Neighborhood Poverty Histories and Physical Activity Among Children: Findings From the Geographic Research on Wellbeing (GROW) Study

Abstract

Purpose:

Examine the association between neighborhood poverty histories and physical activity, and the moderation effect of family poverty and the mediation effect of built environments in such association.

Design:

A cross-sectional study of the Geographic Research on Wellbeing (2012-2013), a follow-up survey of statewide-representative Maternal and Infant Health Assessment (2003-2007).

Setting:

California.

Participants:

A total of 2493 women with children.

Measures:

Outcome measures are (1) daily leisure physical activity and (2) days of physical activity among children. An independent variable is poverty histories of census tract where the child resided. Mediators were mother-perceived social cohesion, mother-perceived neighborhood safety, distance to the closest park, and park acreage within 0.5 miles from the home. A moderator is family poverty.

Analysis:

Weighted regression analysis.

Results:

Family poverty was a significant moderator (P < .05); poor children in neighborhoods with long-term moderate poverty, long-term high poverty, or increasing poverty (vs long-term low poverty) had greater odds of daily physical activity (odds ratio [OR] = 1.46, 1.50, 1.66, respectively). Nonpoor children in neighborhoods with long-term moderate poverty or increasing poverty (vs long-term low poverty) were associated with decreased odds of daily physical activity (OR = 0.61 and 0.44, respectively). Mediation associations were insignificant (P > .05).

Conclusion:

The combined effect of family financial strains and neighborhood economic resources might prevent poor children in neighborhoods with long-term low poverty and decreasing poverty from utilizing health-promoting resources in neighborhoods.

Keywords

neighborhood poverty histories physical activity neighborhood impacts

Physical activity is a health behavior which is strongly related to child obesity.^1

-4 Obese children are at increased risk of being overweight or obese as adults^5
-7 and for negative health outcomes in adulthood.^8

-14 Despite the importance of physical activity in preventing obesity, only 42% of US children aged 6 to 11 years engaged in the recommended amount of moderate-to-vigorous physical activity.^15,16 Over time, children from highly educated, high-income families have presented more physical activity than they did before, whereas children from low educated, low-income families have presented a small improvement or even less physical activity.¹⁷ Researchers have found individual-level factors of child physical activity.^18
-20 However, the individual-level factors could not explain all of the total variance of child physical activity. To address this remaining variance, researchers have identified the role of broader contextual factors in child physical activity, including neighborhood characteristics.^21,22

Theories and studies have investigated associations between neighborhood economic status and child physical activity.^23

-26 Theoretically, 2 mechanisms have been identified to explain the influence of neighborhood economic status on child health behaviors. According to a neighborhood resource model, living in neighborhoods with high economic status (hereafter, nonpoor neighborhoods) may promote health behaviors among children from middle- or high-income families (hereafter, nonpoor children).^27

-30 A neighborhood resource model posits that health-promoting resources, such as favorable physical activity resources, little physical disorder, and trust among neighbors,³¹ may mediate the positive effect of living in nonpoor neighborhoods on health behaviors.^27

-30,32

In contrast, the relative deprivation model posits that living in nonpoor neighborhoods may not be beneficial to children from low-income families (hereafter, poor children). In non-poor neighborhoods, poor children may feel psychologically distressed due to the relatively low economic status of their family to their neighbors’ unlike poor children in neighborhoods with low economic status (hereafter, poor neighborhoods) do not.²⁷ Because distress from relative deprivation may negatively influence health behaviors among poor children and offset the benefits of health-promoting resources in non-poor neighborhoods.^33

-36

The relative deprivation model can be especially notable for child physical activity because most types of child physical activity consist of peer-based activities and children generally socialize with those from the similar family economic status.³⁷ In addition, family economic status determines types of physical activity in which children engage.³⁸ Thus, poor children in non-poor neighborhoods may be less likely than those in poor neighborhoods to be included in a peer group in the neighborhood and to engage in physical activities.³⁹

Empirical literature has examined the effect of living in poor neighborhoods on child physical activity.^{23,26,40

-45} Past studies showed an adverse^23,26,40 or no^41

-45 association between living in poor neighborhoods and physical activity. Aside from differences in samples and methods, there are additional possible reasons for inconsistencies, which have prevented our understanding of neighborhood impacts on physical activity and have limited the utility in policy making and clinical practice: prior literature has not considered (1) histories of neighborhood poverty (vs neighborhood poverty at one point-in-time) and (2) the disparate effect of neighborhood poverty on child physical activity by family poverty status.

