Effect of Perceived Neighborhood Environment on Cannabis Use during Pregnancy among African American Women

Abstract

Environmental context is an important predictor of health behavior. Understanding its effect on cannabis use among pregnant women is yet to be understood. The aim of the study is to assess the impact of perceived neighborhood environment on prenatal cannabis use and explore the mediating role of stress. Data are from the Life-Course Influences on Fetal Environments Study (LIFE), a retrospective cohort of postpartum African American women in Metropolitan Detroit, Michigan (2009–2011). Prenatal cannabis use was defined as self-reported ever use during pregnancy. Three perceived neighborhood scales were considered: social cohesion and trust, social disorder, and danger and safety. Out of 1,369 women, 151 (11.0%) self-reported using cannabis during pregnancy. After adjusting for age, marital status, income, years of education, and general social support scale, the odds of cannabis use significantly increased among the lowest quartiles of all the neighborhood scales suggesting higher cannabis use among women who perceived their neighborhoods to have the worst conditions. Compared to the highest quartile, the odds ratio (OR) for the lowest quartiles for social cohesion and trust, social disorder, and danger and safety were 1.77 (95% confidence interval (CI): 1.04–3.03), 1.83 (95% CI: 1.15–2.91), and 1.93 (95% CI: 1.12–3.31) respectively. Evidence of mediation by perceived stress was only present between the association of perceived levels of safety and danger with cannabis use during pregnancy. Future prospective studies are warranted to understand the causal associations between individual correlates and social and physical environmental factors of prenatal cannabis use.

Introduction

Prenatal cannabis use is a mounting public health concern. With the ongoing legalization of recreational cannabis, an increase in the prevalence, potency, and perceived safety of cannabis use has been documented among pregnant populations [1–3]. Studies have associated prenatal cannabis use with fetal growth restriction, preterm birth, low birth weight, and increased neonatal intensive care admission [4]. Although not yet conclusive, cannabis exposure in utero may also negatively impact the developing brain and can be associated with neurobehavioral outcomes [5].

Multiple levels of influence shape individual health behaviors including intrapersonal, interpersonal, and environmental factors. While intrapersonal and interpersonal correlates of prenatal cannabis have been examined in the literature [6–8], little is known about the impact of environmental and neighborhood context on cannabis use. The majority of the literature is mainly limited to the non-pregnant adult and adolescent populations. Previous research has demonstrated that high rates of perceived neighborhood problems, disorganization, and fear were associated with increased use of substances including cannabis [9–11]. Understanding this association among the pregnant population is yet to be investigated. The limited available research mainly looked at the association between prenatal cannabis use and objective measures of the neighborhood environment. Brown et al. [12] revealed a positive relationship between prenatal cannabis use and neighborhood characteristics including lone-parent household prevalence (adjusted odds ratio (AOR) = 1.01, 95% CI: 1.01–1.02), aboriginal prevalence (AOR = 1.01, 95% CI: 1.00–1.02), high school diploma or less prevalence (AOR = 1.02, 95% CI: 1.01–1.03) and neighborhood low-income measure after-tax (AOR = 1.01, 95% CI: 1.01–1.02). Moreover, Young-Wolff et al. [13] reported an inverse association between prenatal cannabis use and neighborhood deprivation index. They also found that the greater number of cannabis retailers within a 15-min drive was associated with higher odds of cannabis use during pregnancy [13].

The resident’s perceptions about their environment might also provide important and different information. These subjective measures are believed to be more closely associated with health outcomes than objective measures [14]. Perceptions about the challenges and hazards of an individual’s neighborhood may cause psychological distress [15] which in turn can lead to increased odds of substance use [11]. Studies have shown that prenatal tobacco smoking and alcohol drinking are associated with worse perceptions of social cohesion [16] and self-reported neighborhood violence [16, 17]. To our knowledge, no study has previously investigated subjective measures of neighborhood and cannabis use among pregnant women.

