16) 245.851 0.97 (0.28)*** 237.656 0.84 (0.37)* 241.587 0.08 (0.56) 246.688 Protective factors Walkability 0.04 (0.12) 241.001 − 0.02 (0.11) 241.101 0.02 (0.11) 241.090 − 0.01 (0.12) 241.117 NDVI 0.86 (3.30) 242.908 2.54 (3.95) 242.334 2.43 (4.11) 242.470 1.77 (3.86) 242.728 Density of green spaces 0.10 (0.17) 242.730 − 0.04 (0.06) 242.811 − 0.03 (0.05) 242.744 − 0.03 (0.04) 242.436 Density of community organizations 0.04 (0.36) 242.984 − 0.15 (0.18) 242.313 − 0.11 (0.16) 242.264 − 0.05 (0.11) 242.714 Psychological DV victimization Risk factors Density of off-premises alcohol outlets 0.17 (0.13) 460.258 0.06 (0.05) 460.447 0.01 (0.03) 461.867 0.00 (0.02) 462.022 Density of bars 0.01 (0.11) 462.040 − 0.03 (0.03) 461.575 − 0.01 (0.01) 461.414 − 0.01 (0.01) 461.505 Crime rate 0.00 (0.11) 461.795 − 0.01 (0.03) 461.770 − 0.13 (0.35) 461.644 − 0.17 (0.42) 461.612 Protective factors Walkability − 0.05 (0.08) 461.103 − 0.10 (0.09) 460.043 − 0.08 (0.10) 460.752 − 0.01 (0.09) 461.562 NDVI − 2.75 (1.99) 459.792 − 2.88 (2.13) 460.042 − 3.59 (2.13)† 459.257 − 4.24 (2.09)* 458.285 Density of green spaces − 0.22 (0.16) 459.432 − 0.01 (0.05) 462.005 0.00 (0.03) 462.042 − 0.01 (0.03) 461.718 Density of community organizations 0.25 (0.22) 460.124 0.01 (0.11) 462.023 − 0.01 (0.08) 462.007 − 0.02 (0.06) 461.913 Physical or sexual DV victimization Risk factors Density of off-premises alcohol outlets 0.06 (0.13) 403.780 − 0.01 (0.05) 403.894 − 0.03 (0.03) 403.315 0.00 (0.02) 403.891 Density of bars − 0.13 (0.19) 403.313 − 0.02 (0.03) 403.620 − 0.01 (0.02) 403.560 − 0.01 (0.01) 403.618 Crime rate 0.18 (0.09)* 404.269 − 0.02 (0.02) 407.190 0.03 (0.24) 407.190 − 0.13 (0.37) 407.089 Protective factors Walkability − 0.04 (0.08) 401.915 − 0.02 (0.09) 402.165 0.02 (0.10) 402.145 − 0.03 (0.10) 402.123 NDVI 0.35 (1.98) 403.883 0.68 (2.13) 403.817 1.35 (2.15) 403.571 0.57 (2.12) 403.855 Density of green spaces − 0.04 (0.15) 403.830 0.01 (0.07) 403.904 0.04 (0.04) 402.625 0.02 (0.03) 403.306 Density of community organizations 0.07 (0.21) 403.813 − 0.18 (0.13) 401.684 − 0.02 (0.08) 403.879 0.01 (0.06) 403.898 Models were adjusted for individual-level covariates and sociodemographic characteristics at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates and sociodemographic characteristics and walkability at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates, and sociodemographic characteristics, walkability, density of green spaces, density of off-premises alcohol outlets and density of bars at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) *** p < 0.001; **p < 0.01; *p < 0.05; †p < 0.10 R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 13 of 21 Table 3 Associations between neighborhood characteristics and DV among girls 250 m 500 m 750 m 1000 m ß (SE) AIC ß (SE) AIC ß (SE) AIC ß (SE) AIC Psychological DV perpetration Risk factors Density of off-premises alcohol outlets 0.03 (0.15) 494.899 − 0.01 (0.06) 494.933 − 0.01 (0.03) 494.892 − 0.01 (0.02) 494.419 Density of bars − 0.13 (0.14) 494.127 − 0.02 (0.05) 494.763 − 0.02 (0.02) 494.128 − 0.01 (0.01) 494.645 Crime rate − 0.13 (0.13) 497.286 − 0.34 (0.25) 496.838 − 0.73 (0.40)† 495.455 − 0.71 (0.48) 495.676 Protective factors Walkability − 0.10 (0.07) 496.008 − 0.19 (0.07)** 492.949 − 0.20 (0.07)** 494.037 − 0.19 (0.07)* 494.413 NDVI 0.02 (1.56) 494.949 − 0.67 (1.75) 494.820 − 0.29 (1.90) 494.926 − 0.09 (1.95) 494.946 Density of green spaces 0.18 (0.14) 493.110 − 0.02 (0.05) 494.860 0.01 (0.03) 494.827 0.01 (0.02) 494.507 Density of community or ganizations 0.12 (0.22) 494.482 − 0.14 (0.16) 493.350 − 0.05(’− 0.09) 494.567 − 0.04 (0.06) 494.276 Physical or sexual DV perpetration Risk factors Density of off-premises alcohol outlets 0.16 (0.14) 457.140 0.00 (0.07) 458.632 0.04 (0.03) 457.265 0.01 (0.02) 458.404 Density of bars − 0.09 (0.15) 458.269 0.05 (0.06) 457.821 0.00 (0.02) 458.592 − 0.01 (0.01) 458.261 Crime rate − 0.00 (0.05) 460.683 − 0.20 (0.25) 459.964 − 0.66 (0.41) 456.010 − 0.61 (0.44) 458.209 Protective factors Walkability − 0.00 (0.07) 456.774 − 0.05 (0.07) 456.281 − 0.06 (0.07) 456.028 − 0.05 (0.07) 456.243 NDVI 1.47 (1.66) 457.857 3.09 (1.91) 456.057 3.52 (1.99)† 455.581 3.42 (1.93)† 455.795 Density of green spaces 0.03 (0.13) 458.585 0.06 (0.06) 457.536 0.03 (0.03) 457.622 0.02 (0.02) 457.867 Density of community or ganizations 0.02 (0.21) 458.629 − 0.09 (0.13) 458.027 − 0.08 (0.08) 457.717 − 0.05 (0.06) 457.903 Psychological DV victimization Risk factors Density of off-premises alcohol outlets 0.04 (0.12) 582.214 0.01 (0.05) 582.304 − 0.02 (0.03) 581.712 − 0.02 (0.02) 580.185 Density of bars − 0.29 (0.15)* 578.118 − 0.08 (0.06) 580.316 − 0.04 (0.02)† 579.715 − 0.03 (0.01)* 578.322 Crime rate − 0.08 (0.10) 584.017 − 0.14 (0.16) 584.536 − 0.11 (0.32) 585.194 − 0.28 (0.37) 584.684 Protective factors Walkability − 0.06 (0.06) 585.322 − 0.13 (0.06)* 582.121 − 0.17 (0.06)** 580.321 − 0.17 (0.07)* 580.406 NDVI − 0.01 (1.41) 582.321 − 0.3 (1.64) 582.287 0.36 (1.82) 582.276 0.41 (1.86) 582.265 Density of green spaces − 0.01 (0.13) 582.312 − 0.04 (0.05) 581.598 − 0.02 (0.03) 581.934 − 0.01 (0.02) 581.934 Density of community or ganizations 0.06 (0.20) 582.208 − 0.07 (0.11) 581.842 − 0.01 (0.07) 582.315 − 0.04 (0.05) 581.745 Physical or sexual DV victimization Risk factors Density of off-premises alcohol outlets 0.06 (0.12) 487.397 − 0.04 (0.06) 487.201 − 0.01 (0.03) 487.537 − 0.01 (0.02) 487.503 Density of bars − 0.10 (0.18) 487.300 − 0.02 (0.08) 487.494 − 0.01 (0.02) 487.482 − 0.01 (0.01) 487.317 Crime rate − 0.01 (0.04) 488.524 − 0.04 (0.15) 488.502 − 0.05 (0.46) 488.552 0.23 (0.50) 488.303 Protective factors Walkability − 0.15 (0.07)* 485.604 − 0.15 (0.07)* 486.837 − 0.13 (0.07)† 488.062 − 0.08 (0.07) 490.049 NDVI 2.69 (1.56)† 484.771 3.30 (1.79)† 484.218 3.86 (1.81)* 483.195 4.25 (1.82)* 482.368 Density of green spaces 0.02 (0.12) 487.574 0.06 (0.05) 486.369 0.03 (0.03) 486.950 0.03 (0.02) 485.826 Density of community or ganizations 0.42 (0.20)* 482.699 0.13 (0.11) 486.268 0.06 (0.07) 486.763 0.02 (0.05) 487.454 Models were adjusted for individual-level covariates and sociodemographic characteristics at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates and sociodemographic characteristics and walkability at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates, and sociodemographic characteristics, walkability, density of green spaces, density of off-premises alcohol outlets and density of bars at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) *** p < 0.001; **p < 0.01; *p < 0.05; †p < 0.10 Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 14 of 21 Table 4 Correlations between buffers for neighborhood-level (AIC = 482.368) but the difference in AIC between the variables two models were negligible (ΔAIC = 0.827). An increase in density of community organizations within a 250 m 500 m 750 m 1000 m radius (ß = 0.42; SE = 0.20; p = 0.039) was also associated Rho Rho Rho with greater physical/sexual DV victimization. No sig- 250 m nificant associations were found between the density of Density of off-premises alcohol outlets 0.72 0.66 0.63 green spaces and DV. Density of bars 0.60 0.48 0.43 Crime rate 0.71 0.61 0.56 Correlations between buffers for neighborhood‑level Walkability 0.90 0.85 0.82 variables NDVI 0.95 0.89 0.86 Table 4 shows the Spearman’s rho correlation coeffi - Density of green spaces 0.61 0.57 0.47 cients between the four buffers for each neighborhood- Density of community organizations 0.63 0.47 0.38 level variable. All correlations were significant (p < 0.001) 500 m with rho coefficients ranging from 0.38 to 99. There were Density of off-premises alcohol outlets 0.90 0.85 strong correlations across buffers for walkability (from Density of bars 0.80 0.71 0.82 to 0.98) and NDVI (from 0.86 to 0.99). Moderate to Crime rate 0.88 0.81 strong correlations were observed for crime rate (from Walkability 0.96 0.93 0.56 to 0.94), density of green spaces (from 0.47 to 0.93), NDVI 0.97 0.94 density of alcohol outlets (from 0.63 to 0.96), and density Density of green spaces 0.86 0.78 of bars (from 0.43 to 0.89). Finally, there were weak to Density of community organizations 0.77 0.65 strong correlations between buffers for density of com - 750 m munity organizations (from 0.38 to 0.86). Density of off-premises alcohol outlets 0.96 Density of bars 0.89 Discussion Crime rate 0.94 The current study aimed to contribute to the body of Walkability 0.98 research on the determinants of DV by analyzing the NDVI 0.99 effects of several neighborhood characteristics (crimi - Density of green spaces 0.93 nality, greenness, walkability, density of green spaces, Density of community organizations 0.86 alcohol outlets, and community organizations) using a All p values < 0.001 multi-scalar approach. To our knowledge, no other study had simultaneously assessed various forms of DV vic- timization and perpetration in relation to multiple neigh- Neighborhood risk factors and DV borhood risk and protective factors. Results of this study The effects of neighborhood risk factors on DV were suggest that several neighborhood characteristics could mainly observed for DV perpetration among boys. There influence DV. These factors would be active at different was an association between crime rate within a buffer of scales, and their effects would be modified by gender 500 m and 750 m and the perpetration of physical/sexual with a varying amplitude depending on the form of vio- violence for boys only. Although there was a strong cor- lence considered. relation for crime rate between the buffer of 500 m and Our results showed that associations between neigh- 750 m, the effect of this characteristic was stronger in borhood factors and DV varied across buffers. In addi - the model using a buffer of 500 m (substantially lower tion to these findings, correlations between the buffers AIC). A positive association was also observed among of 500 m, 750 m, and 1000 m were strong for most vari- boys between the crime rate within a 250 m radius and ables, but correlations between the buffer of 250 m and physical/sexual DV victimization. To date, only stud- the other buffers were weaker, suggesting that processes ies that have used measures of perceived neighborhood acting at different scales could be captured. Criminality, disorder have found significant associations with DV density of green spaces, alcohol outlets, and community [21–24]. However, perceptual measures are subject to organizations were specifically sensitive to the influence same-source bias [79]. This bias occurs when outcomes of scale. The sensitivity to scale in our findings is indica - and neighborhood characteristics are self-reported and tive that not all neighborhood determinants of DV act at can be interrelated. In our case, having experienced DV the same scale. Reflecting on the choice of an appropriate could influence the perception of neighborhood disorder scale to analyze associations between neighborhood fac- and vice versa. Studies using police data to describe the tors and DV is essential to reduce the risk of bias in esti- effect of crime have found no significant association with mating the risks associated with these factors. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 15 of 21 DV [25, 26]. However, crime rate was measured at the between the buffer of 500 m, 750 m, and 1000 m were census tract level, which corresponds to the average size strong, but the effect of this characteristic on DV is only of population and area of a 750 m buffer. Our results sug - observed for the 500 m buffer, suggesting that the density gest that this variable may act more locally (250 m and of these retails in the immediate neighborhood may have 500 m). Exposure to neighborhood crime could lead to a more significant influence on consumption. Finally, the the normalization of violence, which may influence the effect of the density of off-premises alcohol outlets was adoption of violent behavior in intimate relationships [9, only observed for boys. This gender sensitivity may be 80]. These processes could be mainly local. Crime tends explained by lower parental monitoring for boys than to be concentrated in some microenvironments (e.g., girls [55]. Furthermore, our results suggested a protec- blocks, street segments) [81] and adolescents living in tive effect of the density of bars (250 m and 1000 m) on these microenvironments could be more impacted by the psychological DV victimization among girls. It is possi- negative effects of criminality. Let us note that the cor - ble that the density of bars was confounded by the den- relation between the buffers of 500 m and 750 m was sity of other services that have not been considered in strong but, a weaker correlation was found between the this study. In particular, cafés and malls could contribute buffers of 500 m and 1000 m. In addition, crime rate in to social interaction and social cohesion in neighbor- a 250 m buffer was only moderately correlated with that hoods [84], a potential protective factor for DV [42, 43]. in the other buffers. Smaller buffers (especially the 250 m Moreover, let us note that the correlation between the buffer) may better capture the concentration of criminal - 250 m and 1000 m buffers for density of bars was moder - ity, while larger buffers could more adequately describe ate (rho = 0.43), suggesting that different processes could processes occurring at larger scales. In addition to these have been captured at each scale. The 1000 m buffer differences in the spatial scale of analysis, adjusting for could reflect the density and diversity in the neighbor - the level of greenness, walkability and density of bars, off- hoods, while the 250 m buffer could better describe the premises alcohol outlets and green spaces resulted in a immediate proximity to services. In some residential suppression effect. Sensitivity analyses in which all these areas where services may be highly concentrated (e.g., on variables were removed found no association between a few streets), the 250 m buffer zone may better reflect crime rate and physical/sexual DV perpetration and vic- the proximity of these places, which would be masked by timization (results available upon request). Suppression the use of larger buffers. occurs when the inclusion of one or more variables in a model increases the effect of another variable by remov - Neighborhood protective factors and DV ing irrelevant variability in the predictors, leading to a Among the protective factors analyzed in this study, more robust estimate of the effect [82]. The adjustment walkability was found to be linked to several forms of DV for a set of neighborhood-level variables in our models for girls but not for boys. Increased walkability within a allowed to isolate the effect of crime that was not related buffer of 500 m, 750 m, and 1000 m around participants’ to the physical characteristics of the environment. These homes was negatively associated with psychological DV results highlight the relevance of our DAG to investi- perpetration and victimization. An increase of walkabil- gate the links between crime and DV. Therefore, further ity in a 250 m and 500 m buffer could also reduce the research should consider contextual factors when analyz- risk of physical/sexual DV victimization. As our theo- ing these associations. retical framework suggests, this physical characteristic of Access to alcohol outlets could also be a potential risk neighborhoods may encourage active behavior and phys- factor for boys. The density of off-premises alcohol out - ical activity [49, 71, 72] as well as contribute to greater lets within a 500 m radius was positively associated with social interactions and a sense of belonging [11, 50], psychological DV perpetration among boys. Several which could positively affect adolescent behaviors, such studies exploring the links between the density of alcohol as reducing violent behavior [52]. These conditions could outlets and intimate partner violence have also identified promote the development of social ties and the presence significant associations [27–29]. Selling alcohol to minors of witnesses who could act to prevent such behavior or is prohibited, which may make it difficult for adoles - intervene. Girls, in particular, would benefit from the cents to visit bars. However, off-premises alcohol outlets positive effects of walkability. The effect of this charac - could be used to obtain alcohol through older relatives teristic of the neighborhoods on psychological violence (e.g., friends, brothers, sisters) [83]. The density of off- was observed up to 1000 m, an acceptable walking dis- premises alcohol outlets may thus influence consumption tance for adolescents [85]. The small difference between among adolescents [32], which may relate to an increased the buffer zones in the estimates of the effect of walkabil - risk of perpetrating DV [7]. According to our results, ity and models’ fit suggests a small influence of the spa - correlations for density of off-premises alcohol outlets tial scale of analysis for this variable. This minor scaling Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 16 of 21 effect was also observed when analyzing correlations among boys. Conversely, an increase in greenness within between buffers. All buffers were strongly correlated (rho a buffer of 750 m and 1000 m was associated with a ranging from 0.82 to 0.98). However, walkability could greater risk of physical/sexual DV victimization among have a more local influence (250 m and 500 m) for the girls. Although no study has evaluated the effect of more severe forms of DV (physical/sexual). Walkability greenness on DV, this feature of neighborhoods could could foster collective efficacy [11], which may depend reduce aggressive behavior [37], a potential risk factor for on social interactions and proximity between residents, DV victimization [89]. Our results showed that boys are as other social processes [86]. Walkability of the local more sensitive to the protective effects of greenness than environment could affect these social processes, while girls. For girls, higher greenness would have a deleterious walkability on a larger scale could have a greater influ - effect on DV victimization, suggesting that this link could ence on active behavior. Future studies should investi- involve other variables that were not measured, such as gate these potential mediation effects to provide a better parental supervision and social control. The characteris - understanding of the mechanisms through which walk- tics of the physical environment may influence parents’ ability acts on DV. perception of the neighborhood, and positive attributes The density of community organizations in a buffer could lead to a more permissive parenting style regard- of 250 m was positively associated with physical/sexual ing outdoor activities [90]. In neighborhoods with a high DV victimization among girls. This result seems coun - level of greenness, girls may have more freedom to go ter-intuitive as these resources tend to positively affect out alone, without parental supervision. However, such adolescents by offering them supervised activities and neighborhoods could be associated with the presence of opportunities to develop their social ties [10, 44]. How- shielded areas with a limited presence of adults resulting ever, some studies suggest that organizations with lit- in opportunities for unsupervised activities [91]. Such tle structured activities may have deleterious effects on contexts could, in turn, increase the risk of DV victimiza- adolescents by promoting interactions between at-risk tion for girls. Furthermore, correlations between buffers individuals [87]. The presence of these services in the showed a minor scaling effect for greenness (rho ranging immediate environment of girls may increase their risk from 0.86 to 0.99). The variation in the level of green - of physical or sexual victimization due to a greater pres- ness could be mainly explained at larger scales, as there ence of adolescents adhering to norms and attitudes would be little local variation in this factor. The buffer that encourage violence. To test this hypothesis, future of 750 m and 1000 m, for which the effects of greenness studies should identify the effect of community organi - were observed, may better capture these processes. These zations by distinguishing between the different types of results are consistent with those reported by Younan et al. organizations and considering the services offered by [37] in that the effects of greenness are observed at larger these organizations. These studies should also consider scales only (1000 m for boys and 750 m and 1000 m for the influence of the spatial scale of analysis. The effect of girls). These radii could correspond to the areas used by density of community organizations was found only for adolescents. The 1000 m distance corresponds to about the buffer of 250 m. In addition, correlations between the 15 min of walking and is consistent with the travel dis- 250 m buffer and the other buffers ranged from moder - tance by walking of adolescents [85]. ate to weak. The scale sensitivity for this variable could Finally, no association between the density of green be explained by the small number of community organi- spaces and any form of DV was found. The density of zations (n = 423) and suggest the importance of choos- these resources is not necessarily associated with greater ing the scale of analysis carefully. Finally, let us note that use. Some characteristics of green spaces, such as facili- no effect was observed for boys and other forms of vio - ties (e.g., sports fields, picnic tables), could be more pre - lence for girls. The availability of these resources is not dictive of adolescents’ use [92]. Further studies should always associated with their use by adolescents, which examine the effect of access to green spaces on DV, con - may depend on other conditions such as the interest of sidering their characteristics. The influence of the spa - adolescents and the cost of the activities [88]. Therefore, tial scale of analysis should also be considered because it would be relevant to further analyze the effect of ado - a potential scaling effect for the density of green spaces lescents’ use of community resources using measures may exist. Our results showed that correlations between of social participation in order to adequately assess the the buffer of 500 m, 750 m, and 1000 m for the density effect of these environments. of green spaces were strong. However, these buffers were Associations between greenness and DV were also only moderately correlated with the buffer of 250 m. observed but, the direction of this relationship varies by The 250 m buffer may better capture proximity to this gender. Greenness in the 1000 m radius buffer could have resource, while larger buffers would be more representa - a protective effect on psychological DV victimization tive of density within neighborhoods. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 17 of 21 Strengths and limitations knowledge, no study has assessed the effect of collective This study contributes to the research on neighbor - efficacy using small spatial units as a proxy of the local hoods’ factors of DV and demonstrates the importance environment. Therefore, future research should develop of considering scale when analyzing these associations. methods to estimate the characteristics of the social envi- The effects of some neighborhood characteristics related ronment at fine scales and further investigate the influ - to social disorganization theory that had already been ence of these factors on DV by considering scale effects. studied in the United States were analyzed, as well as Small area estimation methods [93] are promising tools other characteristics of the physical environment that for estimating components of the social environment had not yet been documented. This study also provides in small spatial units from survey data. Lastly, logistic a multi-scalar understanding of neighborhood-level pro- regressions performed in this study did not consider the cesses and highlights the importance of neighborhood spatial dependency of the data. However, spatial pat- definition. tern in neighborhoods’ characteristics as well as health Some limitations should be considered. Firstly, self- outcomes have been observed in previous research. For reported data for DV are at risk of social desirability bias. example, a study found a spatial patterning of neighbor- This bias could not be controlled due to the absence of hood physical characteristics and depressive symptoms a social desirability measure in the QHSHSS. Secondly, among adolescents [94]. Authors suggested that depres- physical and sexual DV were assessed jointly as the prev- sive symptoms could be clustered due to social interac- alence of sexual DV perpetration for girls was low (2%). tions (e.g., neighborhood peer effects) and exposure Strong correlations between these forms of violence for and response to the same neighborhood factors. Similar both girls and boys were observed in our data, suggest- processes could also be observed with DV. Future stud- ing that physical and sexual DV may co-occur. Using such ies evaluating associations between neighborhood char- a measure is also consistent with other studies reporting acteristics and DV should assess the potential spatial on the relationship between DV and neighborhood fac- pattern of DV by considering the spatial dependency of tors [25–29]. However, measuring physical and sexual data. Spatial regressions would address this issue by tak- DV jointly did not make it possible to identify factors ing into account spatial autocorrelation (i.e., level of simi- specifically associated with sexual DV. Low-prevalence larity/dissimilarity between neighboring observations) outcomes could be more sensitive to outliers and require and spatial non-stationarity (i.e., spatial variability in the a large sample to ensure a representative sample for amplitude and direction of estimated effects) [95]. How - analyzing the influence of neighborhood-level factors. ever, these methods require an adequate spatial density Future studies on the relationship between neighbor- of participants, which is rarely the case for national sur- hoods’ characteristics and sexual DV should therefore veys. Future studies should therefore conduct large sam- focus on sexual DV. Thirdly, more than 26% of the par - ple surveys to address this objective. ticipants were excluded because they did not provide a valid postal code. This information was required to assign Conclusions to respondents the neighborhood-level variables. Missing Multiple neighborhood characteristics could influence data on postal codes led to a reduction in the sample size, DV at different scales. Social disorganization theory, which reduced statistical power. However, let us note that often referred to for explaining these relationships [9], these missing data should not substantially impact our could only partially explain the neighborhood effects. estimates as they are randomly distributed for all covari- Our results suggest that some physical characteristics of ates, except for grade level among girls. Fourthly, social neighborhoods, such as walkability, could have a protec- environment characteristics were not analyzed in this tive effect on DV. Future studies should investigate the study but could influence DV. Some studies have focused effect of the physical environment on these behaviors on the associations between DV and collective efficacy by analyzing underlying processes (e.g., social support, [23, 43, 53, 54], a concept related to social disorganization social participation). Furthermore, our study showed theory [17]. However, collective efficacy was not assessed that there is no single scale to assess the effect of neigh - in the QHSHSS, and the spatial density of respondents borhoods’ characteristics. Therefore, the comparison of was too low to derive neighborhood-level variables from scales should be systematic to consider the multiscale individual responses. Studies in urban areas most often effects of neighborhoods’ characteristics. used measurements at the individual level [23, 43, 53, 54], Our results regarding the association between neigh- which are limited in neighborhood-level processes. Roth- borhood characteristics and DV could be observed in man et al. [43] measured collective efficacy at the neigh - several geographic contexts but may not be applicable to borhood level, but neighborhoods were operationalized urban contexts of the Global South. In effect, studies on with larger spatial units than the census tracts. To our neighborhoods’ effects on DV conducted outside of the Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 18 of 21 United States and Canada are still rare. Specifically, cit - of neighborhoods that are modifiable and reach large ies in the Global South are different from those in North numbers of people [101]. Urban planning and pub- America (e.g., presence of informal settlements, larger lic policies could represent new avenues in efforts to socioeconomic and health inequalities, and extreme efficiently prevent DV. Interventions implemented urban growth). Such differences may cause the neigh - around the world to improve the physical local envi- borhood-level factors analyzed in our study to influ - ronment (e.g., greening, increasing density of facilities, ence DV differently in the Global South, namely, if other streetscape improvements) showed positive effects on neighborhood-level factors contribute to the incidence adolescent behaviors and health outcomes [102, 103]. of DV. Despite these issues, neighborhood effects have Improving the physical environment would also pro- been observed in several countries. For example, empiri- mote social cohesion and residents’ engagement in cal studies found associations between some neighbor- community life, which could benefit individuals [104]. hood factors, such as sociodemographic characteristics, In the United States, an intervention that aimed at neighborhood disorder and crime, and intimate partner greening neighborhoods and improving social cohe- violence among adults in Europe [96], Asia [97] as well as sion through residents’ engagement has shown promis- Africa [98]. Associations between neighborhoods’ char- ing results in preventing violence and crime [105]. Such acteristics and DV among adolescents may therefore be programs could be effective in reducing DV and could observed in different regions of the world, but local spe - also have positive effects on adolescents’ health and cificities could lead to differences in the effects of the fac - well-being. tors analyzed in this study. Although the relationship between neighborhood Supplementary Information factors and DV may depend on geographic context, the The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12942- 022- 00306-3. influence of spatial scale of analysis is likely to affect most types of local environments, suggesting that our Additional file 1: Directed Acyclic Graph. results on scale sensitivity are not limited to the North American context. In effect, the neighborhood factors Acknowledgements analyzed in the current studies are likely to influence The authors thank the Québec Population Health Research Network (QPHRN) individuals in different regions of the world. Several for its contribution to the financing of this publication. The authors would also like to thank the Québec Statistics Institute for providing the dataset studies around the world have suggested neighbor- from the Québec Health Survey of High School Students. The first author was hood effects on health outcomes, and some have shown supported by a scholarship from the Canada Research Chair in interpersonal potential scale effects for certain factors. For example, traumas and resilience. a literature review reported that greenness and access Author contributions to green spaces could influence health outcomes in dif - PR, MP, and MH contributed to the study conception, the interpretation of the ferent regions of the world, including the Global South, results, and reviewed the manuscript. PR performed analysis and drafted the manuscript. This study was supervised by MP and MH. All authors read and and suggested that these effects could depend on the approved the final manuscript. spatial scale of analysis (e.g., large versus small buffers to operationalize exposure to green space) [99]. Fur- Funding Not applicable. thermore, the egocentric neighborhood is a promising approach that could be easily implemented in various Availability of data and materials geographic settings. Although the definition of admin - Data from the Québec Health Survey of High School Students are provided by Québec Statistics Institute and are not publicly available. Statistics Canada istrative units (size and shape) varies across countries, 2016 Census data at the dissemination area level are publicly available: https:// egocentric neighborhoods ensure consistency in oper- www150. statc an. gc. ca/ n1/ en/ catal ogue/ 98- 401- X2016 044. Data on the popu- ationalizing neighborhoods, which may increase the lation size of the dissemination blocks are also provided by Statistics Canada and are publicly available: https:// www150. statc an. gc. ca/ n1/ en/ catal ogue/ 92- comparability of studies. Finally, measures used in the 163-X. All other data used and analyzed during the current study are available current study to operationalize neighborhood factors from the corresponding author on reasonable request. are easily replicable, and some have been validated in different contexts. For example, assessing greenness by Declarations using NDVI is consistent with several studies around Ethics approval and consent to participate the world [99]. The walkability index has also been used Ethical approval was obtained from the ethics board of the Université du in many countries in the Global North [71–73, 100] as Québec à Montréal and from the ethics board of the Québec Statistics Insti- well as in the Global South [100]. tute. Data from the Québec Health Survey of High School Students used in the current study were anonymous, and no participants were contacted. This study also has practice implications. Our results highlight the importance of physical features Competing interests The authors declare that they have no competing interests. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 19 of 21 Received: 30 March 2022 Accepted: 1 June 2022 22. Champion H, Foley KL, Sigmon-Smith K, Sutfin EL, DuRant RH. Con- textual factors and health risk behaviors associated with date fight - ing among high school students. Women Health. 2008;47(3):1–22. 23. Chang L-Y, Foshee VA, Reyes HLM, Ennett ST, Halpern CT. Direct and indirect effects of neighborhood characteristics on the perpe - tration of dating violence across adolescence. J Youth Adolesc. References 2015;44(3):727–44. 1. Wincentak K, Connolly J, Card N. Teen dating violence: A meta-analytic 24. Reed E, Silverman JG, Raj A, Decker MR, Miller E. Male Perpetration review of prevalence rates. Psychol Violence. 2017;7(2):224–41. of teen dating violence: associations with neighborhood violence 2. Leen E, Sorbring E, Mawer M, Holdsworth E, Helsing B, Bowen E. Preva- involvement, gender attitudes, and perceived peer and neighbor- lence, dynamic risk factors and the efficacy of primary interventions hood norms. J Urban Health. 2011;88(2):226–39. for adolescent dating violence: an international review. Aggress Violent 25. Li Q, Kirby RS, Sigler RT, Hwang S-S, LaGory ME, Goldenberg RL. Beh. 2013;18(1):159–74. A Multilevel Analysis of Individual, Household, and Neighbor- 3. Banyard VL, Cross C. Consequences of teen dating violence: under- hood Correlates of Intimate Partner Violence Among Low-Income standing intervening variables in ecological context. Violence Against Pregnant Women in Jefferson County Alabama. Am J Public Health. Women. 2008;14(9):998–1013. 2010;100(3):531–9. 4. Eaton DK, Davis KS, Barrios L, Brener ND, Noonan RK. Associations of 26. Taylor BG, Mumford EA, Okeke N, Rothman E. Neighborhood violent dating violence victimization with lifetime participation, co-occurrence, crime and adolescent relationship aggression. Aggressive Behav. and early initiation of risk behaviors among us high school students. J 2020;46(1):25–36. Interpers Viole. 2007;22(5):585–602. 27. Iritani BJ, Waller MW, Halpern CT, Moracco KE, Christ SL, Flewelling RL. 5. Taquette SR, Monteiro DLM. Causes and consequences of ado- Alcohol outlet density and young women’s perpetration of violence lescent dating violence: a systematic review. J Inj Violence Res. toward male intimate partners. J Family Viol. 2013;28(5):459–70. 2019;11(2):137–47. 28. Waller MW, Iritani BJ, Christ SL, Tucker Halpern C, Moracco KE, Flewelling 6. Hébert M, Daspe M, Lapierre A, Godbout N, Blais M, Fernet M, Lavoie RL. Perpetration of intimate partner violence by young adult males: the F. A meta-analysis of risk and protective factors for dating violence association with alcohol outlet density and drinking behavior. Health victimization: the role of family and peer interpersonal context. Trauma Place. 2013;21:10–9. Violence Abuse. 2019;20(4):574–90. 29. Waller MW, Iritani BJ, Flewelling RL, Christ SL, Halpern CT, Moracco 7. Vagi KJ, Rothman EF, Latzman NE, Tharp AT, Hall DM, Breiding MJ. KE. Violence victimization of young men in heterosexual relation- Beyond correlates: a review of risk and protective factors for adolescent ships: does alcohol outlet density influence outcomes? Violence Vict. dating violence perpetration. J Youth Adolesc. 2013;42(4):633–49. 2012;4:527–47. 8. Vézina J, Hébert M. Risk factors for victimization in romantic relation- 30. Sampson RJ, Raudenbush SW. Systematic social observation of public ships of young women: a review of empirical studies and implications spaces: a new look at disorder in urban neighborhoods. Am J Sociol. for prevention. Trauma Violence Abuse. 2007;8(1):33–66. 1999;105(3):603–51. 9. Johnson RM, Parker EM, Rinehart J, Nail J, Rothman EF. Neighborhood 31. Sampson RJ, Raudenbush SW. Seeing disorder: neighborhood stigma factors and dating violence among youth: a systematic review. Am J and the social construction of “Broken Windows.” Soc Psychol Quarterly. Prev Med. 2015;49(3):458–66. 2004;67(4):319–42. 10. Gardner M, Brooks-Gunn J. Adolescents’ exposure to community vio- 32. Shih RA, Mullins L, Ewing BA, Miyashiro L, Tucker JS, Pedersen ER, lence: are neighborhood youth organizations protective? J Community Miles JNV, D’Amico EJ. Associations between neighborhood alcohol Psychol. 2009;37(4):505–25. availability and young adolescent alcohol use. Psychol Addict Behav. 11. Prince DM, Hohl B, Hunter BA, Thompson AB, Matlin SL, Hausman AJ, 2015;29(4):950–9. Tebes JK. Collective efficacy as a key context in neighborhood support 33. White RM, Witherspoon DP, Wei W, Zhao C, Pasco MC, Maereg TM. for urban youth. Am J Community Psychol. 2019;63(1–2):179–89. Adolescent development in context: a decade review of neighbor. J Res 12. Vanaken G-J, Danckaerts M. Impact of green space exposure on chil- Adolescence. 2021;31(4):944–65. dren’s and adolescents’ mental health: a systematic review. Int J Environ 34. Wilson EO. Biophilia and the conservation ethic. In: Evolutionary per- Res Public Health. 2018;15(12):2668. spectives on environmental problems. Routledge; 2017: 263–272. 13. Philibert M, Riva M. Contribution de la géographie quantitative à l’étude 35. Bezold CP, Banay RF, Coull BA, Hart JE, James P, Kubzansky LD, Missmer des inégalités sociales de santé : Analyse des effets de milieu. Cahiers SA, Laden F. The association between natural environments and de géographie du Québec. 2011;55(156):343–61. depressive symptoms in adolescents living in the United States. J 14. Chaix B, Merlo J, Evans D, Leal C, Havard S. Neighbourhoods in eco-epi- Adolesc Health. 2018;62(4):488–95. demiologic research: delimiting personal exposure areas. aresponse to 36. Mavoa S, Lucassen M, Denny S, Utter J, Clark T, Smith M. Natural Riva, Gauvin Apparicio and Brodeur. Soc Sci Med. 2009;69(9):1306–10. neighbourhood environments and the emotional health of urban New 15. Hipp JR, Bates CJ. Egohoods. Capturing change in spatial crime pat- Zealand adolescents. Landsc Urban Plan. 2019;191: 103638. terns. In: The Oxford Handbook of Environmental Criminology. Edited by 37. Younan D, Tuvblad C, Li L, Wu J, Lurmann F, Franklin M, Berhane K, Bruinsma GJN, Johnson SD: Oxford University Press; 2018. McConnell R, Wu AH, Baker LA, et al. Environmental determinants of 16. Hipp JR, Boessen A. Egohoods as waves washing across the city: a new aggression in adolescents: role of urban neighborhood greenspace. J measure of “neighborhoods.” Criminology. 2013;51(2):287–327. Am Acad Child Adolesc Psychiatry. 2016;55(7):591–601. 17. Sampson RJ, Raudenbush SW, Earls F. Neighborhoods and vio- 38. Kuo FE, Sullivan WC. Aggression and violence in the inner city: effects of lent crime: a multilevel study of collective efficacy. Science. environment via mental fatigue. Environ Behav. 2001;33(4):543–71. 1997;277(5328):918–24. 39. Lee ACK, Maheswaran R. The health benefits of urban green spaces: a 18. Brunton-Smith I, Jackson J, Sutherland A. Bridging structure and per- review of the evidence. J Public Health. 2010;33(2):212–22. ception: on the neighbourhood ecology of beliefs and worries about 40. Biddle SJH, Ciaccioni S, Thomas G, Vergeer I. Physical activity and mental violent crime. Br J Criminol. 2014;54(4):503–26. health in children and adolescents: an updated review of reviews and 19. Cunradi CB. Neighborhoods, alcohol outlets and intimate partner an analysis of causality. Psychol Sport Exerc. 2019;42:146–55. violence: addressing research gaps in explanatory mechanisms. Int J 41. Jennings V, Bamkole O. The relationship between social cohesion and Environ Res Public Health. 2010;7(3):799–813. urban green space: an avenue for health promotion. Int J Environ Res 20. Gruenewald PJ. The spatial ecology of alcohol problems: niche theory Public Health. 2019;16(3):452. and assortative drinking. Addiction. 2007;102(6):870–8. 42. Garthe RC, Gorman-Smith D, Gregory J, Schoeny EM. Neighborhood 21. Banyard VL, Cross C, Modecki KL. Interpersonal violence in adoles- concentrated disadvantage and dating violence among urban ado- cence: ecological correlates of self-reported perpetration. J Interpers lescents: the mediating role of neighborhood social processes. Am J Violence. 2006;21(10):1314–32. Commun Sychol. 2018;61(34):310–20. Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 20 of 21 43. Rothman EF, Johnson RM, Young R, Weinberg J, Azrael D, Molnar BE. 68. Bow CJD, Waters NM, Faris PD, Seidel JE, Galbraith PD, Knudtson ML, Neighborhood-level factors associated with physical dating violence Ghali WA, The AI. Accuracy of city postal code coordinates as a proxy perpetration: results of a representative survey conducted in Boston. for location of residence. Int J Health Geogr. 2004;3(1):5–5. MA J Urban Health. 2011;88(2):201–13. 69. Rhew IC, Vander Stoep A, Kearney A, Smith NL, Dunbar MD. Valida- 44. Quane JM, Rankin BH: Does it pay to participate? Neighborhood-based tion of the normalized difference vegetation index as a measure of organizations and the social development of urban adolescents. 