First, previous research has measured neighborhood poverty at only one point-in-time (e.g., neighborhood poverty rate in 2000). Cumulative dynamic social, economic, and political forces develop and change a neighborhood’s poverty status over decades,⁴⁶ especially since the 1970s as gentrification and urban development projects have occurred.^47
-49 Using one point-in-time measure fails to consider heterogeneous histories among neighborhoods that may have an equivalent poverty status at one point-in-time, but different histories.⁵⁰ For example, the high-poverty category of neighborhood poverty measure in the American Community Survey 2006 to 2010 (i.e., one point-in-time measure) could include neighborhood types of long-term high poverty and poverty increase when using a measure of neighborhood poverty histories between 1970 and 2006 to 2010.⁵¹

Further, a past study showed significant associations between neighborhood poverty histories and adult health outcomes when using a measure of neighborhood poverty histories but not when using one point-in-time measure.⁵¹ To date, whether neighborhood poverty histories affect child physical activity is unknown—as distinct from literature measuring residential mobility histories based on multiple point-in-time measures.⁵²

Second, although 2 different theoretical mechanisms have been identified for poor and non-poor children, the interaction effect of neighborhood poverty and family poverty on physical activity has been little empirically examined, an exception being Kim and Cubbin⁵³ which found different associations between neighborhood poverty and child physical activity by family poverty status. Examining the association between neighborhood poverty and physical activity separately for poor and non-poor children would aid understanding neighborhood impacts on child health.

The Present Study

The present study tests 2 theoretical mechanisms (i.e., neighborhood resource model and relative deprivation model) that highlight family poverty as a moderator of the associations between living in poor neighborhoods and individual health. We also compare associations of one point-in-time measure and a historical measure of neighborhood poverty with child physical activity. Although current scholarship in child health behaviors generally neglects historical changes of neighborhood poverty, our consideration of neighborhood poverty histories is informed by empirical literature examining neighborhood impacts on adult health.^50,51 Considering heterogeneous histories of neighborhoods over 40 years would provide a more rigorous measurement of neighborhood poverty.

In addition, conceptual models of neighborhood impacts on child health posit that neighborhood-level food and physical activity resources, social cohesion, and safety are potential mediators in the effect of living in poor neighborhoods on child health behaviors.^27,29 However, few studies have tested them empirically.^41,53 Given the limited empirical evidence, further exploration of these potential mediators would improve understanding of how neighborhood poverty affects child physical activity. Thus, we tested the mediating roles of neighborhood resources, safety, and social cohesion in such associations.

The following hypotheses guided the study: (1) living in historically high-poverty neighborhoods will be associated with less physical activity among non-poor children; (2) living in historically low-poverty neighborhoods will be associated with less physical activity among poor children; (3) neighborhood resources, safety, and social cohesion will mediate in such associations.

Methods

Data Design and Sample

Data were drawn from the Geographic Research on Wellbeing (GROW) study (2012-2013). The GROW is a follow-up survey of mothers in California who participated in the Maternal and Infant Health Assessment during 2003 to 2007, designed to examine neighborhood effects on health among women and their children. The GROW was collected from 3016 mothers. The GROW study was approved by the institutional review boards of the participating institutions.⁵⁴ Neighborhood poverty history data were drawn from the Neighborhood Change Database for years 1970 through 2010.

The analytic sample included children who lived in California at the time of GROW data collection (excluding 134 cases) and who lived with their mother who participated in GROW (excluding 39 cases). We deleted cases with missing data on an outcome variable (n = 132), neighborhood poverty histories (n = 77), covariates (n = 54), perceived social cohesion and safety (n = 72), and the duration of living in the current neighborhood (n = 2). We also excluded children who moved to their current neighborhood in the past month (n = 13), resulting in 2493 cases.

Measures

The dependent variables were (1) daily leisure physical activity and (2) days of physical activity among children. Both variables were based on the mother’s response on how often the child plays outdoors or does sports activities for at least an hour. For the first dependent variable, children whose mother reported participation in leisure physical activity every day were coded as yes (otherwise no). We use the original continuous variable of physical activity for the second dependent variable.

Neighborhood poverty histories were measured using proportions of families with income below the federal poverty level and linked to the census tract in which the child resided at GROW. First, census tracts were classified as low-,moderate-, and high-poverty by tertiles of poverty rates for each period (1970, 1980, 1990, 2000, and 2006-2010). Second, using poverty categories (low, moderate, high) for each period, census tracts were categorized into 5 types of poverty histories using the a priori method that Margerison-Zilko⁵¹ developed. The 5 types of neighborhood poverty history are as follows: (1) long-term low: all time periods were either low or a combination of low and moderate with no discernible pattern; (2) long-term moderate: 3 time periods or more were moderate with no discernible pattern; (3) long-term high: all time periods were either high or a combination of high and moderate with no discernible pattern; (4) poverty increase: Census tracts were low or moderate and then became and remained moderate or high; and (5) poverty decrease: Census tracts were moderate or high and then became and remained low or moderate.

Mediators were: (1) mothers’ perception of social cohesion; (2) mother’s perception of neighborhood safety; (3) distance to the closest park; and (4) park acreage within 0.5 miles from the child’s residence. For perceived social cohesion, mothers responded to a modified 5-item scale that Sampson et al³² developed based on 4 response options (strongly agree to strongly disagree). The 5 items were then averaged (range 1-4 points; α = .90). Perceived neighborhood safety consisted of the 3 questions about safety in daytime, in nighttime, and from crime based on 4 response options (strongly agree/very safe to strongly disagree/very unsafe). Responses to the 3 items were averaged (range 1-4 points; α = .77).