Given the changing legal landscape around cannabis, it is imperative to expand our knowledge about the correlates of cannabis use among pregnant people. Nevertheless, there is a current dearth of information regarding the impact of neighborhood perceptions on prenatal cannabis use and a lack of research exploring mediators of this association. Towards this end, this study aims to assess the association between perceived neighborhood environment on cannabis use during pregnancy among African American women and to examine whether stress levels mediate this association.

Material and Methods

Study Design

The analysis of the study was based on the Life-Course Influences on Fetal Environments Study (LIFE), a retrospective cohort of postpartum African American women in a suburban hospital in Oakland County, Michigan. Participants who self-identified as African American women, were at least 18 years of age, and had a singleton birth were recruited in 2009–2011 from the labor and delivery and postpartum hospital units. Data were collected from medical records abstractions and in-person interviews conducted by trained personnel that were race- and gender-matched to the participant. All study participants gave written informed consent. The study protocol was approved by relevant institutional review boards. Information about the LIFE study has been previously published [18].

Study Outcome

The main dependent variable is prenatal cannabis use, defined as ever use during pregnancy (yes/no). Due to the sensitivity of the topic, women were initially asked if they had ever used marijuana, hashish, pot, and/or grass. Participants with positive responses were then asked about the date of last use (month, day, and year). If the last self-reported date of cannabis use was on or after the mother’s date of last menstrual period, prenatal cannabis use was determined to be positive. Mothers who responded negatively to ever use or whose date of last use occurred before the mother’s date of last menstrual period were classified as never users. The prevalence of tobacco smoking (yes/no) and alcohol use (yes/no) during pregnancy was also examined by cannabis status. Tobacco smoking was calculated from four different questions inquiring about ever-smoking status, smoking status in the past year, number of cigarettes used in the past year, and amount of cigarette use during the first trimester. Alcohol use, on the other hand, was assessed as the frequency of drinking a beer, glass of wine, or liquor during the pregnancy. All women responding to “never” use were classified in the “no” category.

Exposure Variables

The main exposure variable is the women’s perception on a series of statements regarding the physical and social environment of her current neighborhood, defined as the “area around where you live and around your house… include places you shop, religious or public institutions, or a local business district…. general area around your house where you might perform routine tasks, such as shopping, going to the park, or visiting with neighbors.” Individual scores for each response were summed up to calculate three valid and reliable scales: (i) social disorder [19], (ii) social cohesion and trust [19, 20], and (iii) danger and safety [19, 20]. Detailed information about the scales is available in Appendix 1 (Table 4). In general, higher scores for the social cohesion and trust scale and danger and safety scale imply a better perception of the environment, whereby higher scores for the social disorder scale correspond to more disorder. The neighborhood scale variables were also analyzed as categorical variables by rescaling the scores into their interquartile range.

Mediator

Perceived stress was hypothesized to be a potential mediator in the association between perceived neighborhood environment and prenatal cannabis use. Perceived stress was measured using the 14-item Cohen’s perceived stress scale [21]. The reference period for the scale was the past month and higher values represented higher levels of stress (Appendix 2 (Table 5)).

Covariates

Potential socio-demographic covariates associated with cannabis use and neighborhood environment were considered in the analysis. These covariates included maternal age at enrollment, current marital status, family annual income, total years of education, and general social support scale. The social support scale is based on the Social Support Survey developed for the Medical Outcomes Study (MOS) [22]. It is composed of 11-item statements about the availability of social support (Appendix 2). The composite score of the individual responses was rescaled into a 2-level categorical variable based on its median value.

Data missingness for the covariates ranged from 0.3 to 1.3%, except for income with 10.9% of responses missing. Missing income was hence imputed through multiple imputations using the Markov chain Monte Carlo (MCMC) method on SPSS (Statistical Package for the Social Sciences) to generate 20 imputed datasets. The imputation model included all the study covariates as well as other auxiliary variables that might help predict the income variable. Multiple imputation was also conducted in RStudio using the MICE package and yielded similar results to SPSS.

