It is firmly established that regular physical activity (PA) reduces the risk of many chronic diseases and increases longevity (1). A joint statement by the Centers for Disease Control and Prevention (CDC) and American College of Sports Medicine concluded that moderate-intensity PA performed for 30 minutes on most days of the week will confer significant physical and mental health benefits (2). Despite the positive relationship between PA and several health outcomes, recent national studies indicate that more than 50% of the U.S. adult population does not achieve recommended PA levels and that white adults have a higher rate of regular PA than African American adults (3). The disparity in rates of PA may then contribute to the health disparities that exist between white and African American adults. This idea has led to an increased interest in gaining a greater understanding about the determinants and mediating factors of PA behaviors, including among racial and ethnic populations (4-6).

The ecological model for health promotion emphasizes multiple levels of influence upon individual behavior — intrapersonal, interpersonal and social, organizational, institutional, community, and policy (7,8). Expanded views of this model include environmental attributes that may play an important role in shaping health behaviors such as PA (9). Physical environmental attributes such as traffic, distance, sidewalks, and aesthetics are examples of factors that may influence PA. Social environmental attributes of interest include trust in neighbors, community norms of PA, and social networks. To gain a better understanding in these emerging areas of interest, research has increasingly emphasized the evaluation of individuals' perceptions of social and physical environmental attributes that may support or hinder their PA behavior. There is also speculation that perceptions of different sets of social and physical environmental attributes may influence the PA behavior of different individuals, including people of varying sex, age, race, and ethnicity (10). 

A number of investigators have conducted qualitative studies (i.e., focus groups) to further understanding of perceptions about social and safety-related environmental supports for PA among racial and ethnic populations (11-17). These studies consistently report safety-related environmental features to be influential on PA, including concerns about unsafe facilities, stray dogs, crime, and motorized traffic. In addition, a common theme expressed by people in these studies pertains to social environmental factors: community or neighborhood social cohesion, group participation, and support from family, friends, and neighbors.

Research reporting on the associations between social and safety-related environmental supports and overall PA is rapidly accumulating (18-21). However, relatively few studies have documented the association between such supports and walking (22-27) and even fewer have included more than one racial or ethnic population (22,28-30). The purpose of the present study was to examine the relationship between race and perceptions of social and safety-related environmental supports for recommended levels of PA and walking in white and African American adult residents of a rural county in the southeastern United States.

Of particular interest are the findings from this study indicating that, after adjusting for age, education, and sex of the respondents, no differences in perceptions of social or safety-related environmental factors were found between African American adults who reported meeting the recommended level of PA or walking and African American adults who reported not meeting the recommendation. The lack of association is surprising because the factors included in this study — such as safety factors related to traffic and crime — have been previously mentioned in several focus groups by African American adults as influences on PA level (11-17). It has been suggested, however, that small group discussions such as focus groups may elicit interaction and more thought about a topic among participants than the objective-response choices in a quantitative survey (33). Inherent differences may exist between qualitative and quantitative research techniques; investigators need to consider these differences when constructing quantitative data-collection instruments based on findings from qualitative studies and when comparing results from investigations using these methods (29).

Previous cross-sectional studies have substantiated an association between perceptions of safety related to traffic and crime and other environmental factors (e.g., streetlight quality, unattended dogs, trust in neighbors) and PA and walking in African American and other adult populations, including people living in rural locations (23,28,30). However, some studies have failed to observe such associations (25). For example, Eyler et al (29) found few similarities in social and environmental correlates of PA within a sample of women from the same racial and ethnic group. An Australian study also reported that adults living in high-walkable and low-walkable neighborhoods did not differ in their perceptions of safety related to traffic or crime (34). The lack of consistent results across studies and within groups indicates further research in this area is needed to better understand the environmental factors associated with PA and walking in people of varying races, ethnicities, and geographic locations. Furthermore, to determine the relative importance to PA and walking, the strength of these associations needs to be clarified and contrasted with other variables (e.g., self-efficacy, perceived health, anticipated benefits) that are consistently related to PA and walking (4-6,24,27).

White adults who perceived their neighbors as physically active were more likely to report meeting recommendations for PA and walking than white adults not perceiving their neighbors as physically active. In addition, white adults who perceived less crime in their neighborhood were more likely to report meeting recommendations for walking than those who perceived more crime in the neighborhood. It is not clear why perceptions of these two social and safety-related environmental factors were more strongly associated with PA and walking in white rather than African American adults. It has been postulated that differences may exist between how African American and white adults perceive features of the social and safety environments in their neighborhoods and this may potentially account for such results (10,22,31). For instance, Wilson et al (22) discovered that objective crime data did not support the reported differences in perceived crime between adults categorized as low and high income. We have also found that African American adults are more accurately aware of the presence of crime than white adults (D. Wilson, unpublished data, 2005). In addition, Boslaugh et al (10) reported that African American adults perceived their neighborhoods as less safe and less pleasant for PA than did white adults regardless of the racial composition of the neighborhood. Future studies will be required to determine whether perceived or objectively measured environmental factors are most influential on PA and walking in specific racial and ethnic groups, as well as what factors mediate or moderate their influence on PA-related behaviors.

Another possible explanation may be that race and ethnicity are simply serving as proxies for differing levels and types of PA among African American and white adults. A large proportion of white adults report engaging in structured and planned PA whereas African American adults, particularly women, equate PA with being busy at home or work or caring for children (12,13,35). It has been suggested that active people may be more aware of their environment; this explains the stronger association between perceptions of environmental attributes and PA found among physically active persons (25), but this association would likely hold true only for people who are physically active outdoors. One could then speculate that white adults engage more often in outdoor PA than African American adults, thus contributing to differing perceptions of environmental attributes. It has also been reported that perceptions of environmental attributes differ across levels of PA (31,36) and walking (36). This should not have been a factor in the current study because there were no differences between proportions of white and African American adults meeting PA and walking recommendations. Age and sex have also been shown to influence the association between perceptions of environmental attributes and walking (36), but these variables, along with level of education, were controlled for in our analyses. However, African American adults may accumulate daily walking differently than white adults (i.e., at work or home rather than during leisure time) (12,35). This difference is potentially important because different environmental attributes have been associated with walking for different purposes (e.g., exercise, pleasure) (36) as well as PAs of differing intensity (27). Therefore, variations in preferences and purposes for PA and types of PA may create varying perceptions of social and safety-related environmental factors among people from different racial and ethnic groups, even if they live in the same neighborhood. Further studies exploring the precise meaning of PA among adults from different racial and ethnic groups and neighborhoods and studies using more objective and comprehensive measures of daily PA are needed to help clarify these issues.

One finding that seemed counterintuitive was that white adults who perceived moderate traffic in their neighborhood were about one half as likely to meet the recommendation for walking than white adults who perceived heavy traffic in their neighborhood. However, in agreement with the current results, other studies have also noted an unexpected association between perceptions of traffic safety and PA or walking (25,26,34). For example, Humpel et al (26) reported that Australian adults who perceived traffic as not being a problem were 55% less likely to walk in their neighborhood. While the direction of the association in the present study was not necessarily anticipated, the odds ratio results provide evidence that elements of traffic in neighborhoods may have varying degrees of influence on walking behavior in white adults. The relative strength of this influence may be questioned because there were no associations between perceived traffic level and PA among either white or African American adults or between perceived traffic level and walking among African American adults. Other studies (12,24,25,28,30) using qualitative and quantitative approaches have documented the potential impact of perceived traffic on PA; this environmental attribute deserves additional investigation.

Despite published studies (12,23,30,37,38) that cite inadequate streetlights, unattended dogs, and unsafe recreational facilities as deterrents, the current study found no association among African American or white adults between these factors and meeting PA or walking recommendations. In fact, both groups rated these neighborhood components favorably, with more than one half noting the presence of adequate streetlight quality, active neighbors, safe recreation facilities, and a lack of serious problems with unattended dogs. Therefore, in the present study, a relative lack of variability in the perception of these attributes by both African American and white adults may have limited the likelihood of finding a significant association between these attributes and PA or walking. The processes used in this study do not allow us to draw conclusions on whether the relative lack of variability was attributable to the design of the survey tool, strong similarities in environmental factors near respondents' homes, a combination of both, or other factors.

Limitations to the present study include the use of a cross-sectional design and self-reported measures, similar to limitations noted in recent reviews of studies investigating similar research questions (18-20,24,25). Although the survey response (54%) in this study was modest, the rate was consistent with previous studies using similar methodology (3,22,32), and the final study sample closely mirrored the racial and ethnic composition of the county. Generalization of the results may be somewhat limited because the survey was conducted during the winter in a predominantly rural county with several small towns and only one metropolitan area. Another potential limitation of the current study was that some measures demonstrated low to fair validity (? = ?0.02 to 0.28) (39). However, the "gold standard" objective indices for most perceptual constructs of the social and safety-related environmental attributes included in this and similar PA studies have not been identified or agreed upon (25,31). This is particularly true for the social environmental constructs of perceiving a neighbor as trustworthy or physically active; to our knowledge, these constructs have not been validated against any truly objective indices. In addition, the level of ? constituting adequate agreement between perceptions of environmental supports for PA and various objective measures of these supports has not been established. Researchers should continue to refine survey instruments and other data collection methods to be used in this line of research.  Indeed, additional work in this area is ongoing (40).

The factors included in this study also did not represent the full domain of possible social and safety-related or built environmental influences on PA (e.g., access, convenience, sidewalks, trails, aesthetics) (21,25,27).  Furthermore, concepts of social cohesion, social support, and trust are multifaceted and complex, and perhaps the extent of the influences of these factors on PA and walking among African American adults was not fully captured in our assessment (29). As noted earlier in this paper, PA behavior is impacted by many variables and disentangling the most influential perceptual or environmental factors is challenging (4,5). This is particularly true when considering the many types of people, PAs, locations, and circumstances that need to be included in a comprehensive research agenda or PA promotion program (6,25,40).

In summary, this study documented that white adults who reported perceiving their neighbors as being trustworthy and physically active were more likely to report meeting recommendations for PA or walking. No association between perceptions of neighborhood social and safety-related environmental factors and PA or walking was observed in African American adults. Further quantitative research is needed to identify perceptions of social and safety attributes of the neighborhood environment that strongly impact PA and walking levels of white and African American adults to build upon findings from qualitative research and to design effective interventions to promote PA.