Social capital predicts accelerometry-measured physical activity among older adults in the U.S.: a cross-sectional study in the National Social Life, Health, and Aging Project

Table 3 Multivariate linear regression: social network and social engagement predictors of wrist accelerometer-measured physical activity levels

				Minimally Adjusted Regression Models^a		Fully Adjusted Regression Models^b
Model	Independent Variable		n	β Coefficient (CI)	p value	β Coefficient (CI)	p value
Social Network Structure
1	Network Size		662	6.16 (0.51–11.82)	0.033	4.77 (0.18–9.35)	0.042
2	Network Proportion Friends		662	36.91 (3.9–69.91)	0.029	35.81 (7.97–63.66)	0.013
Social Engagement Frequency
3	Socializing^c		596	12.35 (3.73–20.97)	0.006	8.73 (1–16.46)	0.028
	3a	Socializing with Friends/Relatives^d	596	12.79 (− 2.33–27.9)	0.095	5.72 (− 6.35–17.8)	0.35
	3b	Visiting Neighbors^e	604	17.87 (7.04–28.69)	0.002	14.62 (3.86–25.37)	0.009
4	Community Involvement^f		590	4.5 (0.38–8.63)	0.033	3.33 (− 0.35–7.02)	0.075
	4a	Attending Organized Group Meetings^d	597	11 (1.84–20.17)	0.020	9.04 (0.66–17.41)	0.035
	4b	Attending Religious Services^d	659	3.59 (− 6.4–13.58)	0.47	3.57 (− 4.37–11.51)	0.37
	4c	Volunteering^d	595	10.39 (0.45–20.32)	0.041	5.97 (− 3.53–15.47)	0.21

^aEach minimally adjusted linear regression model is controlled for accelerometer number of hours worn, number of weekend days worn, and month of wear
^bEach fully adjusted linear regression model is adjusted for age, gender, education, ethnic group, net worth, marital status, job status, self-rated health, functional health, comorbidities, depressive symptoms, and cognitive status in addition to accelerometer number of hours worn, number of weekend days worn, and month of wear
^cSocializing refers to the composite scale comprised of frequencies of socializing with friends/relatives and visiting with neighbors
^dFrequency with which respondent participates in the specified social engagement activity in the past 12 months
^eFrequency with which respondent or his/her neighbors visit with one another in homes or on the streets
^fCommunity involvement refers to the composite scale comprised of frequencies of attending organized group meetings, attending religious services, and volunteering

ISSN: 1471-2458