Statistical Analysis

To describe the data and the study population, univariate and bivariate analyses were performed by using chi-square tests and independent sample t-tests for categorical and continuous variables, respectively. Our three main independent variables were considered in the analyses as categorical variables (quartiles) and continuous variables (scales). Logistic regression models were used to determine the unadjusted and adjusted odds ratios (OR) of prenatal cannabis use with 95% confidence intervals (CI) for each of the neighborhood quartile variables separately. The adjusted models accounted for all the covariates listed above.

We tested the mediation effect of stress levels on the association between perceived neighborhood scales and prenatal cannabis use by using multivariate regression-based path analysis. For each independent variable, three equations were calculated to estimate the indirect/mediation effect. In detail, let $X$ denote the independent variable (perceived neighborhood scale), $Y$ denote the dependent variable (prenatal cannabis use), and $M$ denote the mediator (stress level). First, we tested the total effect of the perceived neighborhood scale on prenatal cannabis use using the model $g\left(E,\left(Y\right)\right)=i_1+cX+e_1$. Second, we examined the association between perceived neighborhood scale and stress level by $g\left(E,\left(M\right)\right)=i_2+aX+e_2$. Last, we simultaneously tested the effects of perceived neighborhood scale and stress level on prenatal cannabis use using the model $g\left(E,\left(Y\right)\right)=i_3+c^{\prime }X+bM+e_{3}3$. Here, $i_1$, $i_2$, and $i_3$ represent the intercepts of each model, and $g\left(\bullet \right)$ is the link function, which is identity link for continuous outcome and logit link for dichotomous outcome. Each regression model was adjusted for all the covariates. To test for the significance of the indirect effect (a × b), the Sobel test (Aroian version) was calculated using the Preacher and Leonardelli online interactive tool [23].

All parameter estimates with two-sided p < 0.05 were considered significant. The analyses were conducted using SPSS (version 28.0) and RStudio (version 2023.06.1 Build 524).

Results

Out of the 1409 women in the study sample, 1369 (~ 97.2%) had complete information on cannabis use during pregnancy. No significant differences in the socio-demographic characteristics as well as the perceived neighborhood variables were present between women with cannabis use information and those who had missing information. Given missingness is less than 5%, list-wise deletion was used for missing cannabis information.

A total of 151 women (11.0%) self-reported prenatal cannabis use. The prevalence of prenatal cannabis use was lower than alcohol drinking (17.2%) in the LIFE population but was as common as prenatal tobacco smoking (11.0%). Figure 1 illustrates the concurrence of these substances during pregnancy. Around half of the women who used cannabis during pregnancy did so along with tobacco smoking, alcohol drinking, or both.

Fig. 1.

Concurrence of prenatal cannabis use with other substance uses during pregnancy in the LIFE population

Table 1 summarizes the general characteristics of women using prenatal cannabis in the study population. Young age, single/cohabitating status, lower annual household income, lower years of education, and lower social support were associated with increased prenatal cannabis use. Descriptive analysis of the perceived neighborhood scales and prenatal cannabis use are presented in Table 2. Comparing the means with a t-test, the means for all the neighborhood scales were significantly different among women with and without prenatal cannabis use. Women who reported cannabis use had worse neighborhood scores than their counterparts.

Table 1.

Descriptive characteristics of prenatal cannabis use among women in the LIFE study

Variables	Total sample	Prenatal cannabis use
	N	N (%)	OR (95% CI)	p-value
Age				0.020
< 20 years	115	19 (16.5%)	2.20 (1.22–3.97)
20–29 years	782	93 (11.9%)	1.50 (1.01–2.22)
≥ 30 years	472	39 (8.3%)	1
Marital status				< 0.001
Married	385	21 (5.5%)	1
Single/cohabitating	980	130 (13.3%)	2.65 (1.65–4.27)
Income				< 0.001
< $35,000	644	95 (14.8%)	3.03 (1.99–4.62)
≥ $35,000	574	31 (5.4%)	1
Education years				< 0.001
≤ 12 years	435	68 (15.6%)	1.89 (1.34–2.66)
> 12 years	929	83 (8.9%)	1
General social support scale (median)				0.005
≤ 50	650	88 (13.5%)	1.64 (1.16–2.32)
> 50	700	61 (8.7%)	1

N number, OR odds ratio, CI confidence interval

Table 2.

Association between perceived neighborhood scales and prenatal cannabis use

Scale	Range of scores*	Mean (SD)		p-value
		Non-cannabis users (N = 1,218)	Cannabis users (N = 151)
Social cohesion and trust scale	7–35	24.3 (4.8)	23.0 (4.5)	0.002
Social disorder	8–24	11.5 (4.5)	13.0 (5.1)	0.001
Danger and safety	6–30	21.4 (5.5)	19.8 (5.3)	0.000

SD standard deviation

*Higher scores reflect a better perceived environment except for social disorder (higher values reflect more disorder)

Table 3 presents the unadjusted and adjusted associations between prenatal cannabis use and the three perceived neighborhood scales. At the bivariate level, the odds of cannabis use significantly increased among the lowest quartiles of all the neighborhood scales suggesting higher cannabis use among women who had the worst perceptions about their neighborhoods. The association between neighborhood scales and cannabis use remained significant after adjusting for age, marital status, imputed income, years of education, and general social support scale. Women with the lowest perception about the safety of their neighborhood were about two times more likely to use cannabis during pregnancy. Similarly, the worst perceptions about the neighborhood’s social disorder and social cohesion environment were significantly associated with increased prenatal cannabis use (OR = 1.83, 95% CI: 1.15–2.91 and OR = 1.77, 95% CI: 1.04–3.03, respectively).

Table 3.

Unadjusted and adjusted associations between perceived neighborhood scales and prenatal cannabis use

	Prenatal Cannabis Use
	Unadjusted				Adjusted*
	Total sample (N)	N (%)	OR	95% CI	Total sample (N)a	OR	95% CI
Social cohesion and trust scale					1,271
Q1	366	56 (15.3%)	2.15	1.30–3.57		1.77	1.04–3.03
Q2	291	28 (9.6%)	1.27	0.72–2.24		1.16	0.64–2.11
Q3	330	38 (11.5%)	1.55	0.91–2.65		1.47	0.84–2.58
Q4	310	24 (7.7%)	1			1
Social disorder scaleb					1,309
Q1	506	39 (7.7%)	1	1		1
Q2	159	17 (10.7%)	1.43	0.79–2.61		1.53	0.83–2.83
Q3	355	43 (12.1%)	1.65	1.05–2.61		1.53	0.96–2.45
Q4	316	50 (15.8%)	2.25	1.44–3.51		1.83	1.15–2.91
Danger and safety scale					1,315
Q1	334	50 (15.0%)	2.35	1.40–3.95		1.93	1.12–3.31
Q2	378	55 (14.6%)	2.27	1.36–3.79		2.05	1.21–3.48
Q3	300	22 (7.3%)	1.06	0.58–1.94		1.03	0.56–1.92
Q4	330	23 (7.0%)	1	1		1

N number, OR odds ratio, CI confidence interval, Q1 first quartile, Q2 second quartile, Q3 third quartile, Q4 fourth quartile

*Adjusted for age, marital status, imputed annual income, years of education, and general social support scale

^aFinal sample size of the adjusted models

^bHigher values reflect more disorder

Figures 2, 3, and 4 represent the results of the mediation effects of stress levels on the association of each of the three perceived neighborhood scales with prenatal cannabis use while taking into consideration the confounding effect of all the covariates. Evidence of mediation was present between the association of perceived levels of safety and danger with prenatal cannabis use. No evidence of mediation was present with the other two perceived neighborhood scales. The results suggest that with worse perception of safety and danger, pregnant women experience higher levels of stress which is associated with increased use of cannabis. The proportion of the effect of the perceived safety and danger scale on prenatal cannabis use that goes through stress levels is 13.75% (indirect effect/total effect).

Fig. 2.

Mediation analysis results for the effect of stress levels as a potential mediator on the association between perceived neighborhood social cohesion and trust scale and prenatal cannabis use

Fig. 3.