2006; neighborhood greenness. Ann Epidemiol. 2011;21(12):946–52. 2006. 70. Shannon CE. A mathematical theory of communication. Bell Syst 45. Browning CR, Soller B. Moving beyond neighborhood: activity spaces Tech J. 1948;27(3):379–423. and ecological networks as contexts for youth development. Cityscape. 71. Frank LD, Sallis JF, Saelens BE, Leary L, Cain K, Conway TL, Hess PM. 2014;16(1):165–96. The development of a walkability index: application to the Neighbor- 46. Gottfredson DC, Gerstenblith SA, Soulé DA, Womer SC, Lu S. Do after hood Quality of Life Study. Br J Sports Med. 2010;44(13):924–33. school programs reduce delinquency? Prev Sci. 2004;5(4):253–66. 72. Sundquist K, Eriksson U, Kawakami N, Skog L, Ohlsson H, Arvidsson 47. Durlak JA, Weissberg RP, Pachan M. A meta-analysis of after-school D. Neighborhood walkability, physical activity, and walking behavior: programs that seek to promote personal and social skills in children the Swedish Neighborhood and Physical Activity (SNAP) study. Soc and adolescents. Am J Community Psychol. 2010;45(3–4):294–309. Sci Med. 2011;72(8):1266–73. 48. Molnar BE, Cerda M, Roberts AL, Buka SL. Eec ff ts of neighborhood 73. Mayne DJ, Morgan GG, Willmore A, Rose N, Jalaludin B, Bambrick resources on aggressive and delinquent behaviors among urban H, Bauman A. An objective index of walkability for research and youths. Am J Public Health. 2008;98(6):1086–93. planning in the Sydney metropolitan region of New South Wales, 49. Frank LD, Engelke P, Schmid T. Health and community design: The Australia: an ecological study. Int J Health Geogr. 2013;12:61–61. impact of the built environment on physical activity: Island Press; 2003. 74. Inglehart R, Welzel C. Changing mass priorities: the link between 50. Lund H. Pedestrian environments and sense of community. J Plan Educ modernization and democracy. Perspect Polit. 2010;8(2):551–67. Res. 2002;21(3):301–12. 75. Sas Institute: SAS/STAT 9.3 User’s Guide: Survey Data Analysis (book 51. Culyba AJ, Jacoby SF, Richmond TS, Fein JA, Hohl BC, Branas CC. Modifi- Excerpt): SAS institute. 2011. able neighborhood features associated with adolescent homicide. 76. Pearl J. Causal Diagrams for Empirical Research. Biometrika. JAMA Pediatr. 2016;170(5):473–80. 1995;82(4):669–88. 52. Bushover B, Miller E, Bair-Merritt M, Abebe K, Culyba A. Physical environ- 77. Schisterman EF, Cole SR, Platt RW. Overadjustment bias and ment and violence perpetration among male youth in Pittsburgh: a unnecessary adjustment in epidemiologic studies. Epidemiology. spatial analysis. Inj Prev. 2020;26(6):588–92. 2009;20(4):488–95. 53. Jain S, Buka SL, Subramanian SV, Molnar BE. Neighborhood predictors 78. Burnham KP, Anderson DR. Model selection and multimodel of dating violence victimization and perpetration in young adulthood: inference-2nd ed: a practical information-theoretic approach. New a multilevel study. Am J Public Health. 2010;100(9):1737–44. york: Springer-verlag new york Inc. 2002. 54. Schnurr MP, Lohman BJ. The impact of collective efficacy on risks for 79. Diez Roux AV. Neighborhoods and health: where are we and were do adolescents perpetration of dating violence. J Youth Adolescence J we go from here? Rev Epidemiol Sante Publique. 2007;55(1):13–21. Youth Adolescence A Multidiscipl Res Publicat. 2013;42(4):518–35. 80. Raghavan C, Mennerich A, Sexton E, James SE. Community violence 55. Axinn WG, Young-Demarco L, Ro MC. Gender double standards in and its direct, indirect, and mediating effects on intimate partner parenting attitudes. Soc Sci Res. 2011;40(2):417–32. violence. Violence Against Women. 2006;12(12):1132–49. 56. Kupersmidt JB, Griesler PC, DeRosier ME, Patterson CJ, Davis PW. 81. Braga AA, Papachristos AV, Hureau DM. The concentration and stabil- Childhood aggression and peer relations in the context of family and ity of gun violence at micro places in Boston, 1980–2008. J Quant neighborhood factors. Child Dev. 1995;66(2):360–75. Criminol. 2010;26(1):33–53. 57. Leventhal T, BrooksGunn J. The neighborhoods they live in: the effects 82. MacKinnon DP, Krull JL, Lockwood CM. Equivalence of the mediation, of neighborhood residence on child and adolescent outcomes. Psychol confounding and suppression effect. Prev Sci. 2000;1(4):173–81. Bullet. 2000;126(2):309–37. 83. Friese B, Grube JW, Moore RS. Youth acquisition of alcohol and drink- 58. Fotheringham AS, Wong DWS. The Modifiable Areal Unit Problem ing contexts: an in-depth look. J Drug Educ. 2013;43(4):385–403. in Multivariate Statistical Analysis. Environ Plann A Economy Space. 84. Oldenburg R. The great good place: Cafes, coffee shops, bookstores, 1991;23(7):1025–44. bars, hair salons, and other hangouts at the heart of a community: 59. Openshaw S. The modifiable areal unit problem. Norwich: Geo Da Capo Press; 1999. Abstracts Univ. of East Anglia; 1984. 85. Colabianchi N, Coulton CJ, Hibbert JD, McClure SM, Ievers-Landis 60. Kwan M-P. The uncertain geographic context problem. Ann Assoc Am CE, Davis EM. Adolescent self-defined neighborhoods and activity Geogr. 2012;102(5):958–68. spaces: Spatial overlap and relations to physical activity and obesity. 61. Hipp JR, Perrin AJ. The simultaneous effect of social distance and physi- Health Place. 2014;27:22–9. cal distance on the formation of neighborhood ties. City Community. 86. Hipp JR, Boessen A. Neighborhoods, networks, and crime. In: Chal- 2009;8(1):5–25. lenging Criminological Theory: The Legacy of Ruth Rosner Kornhauser. 62. Apparicio P, Gelb J, Dubé A-S, Kingham S, Gauvin L, Robitaille É. The Edited by Cullen FT: Taylor & Francis; 2017. approaches to measuring the potential spatial access to urban health 87. Mahoney JL, Stattin H, Lord H. Unstructured youth recreation centre services revisited: distance types and aggregation-error issues. Int J participation and antisocial behaviour development: Selection influ- Health Geogr. 2017;16(1):32–32. ences and the moderating role of antisocial peers. Int J Behav Dev. 63. Hewko J, Smoyer-Tomic KE, Hodgson MJ. Measuring neighbourhood 2004;28(6):553–60. spatial accessibility to urban amenities: does aggregation error matter? 88. Bodilly SJ, Beckett MK. Making Out-of-School-Time Matter: Evidence Environ Plann A Econ Space. 2002;34(7):1185–206. for an Action Agenda. In. Santa Monica, CA: RAND; 2005. 64. Ursache A, Regan S, De Marco A, Duncan DT. The Family Life Project 89. O’Donnell L, Stueve A, Myint-U A, Duran R, Agronick G, Wilson-Simmons Key I: Measuring neighborhood deprivation for childhood health and R. Middle school aggression and subsequent intimate partner physical development - scale implications in rural and urban context. Geospat violence. J Youth Adolesc. 2006;35(5):693–703. Health. 2021. https:// doi. org/ 10. 4081/ gh. 2021. 4926. 90. Vlaar J, Brussoni M, Janssen I, Mâsse LC. Roaming the neighbourhood: 65. Oliver LN, Schuurman N, Hall AW. Comparing circular and network influences of independent mobility parenting practices and parental buffers to examine the influence of land use on walking for leisure perceived environment on children’s territorial range. Int J Environ Res and errands. Int J Health Geogr. 2007;6(1):41. Public Health. 2019;16(17):3129. 66. Statistics Canada: Census Profile. 2016 Census. In: Census Profile. 91. Weeland J, Laceulle OM, Nederhof E, Overbeek G, Reijneveld SA. The Statistics Canada; 2017. greener the better? Does neighborhood greenness buffer the effects 67. Straus MA, Hamby SL, Boney-McCoy SUE, Sugarman DB. The revised of stressful life events on externalizing behavior in late adolescence? conflict tactics scales (CTS2): development and preliminary psycho - Health Place. 2019;58: 102163. metric data. J Fam Issues. 1996;17(3):283–316. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 21 of 21 92. Baran PK, Smith WR, Moore RC, Floyd MF, Bocarro JN, Cosco NG, Dan- ninger TM. Park use among youth and adults: examination of individual, social, and urban form factors. Environ Behav. 2013;46(6):768–800. 93. Ghosh M, Rao JN. Small area estimation: an appraisal. Stat Sci. 1994;9(1):55–76. 94. Duncan DT, Piras G, Dunn EC, Johnson RM, Melly SJ, Molnar BE. The built environment and depressive symptoms among urban youth: a spatial regression study. Spat Spatiotemporal Epidemiol. 2013;5:11–25. 95. Anselin L. Spatial Econometrics: Methods and Models. Dordrecht: Springer, Netherlands; 2013. 96. Gracia E, López-Quílez A, Marco M, Lladosa S, Lila M. Exploring neigh- borhood influences on small-area variations in intimate partner vio - lence risk: a Bayesian random-effects modeling approach. Int J Environ Res Public Health. 2014;11(1):866–82. 97. Koenig MA, Stephenson R, Ahmed S, Jejeebhoy SJ, Campbell J. Individual and contextual determinants of domestic violence in North India. Am J Public Health. 2006;96(1):132–8. 98. Uthman OA, Moradi T, Lawoko S. The independent contribution of individual-, neighbourhood-, and country-level socioeconomic position on attitudes towards intimate partner violence against women in sub- Saharan Africa: a multilevel model of direct and moderating effects. Soc Sci Med. 2009;68(10):1801–9. 99. Rigolon A, Browning MHEM, McAnirlin O, Yoon HV. Green space and health equity: a systematic review on the potential of green space to reduce health disparities. Int J Environ Res Public Health. 2021;18(5):2563. 100. Adams MA, Frank LD, Schipperijn J, Smith G, Chapman J, Christiansen LB, Coffee N, Salvo D, du Toit L, Dygrýn J, et al. International variation in neighborhood walkability, transit, and recreation environments using geographic information systems: the IPEN adult study. Int J Health Geogr. 2014;13(1):43. 101. Diez Roux AV. Investigating neighborhood and area effects on health. Am J Public Health. 2001;91(11):1783–9. 102. Hunter RF, Cleland C, Cleary A, Droomers M, Wheeler BW, Sinnett D, Nieuwenhuijsen MJ, Braubach M. Environmental, health, wellbeing, social and equity effects of urban green space interventions: a meta- narrative evidence synthesis. Environ Int. 2019;130:104923. 103. Smith M, Hosking J, Woodward A, Witten K, MacMillan A, Field A, Baas P, Mackie H. Systematic literature review of built environment effects on physical activity and active transport - an update and new findings on health equity. Int J Behav Nutr Phys Act. 2017;14(1):158–158. 104. Aiyer SM, Zimmerman MA, Morrel-Samuels S, Reischl TM. From Broken Windows to Busy Streets: A Community Empowerment Perspective. Health Educ Behav. 2014;42(2):137–47. 105. Heinze JE, Krusky-Morey A, Vagi KJ, Reischl TM, Franzen S, Pruett NK, Cunningham RM, Zimmerman MA. Busy streets theory: the effects of community-engaged greening on violence. Am J Community Psychol. 2018;62(1–2):101–9. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your ﬁeld rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Health Geographics Springer Journals http://www.deepdyve.com/lp/springer-journals/associations-between-neighborhood-characteristics-and-dating-violence-OUBLer4t4v

Loading next page...

References (124)

John Hipp, A. Perrin (2009)
The Simultaneous Effect of Social Distance and Physical Distance on the Formation of Neighborhood Ties
City & Community, 8
J. Pearl (1995)
Causal diagrams for empirical research
Biometrika, 82
D. Duncan, G. Piras, E. Dunn, Renee Johnson, S. Melly, B. Molnar (2013)
The built environment and depressive symptoms among urban youth: A spatial regression study.
Spatial and spatio-temporal epidemiology, 5
H. Lund (2002)
Pedestrian Environments and Sense of Community
Journal of Planning Education and Research, 21
B. Fischer, S. Dash, D. Berrigan (2005)
Health and community design: the impact of the built environment on physical activity
Journal of Epidemiology and Community Health, 59
F. Kuo, W. Sullivan (2001)
Aggression and Violence in the Inner City
Environment and Behavior, 33
R. Sampson, S. Raudenbush (1999)
Systematic Social Observation of Public Spaces: A New Look at Disorder in Urban Neighborhoods1
American Journal of Sociology, 105
Jin Shin, Sang Kim (2006)
A Mathematical Theory of Communication
K. Vagi, E. Rothman, N. Latzman, A. Tharp, D. Hall, M. Breiding (2013)
Beyond Correlates: A Review of Risk and Protective Factors for Adolescent Dating Violence Perpetration
Journal of Youth and Adolescence, 42
(Statistics Canada: Census Profile. 2016 Census. In: Census Profile. Statistics Canada; 2017.)
Statistics Canada: Census Profile. 2016 Census. In: Census Profile. Statistics Canada; 2017.
Statistics Canada: Census Profile. 2016 Census. In: Census Profile. Statistics Canada; 2017., Statistics Canada: Census Profile. 2016 Census. In: Census Profile. Statistics Canada; 2017.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations
Carol Cunradi (2010)
Neighborhoods, Alcohol Outlets and Intimate Partner Violence: Addressing Research Gaps in Explanatory Mechanisms
International Journal of Environmental Research and Public Health, 7
(2017)
Neighborhoods, networks, and crime
Brady Bushover, E. Miller, M. Bair-Merritt, Kaleab Abebe, Alison Culyba (2019)
Physical environment and violence perpetration among male youth in Pittsburgh: a spatial analysis
Injury Prevention, 26
S. Taquette, D. Monteiro (2019)
Causes and consequences of adolescent dating violence: a systematic review
Journal of Injury and Violence Research, 11
A. Lee, R. Maheswaran (2011)
The health benefits of urban green spaces: a review of the evidence.
Journal of public health, 33 2
(Hipp JR, Bates CJ. Egohoods. Capturing change in spatial crime patterns. In: The Oxford Handbook of Environmental Criminology. Edited by Bruinsma GJN, Johnson SD: Oxford University Press; 2018.)
Hipp JR, Bates CJ. Egohoods. Capturing change in spatial crime patterns. In: The Oxford Handbook of Environmental Criminology. Edited by Bruinsma GJN, Johnson SD: Oxford University Press; 2018.
Hipp JR, Bates CJ. Egohoods. Capturing change in spatial crime patterns. In: The Oxford Handbook of Environmental Criminology. Edited by Bruinsma GJN, Johnson SD: Oxford University Press; 2018., Hipp JR, Bates CJ. Egohoods. Capturing change in spatial crime patterns. In: The Oxford Handbook of Environmental Criminology. Edited by Bruinsma GJN, Johnson SD: Oxford University Press; 2018.
(2011)
User's Guide: Survey Data Analysis (book Excerpt): SAS institute
A. Ursache, S. Regan, Allison Marco, D. Duncan, The Investigators (2021)
Measuring neighborhood deprivation for childhood health and development - scale implications in rural and urban context.
Geospatial health, 16 1
Mathieu Philibert, Mylène Riva (2012)
Contribution de la géographie quantitative à l’étude des inégalités sociales de santé : Analyse des effets de milieu
, 55
D. Mackinnon, J. Krull, C. Lockwood (2000)
Equivalence of the Mediation, Confounding and Suppression Effect
Prevention Science, 1
A. Rigolon, M. Browning, Olivia McAnirlin, H. Yoon (2021)
Green Space and Health Equity: A Systematic Review on the Potential of Green Space to Reduce Health Disparities
International Journal of Environmental Research and Public Health, 18
K. Stanilov (2004)
Health and community design: The impact of the built environment on physical activity
Journal of Planning Education and Research
W. Axinn, L. Young-DeMarco, Meeso Ro (2011)
Gender Double Standards in Parenting Attitudes.
Social science research, 40 2
D. Younan, C. Tuvblad, Lianfa Li, Jun Wu, F. Lurmann, M. Franklin, K. Berhane, R. McConnell, A. Wu, L. Baker, J. Chen (2016)
Environmental Determinants of Aggression in Adolescents: Role of Urban Neighborhood Greenspace.
Journal of the American Academy of Child and Adolescent Psychiatry, 55 7
G. Vanaken, M. Danckaerts (2018)
Impact of Green Space Exposure on Children’s and Adolescents’ Mental Health: A Systematic Review
International Journal of Environmental Research and Public Health, 15
L. O'donnell, A. Stueve, Athi Myint-U, R. Duran, G. Agronick, Renée Wilson-Simmons (2006)
Middle School Aggression and Subsequent Intimate Partner Physical Violence
Journal of Youth and Adolescence, 35
Perver Baran, W. Smith, R. Moore, M. Floyd, J. Bocarro, N. Cosco, Tom Danninger (2014)
Park Use Among Youth and Adults
Environment and Behavior, 46
J. Mahoney, H. Stattin, H. Lord (2004)
Unstructured youth recreation centre participation and antisocial behaviour development: Selection influences and the moderating role of antisocial peers
International Journal of Behavioral Development, 28
(Bodilly SJ, Beckett MK. Making Out-of-School-Time Matter: Evidence for an Action Agenda. In. Santa Monica, CA: RAND; 2005.)
Bodilly SJ, Beckett MK. Making Out-of-School-Time Matter: Evidence for an Action Agenda. In. Santa Monica, CA: RAND; 2005.
Bodilly SJ, Beckett MK. Making Out-of-School-Time Matter: Evidence for an Action Agenda. In. Santa Monica, CA: RAND; 2005., Bodilly SJ, Beckett MK. Making Out-of-School-Time Matter: Evidence for an Action Agenda. In. Santa Monica, CA: RAND; 2005.
O. Uthman, T. Moradi, S. Lawoko (2009)
The independent contribution of individual-, neighbourhood-, and country-level socioeconomic position on attitudes towards intimate partner violence against women in sub-Saharan Africa: a multilevel model of direct and moderating effects.
Social science & medicine, 68 10
E. Rothman, Renee Johnson, Robin Young, J. Weinberg, D. Azrael, B. Molnar (2011)
Neighborhood-Level Factors Associated with Physical Dating Violence Perpetration: Results of a Representative Survey Conducted in Boston, MA
Journal of Urban Health, 88
Sophie Aiyer, M. Zimmerman, S. Morrel-Samuels, T. Reischl (2015)
From Broken Windows to Busy Streets
Health Education & Behavior, 42
D. Eaton, Kristen Davis, L. Barrios, N. Brener, R. Noonan (2007)
Associations of Dating Violence Victimization With Lifetime Participation, Co-Occurrence, and Early Initiation of Risk Behaviors Among U.S. High School Students
Journal of Interpersonal Violence, 22
Christopher Browning, B. Soller (2014)
Moving Beyond Neighborhood: Activity Spaces and Ecological Networks As Contexts for Youth Development.
Cityscape, 16 1
P. Apparicio, Jérémy Gelb, A. Dubé, S. Kingham, L. Gauvin, É. Robitaille (2017)
The approaches to measuring the potential spatial access to urban health services revisited: distance types and aggregation-error issues
International Journal of Health Geographics, 16
S. Openshaw (2008)
Modifiable Areal Unit Problem
M. Ghosh, J. Rao (1994)
Small Area Estimation: An Appraisal
Statistical Science, 9
J. Vlaar, M. Brussoni, I. Janssen, L. Mâsse (2019)
Roaming the Neighbourhood: Influences of Independent Mobility Parenting Practices and Parental Perceived Environment on Children’s Territorial Range
International Journal of Environmental Research and Public Health, 16
V. Banyard, Charlotte Cross, K. Modecki (2006)
Interpersonal Violence in Adolescence
Journal of Interpersonal Violence, 21
K. Sundquist, U. Eriksson, N. Kawakami, Lars Skog, H. Ohlsson, D. Arvidsson (2011)
Neighborhood walkability, physical activity, and walking behavior: the Swedish Neighborhood and Physical Activity (SNAP) study.
Social science & medicine, 72 8
Viniece Jennings, Omoshalewa Bamkole (2019)
The Relationship between Social Cohesion and Urban Green Space: An Avenue for Health Promotion
International Journal of Environmental Research and Public Health, 16
B. Molnar, M. Cerdá, A. Roberts, S. Buka (2008)
Effects of neighborhood resources on aggressive and delinquent behaviors among urban youths.
American journal of public health, 98 6
M. Straus, S. Hamby, Sue Boney-McCoy, D. Sugarman (1996)
The Revised Conflict Tactics Scales (CTS2)
Journal of Family Issues, 17
M. Adams, L. Frank, J. Schipperijn, Graham Smith, James Chapman, L. Christiansen, N. Coffee, D. Salvo, Lorinne Toit, J. Dygrýn, A. Hino, P. Lai, S. Mavoa, J. Pinzón, N. Weghe, E. Cerin, R. Davey, D. Macfarlane, N. Owen, J. Sallis (2014)
International variation in neighborhood walkability, transit, and recreation environments using geographic information systems: the IPEN adult study
International Journal of Health Geographics, 13
M. Koenig, R. Stephenson, Saifuddin Ahmed, S. Jejeebhoy, Jacquelyn Campbell (2006)
Individual and contextual determinants of domestic violence in North India.
American journal of public health, 96 1
(Quane JM, Rankin BH: Does it pay to participate? Neighborhood-based organizations and the social development of urban adolescents. 2006; 2006.)
Quane JM, Rankin BH: Does it pay to participate? Neighborhood-based organizations and the social development of urban adolescents. 2006; 2006.
Quane JM, Rankin BH: Does it pay to participate? Neighborhood-based organizations and the social development of urban adolescents. 2006; 2006., Quane JM, Rankin BH: Does it pay to participate? Neighborhood-based organizations and the social development of urban adolescents. 2006; 2006.
Ling-Yin Chang, V. Foshee, H. Reyes, S. Ennett, C. Halpern (2015)
Direct and Indirect Effects of Neighborhood Characteristics on the Perpetration of Dating Violence Across Adolescence
Journal of Youth and Adolescence, 44
Renee Johnson, E. Parker, Jenny Rinehart, Jennifer Nail, E. Rothman (2015)
Neighborhood Factors and Dating Violence Among Youth: A Systematic Review.
American journal of preventive medicine, 49 3
E. Wilson (2017)
Biophilia and the Conservation Ethic
SR Taquette (2019)
137
J Inj Violence Res, 11
E. Gracia, A. López-Quílez, Miriam Marco, S. Lladosa, M. Lila (2014)
Exploring Neighborhood Influences on Small-Area Variations in Intimate Partner Violence Risk: A Bayesian Random-Effects Modeling Approach
International Journal of Environmental Research and Public Health, 11
Melissa Schnurr, Brenda Lohman (2013)
The Impact of Collective Efficacy on Risks for Adolescents’ Perpetration of Dating Violence
Journal of Youth and Adolescence, 42
Margo Gardner, J. Brooks-Gunn (2009)
ADOLESCENTS' EXPOSURE TO COMMUNITY VIOLENCE: ARE NEIGHBORHOOD YOUTH ORGANIZATIONS PROTECTIVE?