Family income was measured by mother’s report of the previous calendar year’s income from all sources and was categorized by percent of the federal poverty thresholds (≤200%, >200%). Missing income values were imputed using sociodemographic variables. Child’s age, gender, and race/ethnicity, and mother’s marital status and educational attainment were included as covariates.

Statistical Analysis

We tested for interactions between family poverty status and neighborhood poverty histories as correlates of each dependent variable by using PROC SURVEYLOGISTIC and PROC SURVEYREG. Where interaction terms had P values of less than .05, we stratified the sample by individual poverty status and conducted a series of logistic or linear regression models to estimate coefficients for daily leisure physical activity and days of physical activity: (model 1) unadjusted models; (model 2) a “sociodemographic” model; and (model 3) a “neighborhood poverty histories” model, in which neighborhood poverty histories were added to the sociodemographic model. We conducted sensitivity analyses in 2 ways. First, we repeated the regression analyses after controlling for the duration of living at the current address to test whether length of time in the home impacted the results. Second, we assessed inverse probability of treatment weighted Poisson regression models to account for possible bias of treatment selection. Finally, mediation tests were conducted according to Baron and Kenny’s guidelines if neighborhood poverty histories were significantly associated with a dependent variable. The association between each potential mediator and an independent variable was first tested in an unadjusted model. Then, each potential mediator was added to model 3 one at a time. All analyses accounted for the complex sample design and weighting by using SAS version 9.4. Formal multilevel modeling was not used because the data were not highly clustered (98% of tracts contained 3 or fewer sampled children).

Results

As shown in Table 1, 28% of children participated in daily physical activity, and children engage in physical activity 3.4 days a week on average. Figure 1 presents neighborhood characteristics by neighborhood-level and family-level poverty status. Neighborhoods with long-term high poverty tended to have a higher level of population density and greater accessibility to parks, but park were smaller than neighborhoods with other poverty histories. Furthermore, within each neighborhood poverty history category, with few exceptions, poor residents had greater access to—but smaller—parks and lower perceptions of social cohesion and safety than nonpoor residents.

Table 1.

Sample Characteristics, Geographic Research on Wellbeing Study, California, United States, 2012 to 2013, N = 2493.

Characteristics	Total
Total	100.0
Child’s age
4-5 years	22.1
6-7 years	37.8
8-10 years	40.1
Child’s gender
Boy	51.7
Girl	48.3
Child’s race/ethnicity
Asian/Pacific Islander	11.1
Black	4.9
Latinx	57.5
White	19.4
Multiracial	7.1
Children who are physically active every day	28.3
Number of days of physical activity (mean)	3.4
Mother’s marital status
Married or living together	84.1
Unmarried	15.8
Mother’s education
Less than high school	20.7
High school graduate/GED	45.2
Some college or above	34.0
Family income (% of federal poverty level)
≤100%	32.1
101%-200%	20.5
201%-300%	10.8
301%-400%	7.9
>400%	28.8
Neighborhood poverty history
Long-term low	19.0
Long-term moderate	22.5
Long-term high	17.6
Increase	27.8
Decrease	13.1

Abbreviation: GED, General Educational Development.

Figure 1.

Neighborhood characteristics by neighborhood-level and family-level poverty status, geographic research on wellbeing study, California, United States, 2012 to 2013, N = 2493.

Significant interactions existed between family poverty status and neighborhood poverty histories as correlates of both dependent variables at α .05 (results not shown). We accordingly stratified the sample by family poverty and conducted logistic and regression analyses. As shown in Table 2, poor children in neighborhoods with long-term moderate poverty, high poverty, or poverty increase had higher odds of daily physical activity than children in neighborhoods with long-term low poverty (OR = 1.46, 1.50, and 1.66, respectively). Non-poor children in neighborhoods with long-term moderate poverty or poverty increase had lower odds of engaging in daily physical activity than those in neighborhoods with long-term low poverty (OR = 0.61 and 0.44, respectively).

Table 2.

Associations Between Neighborhood Poverty History and Daily Physical Activity by Family Poverty Status, Geographic Research on Wellbeing Study, California, United States, 2012 to 2013, N = 2493.^a

Characteristic	Poor children						Nonpoor children
	Model 1		Model 2		Model 3		Model 1		Model 2		Model 3
	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Child’s age	0.88^b	0.81-0.97	0.88^b	0.81-0.97	0.88 ^b	0.80-0.96	1.00	0.91-1.11	0.99	0.89-1.10	0.98	0.89-1.09
Child’s gender
Boy	1.00		1.00		1.00		1.00		1.00		1.00
Girl	0.96	0.74-1.25	0.97	0.75-1.26	0.99	0.76-1.30	0.63 ^b	0.47-0.84	0.62 ^b	0.46-0.84	0.64 ^b	0.47-0.87
Child’s race/ethnicity
Asian/Pacific Islander	0.63	0.23-1.71	0.59	0.21-1.65	0.59	0.21-1.67	0.32^c	0.18-0.55	0.32^c	0.18-0.56	0.30^c	0.17-0.52
Black	0.56	0.30-1.05	0.51^d	0.26-0.99	0.47^d	0.24-0.94	0.41 ^b	0.23-0.75	0.39 ^b	0.21-0.73	0.35 ^b	0.18-0.68
Latinx	0.68	0.41-1.12	0.61	0.36-1.04	0.57^d	0.33-0.98	0.70^d	0.50-0.98	0.66^d	0.46-0.94	0.65^d	0.45-0.94
White	1.00		1.00		1.00		1.00		1.00		1.00
Multiracial	0.66	0.29-1.51	0.64	0.28-1.47	0.59	0.25-1.40	0.78	0.46-1.31	0.78	0.46-1.33	0.76	0.45-1.30
Mother’s marital status
Married or living together	0.85	0.63-1.15	0.80	0.58-1.11	0.81	0.58-1.12	0.95	0.57-1.57	1.06	0.63-1.78	1.04	0.61-1.76
Unmarried or separated	1.00		1.00		1.00		1.00		1.00		1.00
Mother’s education
Less than high school	1.00		1.00		1.00		1.00		1.00		1.00
High school graduate/GED	0.80	0.61-1.06	0.77	0.57-1.02	0.77	0.58-1.03	0.97	0.43-2.20	0.84	0.37-1.90	0.92	0.41-2.05
College graduate or ^b bove	0.78	0.47-1.27	0.71	0.42-1.20	0.72	0.43-1.22	0.84	0.38-1.84	0.68	0.30-1.53	0.76	0.34-1.71
Neighborhood poverty history
Long-term low	1.00				1.00		1.00				1.00
Long-term moderate	1.48^d	1.04-2.10			1.46^d	1.02-2.08	0.63^d	0.43-0.92			0.61 ^b	0.42-0.88
Long-term high	1.45^d	1.04-2.02			1.50^d	1.07-2.11	1.28	0.67-2.47			1.44	0.74-2.80
Increase	1.59	0.99-2.57			1.66^d	1.02-2.71	0.45 ^b	0.26-0.76			0.44 ^b	0.26-0.76
Decrease	1.36	0.74-2.53			1.43	0.76-2.68	0.88	0.46-1.69			0.77	0.39-1.51

Abbreviation: OR, odds ratio.

^a Boldface indicates statistical significance.

^b P < .05,

^c P < .001.

^d P < .01.

Table 3 shows the results of linear regression analysis to examine the association between neighborhood poverty histories and the number of days of physical activity. Poor children in neighborhoods with long-term high poverty or poverty increasing engaged in physical activity more frequently than those in neighborhoods with long-term low poverty (b = 0.22 and 0.38, respectively). Non-poor children in neighborhoods with long-term moderate poverty or poverty increase engaged in physical activity less frequently than those in neighborhoods with long-term low poverty (b = −0.16 and −0.28, respectively).

Table 3.

Associations Between Neighborhood Poverty History and Average Days of Physical Activity by Family Poverty Status, Geographic Research on Wellbeing Study, California, United States, 2012 to 2013, N = 2493.^a

Characteristic	Poor children						Nonpoor children
	Model 1		Model 2		Model 3		Model 1		Model 2		Model 3
	b	SE	b	SE	b	SE	b	SE	b	SE	b	SE
Child’s age	−0.09^b	0.03	−0.09^b	0.03	−0.09^b	0.03	0.03	0.03	0.02	0.02	0.02	0.02
Child’s gender
Boy (reference)
Girl	−0.12	0.08	−0.11	0.08	−0.10	0.08	−0.18^c	0.07	−0.17^c	0.07	−0.16 ^c	0.07
Child’s race/ethnicity
Asian/Pacific Islander	−0.44	0.31	−0.43	0.31	−0.44	0.31	−0.77^d	0.11	−0.77^d	0.11	−0.79^d	0.11
Black	−0.39	0.21	−0.46^c	0.21	−0.50^c	0.21	−0.41^b	0.14	−0.42^b	0.14	−0.48^d	0.14
Latinx	−0.24	0.16	−0.26	0.16	−0.30	0.16	−0.21^b	0.08	−0.19^c	0.08	−0.20^c	0.08
White (reference)
Multiracial	−0.08	0.23	−0.13	0.23	−0.15	0.23	−0.20	0.12	−0.20	0.12	−0.21	0.12
Mother’s marital status
Married or living together	−0.12	0.10	−0.14	0.10	−0.14	0.10	−0.11	0.11	−0.03	0.11	−0.03	0.11
Unmarried or separated (reference)
Mother’s education
Less than high school(reference)
High−school graduate/GED	0.05	0.09	0.03	0.09	0.04	0.09	−0.10	0.19	−0.13	0.18	−0.06	0.18
College graduate or above	−0.06	0.16	−0.08	0.17	−0.07	0.17	−0.07	0.18	−0.06	0.18	0.01	0.18
Neighborhood poverty history
Long-term low (reference)
Long-term moderate	0.17	0.11			0.14	0.11	−0.14	0.09			−0.16^c	0.08
Long-term high	0.20^c	0.10			0.22^c	0.10	0.25	0.17			0.31	0.19
Increase	0.37^c	0.14			0.38^b	0.15	−0.26^c	0.11			−0.28^b	0.11
Decrease	−0.01	0.21			−0.01	0.21	0.12	0.15			−0.01	0.14

Abbreviations: b, unstandardized coefficient; GED, General Educational Development.

^a Boldface indicates statistical significance.

^b P < .01.

^c P < .05.

^d P < .001.

When a point-in-time measure of neighborhood poverty in 2006 to 2010 was substituted for neighborhood poverty history in the model, there was no significant association between neighborhood poverty and the outcome variables (results not shown). Sensitivity analyses, controlling for duration of living at the current address (see Supplemental Appendix Table 1), or inverse probability of treatment-weighted Poisson regression models (see Supplemental Appendix Table 2), showed consistent results.

Finally, potential mediators were added to the multivariate model one at a time (see Table 4). For example, multivariate model in Table 4 is simply the ORs (1.43-1.66) of poor children’s physical activity associated with neighborhood poverty histories that were presented in Table 2, model 3. Adding a measure of distance to nearest park to that multivariate model, distance to nearest park was not associated with daily physical activity (OR = 0.89), and the ORs of daily physical activity associated with neighborhood poverty histories barely decreased or even increased. Similar findings were found for a continuous dependent variable, the number of days of physical activity (results not shown). In summary, there was no substantive evidence of mediation; no mediator was associated with physical activity, or the coefficient for physical activity associated with neighborhood poverty histories did not significantly change when the mediator was included.

Table 4.

Mediating Effects on the Pathway Between Neighborhood Poverty History and Daily Physical Activity, Geographic Research on Wellbeing Study, California, United States, 2012 to 2013, N = 2493.^a

	Mediator, adjusted model OR (95% CI)	Neighborhood poverty history (ref = long-term low poverty)
	Mediator, adjusted model OR (95% CI)	Long-term moderate OR (95% CI)	Long-term high OR (95% CI)	Poverty increase OR (95% CI)	Poverty decrease OR (95% CI)
Poor children
Multivariate model (from Table 2)		1.46^b (1.02-2.08)	1.50^b (1.07-2.11)	1.66^b (1.02-2.71)	1.43 (0.76-2.68)
Distance to nearest park (per 1000 meters)	0.89 (0.61-1.30)	1.47^b (1.03-2.10)	1.50^b (1.07-2.10)	1.67^b (1.03-2.72)	1.43 (0.77-2.66)
Park acreage within 0.5 miles (per 100 acres)	1.59 (0.90-2.80)	1.49^b (1.05-2.13)	1.52^b (1.08-2.14)	1.68 ^b (1.03-2.75)	1.44 (0.77-2.68)
Perceived social cohesion	1.15 (0.91-1.45)	1.46^b (1.02-2.08)	1.53^b (1.09-2.16)	1.68^b (1.03-2.73)	1.42 (0.76-2.65)
Perceived neighborhood safety	1.13 (0.91-1.41)	1.46^b (1.02-2.08)	1.57^c (1.11-2.22)	1.67^b (1.03-2.72)	1.47 (0.78-2.76)
Nonpoor children
Multivariate model (from Table 2)		0.61^c (0.42-0.88)	1.44 (0.74-2.80)	0.44^c (0.26-0.76)	0.77 (0.39-1.51)
Distance to nearest park (per 1000 meters)	0.93 (0.59-1.48)	0.62^b (0.42-0.90)	1.52 (0.79-2.95)	0.44^c (0.26-0.76)	0.77 (0.40-1.51)
Park acreage within 0.5 miles (per 100 acres)	0.95 (0.70-1.29)	0.61^b (0.42-0.89)	1.53 (0.79-2.96)	0.44^c (0.25-0.76)	0.77 (0.39-1.50)
Perceived social cohesion	1.76^d (1.29-2.41)	0.63^b (0.43-0.92)	1.75 (0.89-3.45)	0.44^c (0.25-0.76)	0.77 (0.40-1.50)
Perceived neighborhood safety	1.64^c (1.20-2.24)	0.65^b (0.45-0.96)	2.04^b (1.02-4.07)	0.46^c (0.26-0.79)	0.77 (0.39-1.52)

Abbreviation: OR, odds ratio.

^a Boldface indicates statistical significance.

^bP < .05.

^c P < .01.

^d P < .001.

Discussion

This study examined the effect of neighborhood poverty histories on child physical activity. While past studies have observed child physical activity to be associated with living in poor neighborhoods, the results have been inconsistent.^{23,26,40

-44} We suggest that these inconsistent findings may be because of not considering neighborhood poverty histories, the disparate effect of neighborhood poverty by family poverty status, different study populations and analytic approaches, or some combination of these reasons.

Importance of Neighborhood Poverty Histories

We can measure the historical process of neighborhood poverty. Since 1970s, geographic concentration of poverty and affluence has increased with gentrification.^47
-49 The historical process of neighborhood poverty changes could plausibly determine the development and renewal of neighborhood services, amenities, and social capital using their social, economic, and political power.²⁹ When neighborhood research uses a point-in-time measure of neighborhood poverty, historical economic differences could be obscured: In this study, for instance, the high-poverty category using the American Community Survey 2006-2010 (i.e., a point-in-time measure) consisted of both long-term high poverty (55%) and increasing poverty neighborhoods (45%; Supplemental Appendix Table 3). We found significant associations between neighborhood poverty histories and child physical activity. However, no relationship was found when a point-in-time measure of neighborhood poverty was used, which is consistent with past literature using neighborhood poverty histories.⁵¹

Role of Family Poverty in Neighborhood Impact on Physical Activity

Our findings highlight the importance of family poverty status in understanding the neighborhood-physical activity association. The effect of neighborhood poverty on children may theoretically differ by family poverty status. On the one hand, according to the relative deprivation model,^27,33,34 living in non-poor neighborhoods may not benefit poor children due to feeling relatively deprived and having fewer peers of similar economic status in the neighborhood. Further, “health-promoting” environments in nonpoor neighborhoods tend to be more concentrated where non-poor children reside rather than where poor children reside even in the same neighborhood.⁵³ The current study showed that poor children have lower access to parks, and their mothers perceive their neighborhoods as less safe and cohesive, than nonpoor children in those same neighborhood types (see Figure 1).

On the other hand, living in poor neighborhoods may offer more opportunities for poor children to engage in physical activity because physical activity which could be often done with peers is significantly affected by peer interactions and relationship.³⁹ Our findings that poor children in neighborhoods with long-term moderate or high poverty, or poverty increase (vs long-term low-poverty neighborhoods) were more likely to engage in physical activity supported that the relative deprivation model applies to poor children.

For non-poor children in poor neighborhoods, a neighborhood resource model is at play. Our findings show that non-poor children in neighborhoods characterized by long-term moderate poverty or poverty increase were less likely to engage in physical activity compared with long-term low-poverty neighborhoods. This is possibly because historically low-poverty neighborhoods have greater health-promoting resources, which positively influence physical activity.^27

-32

Limitations

This study has notable limitations. First, the dependent variables were measured using a single, mother-reported item, which is prone to recall bias and social desirability and did not capture duration and types of physical activity. Future research needs to use an objective measure (such as via an accelerometers). Third, mediation mechanisms were little elucidated in this study. Cross-sectional mediation approaches ignored time sequence and lagged effects of neighborhood characteristics on health behaviors,⁵² which may bias estimates of longitudinal parameters in causal relations. Longitudinal research should be conducted to test causal effects of neighborhood economic context on child physical activity and theoretically plausible mediators.

Conclusions

Despite these limitations, this study has several strengths. First, the sample was large and ethnically diverse. Second, to our knowledge, it is the first study to examine neighborhood poverty histories as correlate of child physical activity and the disparate association of neighborhood poverty with health behavior between poor and nonpoor children. Neighborhood poverty histories were estimated based on knowledge that neighborhood conditions have changed over decades and that neighborhood poverty histories shape neighborhood resources, social cohesion, and safety.

Our findings suggest practice implications for poor children and nonpoor children. To promote physical activity of poor children, public health professionals and organizations need to recognize a feeling of relative deprivation that poor children might have in neighborhoods with long-term low poverty or decreasing poverty. The feeling of relative deprivation might prevent poor children in those neighborhoods from utilizing healthy and high-quality resources in neighborhoods. Informational outreach activities for active lifestyles and increasing access to places for physical activity would help promote poor children in non-poor neighborhoods to use neighborhood resources and engage in physical activity. For non-poor children in poor neighborhoods, community development policies are required to enhance neighborhood resources such as safe parks, playgrounds, social cohesion, and neighborhood disorder. Special attention to shortage and unbalanced distribution of resources across neighborhoods would lead to reduce disparities in child physical activity by neighborhood poverty.

So What? What Is Already Known on this Topic?

What is already known on this topic?

Literature has examined the effect of neighborhood-level poverty on child physical activity, but typically measuring neighborhood poverty from a single point in time and not investigating the effect of intersection between family poverty and neighborhood poverty on physical activity.

What does this article add?

We extend past literature by examining associations between neighborhood-level poverty histories and physical activity, the moderation effect of family-level poverty in such associations, and the mediation effect of built and social environments in such associations.

What are the implications for health promotion practice or research?

When designing policies and programs of physical activity promotion, we need to consider a feeling of relative deprivation for poor children in non-poor neighborhoods and lack of neighborhood resources for nonpoor children in poor neighborhoods.

Supplemental Material

Supplemental Material, sj-pdf-1-ahp-10.1177_0890117120923948 - Neighborhood Poverty Histories and Physical Activity Among Children: Findings From the Geographic Research on Wellbeing (GROW) Study

Supplemental Material, sj-pdf-1-ahp-10.1177_0890117120923948 for Neighborhood Poverty Histories and Physical Activity Among Children: Findings From the Geographic Research on Wellbeing (GROW) Study by Yeonwoo Kim and Catherine Cubbin in American Journal of Health Promotion

Footnotes

Authors’ Note

This research was approved by the Institutional Review Boards (IRB) at the University of Texas at Austin, the University of California, San Francisco, and the California Department of Public Health.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: This manuscript is part of the first author’s dissertation.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Preparation of this manuscript was supported through a grant from the American Cancer Society (RSGT-11-010-01-CPPB) and a population center grant from the National Institute of Child Health and Human Development (R24 HD42849). The American Cancer Society and Population Research Center had no role in the design, analysis or writing of this article.