Mediation analysis results for the effect of stress levels as a potential mediator on the association between perceived neighborhood social disorder scale and prenatal cannabis use

Fig. 4.

Mediation analysis results for the effect of stress levels as a potential mediator on the association between perceived neighborhood danger and safety scale and prenatal cannabis use

Discussion

To our knowledge, this is the first study that examined the association between subjective neighborhood context and cannabis use among pregnant women and explored the mediating effect of stress levels. Our results suggest that all three measures of perceived neighborhood context including social disorder, social cohesion and trust, and danger and safety, were significantly associated with prenatal cannabis use. Worse perceptions of neighborhood environment increased the odds of prenatal cannabis use even after adjustment of age, income, marital status, education level, and social support. Moreover, stress levels mediated the association between perceived neighborhood danger and safety scale and prenatal cannabis use. In other words, women with negative perceptions about their neighborhood safety experienced higher stress levels which was associated with increased use of cannabis during pregnancy.

Study findings are consistent with the literature on neighborhood social cohesion and cannabis use among non-pregnant populations and with the literature on other substance use among pregnant populations. Similar to our study, living in neighborhoods with no social cohesion was associated with 48% increased odds of tobacco smoking among pregnant women in Brazil (OR = 1.48, 95 CI%: 1.07–2.05) [16]. Patterson et al. [24] also demonstrated that higher social cohesion measured at the ecological level (OR = 0.85, 95% CI: 0.74–0.98) and at the individual level (OR = 0.96, 95% CI: 0.92–0.99) was associated with lower likelihoods of smoking among adults living in Minnesota. Likewise, results from a California study on adults living with children revealed a reduction in the odds of being a current smoker (AOR = 0.92; 95% CI: 0.85–0.99) with every increase in one unit in the social cohesion scale [25].

Additionally, neighborhood social disorder was found to be positively associated with prenatal cannabis use in our study. The literature on neighborhood social disorder and other substance use does not include pregnant women but is focused on the same racial group as the LIFE study. Among African American young women, Floyd [9] revealed a three-fold increase in the odds of heavy marijuana in the highest quartile of neighborhood disorder as compared to the lowest quartile (AOR = 3.3; 95% CI: 1.06–10.29). Similarly, negative perceptions of neighborhood disorder (AOR = 3.20, 95%CI: 1.46–7.01) were associated with an increased risk of transitioning from non-cannabis use to frequent or problematic use among African American high school students [10]. Furr-Holden et al. [26] also revealed a positive direct effect between objective measures of neighborhood physical disorder and marijuana use among young African American adults.

In the present study, the lowest two quartile of neighborhood danger and safety scale were associated with increased prenatal cannabis use. Consistent with our findings, Patterson et al. [17] reported a positive association between self-reported neighborhood violence and tobacco use among urban (92% Black) pregnant women (AOR: 1.38, 95% CI: 1.07–1.60). Also, pregnant women living in self-reported violent neighborhoods in Brazil reported 42% increased odds of smoking than their counterparts (OR = 1.42, 95% CI: 1.02–1.98) [16]. Another study among African American high school students found that community violence exposure (AOR = 4.54, 95% CI: 1.94–10.60) was correlated with an increased risk of transitioning from no cannabis use to frequent marijuana use [10].

The association between neighborhood environment and cannabis use may be influenced by the interplay of several environmental and social factors [9, 26, 27]. Social support in a neighborhood may lead to decreased substance use by increasing self-efficacy and promoting adherence to neighborhood norms [25]. Conversely, neighborhoods with fewer resources, which are also more likely to be inhabited by minority community members due to the impacts of structural racism, may not be able to fund the creation of public spaces that foster social connection, such as schools, parks, or libraries [28]. Furthermore, the absence of community and health resources that serve to educate and discourage substance use can also contribute to more substance availability and social acceptance [27]. It has also been shown that cannabis retailers are usually concentrated in historically disadvantaged neighborhoods which subjects its residents to disproportionate marketing and increased substance use [29]. Stressful neighborhood environments are further perpetuated by generational cycles of early substance use, neighborhood disorder, and structural racism. Communities that lack monetary and temporal resources are less likely to be able to lobby or fundraise for neighborhood improvements or monitor youth activity in the community [28]. Children born into these stressful, low-resource environments normalize substance use [29] and are more likely to use substances during adolescence [28]. Racial inequities and social determinants of health further compound these issues and may exacerbate stressors. High rates of neighborhood unemployment [30] and abandoned homes [26] are linked to higher rates of young adult cannabis use. Racial inequities that accumulate over generations also increase stress reactivity, impair recovery from stressors, prevent the accumulation of social and material capital needed to respond to threats, and increase morbidity and mortality further eroding the loss of social and familial structures that may help mitigate the impact of these stressors [27]. Furthermore, residents of disadvantaged communities have few avenues to escape the conditions they perceive as disordered and dangerous which leads to a chronic stress response [9, 11, 27].

Cannabis use in prenatal populations may also be a coping mechanism to relieve negative emotions and deal with chronic exposure to dangerous neighborhoods [9, 28]. This finding aligns with previous research that suggest an association between stress, anxiety, depression, and trauma on cannabis utilization among pregnant women [31, 32]. Although women report using cannabis to treat nausea, vomiting, and morning sickness during pregnancy, the highest proportion of pregnant women report using cannabis to relieve anxiety and stress [8]. They believe it is a natural remedy and might be less harmful than the use of prescription drugs or other substances [33].

Given the fact that physiological factors can be an underlying mechanism that links neighborhood environment and cannabis use, we investigated the mediating role of stress. We found that stress partially mediated the correlation between prenatal cannabis use in women who had worse perceptions on the danger and safety scale. Along the same lines, Reboussin et al. [10] found that depressed mood partially mediated the association between transitioning to cannabis use and perceived neighborhood disorder, drug activity and sales in the neighborhood, and community violence exposure among African American high school students. Similarly, the relationship between social cohesion and substance use disorders in Swiss men was partially mediated by depression [34]. In our study, on the other hand, no evidence of mediation was detected with the other two scales, perceived social disorder and social cohesion. While social cohesion and neighborhood order are important factors that may help explain the environment of the neighborhood, they might not be as fundamental to our existence and well-being as perceived safety.

The findings of the study should be interpreted within the scope of its limitations. The LIFE study was retrospective in design and does not allow inferences about causality. Participants might have also lived in different neighborhoods prior to the current pregnancy and might have experienced different stressors that were not captured. However, we found that few women had moved in the 2 years prior and no evidence that those neighborhoods were better or worse. Moreover, cannabis use was self-reported and not ascertained with toxicology measures. Rates, therefore, might have been underestimated due to social desirability bias and fear of legal consequences especially that the LIFE study (2009–2011) was conducted before the adult legalization of recreational cannabis in Michigan. However, the study rates are higher than the 4.4% (95% CI: 3.6–5.5) national rate reported in 2010–2013 among pregnant women [35]. In addition, the measurement of cannabis was limited to ever use. More information about the frequency, timing by trimester, reason, and mode of use should be collected in future studies. Finally, since recruitment was specific to one hospital site, the study findings should be replicated in other populations to ensure generalizability of the results. Yet, we note that the LIFE study was comprised of African American women who had a high rate of adverse birth outcomes and is therefore an important population to study in this regard [18].

Conclusions

This research study builds on previous existing literature on neighborhood context and substance use but specifically examines cannabis use, an understudied outcome, among the pregnant population and specifically among African American women, who are more likely to reside in disadvantaged neighborhoods. It is the first of its kind to examine the relationship between subjective neighborhood context and prenatal cannabis use and explore the mediating role of stress. The data suggest that negative perceptions about neighborhoods are related to increased use of cannabis in the prenatal period and that this association operates, to a certain extent, through stress. Future prospective studies should further examine the causal associations of social and physical environmental factors as well as individual factors of cannabis use to potentially reduce adverse outcomes and health disparities that result from its use during pregnancy. Research should also focus on other factors that may be concomitantly mediating the association between prenatal cannabis use and neighborhood environment. Finally, multi-level intervention studies promoting social cohesion and trust and improving physical environment and resources are warranted to help reduce cannabis availability and acceptability with the overall goal of reducing cannabis use among pregnant populations.

Abbreviations

AOR

Adjusted odds ratio

CI

Confidence interval

LIFE

Life-Course Influences on Fetal Environments Study

MCMC

Markov chain Monte Carlo

MOS

Medical Outcomes Study

OR

Odds ratio

Q1

First quartile

Q2

Second quartile

Q3

Third quartile

Q4

Fourth quartile

SD

Standard deviation

SPSS

Statistical Package for the Social Sciences

Appendix 1

Table 4.

Perceived neighborhood scales

Scale	Measurement items	Scoring	Range of scores	Cronbach’s α
Social cohesion and trust	1. I live in a close-knit neighborhood 2. People in my neighborhood are willing to help their neighbors 3. People in my neighborhood generally don’t get along with each other 4. People in my neighborhood do not share the same values 5. People in my neighborhood can be trusted 6. People in this neighborhood are willing to help women if they are in trouble 7. People in this neighborhood are willing to help pregnant women	5-point Likert scale: - Strongly agree - Agree - Neither agree nor disagree - Disagree - Strongly disagree	7–35	0.84
Social disorder	1. How much of a problem is litter, broken glass, or trash on the sidewalks and streets? 2. How much of a problem is graffiti on buildings and walls? 3. How much of a problem are vacant or deserted houses or storefronts? 4. How much of a problem is drinking in public? 5. How much of a problem is people selling or using drugs? 6. How much of a problem is groups of teenagers or adults hanging out in the neighborhood and causing trouble? 7. How much of a problem is noise in the neighborhood? 8. How much of a problem is yelling and fighting?	3-point Likert scale: - A big problem - Somewhat of a problem - Not a problem	8–24	0.93
Danger and safety	1. Many people in your neighborhood are afraid to go out at night 2. There are areas of this neighborhood where everyone knows “trouble” is expected 3. You’re taking a big chance if you walk in this neighborhood alone after dark 4. I feel safe walking in my neighborhood 5. Violence is a problem in my neighborhood 6. I feel very safe from crime in my neighborhood	5-point Likert scale: - Strongly agree - Agree - Neither agree nor disagree - Disagree - Strongly disagree	6–30	0.90

Appendix 2

Table 5.

Stress and social support scales

Scale	Measurement items	Scoring	Range of scores	Cronbach’s α
Cohen’s perceived stress scale	In the last month, how often have you: 1. been upset because of something that happened unexpectedly? 2. felt that you were in control of your life? 3. felt nervous and stressed out? 4. dealt successfully with daily hassles? 5. felt that you were able to successfully handle the important changes occurring in your life? 6. felt able to handle your personal problems? 7. felt that things were going your way? 8. found that you could not deal with all the things that you had to do? 9. been able to control hassles in your life? 10. felt that you were on top of things? 11. gotten angry because of things that happened that were outside of your control? 12. found yourself thinking about things you need to do? 13. been able to control the way you spend your time? 14. felt that you had so many problems that you could not deal with them?	5-point Likert scale: - Never - Almost never - Sometimes - Fairly often - Very often	14–70	0.87
General social support	How often is each of the following kinds of support available to you if you need it? 1. Someone to help you if you were confined to bed 2. Someone you can count on to listen when you need to talk 3. Someone to give you good advice about a crisis 4. Someone to take you to the doctor if you needed it 5. Someone who shows you love and affection 6. Someone to have a good time with 7. Someone to give you information to help you understand a situation 8. Someone to confide in or talk to about yourself or your problems 9. Someone who hugs you 10. Someone to get together with for relaxation 11. Someone to prepare your meals if you were unable to do it yourself	5-point Likert scale: - None of the time - A little of the time - Some of the time - Most of the time - All of the time	11–55	0.90

Data Availability

Data will be made available on request.