Journal of community psychology, 37 4
(2018)
Egohoods. Capturing change in spatial crime patterns. In: The Oxford Handbook of Environmental Criminology. Edited by Bruinsma GJN, Johnson SD
Martha Waller, B. Iritani, R. Flewelling, S. Christ, C. Halpern, Kathryn Moracco (2012)
Violence Victimization of Young Men in Heterosexual Relationships: Does Alcohol Outlet Density Influence Outcomes?
Violence and Victims, 27
J. Vézina, M. Hébert (2007)
Risk Factors for Victimization in Romantic Relationships of Young Women
Trauma, Violence, & Abuse, 8
J. Weeland, O. Laceulle, E. Nederhof, G. Overbeek, S. Reijneveld (2019)
The greener the better? Does neighborhood greenness buffer the effects of stressful life events on externalizing behavior in late adolescence?
Health & place, 58
A. Roux (2007)
Neighborhoods and health: where are we and were do we go from here?
Revue d'epidemiologie et de sante publique, 55 1
C. Bezold, Rachel Banay, B. Coull, J. Hart, Peter James, L. Kubzansky, S. Missmer, F. Laden (2018)
The Association Between Natural Environments and Depressive Symptoms in Adolescents Living in the United States.
The Journal of adolescent health : official publication of the Society for Adolescent Medicine, 62 4
B. Chaix, J. Merlo, David Evans, C. Leal, S. Havard (2009)
Neighbourhoods in eco-epidemiologic research: delimiting personal exposure areas. A response to Riva, Gauvin, Apparicio and Brodeur.
Social science & medicine, 69 9
Martha Waller, B. Iritani, S. Christ, C. Halpern, Kathryn Moracco, R. Flewelling (2013)
Perpetration of intimate partner violence by young adult males: the association with alcohol outlet density and drinking behavior.
Health & place, 21
D. Gottfredson, Stephanie Gerstenblith, D. Soulé, Shannon Womer, Shaoli Lu (2004)
Do After School Programs Reduce Delinquency?
Prevention Science, 5
R. Sampson, S. Raudenbush (2004)
Seeing Disorder: Neighborhood Stigma and the Social Construction of “Broken Windows”
Social Psychology Quarterly, 67
(Burnham KP, Anderson DR. Model selection and multimodel inference-2nd ed: a practical information-theoretic approach. New york: Springer-verlag new york Inc. 2002.)
Burnham KP, Anderson DR. Model selection and multimodel inference-2nd ed: a practical information-theoretic approach. New york: Springer-verlag new york Inc. 2002.
Burnham KP, Anderson DR. Model selection and multimodel inference-2nd ed: a practical information-theoretic approach. New york: Springer-verlag new york Inc. 2002., Burnham KP, Anderson DR. Model selection and multimodel inference-2nd ed: a practical information-theoretic approach. New york: Springer-verlag new york Inc. 2002.
V. Banyard, Charlotte Cross (2008)
Consequences of Teen Dating Violence
Violence Against Women, 14
S. Mavoa, Mathijs Lucassen, S. Denny, J. Utter, T. Clark, Melody Smith (2019)
Natural neighbourhood environments and the emotional health of urban New Zealand adolescents
Landscape and Urban Planning
Dana Prince, Bernadette Hohl, Bronwyn Hunter, Azure Thompson, Samantha Matlin, A. Hausman, J. Tebes (2019)
Collective Efficacy as a Key Context in Neighborhood Support for Urban Youth.
American journal of community psychology, 63 1-2
Katherine Wincentak, J. Connolly, Noel Card (2017)
Teen Dating Violence: A Meta-Analytic Review of Prevalence Rates
Psychology of Violence, 7
J. Quane, B. Rankin (2006)
Does it pay to participate? Neighborhood-based organizations and the social development of urban adolescents
Children and Youth Services Review, 28
Bruce Taylor, E. Mumford, N. Okeke, E. Rothman (2019)
Neighborhood violent crime and adolescent relationship aggression.
Aggressive behavior
E. Reed, J. Silverman, A. Raj, M. Decker, E. Miller (2011)
Male Perpetration of Teen Dating Violence: Associations with Neighborhood Violence Involvement, Gender Attitudes, and Perceived Peer and Neighborhood Norms
Journal of Urban Health, 88
Jared Hewko, K. Smoyer-Tomic, M. Hodgson (2002)
Measuring Neighbourhood Spatial Accessibility to Urban Amenities: Does Aggregation Error Matter?
Environment and Planning A, 34
Sonia Jain, S. Buka, S. Subramanian, B. Molnar (2010)
Neighborhood predictors of dating violence victimization and perpetration in young adulthood: a multilevel study.
American journal of public health, 100 9
(Oldenburg R. The great good place: Cafes, coffee shops, bookstores, bars, hair salons, and other hangouts at the heart of a community: Da Capo Press; 1999.)
Oldenburg R. The great good place: Cafes, coffee shops, bookstores, bars, hair salons, and other hangouts at the heart of a community: Da Capo Press; 1999.
Oldenburg R. The great good place: Cafes, coffee shops, bookstores, bars, hair salons, and other hangouts at the heart of a community: Da Capo Press; 1999., Oldenburg R. The great good place: Cafes, coffee shops, bookstores, bars, hair salons, and other hangouts at the heart of a community: Da Capo Press; 1999.
S. Biddle, Simone Ciaccioni, G. Thomas, I. Vergeer (2019)
Physical activity and mental health in children and adolescents: An updated review of reviews and an analysis of causality
Psychology of Sport and Exercise, 42
A Ursache, S Regan, A De Marco, DT Duncan (2021)
The Family Life Project Key I: Measuring neighborhood deprivation for childhood health and development - scale implications in rural and urban context
Geospat Health
R. Inglehart, C. Welzel (2010)
Changing Mass Priorities: The Link between Modernization and Democracy
Perspectives on Politics, 8
(2016)
Statistics Canada: Census Profile
(Hipp JR, Boessen A. Neighborhoods, networks, and crime. In: Challenging Criminological Theory: The Legacy of Ruth Rosner Kornhauser. Edited by Cullen FT: Taylor & Francis; 2017.)
Hipp JR, Boessen A. Neighborhoods, networks, and crime. In: Challenging Criminological Theory: The Legacy of Ruth Rosner Kornhauser. Edited by Cullen FT: Taylor & Francis; 2017.
Hipp JR, Boessen A. Neighborhoods, networks, and crime. In: Challenging Criminological Theory: The Legacy of Ruth Rosner Kornhauser. Edited by Cullen FT: Taylor & Francis; 2017., Hipp JR, Boessen A. Neighborhoods, networks, and crime. In: Challenging Criminological Theory: The Legacy of Ruth Rosner Kornhauser. Edited by Cullen FT: Taylor & Francis; 2017.
P. Gruenewald (2007)
The spatial ecology of alcohol problems: niche theory and assortative drinking.
Addiction, 102 6
N. Colabianchi, C. Coulton, J. Hibbert, S. McClure, C. Ievers-Landis, E. Davis (2014)
Adolescent self-defined neighborhoods and activity spaces: spatial overlap and relations to physical activity and obesity.
Health & place, 27
Eline Leen, E. Sorbring, Matt Mawer, E. Holdsworth, Bo Helsing, E. Bowen (2013)
Prevalence, dynamic risk factors and the efficacy of primary interventions for adolescent dating violence: An international review
Aggression and Violent Behavior, 18
Rebecca White, Dawn Witherspoon, Wei Wei, Chang Zhao, Michelle Pasco, Tiyobista Maereg (2021)
Adolescent Development in Context: A Decade Review of Neighborhood and Activity Space Research.
Journal of research on adolescence : the official journal of the Society for Research on Adolescence, 31 4
R. Sampson, S. Raudenbush, F. Earls (1997)
Neighborhoods and violent crime: a multilevel study of collective efficacy.
Science, 277 5328
I. brunton-smith, J. Jackson, A. Sutherland (2014)
Bridging Structure and Perception On the Neighbourhood Ecology of Beliefs and Worries About Violent Crime
British Journal of Criminology, 54
A. Braga, A. Papachristos, David Hureau (2010)
The Concentration and Stability of Gun Violence at Micro Places in Boston, 1980–2008
Journal of Quantitative Criminology, 26
H. Champion, K. Foley, Karen Sigmon-Smith, E. Sutfin, R. Durant (2008)
Contextual Factors and Health Risk Behaviors Associated with Date Fighting Among High School Students
Women & Health, 47
T. Leventhal, J. Brooks-Gunn (2000)
The neighborhoods they live in: the effects of neighborhood residence on child and adolescent outcomes.
Psychological bulletin, 126 2
J. Kupersmidt, Pamela Griesler, M. DeRosier, C. Patterson, P. Davis (1995)
Childhood aggression and peer relations in the context of family and neighborhood factors.
Child development, 66 2
M. Kwan (2012)
The Uncertain Geographic Context Problem
Annals of the Association of American Geographers, 102
H. Togt (2003)
Publisher's Note
J. Netw. Comput. Appl., 26
Melody Smith, Jamie Hosking, A. Woodward, K. Witten, A. Macmillan, A. Field, P. Baas, H. Mackie (2017)
Systematic literature review of built environment effects on physical activity and active transport – an update and new findings on health equity
The International Journal of Behavioral Nutrition and Physical Activity, 14
Qing Li, R. Kirby, R. Sigler, Sean-Shong Hwang, M. Lagory, R. Goldenberg (2010)
A multilevel analysis of individual, household, and neighborhood correlates of intimate partner violence among low-income pregnant women in Jefferson county, Alabama.
American journal of public health, 100 3
S. Bodilly, M. Beckett (2005)
Making Out-of-School-Time Matter: Evidence for an Action Agenda. Summary.
The RAND Corporation
(2002)
Model selection and multimodel inference-2nd ed: a practical information-theoretic approach
I. Rhew, A. Stoep, A. Kearney, N. Smith, M. Dunbar (2011)
Validation of the normalized difference vegetation index as a measure of neighborhood greenness.
Annals of epidemiology, 21 12
V. Banyard, Charlotte Cross, K. Modecki (2006)
Ecological Correlates of Self-Reported Perpetration
Justin Heinze, Allison Krusky-Morey, K. Vagi, T. Reischl, Susan Franzen, Natalie Pruett, R. Cunningham, M. Zimmerman (2018)
Busy Streets Theory: The Effects of Community-engaged Greening on Violence.
American journal of community psychology, 62 1-2
J. Durlak, R. Weissberg, Molly Pachan (2010)
A Meta-Analysis of After-School Programs That Seek to Promote Personal and Social Skills in Children and Adolescents
American Journal of Community Psychology, 45
R. Shih, L. Mullins, Brett Ewing, Lisa Miyashiro, J. Tucker, E. Pedersen, J. Miles, E. D’Amico (2015)
Associations between neighborhood alcohol availability and young adolescent alcohol use.
Psychology of addictive behaviors : journal of the Society of Psychologists in Addictive Behaviors, 29 4
A. Fotheringham, D. Wong (1991)
The Modifiable Areal Unit Problem in Multivariate Statistical Analysis
Environment and Planning A, 23
R. Oldenburg (1999)
The great good place : cafés, coffee shops, bookstores, bars, hair salons, and other hangouts at the heart of a community
M. Hébert, Marie-Ève Daspe, A. Lapierre, N. Godbout, M. Blais, Mylène Fernet, F. Lavoie (2019)
A Meta-Analysis of Risk and Protective Factors for Dating Violence Victimization: The Role of Family and Peer Interpersonal Context
Trauma, Violence, & Abuse, 20
Rachel Garthe, D. Gorman-Smith, Joshua Gregory, Michael Schoeny (2018)
Neighborhood Concentrated Disadvantage and Dating Violence among Urban Adolescents: The Mediating Role of Neighborhood Social Processes
American Journal of Community Psychology, 61
C. Raghavan, Amy Mennerich, Ellen Sexton, Susan James (2006)
Community Violence and Its Direct, Indirect, and Mediating Effects on Intimate Partner Violence
Violence Against Women, 12
R. Hunter, C. Cleland, A. Cleary, M. Droomers, B. Wheeler, Danielle Sinnett, M. Nieuwenhuijsen, M. Braubach (2019)
Environmental, health, wellbeing, social and equity effects of urban green space interventions: A meta-narrative evidence synthesis.
Environment international, 130
A. Roux (2001)
Investigating neighborhood and area effects on health.
American journal of public health, 91 11
Bettina Friese, J. Grube, Roland Moore (2013)
Youth Acquisition of Alcohol and Drinking Contexts: An In-Depth Look
Journal of Drug Education, 43
(2018)
Capturing change in spatial crime patterns
Census Profile . 2016 Census
L. Oliver, N. Schuurman, Alexander Hall (2007)
Comparing circular and network buffers to examine the influence of land use on walking for leisure and errands
International Journal of Health Geographics, 6
C. Bow, Nigel Waters, Peter Faris, Judy Seidel, P. Galbraith, M. Knudtson, William Ghali, the Investigators (2004)
Accuracy of city postal code coordinates as a proxy for location of residence
International Journal of Health Geographics, 3
(Frank LD, Engelke P, Schmid T. Health and community design: The impact of the built environment on physical activity: Island Press; 2003.)
Frank LD, Engelke P, Schmid T. Health and community design: The impact of the built environment on physical activity: Island Press; 2003.
Frank LD, Engelke P, Schmid T. Health and community design: The impact of the built environment on physical activity: Island Press; 2003., Frank LD, Engelke P, Schmid T. Health and community design: The impact of the built environment on physical activity: Island Press; 2003.
Alison Culyba, Sara Jacoby, T. Richmond, J. Fein, Bernadette Hohl, C. Branas (2016)
Modifiable Neighborhood Features Associated With Adolescent Homicide.
JAMA pediatrics, 170 5
(Sas Institute: SAS/STAT 9.3 User's Guide: Survey Data Analysis (book Excerpt): SAS institute. 2011.)
Sas Institute: SAS/STAT 9.3 User's Guide: Survey Data Analysis (book Excerpt): SAS institute. 2011.
Sas Institute: SAS/STAT 9.3 User's Guide: Survey Data Analysis (book Excerpt): SAS institute. 2011., Sas Institute: SAS/STAT 9.3 User's Guide: Survey Data Analysis (book Excerpt): SAS institute. 2011.
E. Schisterman, S. Cole, R. Platt (2009)
Overadjustment Bias and Unnecessary Adjustment in Epidemiologic Studies
Epidemiology, 20
John Hipp, Adam Boessen (2013)
EGOHOODS AS WAVES WASHING ACROSS THE CITY: A NEW MEASURE OF “NEIGHBORHOODS”
Criminology, 51
JB Kupersmidt, PC Griesler, ME DeRosier, CJ Patterson, PW Davis (1995)
Childhood aggression and peer relations in the context of family and neighborhood factors
Child Dev, 66
L. Frank, J. Sallis, B. Saelens, Lauren Leary, K. Cain, T. Conway, P. Hess (2009)
The development of a walkability index: application to the Neighborhood Quality of Life Study
British Journal of Sports Medicine, 44
F. Kuo (2001)
AGGRESSION AND VIOLENCE IN THE INNER CITY Effects of Environment via Mental Fatigue
(Wilson EO. Biophilia and the conservation ethic. In: Evolutionary perspectives on environmental problems. Routledge; 2017: 263–272.)
Wilson EO. Biophilia and the conservation ethic. In: Evolutionary perspectives on environmental problems. Routledge; 2017: 263–272.
Wilson EO. Biophilia and the conservation ethic. In: Evolutionary perspectives on environmental problems. Routledge; 2017: 263–272., Wilson EO. Biophilia and the conservation ethic. In: Evolutionary perspectives on environmental problems. Routledge; 2017: 263–272.
B. Iritani, Martha Waller, C. Halpern, Kathryn Moracco, S. Christ, R. Flewelling (2013)
Alcohol Outlet Density and Young Women’s Perpetration of Violence Toward Male Intimate Partners
Journal of Family Violence, 28
D. Mayne, G. Morgan, A. Willmore, Nectarios Rose, B. Jalaludin, H. Bambrick, A. Bauman (2013)
An objective index of walkability for research and planning in the Sydney Metropolitan Region of New South Wales, Australia: an ecological study
International Journal of Health Geographics, 12

Publisher: Springer Journals
Copyright: Copyright © The Author(s) 2022
eISSN: 1476-072X
DOI: 10.1186/s12942-022-00306-3
Publisher site: See Article on Publisher Site

Abstract

Background: Dating violence (DV ) is a public health problem that could have serious repercussions for the health and well-being of a large number of adolescents. Several neighborhood characteristics could influence these behav- iors, but knowledge on such influences is still limited. This study aims at (1) evaluating the associations between neighborhood characteristics and DV, and (2) assessing how spatial scale influences the estimations of the latter associations. Methods: The Québec Health Survey of High School Students (2016–2017) was used to describe DV. Neighborhoods were operationalized with polygon-based network buffers of varying sizes (ranging from 250 to 1000 m). Multiple data sources were used to describe neighborhood characteristics: crime rate, alcohol outlet density (on-premises and off-premises), walkability, greenness, green spaces density, and youth organizations density. Gendered-stratified logistic regressions were used for assessing the association between neighborhood characteristics and DV. Results: For boys, off-premises alcohol outlet density (500 m) is associated with an increase in perpetrating psy- chological DV. Crime rate (500 m) is positively associated with physical or sexual DV perpetration, and crime rate (250 m) is positively associated with physical or sexual DV victimization. Greenness (1000 m) has a protective effect on psychological DV victimization. For girls, walkability (500 m to 1000 m) is associated with a decrease in perpetrat- ing and experiencing psychological DV, and walkability (250 m) is negatively associated with physical or sexual DV victimization. Conclusions: Several neighborhood characteristics are likely to influence DV, and their effects depend on the form of DV, gender, and spatial scale. Public policies should develop neighborhood-level interventions by improving neigh- borhood living conditions. and 9% (ranging from < 1 to 54%) for sexual victimization Introduction [1]. Psychological violence was not assessed in this meta- Teen dating violence (DV), which can be described as analysis, but it appears to be the most common form of psychologically, physically, or sexually abusive behaviors DV. A systematic review suggests that the prevalence of from a dating partner, is a major public health problem. psychological violence ranges from 17 to 88% [2]. In addi- A recent meta-analysis estimated an overall prevalence tion to its high prevalence, DV is associated with negative of 20% (ranging from 1 to 61%) for physical victimization repercussions on the health and well-being of victims, such as anxiety, depression, post-traumatic stress symp- toms, as well as suicidal ideation [3–5]. *Correspondence: philibert.mathieu@uqam.ca Research on determinants of DV has mainly focused on Département de Sexologie, Université du Québec à Montréal, Succursale individual (e.g., antisocial or risky behaviors), family (e.g., Centre-Ville, Case postale 8888, Montréal, Québec H3C 3P8, Canada © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 2 of 21 exposure to family violence, lack of parental supervision), positive associations between perceived neighborhood and peer (e.g., affiliation with deviant peer) factors [6–8]. disorder and DV [21–24], while others reported no Empirical analyses of the association between neighbor- significant associations between crime rate, a dimen - hood characteristics and DV are scarce. Among studies sion of neighborhood disorder, and DV [25, 26]. The exploring such an association, most studies analyzed the association between alcohol outlet density and DV has effect of neighborhoods’ characteristics such as the soci - been observed among young adults [27–29], but studies odemographic composition, neighborhood disorder (i.e., have yet to assess this relationship among adolescents. visible social (e.g., crime) and physical (e.g., vandalism) However, alcohol outlet density could have a different signs of decay), collective efficacy and access to alcohol effect on this population as selling alcohol to minors is outlets, and led to inconsistent findings [9]. Yet, charac - prohibited. teristics of the physical environment are seldom analyzed Different mechanisms could explain the effect of neigh - in relation to DV despite reports of associations with borhood disorder and access to alcohol outlets on DV. adolescents’ behaviors, possibly reflecting socioenviron - Neighborhood disorder is often incorporated into social mental and psychobehavioral mechanisms. For example, disorganization theory, which describes it as a marker of neighborhood greenness, walkability, access to green the lack of order in the community and the inefficiency spaces, and access to community organizations could of social control [30, 31]. Disorder may refer to minor enhance social cohesion and reduce adolescents’ aggres- offences (e.g. graffiti, vandalism) which may lead to sion [10–12], which could positively affect DV [7–9]. more serious crimes and create fear and distrust among However, to our knowledge, the influence of these physi - residents [31]. Such conditions could encourage violent cal environment factors on DV has never been assessed. behaviors, including DV [9]. Exposure to neighborhood Furthermore, previous studies on the neighborhood violence, a component of disorder sometimes used to effects on DV have used census tracts to operational - measure the effect of neighborhood disorder, may also ize neighborhoods. There is little discussion of how to come to foster the normalization of violence and increase operationalize neighborhoods and the scale of analysis, frustration and anger in adolescents, which may influ - but these choices could potentially influence the estima - ence the adoption of such behaviors [9]. tion of the effect of neighborhood factors [13]. In addi - The density of alcohol outlets in neighborhoods is a tion, administrative units, such as census tracts, may not potential risk factor for many violent behaviors, includ- accurately reflect individual experience of space. Egocen - ing interpersonal violence. The presence of alcohol out - tric neighborhoods, defined as a buffer around a location, lets provides opportunities for consumption [32], which such as individuals’ homes, could better address these could result in risky drinking behaviors [19]. Alcohol limitations [14–16], but this approach is little used in abuse could, in turn, increase the risk of interpersonal research on DV. Against this backdrop, the current study violence victimization [8] and perpetration [7]. In socially aimed to analyze a range of neighborhood characteristics disorganized neighborhoods, alcohol outlet density may possibly related to DV, many of which have not yet been also enhance neighborhood disorder by encouraging the explored, and explore the effects of the spatial scale of gathering of people likely to adhere to norms and atti- analysis. tudes conducive to alcohol consumption and violence [19]. Context Neighborhood risk factors and DV Neighborhood protective factors and DV The relationship between neighborhood characteristics While neighborhood disorder and alcohol outlet density and DV has been explored through social disorganization could negatively impact adolescents’ behaviors, some theory focusing on neighborhood-level risk factors [9]. environmental and institutional resources could be asso- This theory posits that violence and criminality are more ciated with positive effects [33]. In particular, the level likely to occur in socially disorganized neighborhoods of greenness, the density of green spaces, the density of due to the community’s inability to collectively manage community organizations and walkability could prevent problems within the neighborhood [17]. Such context DV through several mechanisms. could also foster neighborhood disorder [17, 18] and Greenness and accessibility to green spaces could exacerbate negative consequences of alcohol outlets (e.g., affect some physiological and psychosocial processes, criminality, alcohol abuse) [19, 20], thereby influencing which in turn could influence DV. According to the bio - individuals’ violent behaviors. philia hypothesis, exposure to nature may improve men- Several empirical analyses assessed the effects of tal health by restoring cognitive functions and reducing neighborhood disorder and density of alcohol out- mental fatigue [34]. A higher level of greenness in lets on DV, but results are mixed. Some studies found neighborhoods may also be associated with a lower risk R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 3 of 21 of depressive symptoms [35, 36] and aggressivity [37] 26] and the density of alcohol outlets [27–29], but this in adolescents, both of which have been identified as choice remains little discussed, despite being known to potential risk factors of DV perpetration [7] as well as geographers to affect statistical analyses. victimization [8]. A study on intimate partner violence Two biases are related to neighborhoods’ operation- also found that rates of aggression were lower in neigh- alization: the Modifiable Areal Unit Problem (MAUP) borhoods characterized by a high level of greenness [38]. and the Uncertain Geographic Context Problem Access to green spaces, such as public parks, could pro- (UGCoP). The MAUP suggests that the shape and size mote participation in physical and recreational activities of spatial units could lead to variations in the estimated and encourage social interactions [39]. Physical activity effects of neighborhood characteristics [58, 59]. The may improve self-esteem and have a protective effect on UGCoP argues that the estimation of neighborhood the onset of depression and anxiety [40], all three deter- effects may be affected by the definition of neighbor - minants of DV [7, 8]. Social interactions may increase hoods and the use of inappropriate spatial units [60]. social cohesion in neighborhoods [41], which is likely to Both biases could limit the ability to observe associa- reduce the risk of DV [42, 43]. tions between neighborhood characteristics and a given Geographic access to community organizations (e.g., outcome. For example, census tracts were used in pre- social clubs, sports clubs) may promote adolescents’ par- vious studies on DV even though they may not be an ticipation in structured and supervised activities [10, 44]. adequate representation of neighborhood-level influ - Such resources provide opportunities to develop social ences. The MAUP and UGCoP could also partly explain ties between adolescents as well as between residents why the effect of certain factors, such as crime [25, 26], of the neighborhood (e.g., their parents) [10]. Commu- has not yet been observed. nity organizations are also safe places that could reduce In effect, administrative units, such as census tracts, exposure to the negative aspects of neighborhoods, such may not adequately represent individuals’ actual expo- as disorder [10, 45] and foster participation in supervised sure to their neighborhood. These units have artificial activities. Thus, adolescents attending these facilities boundaries, which implicitly assume that individuals’ could have a lower risk of risky behaviors, such as sub- environments are limited to the corresponding spatial stance use [10, 46] and delinquency [46–48], which are units, regardless of their real location inside the spatial determinants of DV [7, 8]. unit. However, individuals living closely to each other Walkability refers to the attributes of urban design (e.g., but in two different spatial units could have more simi - road network, land use) promoting walking [49]. Walk- lar exposures to neighborhood factors than two indi- ability is likely to promote social interactions, which viduals living further away but in the same spatial unit. could contribute to a better sense of belonging [11, 50] Neighborhood-level processes, such as social interac- and collective efficacy (i.e., the ability of the community tions or access to resources, may depend on geographical to act collectively to regulate deviant behaviors) [11]. proximity and can be observed across these boundaries These concepts are central to social disorganization the - [14, 16, 61]. In consequence, the accuracy of measures of ory and could influence DV [9]. To our knowledge, the exposure to neighborhood factors could vary depending effect of walkability on DV has not been described. How - on the location of individuals within a spatial unit. These ever, empirical studies suggest that increased walkability measures could be less accurate for people living fur- is associated with a decrease in homicide [51] and violent ther away from the centroid of the spatial unit [62–64]. behaviors [52] among adolescents. When geographic access to resources is assessed, using Finally, it is essential to note that the associations large spatial units such as census tracts could also lead to between neighborhood characteristics and DV may be aggregation error [62, 63]. In response to such a caveat, modified by gender. Most studies suggest that neighbor - egocentric neighborhoods were proposed as operational hood effects are stronger for boys than for girls [53, 54]. forms of neighborhoods which allow for integrating prox- Boys may have less parental supervision than girls [55], imity. They provide a person-centered approach and refer leading to greater exposure to their neighborhood [56, to the set of places located within a given distance from 57]. the individuals’ homes [15, 16, 65]. Egocentric neighbor- hoods are defined specifically for each individual and allow for overlapping neighborhoods, suggesting that Defining neighborhoods some individuals share common exposures to neighbor- In addition to the lack of knowledge about neighbor- hood factors depending on their proximity. An egocen- hood effects on DV, previous studies have not explored tric definition of neighborhoods could more adequately the impact of neighborhoods’ operationalization and reflect the real use of spaces by residents and their per - spatial scale of the estimated effects. Most studies on ception of the residential context [14, 16, 61]. DV used census tracts to measure the level of crime [25, Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 4 of 21 Objectives Québec, Canada) and is composed of 15 municipalities The current study had two main objectives: (1) to analyze (Fig. 1) [66]. About 85% of the island’s population lives in the association between several contextual characteris- the city of Montréal, the largest city in Québec and the tics (crime, density of off-premises alcohol outlets, den - second largest in Canada. sity of bars, density of community organizations, level of greenness, density of green spaces and walkability) and Dependent variables DV (victimization and perpetration) and (2) to assess Four variables describing different forms of DV were scale effects in studying these relationships. used. Two variables referred to victimization (e.g., “He/ she (…) me”), and two referred to perpetration (e.g., “I Data and method (…) him/her). Participants Psychological DV was assessed by items derived from The current study used data from the Québec Health two questions. One item referred to verbal violence Survey of High School Students (QHSHSS) 2016–17, (“I criticized him/her viciously about his/her physical a cross-sectional survey of secondary-school Québec appearance; I insulted him/her in front of people; I put youth (Grades 7 to 11). A three-stage stratified cluster him/her down.” / “He/she viciously criticized my physi- sampling was used to recruit participants in this survey. cal appearance; he/she insulted me in front of people; he/ Schools were randomly selected for each grade level and she put me down.”), while the other related to controlling each health region separately. Classes were randomly behaviors (“I controlled his/her outings, email conversa- selected from the selected schools. The QHSHSS pro - tions or cell phone; I prevented him/her from seeing his/ vides a representative sample for the province of Québec her friends.” / “He/she controlled my outings, my email and each health region. Only participants whose postal conversations or cell phone; he/she prevented me from code of residence was located in the island of Montréal seeing my friends”). The response scale ranged from 0 were included in this study (n = 2,687). The island of (never during the past 12 months) to 3 (three times or Montréal had a population of 1,942,040 inhabitants in more during the past 12 months). A dichotomous meas- 2016 (about 23.8% of the population of the province of ure was obtained by distinguishing participants who Fig. 1 Study area-island of Montréal R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 5 of 21 had experienced one of these events at least once (one theft, and robbery. First, the geolocation of crime events or more) from those who reported having never experi- allowed their number to be calculated for each egocen- enced these situations. tric neighborhood (buffer zone). Secondly, the popula - To assess physical violence, four items from the Con- tion size of each egocentric neighborhood was estimated ict T fl actics Scale [67] (e.g., “I slapped him/her”/ “He/She by summing the population of dissemination blocks (DB) slapped me”) were used. Sexual violence was assessed by included in the neighborhood. DBs are the smallest spa- two items referring to experiences of sexual activity with- tial units for which data on population size are dissemi- out consent (e.g., “I forced him/her to have sexual con- nated by Statistics Canada. If only part of the DBs was tact or sexual intercourse with me when he/she didn’t included in a buffer, only the proportion of its population want to” / “He/she forced me to have sexual contact or corresponding to the proportion of area included in the sexual intercourse when I didn’t want to.”). Due to the neighborhood was considered for summing the popula- small number of participants who reported sexual vio- tion. The crime rate is the number of crimes divided by lence perpetration, physical and sexual DV were merged. twice the population size (the number of crimes was esti- Dichotomous measures were derived by differentiating mated for two years) multiplied by 100. participants who had experienced physical or sexual vio- lence at least once from those who had never experienced Greenness and access to green space these situations. The Normalized Difference Vegetation Index (NDVI) is commonly used in epidemiological studies and provides Neighborhood‑level variables an objective and accurate measure of overall greenness Egocentric neighborhoods [69]. This index was first produced at a 30 × 30 m spatial Egocentric neighborhood refers to the area within a spe- resolution from Landsat 8 satellite data (United States cific radius around the individual’s home and is often Geological Survey, 2015–2016). The greenness of the operationalized using buffer zones [15, 16, 65]. In this egocentric neighborhoods was estimated using the aver- study, the participants’ place of residence was estimated age NDVI in the area covered by a buffer. from the centroid of the postal code area. Postal codes To assess access to green spaces, a map of public parks are managed by Canada Post for mail delivery and have and green spaces was first obtained for the island of an average of 14.5 dwellings in Montréal. Their cen - Montréal by cross-referencing information provided by troids provide a good approximation of the exact address the municipalities. The municipalities of the City of Mon - of participants [68]. Egocentric neighborhoods were tréal, Dollard-des-Ormeaux, Pointe-Claire, Kirkland, and operationalized from polygon-based network buffers East Montréal provided the location of parks and green around the participants’ place of residence, a method spaces as open access data. A map of parks and green of buffering providing an accurate representation of the spaces was created for the remaining cities using land spatial area used by individuals [65]. In this study, the use data at the parcel-level from the property assessment government of Québec’s official road network database roll provided by the Montréal Metropolitan Community (Addresses Québec) and the Network Analyst extension (2016) and information available on the municipalities’ of ArcGIS Pro 2.7.0 were used to generate all routes from websites. The parcels on the property assessment roll participants’ place of residence to a specific distance. were manually identified as green space based on the Polygon-based network buffers were then computed documents and maps available on the websites. A total of by connecting the endpoints of these routes, result- 1389 green spaces were identified for the island of Mon - ing in irregular polygons (Fig. 2). For each participant, tréal. Access to green spaces was estimated by the num- four buffer zones were created using different distances: ber of green spaces that intersected each buffer. 250 m, 500 m, 750 m, and 1000 m. Figure 2 provides an example of polygon-based network buffers for two postal Access to alcohol outlets codes. Alcohol outlets were located using data on private out- lets licensed by the Régie des alcools, des courses et des Criminality jeux (RACJ, 2016) and outlets administered by the Société The level of crime was measured using data from Ser - des alcools du Québec (SAQ, 2016). A total of 1432 bars vice de police de la Ville de Montréal (SPVM) for the and 1842 off-premises alcohol outlets (e.g., convenience years 2016 and 2017. This database provides the date stores, grocery shops and SAQ outlets) were identified and the location of crime events at the nearest inter- from these data. Using the same method as for green section. Events pertaining to one of the following six spaces, density measures have been used to estimate categories were retained: offences resulting in death, access to bars and access to off-premises alcohol outlets intrusion, mischief, theft in/on a vehicle, motor vehicle separately. Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 6 of 21 Fig. 2 Egocentric neighborhoods Access to community organizations (2016) was used to assess the land use mix. An entropy The Directory of Community Organisations from 211, index [70] was elaborated to describe the level of het- an information and referral public service, was used to erogeneity in land use considering four categories of identify and locate community organizations for young land use: residential, commercial, services and cul- people (e.g., youth centers, YMCA, sports associations). tural, recreational and leisure. Secondly, data from the Lists of community organizations in the different munici - property assessment roll were used to calculate the net palities of the island of Montréal, which are available on residential density. This variable refers to the number the corresponding websites, supplemented these data. of dwellings per hectare of the residential area included A total of 423 community organizations offering ser - in each buffer. Thirdly, data from Addresses Québec vices for young people were identified. To assess access (2016) were used to calculate the density of intersec- to community organizations, density measures were used tions with three or more segments of the road network by calculating the number of community organizations (excluding highways). The walkability measure used within each buffer. in this study is based on several studies reporting on the development of this index [71, 72]. It was obtained Walkability by the sum of the z-scores of all variables described For each egocentric neighborhood, walkability was above: z-score(land use mix) + z-score(residential den- measured using three variables: land use mix, residen- sity) + 1.5(z-score(intersection density)). In its complete tial density, and intersection density. Firstly, data from form, the walkability index also requires the retail floor the property assessment roll (parcel-level land use) area ratio [71], which could not be obtained as is the provided by the Montréal Metropolitan Community case for many studies. The weight of 1.5 for intersection R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 7 of 21 density (replacing a weight of 2 in the complete index) corresponding buffer. Median income was used as it is was proposed by Sundquist et al. [72] as a way to com- less sensitive to extreme values than average income and pensate for the lack of data on retail floor area. This walk - may therefore better capture the spatial distribution of ability index has been validated and is a good predictor income. of walking behaviors [72]. A study also found a strong The percentage of single-parent dwellings was used to correlation between the three-components index and the assess single parenthood. The number of dwellings and four-components index, and both were predictors of util- the number of single-parent dwellings were estimated for itarian walking [73], confirming the validity of using the each egocentric neighborhood by summing the frequen- three-components walkability index as a surrogate for the cies of the two variables at the DB-level. four-components index. Residential instability was measured using the percent- age of residents living in the egocentric neighborhood for Covariates five years or less. For each egocentric neighborhood, the Individual‑level covariates number of residents living there for less than five years Several individual-level covariates were used: gender (girl and the population size were estimated by summing the or boy), high school grade level (Grades 7 or 8, 9, 10, and frequencies of the DBs considered as part of the corre- 11), the highest level of parental education (high school sponding buffer. or less, college or professional training, university), fam- Finally, to assess ethnocultural diversity, a language ily structure (two parents, blended family, or shared cus- diversity index was developed. Languages spoken at tody, living with one parent or other family structure), home were divided into 16 classes based on the World and parental country of birth (two parents born in Can- Value Survey classification [74]. Frequencies of each class ada, at least one parent born outside Canada). were first obtained for each egocentric neighborhood by summing the frequencies of the corresponding DBs. Neighborhood sociodemographic characteristics covariates Shannon entropy index [70] was then calculated for each Five potentially confounding neighborhood sociodemo- egocentric neighborhood to estimate the level of hetero- graphic characteristics were measured using data from geneity/homogeneity of languages spoken at home. the 2016 Canadian census. Population density corresponded to the number of Statistical analysis inhabitants per hectare and was estimated from the pop- Data at the neighborhood level were matched with data ulation size and the area of the buffers. The population from the QHSHSS at the individual level using partici- size of each egocentric neighborhood was obtained by pants’ postal codes. Only adolescents who reported being summing the population size of the DBs (i.e., the smallest involved in a romantic relationship in the last 12 months spatial units for which population size is available) within and had no missing values for the DV measures were the corresponding buffer zone weighted by the propor - included (37% of participants with a postal code in Mon- tion of the area of the included DBs (a weight of 1 was tréal). Among them, 121 had at least one missing value used for DBs completely located in the buffer). for covariates and were excluded (12% of participants Although population size data are available at the DB- who reported a romantic relationship). The final sample level, no sociodemographic data are disseminated by consisted of 879 adolescents (Fig. 3). Statistics Canada at this scale. For this reason, the four The associations between neighborhood character - remaining sociodemographic variables were derived istics and DV were estimated with logistic regressions from variables at the dissemination area (DA) level, the using the SURVEYLOGISTIC procedure of SAS Enter- smallest spatial units for which sociodemographic data prise Guide 8.5 [75]. All models considered the sam- are available. As DBs are embedded into DAs (see Fig. 2), pling design and used bootstrap weights. They were the sociodemographic measures were assigned to the carried out separately for girls and boys to account for corresponding DBs, assuming a homogeneous distribu- gender differences. A Directed Acyclic Graph (DAG) tion within each DA. Similar to the method used to esti- was developed to identify confounding variables (Addi- mate population density, each egocentric neighborhood tional file 1). A DAG is a graphical representation of was composed of the DBs included in the corresponding causal assumptions regarding a set of variables and can buffer, weighted by the proportion of the area of included be used as a tool to identify confounders [76, 77]. This DBs. approach can also limit the risk of overadjustment bias Socioeconomic status (SES) was measured by [77]. This bias occurs when a model controls for vari - median income. For each egocentric neighborhood, ables that are not confounders. Assessing neighbor- the median income was estimated by calculating the hood effects could be subject to overadjustment bias population-weighted median of the DBs included in the due to the complex relationships and high correlations Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 8 of 21 Fig. 3 Flowchart for sample selection between neighborhood factors. The DAG used in the of crime were modelled controlling for all the variables current study thus provided a parsimonious approach described above, except for the density of community and allowed the identification of confounding factors organizations. tailored to each neighborhood factor specifically. Based At each stage and for each dependent variable, four mod- on this DAG, not all models required the same set of els estimated the association with a given neighborhood covariates. Associations between neighborhood char- characteristic at different scales, i.e. using different buffer acteristics and DV variables were estimated using dif- sizes (250 m, 500 m, 750 m, and 1000 m). For each of these ferent models in three stages. (1) Associations between models, covariates were modelled at the scale considered DV and walkability were estimated using models the most appropriate in the previous stage. Choosing the including individual-level variables and neighborhood most appropriate scale for modelling a given neighborhood sociodemographic characteristics as covariates. (2) characteristics was based on the comparison of the mod- Associations between DV and the level of greenness, els’ fit using the Akaike Information Criterion (AIC). The the density of green spaces, the density of community selected scale was the one with the lowest AIC. It should be organizations, the density of off-premises alcohol out - noted that the scale for neighborhood sociodemographic lets, and the density of bars were estimated by con- characteristics used as covariates in all models was simi- trolling for individual-level variables, neighborhood larly identified in preliminary analyses. In these analyses, sociodemographic characteristics as well as walkability. the effect of neighborhoods’ sociodemographic charac - (3) Associations between DV and neighborhoods’ level teristics on the DV measures were estimated separately in R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 9 of 21 models adjusted by individual-level variables only (results Our results suggest an effect of the density of off-prem - available upon request). For example, for perpetration ises alcohol outlets and crime rate on DV. There was a of psychological DV among boys, preliminary analyses positive association between the density of off-premises showed that the most appropriate scale for neighborhood alcohol outlets within a 500 m radius and psychological sociodemographic characteristics (lowest AIC) were DV perpetration (ß = 0.08; SE = 0.04; p = 0.049). Crime 1000 m for median income, 250 m for percentage of single- rate was positively associated with physical/sexual DV parent dwellings, 500 m for residential instability, 250 m for perpetration with radii of 500 m (ß = 0.97; SE = 0.28; ethnocultural diversity, and 1000 m for population density. p < 0.001) and 750 m (ß = 0.84; SE = 0.37; p = 0.024). The These variables were used in Stage 1 to estimate the effect AIC of the model using a 500 m buffer was consider - of walkability. In this stage, models showed that walkabil- ably smaller than that of the model using a 750 m radius ity within 250 m had the best fit (lowest AIC). Variables (ΔAIC = 3.931). There was also an association between selected in preliminary analyses and Stage 1 were then used crime rate measured with a 250 m radius buffer and to estimate the effects of all variables identified in Stage 2. physical/sexual DV victimization (ß = 0.18; SE = 0.09; In Stage 3, to assess the effect of crime rate, the scales used p = 0.050). No significant associations were identified for neighborhood characteristics covariates were based on between the density of bars and DV. results in all previous stages. This method provided parsi - The analysis of possible protective factors revealed only monious estimations of neighborhood factors for each out- a negative association between NDVI within a 1000 m come. Multicollinearity was assessed for all models using radius and psychological DV victimization (ß = − 4.24; Variance Inflation Factors (VIF) and the data showed no SE = 2.09; p = 0.044). multicollinearity problem (VIF < 4). In line with the second objective of this study, the influ - Models of associations between neighborhood ence of the spatial scale of analysis was evaluated by com- characteristics on DV among girls paring the AIC of models estimating associations between Results from logistic regression assessing the associations the same neighborhood-level factors and a given outcome between neighborhood characteristics and DV among across buffers (250 m, 500 m, 750 m, and 1000 m). As a rule girls are shown in Table 3. of thumb, an AIC difference greater than two units sug - Regarding risk factors, results revealed negative asso- gest that the model with the lowest AIC is the most predic- ciations between density of bars within a buffer of 250 m tive [78]. Scale effects were also investigated by calculating (ß = − 0.29; SE = 0.15; p = 0.045) and 1000 m (ß = − 0.03; Spearman’s rho correlation coefficients between all neigh - SE = 0.01; p = 0.049) and psychological DV victimization. borhood-level variables across buffers. Weak correlations The difference in AIC between the two models was small suggested differences in measurements of neighborhood- (ΔAIC = 0.204), suggesting that both spatial scales are level variables across buffers, while strong correlations comparable. No significant effects were observed for the suggested minor differences. Spearman’s correlations were density of off-premises alcohol outlets and crime rate. used instead of Pearson’s correlations because some vari- In analyzing possible protective factors, several asso- ables were skewed. ciations were found between walkability and DV. Walk- ability within a radius of 500 m (ß = − 0.19; SE = 0.07; Results p = 0.007), 750 m (ß = − 0.20; SE = 0.07; p = 0.007), and Sample description 1000 m (ß = − 0.19; SE = 0.07; p = 0.013) were negatively Most participants reported living in a two-parent family associated with psychological DV perpetration. The (64.74%) with at least one parent who has obtained a uni- model using a 500 m radius buffer had the smallest AIC versity degree (68.56%) (Table 1). Psychological violence (AIC = 496.119), but the difference with the other mod - was the most prevalent form of DV: 20.25% of adolescents els was negligible (ΔAIC < 1.464). There were associa - reported perpetration, while 28.78% reported experiences tions between walkability in a buffer of 500 m (ß = − 0.13; of victimization. Perpetration and victimization of physi- SE = 0.06; p = 0.028), 750 m (ß = − 0.17; SE = 0.06; cal/sexual DV were observed respectively by 14.10% and p = 0.008), and 1000 m (ß = − 0.17; SE = 0.07; p = 0.012) 19.94% of participants. For all forms, girls were more likely with psychological DV victimization. The model using a to report an experience of DV. 750 m radius had the lowest AIC (AIC = 580.321) but was comparable with the other two models (ΔAIC < 1.800). Models of associations between neighborhood Finally, a decrease of walkability within a radius of 250 m characteristics on DV among boys (ß = − 0.15; SE = 0.07; p = 0.029), and 500 m (ß = − 0.15; Table 2 summarizes the results from logistic regressions SE = 0.07; p = 0.049) were associated with lower physical/ analyzing the relationship between neighborhood char- sexual DV victimization. The difference between AIC is acteristics and DV among boys. small (ΔAIC = 1.233), suggesting that the two models are Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 10 of 21 Table 1 Descriptive statistics All (n = 879) Girls (n = 452) Boys (n = 427) Individual‑level variables % % % Psychological DV perpetration At least once 20.25 23.76 16.68 Never 79.75 76.24 83.32 Physical or sexual DV perpetration At least once 14.10 20.41 7.67 Never 85.90 79.59 92.33 Psychological DV victimization At least once 28.78 34.13 23.33 Never 71.22 65.87 76.67 Physical or sexual DV victimization At least once 19.94 22.83 17.01 Never 80.06 77.17 82.99 Grade level Grade 7 or 8 29.09 25.09 33.15 Grade 9 18.32 18.06 18.58 Grade 10 23.71 23.93 23.50 Grade 11 28.88 32.92 24.77 Parental country of birth Two parents born in Canada 42.21 37.88 46.61 At least one parent born outside Canada 57.79 62.12 53.39 Family structure Two parents 64.74 61.96 67.57 Blended family or shared custody 16.91 19.13 14.66 Living with one parent or other family structure 18.35 18.92 17.78 Highest level of parental education High school or less 13.06 16.55 9.50 College or professional training 18.38 20.20 16.54 University 68.56 63.25 73.96 Neighborhood‑level variables Mean (SD) Mean (SD) Mean (SD) 250 m Sociodemographic characteristics Median income 58,754.92 (25,196.54) 57,754.87 (22,883.30) 59,772.37 (27,339.35) Single parenthood 20.32 (7.63) 20.72 (7.40) 19.92 (7.84) Residential instability 38.94 (13.13) 39.39 (12.84) 38.48 (13.42) Ethnocultural diversity 1.28 (0.39) 1.30 (0.39) 1.27 (0.39) Population density 9.89 (6.07) 9.90 (5.88) 9.88 (6.25) Risk factors Density of off-premises alcohol outlets 0.75 (1.18) 0.74 (1.12) 0.76 (1.23) Density of bars 0.31 (1.00) 0.35 (1.03) 0.28 (0.97) Crime rate 1.54 (2.18) 1.64 (2.76) 1.44 (1.34) Protective factors Walkability − 0.47 (2.61) − 0.37 (2.57) − 0.58 (2.65) NDVI 0.50 (0.11) 0.49 (0.11) 0.50 (0.11) Density of green spaces 0.87 (0.95) 0.89 (0.92) 0.86 (0.99) Density of community organizations 0.39 (0.65) 0.37 (0.64) 0.40 (0.67) 500 m Sociodemographic characteristics Median income 57,714.61 (22,067.14) 56,744.45 (20,220.35) 58,701.65 (23,782.52) R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 11 of 21 Table 1 (continued) Neighborhood‑level variables Mean (SD) Mean (SD) Mean (SD) Single parenthood 20.33 (6.56) 20.74 (6.43) 19.91 (6.67) Residential instability 39.37 (11.88) 39.67 (11.50) 39.06 (12.26) Ethnocultural diversity 1.29 (0.37) 1.30 (0.38) 1.27 (0.37) Population density 8.61 (4.42) 8.65 (4.41) 8.58 (4.44) Risk factors Density of off-premises alcohol outlets 3.26 (3.91) 3.24 (3.77) 3.27 (4.04) Density of bars 1.50 (3.63) 1.51 (3.04) 1.49 (4.14) Crime rate 1.32 (3.16) 1.24 (1.01) 1.41 (4.37) Protective factors Walkability − 0.53 (2.66) − 0.44 (2.62) − 0.61 (2.69) NDVI 0.50 (0.10) 0.49 (0.10) 0.50 (0.10) Density of green spaces 2.86 (2.35) 2.9 (2.32) 2.82 (2.38) Density of community organizations 1.10 (1.24) 1.06 (1.15) 1.15 (1.31) 750 m Sociodemographic characteristics Median income 56,871.79 (19,941.78) 56,039.25 (18,387.81) 57,718.81 (21,395.60) Single parenthood 20.52 (5.86) 20.87 (5.73) 20.16 (5.97) Residential instability 39.68 (11.01) 39.97 (10.5) 39.40 (11.52) Ethnocultural diversity 1.29 (0.36) 1.31 (0.36) 1.28 (0.36) Population density 7.99 (3.80) 8.09 (3.79) 7.89 (3.81) Risk factors Density of off-premises alcohol outlets 7.32 (7.64) 7.29 (7.56) 7.36 (7.72) Density of bars 3.78 (7.98) 3.83 (7.55) 3.72 (8.41) Crime rate 1.18 (0.64) 1.18 (0.63) 1.19 (0.65) Protective factors Walkability − 0.55 (2.68) − 0.46 (2.63) − 0.65 (2.74) NDVI 0.49 (0.10) 0.49 (0.10) 0.50 (0.10) Density of green spaces 5.75 (4.18) 5.84 (4.29) 5.66 (4.07) Density of community organizations 2.08 (1.89) 2.00 (1.81) 2.16 (1.97) 1000 m Sociodemographic characteristics Median income 56,237.1 (17,931.19) 55,576.33 (16,738.62) 56,909.36 (19,064.11) Single parenthood 20.65 (5.40) 20.95 (5.34) 20.34 (5.45) Residential instability 40.08 (10.44) 40.39 (9.93) 39.76 (10.94) Ethnocultural diversity 1.30 (0.35) 1.31 (0.35) 1.29 (0.34) Population density 7.51 (3.40) 7.60 (3.35) 7.42 (3.45) Risk factors Density of off-premises alcohol outlets 12.69 (12.60) 12.73 (12.72) 12.65 (12.49) Density of bars 6.95 (13.98) 7.05 (13.57) 6.85 (14.40) Crime rate 1.18 (0.58) 1.18 (0.58) 1.18 (0.59) Protective factors Walkability − 0.54 (2.69) − 0.44 (2.63) − 0.64 (2.75) NDVI 0.49 (0.09) 0.49 (0.09) 0.49 (0.09) Density of green spaces 9.42 (6.38) 9.53 (6.60) 9.31 (6.15) Density of community organizations 3.42 (2.75) 3.29 (2.72) 3.54 (2.78) comparable. Significant positive associations were also p = 0.035) and 1000 m (ß = 4.25; SE = 1.82; p = 0.021) observed for NDVI and density of community organi- were related to greater physical/sexual DV victimiza- zations. NDVI in a buffer of 750 m (ß = 3.86; SE = 1.81; tion. Model using a 1000 m buffer had the lowest AIC Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 12 of 21 Table 2 Associations between neighborhood characteristics and DV among boys 250 m 500 m 750 m 1000 m ß (SE) AIC ß (SE) AIC ß (SE) AIC ß (SE) AIC Psychological DV perpetration Risk factors Density of off-premises alcohol outlets 0.15 (0.14) 363.351 0.08 (0.04)* 362.064 0.05 (0.03) 362.621 0.02 (0.02) 363.195 Density of bars − 0.18 (0.17) 363.249 − 0.03 (0.03) 363.620 − 0.02 (0.01) 363.449 − 0.01 (0.01) 363.380 Crime rate − 0.08 (0.13) 362.791 − 0.21 (0.34) 362.638 − 0.56 (0.35) 361.415 − 0.33 (0.39) 362.628 Protective factors Walkability 0.09 (0.08) 362.323 0.10 (0.09) 362.486 0.09 (0.09) 362.629 0.10 (0.09) 362.365 NDVI − 1.92 (2.29) 363.504 − 1.00 (2.37) 364.134 − 0.89 (2.53) 364.187 − 0.45 (2.54) 364.289 Density of green spaces − 0.28 (0.21) 361.705 − 0.07 (0.08) 363.399 − 0.06 (0.05) 362.343 − 0.04 (0.03) 362.239 Density of community organizations 0.24 (0.23) 362.165 − 0.02 (0.12) 364.297 − 0.07 (0.08) 363.611 − 0.03 (0.06) 364.085 Physical or sexual DV perpetration Risk factors Density of off-premises alcohol outlets 0.01 (0.17) 242.997 0.04 (0.06) 242.771 0.00 (0.03) 243.000 0.01 (0.03) 242.950 Density of bars − 0.32 (0.26) 241.920 − 0.03 (0.06) 242.897 − 0.03 (0.02)† 242.115 − 0.02 (0.01) 242.270 Crime rate 0.14 (0.16) 245.851 0.97 (0.28)*** 237.656 0.84 (0.37)* 241.587 0.08 (0.56) 246.688 Protective factors Walkability 0.04 (0.12) 241.001 − 0.02 (0.11) 241.101 0.02 (0.11) 241.090 − 0.01 (0.12) 241.117 NDVI 0.86 (3.30) 242.908 2.54 (3.95) 242.334 2.43 (4.11) 242.470 1.77 (3.86) 242.728 Density of green spaces 0.10 (0.17) 242.730 − 0.04 (0.06) 242.811 − 0.03 (0.05) 242.744 − 0.03 (0.04) 242.436 Density of community organizations 0.04 (0.36) 242.984 − 0.15 (0.18) 242.313 − 0.11 (0.16) 242.264 − 0.05 (0.11) 242.714 Psychological DV victimization Risk factors Density of off-premises alcohol outlets 0.17 (0.13) 460.258 0.06 (0.05) 460.447 0.01 (0.03) 461.867 0.00 (0.02) 462.022 Density of bars 0.01 (0.11) 462.040 − 0.03 (0.03) 461.575 − 0.01 (0.01) 461.414 − 0.01 (0.01) 461.505 Crime rate 0.00 (0.11) 461.795 − 0.01 (0.03) 461.770 − 0.13 (0.35) 461.644 − 0.17 (0.42) 461.612 Protective factors Walkability − 0.05 (0.08) 461.103 − 0.10 (0.09) 460.043 − 0.08 (0.10) 460.752 − 0.01 (0.09) 461.562 NDVI − 2.75 (1.99) 459.792 − 2.88 (2.13) 460.042 − 3.59 (2.13)† 459.257 − 4.24 (2.09)* 458.285 Density of green spaces − 0.22 (0.16) 459.432 − 0.01 (0.05) 462.005 0.00 (0.03) 462.042 − 0.01 (0.03) 461.718 Density of community organizations 0.25 (0.22) 460.124 0.01 (0.11) 462.023 − 0.01 (0.08) 462.007 − 0.02 (0.06) 461.913 Physical or sexual DV victimization Risk factors Density of off-premises alcohol outlets 0.06 (0.13) 403.780 − 0.01 (0.05) 403.894 − 0.03 (0.03) 403.315 0.00 (0.02) 403.891 Density of bars − 0.13 (0.19) 403.313 − 0.02 (0.03) 403.620 − 0.01 (0.02) 403.560 − 0.01 (0.01) 403.618 Crime rate 0.18 (0.09)* 404.269 − 0.02 (0.02) 407.190 0.03 (0.24) 407.190 − 0.13 (0.37) 407.089 Protective factors Walkability − 0.04 (0.08) 401.915 − 0.02 (0.09) 402.165 0.02 (0.10) 402.145 − 0.03 (0.10) 402.123 NDVI 0.35 (1.98) 403.883 0.68 (2.13) 403.817 1.35 (2.15) 403.571 0.57 (2.12) 403.855 Density of green spaces − 0.04 (0.15) 403.830 0.01 (0.07) 403.904 0.04 (0.04) 402.625 0.02 (0.03) 403.306 Density of community organizations 0.07 (0.21) 403.813 − 0.18 (0.13) 401.684 − 0.02 (0.08) 403.879 0.01 (0.06) 403.898 Models were adjusted for individual-level covariates and sociodemographic characteristics at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates and sociodemographic characteristics and walkability at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates, and sociodemographic characteristics, walkability, density of green spaces, density of off-premises alcohol outlets and density of bars at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) *** p < 0.001; **p < 0.01; *p < 0.05; †p < 0.10 R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 13 of 21 Table 3 Associations between neighborhood characteristics and DV among girls 250 m 500 m 750 m 1000 m ß (SE) AIC ß (SE) AIC ß (SE) AIC ß (SE) AIC Psychological DV perpetration Risk factors Density of off-premises alcohol outlets 0.03 (0.15) 494.899 − 0.01 (0.06) 494.933 − 0.01 (0.03) 494.892 − 0.01 (0.02) 494.419 Density of bars − 0.13 (0.14) 494.127 − 0.02 (0.05) 494.763 − 0.02 (0.02) 494.128 − 0.01 (0.01) 494.645 Crime rate − 0.13 (0.13) 497.286 − 0.34 (0.25) 496.838 − 0.73 (0.40)† 495.455 − 0.71 (0.48) 495.676 Protective factors Walkability − 0.10 (0.07) 496.008 − 0.19 (0.07)** 492.949 − 0.20 (0.07)** 494.037 − 0.19 (0.07)* 494.413 NDVI 0.02 (1.56) 494.949 − 0.67 (1.75) 494.820 − 0.29 (1.90) 494.926 − 0.09 (1.95) 494.946 Density of green spaces 0.18 (0.14) 493.110 − 0.02 (0.05) 494.860 0.01 (0.03) 494.827 0.01 (0.02) 494.507 Density of community or ganizations 0.12 (0.22) 494.482 − 0.14 (0.16) 493.350 − 0.05(’− 0.09) 494.567 − 0.04 (0.06) 494.276 Physical or sexual DV perpetration Risk factors Density of off-premises alcohol outlets 0.16 (0.14) 457.140 0.00 (0.07) 458.632 0.04 (0.03) 457.265 0.01 (0.02) 458.404 Density of bars − 0.09 (0.15) 458.269 0.05 (0.06) 457.821 0.00 (0.02) 458.592 − 0.01 (0.01) 458.261 Crime rate − 0.00 (0.05) 460.683 − 0.20 (0.25) 459.964 − 0.66 (0.41) 456.010 − 0.61 (0.44) 458.209 Protective factors Walkability − 0.00 (0.07) 456.774 − 0.05 (0.07) 456.281 − 0.06 (0.07) 456.028 − 0.05 (0.07) 456.243 NDVI 1.47 (1.66) 457.857 3.09 (1.91) 456.057 3.52 (1.99)† 455.581 3.42 (1.93)† 455.795 Density of green spaces 0.03 (0.13) 458.585 0.06 (0.06) 457.536 0.03 (0.03) 457.622 0.02 (0.02) 457.867 Density of community or ganizations 0.02 (0.21) 458.629 − 0.09 (0.13) 458.027 − 0.08 (0.08) 457.717 − 0.05 (0.06) 457.903 Psychological DV victimization Risk factors Density of off-premises alcohol outlets 0.04 (0.12) 582.214 0.01 (0.05) 582.304 − 0.02 (0.03) 581.712 − 0.02 (0.02) 580.185 Density of bars − 0.29 (0.15)* 578.118 − 0.08 (0.06) 580.316 − 0.04 (0.02)† 579.715 − 0.03 (0.01)* 578.322 Crime rate − 0.08 (0.10) 584.017 − 0.14 (0.16) 584.536 − 0.11 (0.32) 585.194 − 0.28 (0.37) 584.684 Protective factors Walkability − 0.06 (0.06) 585.322 − 0.13 (0.06)* 582.121 − 0.17 (0.06)** 580.321 − 0.17 (0.07)* 580.406 NDVI − 0.01 (1.41) 582.321 − 0.3 (1.64) 582.287 0.36 (1.82) 582.276 0.41 (1.86) 582.265 Density of green spaces − 0.01 (0.13) 582.312 − 0.04 (0.05) 581.598 − 0.02 (0.03) 581.934 − 0.01 (0.02) 581.934 Density of community or ganizations 0.06 (0.20) 582.208 − 0.07 (0.11) 581.842 − 0.01 (0.07) 582.315 − 0.04 (0.05) 581.745 Physical or sexual DV victimization Risk factors Density of off-premises alcohol outlets 0.06 (0.12) 487.397 − 0.04 (0.06) 487.201 − 0.01 (0.03) 487.537 − 0.01 (0.02) 487.503 Density of bars − 0.10 (0.18) 487.300 − 0.02 (0.08) 487.494 − 0.01 (0.02) 487.482 − 0.01 (0.01) 487.317 Crime rate − 0.01 (0.04) 488.524 − 0.04 (0.15) 488.502 − 