ORCID iD

Yeonwoo Kim

Supplemental Material

Supplemental material for this article is available online.

References

Berkey

Rockett

Field

, et al. Activity, dietary intake, and weight changes in a longitudinal study of preadolescent and adolescent boys and girls. Pediatrics. 2000;105(4):e56.

Janssen

Katzmarzyk

Boyce

, et al. Comparison of overweight and obesity prevalence in school-aged youth from 34 countries and their relationships with physical activity and dietary patterns. Obes Rev. 2005;6(2):123–132.

Kawabe

Murata

Shibata

, et al. Participation in school sports clubs and related effects on cardiovascular risk factors in young males. Hypertens Res. 2000;23(3):227–232.

Lazzer

Boirie

Bitar

, et al. Assessment of energy expenditure associated with physical activities in free-living obese and nonobese adolescents. Am J Clin Nutr. 2003;78(3):471–479.

Deshmukh-Taskar

Nicklas

Morales

Yang

Zakeri

Berenson

. Tracking of overweight status from childhood to young adulthood: The Bogalusa heart study. Eur J Clin Nutr. 2006;60(1):48–57.

Engeland

Bjørge

Tverdal

Søgaard

. Obesity in adolescence and adulthood and the risk of adult mortality. Epidemiology. 2004;15(1):79–85.

Magarey

Daniels

Boulton

Cockington

. Predicting obesity in early adulthood from childhood and parental obesity. Int J Obes. 2003;27(4):505–513.

Csábi

Török

Jeges

Molnár

. Presence of metabolic cardiovascular syndrome in obese children. Eur J Pediatr. 2000;159(1-2):91–94.

Franks

Hanson

Knowler

, et al. Childhood obesity, other cardiovascular risk factors, and premature death. N Engl J Med. 2010;362(6):485–493.

10.

Freedman

Dietz

Srinivasan

Berenson

. The relation of overweight to cardiovascular risk factors among children and adolescents: the Bogalusa heart study. Pediatrics. 1999;103(6):1175–1182.

11.

Freedman

Mei

Srinivasan

Berenson

Dietz

. Cardiovascular risk factors and excess adiposity among overweight children and adolescents: the Bogalusa heart study. J Pediatr. 2007;150(1):12–17.

12.

Neovius

Sundström

Rasmussen

. Combined effects of overweight and smoking in late adolescence on subsequent mortality: nationwide cohort study. BMJ. 2009;338:b496.

13.

Morrison

Friedman

Gray-McGuire

. Metabolic syndrome in childhood predicts adult cardiovascular disease 25 years later: the Princeton lipid research clinics follow-up study. Pediatrics. 2007;120(2):340–345.

14.

Baker

Olsen

Sørensen

. Childhood body-mass index and the risk of coronary heart disease in adulthood. N Engl J Med. 2007;357(23):2329–2337.

15.

Physical Activity Guidelines Advisory Committee. Physical Activity Guidelines Advisory Committee Report 2008. U.S. Department of Health and Human Services; 2008.

16.

Troiano

Berrigan

Dodd

, et al. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181–188.

17.

Frederick

Snellman

Putnam

. Increasing socioeconomic disparities in adolescent obesity. Proc Natl Acad Sci U S A. 2014;111(4):1338–1342.

18.

Sallis

Prochaska

Taylor

. A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc. 2000;32(5):963–975.

19.

Trost

Sallis

Pate

Patty

Wendell

Marsha

. Evaluating a model of parental influence on youth physical activity. Am J Prev Med. 2003;25(4):277–282.

20.

Stalsberg

Pedersen

. Effects of socioeconomic status on the physical activity in adolescents: a systematic review of the evidence. Scand J Med Sci Sports. 2010;20(3):368–383.

21.

Sallis

Cervero

Ascher

, et al. An ecological approach to creating active living communities. Annu Rev Public Health. 2006;27:297–322.

22.

Sallis

Glanz

. The role of built environments in physical activity, eating, and obesity in childhood. Future Child. 2006;16(1):89–108.

23.

Carroll-Scott

Gilstad-Hayden

Rosenthal

, et al. Disentangling neighborhood contextual associations with child body mass index, diet, and physical activity: the role of built, socioeconomic, and social environments. Soc Sci Med. 2013;95:106–114.

24.

De Meester

Van Dyck

De Bourdeaudhuij

, et al.

Active living neighborhoods: Is neighborhood walkability a key element for Belgian adolescents?

BMC Public Health. 2012;12(1):1–12.

25.

Kim

Liu

Colabianchi

Pate

. The effect of perceived and structural neighborhood conditions on adolescents’ physical activity and sedentary behaviors. Arch Pediatr Adolesc Med. 2010;164(10):935–942.

26.

Pabayo

Belsky

Gauvin

Curtis

. Do area characteristics predict change in moderate-to-vigorous physical activity from ages 11 to 15 years? Soc Sci Med. 2011;72(3):430–438.

27.

Jencks

Mayer

. The social consequences of growing up in a poor neighborhood. In: Lynn

McGeary

MGH

, eds. Inner-City Poverty in the United States. National Academy Press; 1990:111–186.

28.

Massey

. The prodigal paradigm returns: ecology comes back to sociology. In: Booth

Crouter

, eds. Does It Take a Village? Community Effects on Children, Adolescents, and Families. Lawrence Erlbaum Associates; 2001:41–48.

29.

Robert

SA.

Socioeconomic position and health: the independent contribution of community socioeconomic context 1. Annu Rev Sociol. 1999;25(1):489–516.

30.

Shaw

McKay

. Juvenile Delinquency and Urban Areas. University of Chicago Press; 1942.

31.

Franzini

Taylor

Elliott

, et al. Neighborhood characteristics favorable to outdoor physical activity: disparities by socioeconomic and racial/ethnic composition. Health Place. 2010;16(2):267–274.

32.

Sampson

Raudenbush

Earls

. Neighborhoods and violent crime: a multilevel study of collective efficacy. Science. 1997;277(5328):918–924.

33.

Kawachi

Kennedy

. Socioeconomic determinants of health: health and social cohesion: why care about income inequality? BMJ. 1997;314(7086):1037.

34.

Marmot

Wilkinson

. Psychosocial and material pathways in the relation between income and health: a response to Lynch et al. BMJ. 2001;322(7296):1233–1236.

35.

Fulkerson

Sherwood

Perry

Neumark-Sztainer

Story

. Depressive symptoms and adolescent eating and health behaviors: a multifaceted view in a population-based sample. Prev Med. 2004;38(6):865–875.

36.

Motl

Birnbaum

Kubik

Dishman

. Naturally occurring changes in physical activity are inversely related to depressive symptoms during early adolescence. Psychosom Med. 2004;66(3):336–342.

37.

Weinger

Economic status: middle class and poor children’s views. Child Soc. 2000;14(2):135–146.

38.

Kantomaa

Tammelin

Näyhä

Taanila

. Adolescents’ physical activity in relation to family income and parents’ education. Prev Med. 2007;44(5):410–415.

39.

Barkley

Salvy

Roemmich

. The effect of simulated ostracism on physical activity behavior in children. Pediatrics. 2012;129(3):x26.

40.

Boone-Heinonen

Evenson

Song

Gordon-Larsen

. Built and socioeconomic environments: patterning and associations with physical activity in US adolescents. Int J Behav Nutr Phys Act. 2010;7(1):45.

41.

Frost

SS.

Childhood Obesity in Context: Examining the Role of Ecological Factors Related to Childhood Obesity Among West Virginia’s Youth. West Virginia University; 2011.

42.

Laxer

Janssen

. The proportion of youths’ physical inactivity attributable to neighbourhood built environment features. Int J Health Geogr. 2013;12:31.

43.

Leung

Gregorich

Laraia

Kushi

Yen

. Measuring the neighborhood environment: associations with young girls’ energy intake and expenditure in a cross-sectional study. Int J Behav Nutr Phys Act. 2010;7(1):52.

44.

Lovasi

Jacobson

Quinn

, et al.

Is the environment near home and school associated with physical activity and adiposity of urban preschool children?

J Urban Health. 2011;88(6):1143–1157.

45.

Janssen

Boyce

Simpson

Pickett

. Influence of individual-and area-level measures of socioeconomic status on obesity, unhealthy eating, and physical inactivity in Canadian adolescents. Am J Clin Nutr. 2006;83(1):139–145.

46.

Park

Burgess

McKenzie

. The City. University of Chicago Press; 1925.

47.

Fischer

Stockmayer

Stiles

Hout

. Distinguishing the geographic levels and social dimensions of US metropolitan segregation, 1960–2000. Demography. 2004;41(1):37–59.

48.

Massey

Fischer

Dickens

Levy

. The geography of inequality in the United States, 1950-2000 [with comments]. Brookings-Wharton Papers on Urban Affairs. 2003;1–40.

49.

Papayanis

MA.

Sex and the revanchist city: zoning out pornography in New York. Environ Plan D. 2000;18(3):341–353.

50.

DP.

The dynamics of income and neighborhood context for population health: do long-term measures of socioeconomic status explain more of the black/white health disparity than single-point-in-time measures?

Soc Sci Med. 2009;68(8):1368–1375.

51.

Margerison-Zilko

Cubbin

Jun

, et al. Beyond the cross-sectional: neighborhood poverty histories and preterm birth. Am J Public Health. 2015;105(6):1174–1180.

52.

Alvarado

SE.

The Effect of Neighborhood Context on Obesity and Educational Achievement Among Youth. University of Wisconsin–Madison; 2011.

53.

Kim

Cubbin

. The role of neighborhood economic context on physical activity among children: evidence from the geographic research on wellbeing (GROW) study. Prev Med. 2017;101:149–155.

54.

Cubbin

. Survey methodology of the geographic research on wellbeing (GROW) study. BMC Res Notes. 2015;8(1):1.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB