Race and Ethnicity Differences in Walking and Associations with Neighborhood Perceptions among Older Adults in California

Abstract

This paper estimates differences in walking and associations of walking with neighborhood perceptions among older adults in California. We analyzed the restricted-use California Health Interview Survey data merged with the American Community Survey data. Walking at least 10 minutes for leisure in the last 7 days and number of minutes walked were regressed on perceived neighborhood social cohesion scale (PNSCS) and neighborhood socioeconomic characteristics. We estimated ordinary least squares regression models. We found that positive neighborhood perceptions were significantly (p< 0.05) associated with walking among older Hispanic and Asian adults. Older Hispanic adults with above median PNSCS report that they walk 6.3 percentage points more than Non-Hispanic White adults. Older adults' perceptions of their neighbors and neighborhoods are positively correlated with walking. Policymakers can consider ways to improve neighborhood perceptions and relationships to help increase walking among minority older adults especially in low SES neighborhoods.

Keywords: Health behaviors; minority older adults; neighborhood perception; social cohesion; walking.

Figure 1.. Percentage of older adults who walked by perceived neighborhood social cohesion scale (PNSCS) and neighborhood socioeconomic status score (NSES).

Data: CHIS 2011-2017 data (All other variables) and Census ACS 2010-2016 data (for neighborhood SES score) We use the CHIS replicate weights to obtain valid variance estimates and also the final weights to account for sample selection probabilities.

Figure 2.. Predicted mean minutes walked for race/ethnic subgroups by perceived neighborhood social cohesion scale (PNSCS) from OLS regressions

Data: CHIS 2011-2017 data (All other variables) and Census ACS 2010-2016 data (for neighborhood SES score) Note: Controlling for other demographic characteristics, NSES, and year fixed effects. We use the CHIS replicate weights to obtain valid variance estimates and also the final weights to account for sample selection probabilities.