0.05 (0.46) 488.552 0.23 (0.50) 488.303 Protective factors Walkability − 0.15 (0.07)* 485.604 − 0.15 (0.07)* 486.837 − 0.13 (0.07)† 488.062 − 0.08 (0.07) 490.049 NDVI 2.69 (1.56)† 484.771 3.30 (1.79)† 484.218 3.86 (1.81)* 483.195 4.25 (1.82)* 482.368 Density of green spaces 0.02 (0.12) 487.574 0.06 (0.05) 486.369 0.03 (0.03) 486.950 0.03 (0.02) 485.826 Density of community or ganizations 0.42 (0.20)* 482.699 0.13 (0.11) 486.268 0.06 (0.07) 486.763 0.02 (0.05) 487.454 Models were adjusted for individual-level covariates and sociodemographic characteristics at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates and sociodemographic characteristics and walkability at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) Models were adjusted for individual-level covariates, and sociodemographic characteristics, walkability, density of green spaces, density of off-premises alcohol outlets and density of bars at the neighborhood-level. The scale used for each neighborhood-level covariates varies by outcomes (lowest AIC) *** p < 0.001; **p < 0.01; *p < 0.05; †p < 0.10 Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 14 of 21 Table 4 Correlations between buffers for neighborhood-level (AIC = 482.368) but the difference in AIC between the variables two models were negligible (ΔAIC = 0.827). An increase in density of community organizations within a 250 m 500 m 750 m 1000 m radius (ß = 0.42; SE = 0.20; p = 0.039) was also associated Rho Rho Rho with greater physical/sexual DV victimization. No sig- 250 m nificant associations were found between the density of Density of off-premises alcohol outlets 0.72 0.66 0.63 green spaces and DV. Density of bars 0.60 0.48 0.43 Crime rate 0.71 0.61 0.56 Correlations between buffers for neighborhood‑level Walkability 0.90 0.85 0.82 variables NDVI 0.95 0.89 0.86 Table 4 shows the Spearman’s rho correlation coeffi - Density of green spaces 0.61 0.57 0.47 cients between the four buffers for each neighborhood- Density of community organizations 0.63 0.47 0.38 level variable. All correlations were significant (p < 0.001) 500 m with rho coefficients ranging from 0.38 to 99. There were Density of off-premises alcohol outlets 0.90 0.85 strong correlations across buffers for walkability (from Density of bars 0.80 0.71 0.82 to 0.98) and NDVI (from 0.86 to 0.99). Moderate to Crime rate 0.88 0.81 strong correlations were observed for crime rate (from Walkability 0.96 0.93 0.56 to 0.94), density of green spaces (from 0.47 to 0.93), NDVI 0.97 0.94 density of alcohol outlets (from 0.63 to 0.96), and density Density of green spaces 0.86 0.78 of bars (from 0.43 to 0.89). Finally, there were weak to Density of community organizations 0.77 0.65 strong correlations between buffers for density of com - 750 m munity organizations (from 0.38 to 0.86). Density of off-premises alcohol outlets 0.96 Density of bars 0.89 Discussion Crime rate 0.94 The current study aimed to contribute to the body of Walkability 0.98 research on the determinants of DV by analyzing the NDVI 0.99 effects of several neighborhood characteristics (crimi - Density of green spaces 0.93 nality, greenness, walkability, density of green spaces, Density of community organizations 0.86 alcohol outlets, and community organizations) using a All p values < 0.001 multi-scalar approach. To our knowledge, no other study had simultaneously assessed various forms of DV vic- timization and perpetration in relation to multiple neigh- Neighborhood risk factors and DV borhood risk and protective factors. Results of this study The effects of neighborhood risk factors on DV were suggest that several neighborhood characteristics could mainly observed for DV perpetration among boys. There influence DV. These factors would be active at different was an association between crime rate within a buffer of scales, and their effects would be modified by gender 500 m and 750 m and the perpetration of physical/sexual with a varying amplitude depending on the form of vio- violence for boys only. Although there was a strong cor- lence considered. relation for crime rate between the buffer of 500 m and Our results showed that associations between neigh- 750 m, the effect of this characteristic was stronger in borhood factors and DV varied across buffers. In addi - the model using a buffer of 500 m (substantially lower tion to these findings, correlations between the buffers AIC). A positive association was also observed among of 500 m, 750 m, and 1000 m were strong for most vari- boys between the crime rate within a 250 m radius and ables, but correlations between the buffer of 250 m and physical/sexual DV victimization. To date, only stud- the other buffers were weaker, suggesting that processes ies that have used measures of perceived neighborhood acting at different scales could be captured. Criminality, disorder have found significant associations with DV density of green spaces, alcohol outlets, and community [21–24]. However, perceptual measures are subject to organizations were specifically sensitive to the influence same-source bias [79]. This bias occurs when outcomes of scale. The sensitivity to scale in our findings is indica - and neighborhood characteristics are self-reported and tive that not all neighborhood determinants of DV act at can be interrelated. In our case, having experienced DV the same scale. Reflecting on the choice of an appropriate could influence the perception of neighborhood disorder scale to analyze associations between neighborhood fac- and vice versa. Studies using police data to describe the tors and DV is essential to reduce the risk of bias in esti- effect of crime have found no significant association with mating the risks associated with these factors. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 15 of 21 DV [25, 26]. However, crime rate was measured at the between the buffer of 500 m, 750 m, and 1000 m were census tract level, which corresponds to the average size strong, but the effect of this characteristic on DV is only of population and area of a 750 m buffer. Our results sug - observed for the 500 m buffer, suggesting that the density gest that this variable may act more locally (250 m and of these retails in the immediate neighborhood may have 500 m). Exposure to neighborhood crime could lead to a more significant influence on consumption. Finally, the the normalization of violence, which may influence the effect of the density of off-premises alcohol outlets was adoption of violent behavior in intimate relationships [9, only observed for boys. This gender sensitivity may be 80]. These processes could be mainly local. Crime tends explained by lower parental monitoring for boys than to be concentrated in some microenvironments (e.g., girls [55]. Furthermore, our results suggested a protec- blocks, street segments) [81] and adolescents living in tive effect of the density of bars (250 m and 1000 m) on these microenvironments could be more impacted by the psychological DV victimization among girls. It is possi- negative effects of criminality. Let us note that the cor - ble that the density of bars was confounded by the den- relation between the buffers of 500 m and 750 m was sity of other services that have not been considered in strong but, a weaker correlation was found between the this study. In particular, cafés and malls could contribute buffers of 500 m and 1000 m. In addition, crime rate in to social interaction and social cohesion in neighbor- a 250 m buffer was only moderately correlated with that hoods [84], a potential protective factor for DV [42, 43]. in the other buffers. Smaller buffers (especially the 250 m Moreover, let us note that the correlation between the buffer) may better capture the concentration of criminal - 250 m and 1000 m buffers for density of bars was moder - ity, while larger buffers could more adequately describe ate (rho = 0.43), suggesting that different processes could processes occurring at larger scales. In addition to these have been captured at each scale. The 1000 m buffer differences in the spatial scale of analysis, adjusting for could reflect the density and diversity in the neighbor - the level of greenness, walkability and density of bars, off- hoods, while the 250 m buffer could better describe the premises alcohol outlets and green spaces resulted in a immediate proximity to services. In some residential suppression effect. Sensitivity analyses in which all these areas where services may be highly concentrated (e.g., on variables were removed found no association between a few streets), the 250 m buffer zone may better reflect crime rate and physical/sexual DV perpetration and vic- the proximity of these places, which would be masked by timization (results available upon request). Suppression the use of larger buffers. occurs when the inclusion of one or more variables in a model increases the effect of another variable by remov - Neighborhood protective factors and DV ing irrelevant variability in the predictors, leading to a Among the protective factors analyzed in this study, more robust estimate of the effect [82]. The adjustment walkability was found to be linked to several forms of DV for a set of neighborhood-level variables in our models for girls but not for boys. Increased walkability within a allowed to isolate the effect of crime that was not related buffer of 500 m, 750 m, and 1000 m around participants’ to the physical characteristics of the environment. These homes was negatively associated with psychological DV results highlight the relevance of our DAG to investi- perpetration and victimization. An increase of walkabil- gate the links between crime and DV. Therefore, further ity in a 250 m and 500 m buffer could also reduce the research should consider contextual factors when analyz- risk of physical/sexual DV victimization. As our theo- ing these associations. retical framework suggests, this physical characteristic of Access to alcohol outlets could also be a potential risk neighborhoods may encourage active behavior and phys- factor for boys. The density of off-premises alcohol out - ical activity [49, 71, 72] as well as contribute to greater lets within a 500 m radius was positively associated with social interactions and a sense of belonging [11, 50], psychological DV perpetration among boys. Several which could positively affect adolescent behaviors, such studies exploring the links between the density of alcohol as reducing violent behavior [52]. These conditions could outlets and intimate partner violence have also identified promote the development of social ties and the presence significant associations [27–29]. Selling alcohol to minors of witnesses who could act to prevent such behavior or is prohibited, which may make it difficult for adoles - intervene. Girls, in particular, would benefit from the cents to visit bars. However, off-premises alcohol outlets positive effects of walkability. The effect of this charac - could be used to obtain alcohol through older relatives teristic of the neighborhoods on psychological violence (e.g., friends, brothers, sisters) [83]. The density of off- was observed up to 1000 m, an acceptable walking dis- premises alcohol outlets may thus influence consumption tance for adolescents [85]. The small difference between among adolescents [32], which may relate to an increased the buffer zones in the estimates of the effect of walkabil - risk of perpetrating DV [7]. According to our results, ity and models’ fit suggests a small influence of the spa - correlations for density of off-premises alcohol outlets tial scale of analysis for this variable. This minor scaling Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 16 of 21 effect was also observed when analyzing correlations among boys. Conversely, an increase in greenness within between buffers. All buffers were strongly correlated (rho a buffer of 750 m and 1000 m was associated with a ranging from 0.82 to 0.98). However, walkability could greater risk of physical/sexual DV victimization among have a more local influence (250 m and 500 m) for the girls. Although no study has evaluated the effect of more severe forms of DV (physical/sexual). Walkability greenness on DV, this feature of neighborhoods could could foster collective efficacy [11], which may depend reduce aggressive behavior [37], a potential risk factor for on social interactions and proximity between residents, DV victimization [89]. Our results showed that boys are as other social processes [86]. Walkability of the local more sensitive to the protective effects of greenness than environment could affect these social processes, while girls. For girls, higher greenness would have a deleterious walkability on a larger scale could have a greater influ - effect on DV victimization, suggesting that this link could ence on active behavior. Future studies should investi- involve other variables that were not measured, such as gate these potential mediation effects to provide a better parental supervision and social control. The characteris - understanding of the mechanisms through which walk- tics of the physical environment may influence parents’ ability acts on DV. perception of the neighborhood, and positive attributes The density of community organizations in a buffer could lead to a more permissive parenting style regard- of 250 m was positively associated with physical/sexual ing outdoor activities [90]. In neighborhoods with a high DV victimization among girls. This result seems coun - level of greenness, girls may have more freedom to go ter-intuitive as these resources tend to positively affect out alone, without parental supervision. However, such adolescents by offering them supervised activities and neighborhoods could be associated with the presence of opportunities to develop their social ties [10, 44]. How- shielded areas with a limited presence of adults resulting ever, some studies suggest that organizations with lit- in opportunities for unsupervised activities [91]. Such tle structured activities may have deleterious effects on contexts could, in turn, increase the risk of DV victimiza- adolescents by promoting interactions between at-risk tion for girls. Furthermore, correlations between buffers individuals [87]. The presence of these services in the showed a minor scaling effect for greenness (rho ranging immediate environment of girls may increase their risk from 0.86 to 0.99). The variation in the level of green - of physical or sexual victimization due to a greater pres- ness could be mainly explained at larger scales, as there ence of adolescents adhering to norms and attitudes would be little local variation in this factor. The buffer that encourage violence. To test this hypothesis, future of 750 m and 1000 m, for which the effects of greenness studies should identify the effect of community organi - were observed, may better capture these processes. These zations by distinguishing between the different types of results are consistent with those reported by Younan et al. organizations and considering the services offered by [37] in that the effects of greenness are observed at larger these organizations. These studies should also consider scales only (1000 m for boys and 750 m and 1000 m for the influence of the spatial scale of analysis. The effect of girls). These radii could correspond to the areas used by density of community organizations was found only for adolescents. The 1000 m distance corresponds to about the buffer of 250 m. In addition, correlations between the 15 min of walking and is consistent with the travel dis- 250 m buffer and the other buffers ranged from moder - tance by walking of adolescents [85]. ate to weak. The scale sensitivity for this variable could Finally, no association between the density of green be explained by the small number of community organi- spaces and any form of DV was found. The density of zations (n = 423) and suggest the importance of choos- these resources is not necessarily associated with greater ing the scale of analysis carefully. Finally, let us note that use. Some characteristics of green spaces, such as facili- no effect was observed for boys and other forms of vio - ties (e.g., sports fields, picnic tables), could be more pre - lence for girls. The availability of these resources is not dictive of adolescents’ use [92]. Further studies should always associated with their use by adolescents, which examine the effect of access to green spaces on DV, con - may depend on other conditions such as the interest of sidering their characteristics. The influence of the spa - adolescents and the cost of the activities [88]. Therefore, tial scale of analysis should also be considered because it would be relevant to further analyze the effect of ado - a potential scaling effect for the density of green spaces lescents’ use of community resources using measures may exist. Our results showed that correlations between of social participation in order to adequately assess the the buffer of 500 m, 750 m, and 1000 m for the density effect of these environments. of green spaces were strong. However, these buffers were Associations between greenness and DV were also only moderately correlated with the buffer of 250 m. observed but, the direction of this relationship varies by The 250 m buffer may better capture proximity to this gender. Greenness in the 1000 m radius buffer could have resource, while larger buffers would be more representa - a protective effect on psychological DV victimization tive of density within neighborhoods. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 17 of 21 Strengths and limitations knowledge, no study has assessed the effect of collective This study contributes to the research on neighbor - efficacy using small spatial units as a proxy of the local hoods’ factors of DV and demonstrates the importance environment. Therefore, future research should develop of considering scale when analyzing these associations. methods to estimate the characteristics of the social envi- The effects of some neighborhood characteristics related ronment at fine scales and further investigate the influ - to social disorganization theory that had already been ence of these factors on DV by considering scale effects. studied in the United States were analyzed, as well as Small area estimation methods [93] are promising tools other characteristics of the physical environment that for estimating components of the social environment had not yet been documented. This study also provides in small spatial units from survey data. Lastly, logistic a multi-scalar understanding of neighborhood-level pro- regressions performed in this study did not consider the cesses and highlights the importance of neighborhood spatial dependency of the data. However, spatial pat- definition. tern in neighborhoods’ characteristics as well as health Some limitations should be considered. Firstly, self- outcomes have been observed in previous research. For reported data for DV are at risk of social desirability bias. example, a study found a spatial patterning of neighbor- This bias could not be controlled due to the absence of hood physical characteristics and depressive symptoms a social desirability measure in the QHSHSS. Secondly, among adolescents [94]. Authors suggested that depres- physical and sexual DV were assessed jointly as the prev- sive symptoms could be clustered due to social interac- alence of sexual DV perpetration for girls was low (2%). tions (e.g., neighborhood peer effects) and exposure Strong correlations between these forms of violence for and response to the same neighborhood factors. Similar both girls and boys were observed in our data, suggest- processes could also be observed with DV. Future stud- ing that physical and sexual DV may co-occur. Using such ies evaluating associations between neighborhood char- a measure is also consistent with other studies reporting acteristics and DV should assess the potential spatial on the relationship between DV and neighborhood fac- pattern of DV by considering the spatial dependency of tors [25–29]. However, measuring physical and sexual data. Spatial regressions would address this issue by tak- DV jointly did not make it possible to identify factors ing into account spatial autocorrelation (i.e., level of simi- specifically associated with sexual DV. Low-prevalence larity/dissimilarity between neighboring observations) outcomes could be more sensitive to outliers and require and spatial non-stationarity (i.e., spatial variability in the a large sample to ensure a representative sample for amplitude and direction of estimated effects) [95]. How - analyzing the influence of neighborhood-level factors. ever, these methods require an adequate spatial density Future studies on the relationship between neighbor- of participants, which is rarely the case for national sur- hoods’ characteristics and sexual DV should therefore veys. Future studies should therefore conduct large sam- focus on sexual DV. Thirdly, more than 26% of the par - ple surveys to address this objective. ticipants were excluded because they did not provide a valid postal code. This information was required to assign Conclusions to respondents the neighborhood-level variables. Missing Multiple neighborhood characteristics could influence data on postal codes led to a reduction in the sample size, DV at different scales. Social disorganization theory, which reduced statistical power. However, let us note that often referred to for explaining these relationships [9], these missing data should not substantially impact our could only partially explain the neighborhood effects. estimates as they are randomly distributed for all covari- Our results suggest that some physical characteristics of ates, except for grade level among girls. Fourthly, social neighborhoods, such as walkability, could have a protec- environment characteristics were not analyzed in this tive effect on DV. Future studies should investigate the study but could influence DV. Some studies have focused effect of the physical environment on these behaviors on the associations between DV and collective efficacy by analyzing underlying processes (e.g., social support, [23, 43, 53, 54], a concept related to social disorganization social participation). Furthermore, our study showed theory [17]. However, collective efficacy was not assessed that there is no single scale to assess the effect of neigh - in the QHSHSS, and the spatial density of respondents borhoods’ characteristics. Therefore, the comparison of was too low to derive neighborhood-level variables from scales should be systematic to consider the multiscale individual responses. Studies in urban areas most often effects of neighborhoods’ characteristics. used measurements at the individual level [23, 43, 53, 54], Our results regarding the association between neigh- which are limited in neighborhood-level processes. Roth- borhood characteristics and DV could be observed in man et al. [43] measured collective efficacy at the neigh - several geographic contexts but may not be applicable to borhood level, but neighborhoods were operationalized urban contexts of the Global South. In effect, studies on with larger spatial units than the census tracts. To our neighborhoods’ effects on DV conducted outside of the Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 18 of 21 United States and Canada are still rare. Specifically, cit - of neighborhoods that are modifiable and reach large ies in the Global South are different from those in North numbers of people [101]. Urban planning and pub- America (e.g., presence of informal settlements, larger lic policies could represent new avenues in efforts to socioeconomic and health inequalities, and extreme efficiently prevent DV. Interventions implemented urban growth). Such differences may cause the neigh - around the world to improve the physical local envi- borhood-level factors analyzed in our study to influ - ronment (e.g., greening, increasing density of facilities, ence DV differently in the Global South, namely, if other streetscape improvements) showed positive effects on neighborhood-level factors contribute to the incidence adolescent behaviors and health outcomes [102, 103]. of DV. Despite these issues, neighborhood effects have Improving the physical environment would also pro- been observed in several countries. For example, empiri- mote social cohesion and residents’ engagement in cal studies found associations between some neighbor- community life, which could benefit individuals [104]. hood factors, such as sociodemographic characteristics, In the United States, an intervention that aimed at neighborhood disorder and crime, and