(1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | |
Variables | Int=VA | Int=DAYS | Int=UE | Int=EP | Int=ADB | Int=SHB | Int=POP | |
Trust | 0.0761*** | 0.0754*** | 0.114*** | 0.0687*** | 0.130*** | 0.119** | 0.0554*** | 0.0802*** |
(6.90) | (3.83) | (4.09) | (3.63) | (4.16) | (2.08) | (2.63) | (6.82) | |
Membership | −0.00843 | 0.0595* | −0.00717 | −0.0334 | −0.112* | −0.180* | −0.0581 | −0.00728 |
(0.38) | (1.65) | (0.13) | (0.7052) | (1.74) | (1.74) | (1.13) | (0.31) | |
Social isolation | −0.112*** | −0.123*** | −0.128*** | −0.0931*** | −0.149*** | −0.124** | −0.120*** | −0.105*** |
(9.00) | (5.89) | (4.31) | (4.43) | (4.32) | (2.28) | (5.04) | (8.04) | |
Trust*Int | −0.00119 | −0.0810 | 0.135 | −0.325* | −0.344 | 0.0373 | −0.0183** | |
(0.04) | (1.56) | (0.78) | (1.83) | (−0.67) | (1.43) | (1.97) | ||
Membership*Int | 0.124** | −0.00335 | 0.104 | 0.544 | 1.455 | 0.0405 | −0.00334 | |
(2.18) | (0.03) | (0.18) | (1.53) | (1.54) | (0.67) | (0.23) | ||
Social Isolation*Int | −0.0194 | 0.0340 | −0.359* | 0.158 | −0.00603 | −0.00593 | −0.0322*** | |
(0.61) | (0.61) | (1.87) | (0.81) | (0.01) | (0.20) | (3.07) | ||
Observations | 11178 | 11178 | 11178 | 8129 | 8129 | 8129 | 8129 | 11178 |
Number of communities | 366 | 366 | 366 | 256 | 256 | 256 | 256 | 366 |
R-squared | 0.1671 | 0.1674 | 0.1673 | 0.1781 | 0.1782 | 0.1778 | 0.1780 | 0.1672 |
Absolute value of z statistics in parentheses; *significant at 10%; **significant at 5%; ***significant at 1%.
Results reported in the first column correspond to the estimate of the health equation based on the full sample. Columns 2–7 report estimates of the same model augmented by interactions of all social capital indicators with the variable defined as ‘int’ in the column head (in the table trust*int, membership*int, social isolation*int). The set of control variables included is the same as in table 3. The external instruments are, for each individual, the mean level of social capital of all other individuals living in the same community and the interactions of these community averages of social capital with the variable defined as ‘int’.
Source: Living Standards, Lifestyles and Health Survey, 2001.