Table 5. Effects of individual-level and work unit-level social capital on systolic blood pressure of workers, Japan, 2009.
| Cluster-mean centered self-included model | Self-excluded model | ||||
| Estimate | (95% CI) | Estimate | (95% CI) | ||
| Fixed | Intercept | 120.050 | (109.125, 130.975) | 119.959 | (109.336, 130.582) |
| Individual-level variables | Individual-level social capital a | 0.042 | (−0.048, 0.132) | 0.047 | (−0.039, 0.133) |
| Women (vs. men) | −5.286 | (−8.510, −2.062) | −5.283 | (−8.505, −2.061) | |
| Age (year) b | 0.532 | (0.430, 0.634) | 0.532 | (0.430, 0.634) | |
| Work unit-level variable | Work unit-level social capital c | 0.136 | (−0.101, 0.373) | 0.091 | (−0.138, 0.320) |
| Random | Individual-level variance (SE) | 314.193 | (13.852) | 314.176 | (13.851) |
| Work unit-level variance (SE) | 1.791 | (3.311) | 1.789 | (3.309) | |
CI, confidence interval; SE, standard error.
a
The individual-level social capital was assessed on a scale ranging from 0 to 80, with higher values indicating higher social capital.
b
Age was grand-mean centered.
c
Work unit-level social capital was defined as the mean of all workers' scores in the work unit in the self-included model, whereas it was defined as the mean of coworkers' responses in the self-excluded model.