intimate partner greening neighborhoods and improving social cohe- violence among adults in Europe [96], Asia [97] as well as sion through residents’ engagement has shown promis- Africa [98]. Associations between neighborhoods’ char- ing results in preventing violence and crime [105]. Such acteristics and DV among adolescents may therefore be programs could be effective in reducing DV and could observed in different regions of the world, but local spe - also have positive effects on adolescents’ health and cificities could lead to differences in the effects of the fac - well-being. tors analyzed in this study. Although the relationship between neighborhood Supplementary Information factors and DV may depend on geographic context, the The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12942- 022- 00306-3. influence of spatial scale of analysis is likely to affect most types of local environments, suggesting that our Additional file 1: Directed Acyclic Graph. results on scale sensitivity are not limited to the North American context. In effect, the neighborhood factors Acknowledgements analyzed in the current studies are likely to influence The authors thank the Québec Population Health Research Network (QPHRN) individuals in different regions of the world. Several for its contribution to the financing of this publication. The authors would also like to thank the Québec Statistics Institute for providing the dataset studies around the world have suggested neighbor- from the Québec Health Survey of High School Students. The first author was hood effects on health outcomes, and some have shown supported by a scholarship from the Canada Research Chair in interpersonal potential scale effects for certain factors. For example, traumas and resilience. a literature review reported that greenness and access Author contributions to green spaces could influence health outcomes in dif - PR, MP, and MH contributed to the study conception, the interpretation of the ferent regions of the world, including the Global South, results, and reviewed the manuscript. PR performed analysis and drafted the manuscript. This study was supervised by MP and MH. All authors read and and suggested that these effects could depend on the approved the final manuscript. spatial scale of analysis (e.g., large versus small buffers to operationalize exposure to green space) [99]. Fur- Funding Not applicable. thermore, the egocentric neighborhood is a promising approach that could be easily implemented in various Availability of data and materials geographic settings. Although the definition of admin - Data from the Québec Health Survey of High School Students are provided by Québec Statistics Institute and are not publicly available. Statistics Canada istrative units (size and shape) varies across countries, 2016 Census data at the dissemination area level are publicly available: https:// egocentric neighborhoods ensure consistency in oper- www150. statc an. gc. ca/ n1/ en/ catal ogue/ 98- 401- X2016 044. Data on the popu- ationalizing neighborhoods, which may increase the lation size of the dissemination blocks are also provided by Statistics Canada and are publicly available: https:// www150. statc an. gc. ca/ n1/ en/ catal ogue/ 92- comparability of studies. Finally, measures used in the 163-X. All other data used and analyzed during the current study are available current study to operationalize neighborhood factors from the corresponding author on reasonable request. are easily replicable, and some have been validated in different contexts. For example, assessing greenness by Declarations using NDVI is consistent with several studies around Ethics approval and consent to participate the world [99]. The walkability index has also been used Ethical approval was obtained from the ethics board of the Université du in many countries in the Global North [71–73, 100] as Québec à Montréal and from the ethics board of the Québec Statistics Insti- well as in the Global South [100]. tute. Data from the Québec Health Survey of High School Students used in the current study were anonymous, and no participants were contacted. This study also has practice implications. Our results highlight the importance of physical features Competing interests The authors declare that they have no competing interests. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 19 of 21 Received: 30 March 2022 Accepted: 1 June 2022 22. Champion H, Foley KL, Sigmon-Smith K, Sutfin EL, DuRant RH. Con- textual factors and health risk behaviors associated with date fight - ing among high school students. Women Health. 2008;47(3):1–22. 23. Chang L-Y, Foshee VA, Reyes HLM, Ennett ST, Halpern CT. Direct and indirect effects of neighborhood characteristics on the perpe - tration of dating violence across adolescence. J Youth Adolesc. References 2015;44(3):727–44. 1. Wincentak K, Connolly J, Card N. Teen dating violence: A meta-analytic 24. Reed E, Silverman JG, Raj A, Decker MR, Miller E. Male Perpetration review of prevalence rates. Psychol Violence. 2017;7(2):224–41. of teen dating violence: associations with neighborhood violence 2. Leen E, Sorbring E, Mawer M, Holdsworth E, Helsing B, Bowen E. Preva- involvement, gender attitudes, and perceived peer and neighbor- lence, dynamic risk factors and the efficacy of primary interventions hood norms. J Urban Health. 2011;88(2):226–39. for adolescent dating violence: an international review. Aggress Violent 25. Li Q, Kirby RS, Sigler RT, Hwang S-S, LaGory ME, Goldenberg RL. Beh. 2013;18(1):159–74. A Multilevel Analysis of Individual, Household, and Neighbor- 3. Banyard VL, Cross C. Consequences of teen dating violence: under- hood Correlates of Intimate Partner Violence Among Low-Income standing intervening variables in ecological context. Violence Against Pregnant Women in Jefferson County Alabama. Am J Public Health. Women. 2008;14(9):998–1013. 2010;100(3):531–9. 4. Eaton DK, Davis KS, Barrios L, Brener ND, Noonan RK. Associations of 26. Taylor BG, Mumford EA, Okeke N, Rothman E. Neighborhood violent dating violence victimization with lifetime participation, co-occurrence, crime and adolescent relationship aggression. Aggressive Behav. and early initiation of risk behaviors among us high school students. J 2020;46(1):25–36. Interpers Viole. 2007;22(5):585–602. 27. Iritani BJ, Waller MW, Halpern CT, Moracco KE, Christ SL, Flewelling RL. 5. Taquette SR, Monteiro DLM. Causes and consequences of ado- Alcohol outlet density and young women’s perpetration of violence lescent dating violence: a systematic review. J Inj Violence Res. toward male intimate partners. J Family Viol. 2013;28(5):459–70. 2019;11(2):137–47. 28. Waller MW, Iritani BJ, Christ SL, Tucker Halpern C, Moracco KE, Flewelling 6. Hébert M, Daspe M, Lapierre A, Godbout N, Blais M, Fernet M, Lavoie RL. Perpetration of intimate partner violence by young adult males: the F. A meta-analysis of risk and protective factors for dating violence association with alcohol outlet density and drinking behavior. Health victimization: the role of family and peer interpersonal context. Trauma Place. 2013;21:10–9. Violence Abuse. 2019;20(4):574–90. 29. Waller MW, Iritani BJ, Flewelling RL, Christ SL, Halpern CT, Moracco 7. Vagi KJ, Rothman EF, Latzman NE, Tharp AT, Hall DM, Breiding MJ. KE. Violence victimization of young men in heterosexual relation- Beyond correlates: a review of risk and protective factors for adolescent ships: does alcohol outlet density influence outcomes? Violence Vict. dating violence perpetration. J Youth Adolesc. 2013;42(4):633–49. 2012;4:527–47. 8. Vézina J, Hébert M. Risk factors for victimization in romantic relation- 30. Sampson RJ, Raudenbush SW. Systematic social observation of public ships of young women: a review of empirical studies and implications spaces: a new look at disorder in urban neighborhoods. Am J Sociol. for prevention. Trauma Violence Abuse. 2007;8(1):33–66. 1999;105(3):603–51. 9. Johnson RM, Parker EM, Rinehart J, Nail J, Rothman EF. Neighborhood 31. Sampson RJ, Raudenbush SW. Seeing disorder: neighborhood stigma factors and dating violence among youth: a systematic review. Am J and the social construction of “Broken Windows.” Soc Psychol Quarterly. Prev Med. 2015;49(3):458–66. 2004;67(4):319–42. 10. Gardner M, Brooks-Gunn J. Adolescents’ exposure to community vio- 32. Shih RA, Mullins L, Ewing BA, Miyashiro L, Tucker JS, Pedersen ER, lence: are neighborhood youth organizations protective? J Community Miles JNV, D’Amico EJ. Associations between neighborhood alcohol Psychol. 2009;37(4):505–25. availability and young adolescent alcohol use. Psychol Addict Behav. 11. Prince DM, Hohl B, Hunter BA, Thompson AB, Matlin SL, Hausman AJ, 2015;29(4):950–9. Tebes JK. Collective efficacy as a key context in neighborhood support 33. White RM, Witherspoon DP, Wei W, Zhao C, Pasco MC, Maereg TM. for urban youth. Am J Community Psychol. 2019;63(1–2):179–89. Adolescent development in context: a decade review of neighbor. J Res 12. Vanaken G-J, Danckaerts M. Impact of green space exposure on chil- Adolescence. 2021;31(4):944–65. dren’s and adolescents’ mental health: a systematic review. Int J Environ 34. Wilson EO. Biophilia and the conservation ethic. In: Evolutionary per- Res Public Health. 2018;15(12):2668. spectives on environmental problems. Routledge; 2017: 263–272. 13. Philibert M, Riva M. Contribution de la géographie quantitative à l’étude 35. Bezold CP, Banay RF, Coull BA, Hart JE, James P, Kubzansky LD, Missmer des inégalités sociales de santé : Analyse des effets de milieu. Cahiers SA, Laden F. The association between natural environments and de géographie du Québec. 2011;55(156):343–61. depressive symptoms in adolescents living in the United States. J 14. Chaix B, Merlo J, Evans D, Leal C, Havard S. Neighbourhoods in eco-epi- Adolesc Health. 2018;62(4):488–95. demiologic research: delimiting personal exposure areas. aresponse to 36. Mavoa S, Lucassen M, Denny S, Utter J, Clark T, Smith M. Natural Riva, Gauvin Apparicio and Brodeur. Soc Sci Med. 2009;69(9):1306–10. neighbourhood environments and the emotional health of urban New 15. Hipp JR, Bates CJ. Egohoods. Capturing change in spatial crime pat- Zealand adolescents. Landsc Urban Plan. 2019;191: 103638. terns. In: The Oxford Handbook of Environmental Criminology. Edited by 37. Younan D, Tuvblad C, Li L, Wu J, Lurmann F, Franklin M, Berhane K, Bruinsma GJN, Johnson SD: Oxford University Press; 2018. McConnell R, Wu AH, Baker LA, et al. Environmental determinants of 16. Hipp JR, Boessen A. Egohoods as waves washing across the city: a new aggression in adolescents: role of urban neighborhood greenspace. J measure of “neighborhoods.” Criminology. 2013;51(2):287–327. Am Acad Child Adolesc Psychiatry. 2016;55(7):591–601. 17. Sampson RJ, Raudenbush SW, Earls F. Neighborhoods and vio- 38. Kuo FE, Sullivan WC. Aggression and violence in the inner city: effects of lent crime: a multilevel study of collective efficacy. Science. environment via mental fatigue. Environ Behav. 2001;33(4):543–71. 1997;277(5328):918–24. 39. Lee ACK, Maheswaran R. The health benefits of urban green spaces: a 18. Brunton-Smith I, Jackson J, Sutherland A. Bridging structure and per- review of the evidence. J Public Health. 2010;33(2):212–22. ception: on the neighbourhood ecology of beliefs and worries about 40. Biddle SJH, Ciaccioni S, Thomas G, Vergeer I. Physical activity and mental violent crime. Br J Criminol. 2014;54(4):503–26. health in children and adolescents: an updated review of reviews and 19. Cunradi CB. Neighborhoods, alcohol outlets and intimate partner an analysis of causality. Psychol Sport Exerc. 2019;42:146–55. violence: addressing research gaps in explanatory mechanisms. Int J 41. Jennings V, Bamkole O. The relationship between social cohesion and Environ Res Public Health. 2010;7(3):799–813. urban green space: an avenue for health promotion. Int J Environ Res 20. Gruenewald PJ. The spatial ecology of alcohol problems: niche theory Public Health. 2019;16(3):452. and assortative drinking. Addiction. 2007;102(6):870–8. 42. Garthe RC, Gorman-Smith D, Gregory J, Schoeny EM. Neighborhood 21. Banyard VL, Cross C, Modecki KL. Interpersonal violence in adoles- concentrated disadvantage and dating violence among urban ado- cence: ecological correlates of self-reported perpetration. J Interpers lescents: the mediating role of neighborhood social processes. Am J Violence. 2006;21(10):1314–32. Commun Sychol. 2018;61(34):310–20. Rodrigues et al. International Journal of Health Geographics (2022) 21:6 Page 20 of 21 43. Rothman EF, Johnson RM, Young R, Weinberg J, Azrael D, Molnar BE. 68. Bow CJD, Waters NM, Faris PD, Seidel JE, Galbraith PD, Knudtson ML, Neighborhood-level factors associated with physical dating violence Ghali WA, The AI. Accuracy of city postal code coordinates as a proxy perpetration: results of a representative survey conducted in Boston. for location of residence. Int J Health Geogr. 2004;3(1):5–5. MA J Urban Health. 2011;88(2):201–13. 69. Rhew IC, Vander Stoep A, Kearney A, Smith NL, Dunbar MD. Valida- 44. Quane JM, Rankin BH: Does it pay to participate? Neighborhood-based tion of the normalized difference vegetation index as a measure of organizations and the social development of urban adolescents. 2006; neighborhood greenness. Ann Epidemiol. 2011;21(12):946–52. 2006. 70. Shannon CE. A mathematical theory of communication. Bell Syst 45. Browning CR, Soller B. Moving beyond neighborhood: activity spaces Tech J. 1948;27(3):379–423. and ecological networks as contexts for youth development. Cityscape. 71. Frank LD, Sallis JF, Saelens BE, Leary L, Cain K, Conway TL, Hess PM. 2014;16(1):165–96. The development of a walkability index: application to the Neighbor- 46. Gottfredson DC, Gerstenblith SA, Soulé DA, Womer SC, Lu S. Do after hood Quality of Life Study. Br J Sports Med. 2010;44(13):924–33. school programs reduce delinquency? Prev Sci. 2004;5(4):253–66. 72. Sundquist K, Eriksson U, Kawakami N, Skog L, Ohlsson H, Arvidsson 47. Durlak JA, Weissberg RP, Pachan M. A meta-analysis of after-school D. Neighborhood walkability, physical activity, and walking behavior: programs that seek to promote personal and social skills in children the Swedish Neighborhood and Physical Activity (SNAP) study. Soc and adolescents. Am J Community Psychol. 2010;45(3–4):294–309. Sci Med. 2011;72(8):1266–73. 48. Molnar BE, Cerda M, Roberts AL, Buka SL. Eec ff ts of neighborhood 73. Mayne DJ, Morgan GG, Willmore A, Rose N, Jalaludin B, Bambrick resources on aggressive and delinquent behaviors among urban H, Bauman A. An objective index of walkability for research and youths. Am J Public Health. 2008;98(6):1086–93. planning in the Sydney metropolitan region of New South Wales, 49. Frank LD, Engelke P, Schmid T. Health and community design: The Australia: an ecological study. Int J Health Geogr. 2013;12:61–61. impact of the built environment on physical activity: Island Press; 2003. 74. Inglehart R, Welzel C. Changing mass priorities: the link between 50. Lund H. Pedestrian environments and sense of community. J Plan Educ modernization and democracy. Perspect Polit. 2010;8(2):551–67. Res. 2002;21(3):301–12. 75. Sas Institute: SAS/STAT 9.3 User’s Guide: Survey Data Analysis (book 51. Culyba AJ, Jacoby SF, Richmond TS, Fein JA, Hohl BC, Branas CC. Modifi- Excerpt): SAS institute. 2011. able neighborhood features associated with adolescent homicide. 76. Pearl J. Causal Diagrams for Empirical Research. Biometrika. JAMA Pediatr. 2016;170(5):473–80. 1995;82(4):669–88. 52. Bushover B, Miller E, Bair-Merritt M, Abebe K, Culyba A. Physical environ- 77. Schisterman EF, Cole SR, Platt RW. Overadjustment bias and ment and violence perpetration among male youth in Pittsburgh: a unnecessary adjustment in epidemiologic studies. Epidemiology. spatial analysis. Inj Prev. 2020;26(6):588–92. 2009;20(4):488–95. 53. Jain S, Buka SL, Subramanian SV, Molnar BE. Neighborhood predictors 78. Burnham KP, Anderson DR. Model selection and multimodel of dating violence victimization and perpetration in young adulthood: inference-2nd ed: a practical information-theoretic approach. New a multilevel study. Am J Public Health. 2010;100(9):1737–44. york: Springer-verlag new york Inc. 2002. 54. Schnurr MP, Lohman BJ. The impact of collective efficacy on risks for 79. Diez Roux AV. Neighborhoods and health: where are we and were do adolescents perpetration of dating violence. J Youth Adolescence J we go from here? Rev Epidemiol Sante Publique. 2007;55(1):13–21. Youth Adolescence A Multidiscipl Res Publicat. 2013;42(4):518–35. 80. Raghavan C, Mennerich A, Sexton E, James SE. Community violence 55. Axinn WG, Young-Demarco L, Ro MC. Gender double standards in and its direct, indirect, and mediating effects on intimate partner parenting attitudes. Soc Sci Res. 2011;40(2):417–32. violence. Violence Against Women. 2006;12(12):1132–49. 56. Kupersmidt JB, Griesler PC, DeRosier ME, Patterson CJ, Davis PW. 81. Braga AA, Papachristos AV, Hureau DM. The concentration and stabil- Childhood aggression and peer relations in the context of family and ity of gun violence at micro places in Boston, 1980–2008. J Quant neighborhood factors. Child Dev. 1995;66(2):360–75. Criminol. 2010;26(1):33–53. 57. Leventhal T, BrooksGunn J. The neighborhoods they live in: the effects 82. MacKinnon DP, Krull JL, Lockwood CM. Equivalence of the mediation, of neighborhood residence on child and adolescent outcomes. Psychol confounding and suppression effect. Prev Sci. 2000;1(4):173–81. Bullet. 2000;126(2):309–37. 83. Friese B, Grube JW, Moore RS. Youth acquisition of alcohol and drink- 58. Fotheringham AS, Wong DWS. The Modifiable Areal Unit Problem ing contexts: an in-depth look. J Drug Educ. 2013;43(4):385–403. in Multivariate Statistical Analysis. Environ Plann A Economy Space. 84. Oldenburg R. The great good place: Cafes, coffee shops, bookstores, 1991;23(7):1025–44. bars, hair salons, and other hangouts at the heart of a community: 59. Openshaw S. The modifiable areal unit problem. Norwich: Geo Da Capo Press; 1999. Abstracts Univ. of East Anglia; 1984. 85. Colabianchi N, Coulton CJ, Hibbert JD, McClure SM, Ievers-Landis 60. Kwan M-P. The uncertain geographic context problem. Ann Assoc Am CE, Davis EM. Adolescent self-defined neighborhoods and activity Geogr. 2012;102(5):958–68. spaces: Spatial overlap and relations to physical activity and obesity. 61. Hipp JR, Perrin AJ. The simultaneous effect of social distance and physi- Health Place. 2014;27:22–9. cal distance on the formation of neighborhood ties. City Community. 86. Hipp JR, Boessen A. Neighborhoods, networks, and crime. In: Chal- 2009;8(1):5–25. lenging Criminological Theory: The Legacy of Ruth Rosner Kornhauser. 62. Apparicio P, Gelb J, Dubé A-S, Kingham S, Gauvin L, Robitaille É. The Edited by Cullen FT: Taylor & Francis; 2017. approaches to measuring the potential spatial access to urban health 87. Mahoney JL, Stattin H, Lord H. Unstructured youth recreation centre services revisited: distance types and aggregation-error issues. Int J participation and antisocial behaviour development: Selection influ- Health Geogr. 2017;16(1):32–32. ences and the moderating role of antisocial peers. Int J Behav Dev. 63. Hewko J, Smoyer-Tomic KE, Hodgson MJ. Measuring neighbourhood 2004;28(6):553–60. spatial accessibility to urban amenities: does aggregation error matter? 88. Bodilly SJ, Beckett MK. Making Out-of-School-Time Matter: Evidence Environ Plann A Econ Space. 2002;34(7):1185–206. for an Action Agenda. In. Santa Monica, CA: RAND; 2005. 64. Ursache A, Regan S, De Marco A, Duncan DT. The Family Life Project 89. O’Donnell L, Stueve A, Myint-U A, Duran R, Agronick G, Wilson-Simmons Key I: Measuring neighborhood deprivation for childhood health and R. Middle school aggression and subsequent intimate partner physical development - scale implications in rural and urban context. Geospat violence. J Youth Adolesc. 2006;35(5):693–703. Health. 2021. https:// doi. org/ 10. 4081/ gh. 2021. 4926. 90. Vlaar J, Brussoni M, Janssen I, Mâsse LC. Roaming the neighbourhood: 65. Oliver LN, Schuurman N, Hall AW. Comparing circular and network influences of independent mobility parenting practices and parental buffers to examine the influence of land use on walking for leisure perceived environment on children’s territorial range. Int J Environ Res and errands. Int J Health Geogr. 2007;6(1):41. Public Health. 2019;16(17):3129. 66. Statistics Canada: Census Profile. 2016 Census. In: Census Profile. 91. Weeland J, Laceulle OM, Nederhof E, Overbeek G, Reijneveld SA. The Statistics Canada; 2017. greener the better? Does neighborhood greenness buffer the effects 67. Straus MA, Hamby SL, Boney-McCoy SUE, Sugarman DB. The revised of stressful life events on externalizing behavior in late adolescence? conflict tactics scales (CTS2): development and preliminary psycho - Health Place. 2019;58: 102163. metric data. J Fam Issues. 1996;17(3):283–316. R odrigues et al. International Journal of Health Geographics (2022) 21:6 Page 21 of 21 92. Baran PK, Smith WR, Moore RC, Floyd MF, Bocarro JN, Cosco NG, Dan- ninger TM. Park use among youth and adults: examination of individual, social, and urban form factors. Environ Behav. 2013;46(6):768–800. 93. Ghosh M, Rao JN. Small area estimation: an appraisal. Stat Sci. 1994;9(1):55–76. 94. Duncan DT, Piras G, Dunn EC, Johnson RM, Melly SJ, Molnar BE. The built environment and depressive symptoms among urban youth: a spatial regression study. Spat Spatiotemporal Epidemiol. 2013;5:11–25. 95. Anselin L. Spatial Econometrics: Methods and Models. Dordrecht: Springer, Netherlands; 2013. 96. Gracia E, López-Quílez A, Marco M, Lladosa S, Lila M. Exploring neigh- borhood influences on small-area variations in intimate partner vio - lence risk: a Bayesian random-effects modeling approach. Int J Environ Res Public Health. 2014;11(1):866–82. 97. Koenig MA, Stephenson R, Ahmed S, Jejeebhoy SJ, Campbell J. Individual and contextual determinants of domestic violence in North India. Am J Public Health. 2006;96(1):132–8. 98. Uthman OA, Moradi T, Lawoko S. The independent contribution of individual-, neighbourhood-, and country-level socioeconomic position on attitudes towards intimate partner violence against women in sub- Saharan Africa: a multilevel model of direct and moderating effects. Soc Sci Med. 2009;68(10):1801–9. 99. Rigolon A, Browning MHEM, McAnirlin O, Yoon HV. Green space and health equity: a systematic review on the potential of green space to reduce health disparities. Int J Environ Res Public Health. 2021;18(5):2563. 100. Adams MA, Frank LD, Schipperijn J, Smith G, Chapman J, Christiansen LB, Coffee N, Salvo D, du Toit L, Dygrýn J, et al. International variation in neighborhood walkability, transit, and recreation environments using geographic information systems: the IPEN adult study. Int J Health Geogr. 2014;13(1):43. 101. Diez Roux AV. Investigating neighborhood and area effects on health. Am J Public Health. 2001;91(11):1783–9. 102. Hunter RF, Cleland C, Cleary A, Droomers M, Wheeler BW, Sinnett D, Nieuwenhuijsen MJ, Braubach M. Environmental, health, wellbeing, social and equity effects of urban green space interventions: a meta- narrative evidence synthesis. Environ Int. 2019;130:104923. 103. Smith M, Hosking J, Woodward A, Witten K, MacMillan A, Field A, Baas P, Mackie H. Systematic literature review of built environment effects on physical activity and active transport - an update and new findings on health equity. Int J Behav Nutr Phys Act. 2017;14(1):158–158. 104. Aiyer SM, Zimmerman MA, Morrel-Samuels S, Reischl TM. From Broken Windows to Busy Streets: A Community Empowerment Perspective. Health Educ Behav. 2014;42(2):137–47. 105. Heinze JE, Krusky-Morey A, Vagi KJ, Reischl TM, Franzen S, Pruett NK, Cunningham RM, Zimmerman MA. Busy streets theory: the effects of community-engaged greening on violence. Am J Community Psychol. 2018;62(1–2):101–9. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your ﬁeld rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

International Journal of Health Geographics – Springer Journals

Published: Jun 20, 2022

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Associations between neighborhood characteristics and dating violence: does spatial scale matter?

Associations between neighborhood characteristics and dating violence: does spatial scale matter?

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Associations between neighborhood characteristics and dating violence: does spatial scale matter?

Associations between neighborhood characteristics and dating violence: does spatial scale matter?

References (124)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies