Table 3. Independent variables used to estimate average treatment effect of working past retirement age for Japanese men, 1987–2002.
| Independent variable | Men not employed n = 643 | Men in employment n = 645 | Pa,b |
|---|---|---|---|
| Mean equalized household income, × 10 000 yen (SD) | 199 (133) | 323 (250) | < 0.001 |
| Owning own home, mean % (SD) | 82 (38) | 85 (36) | 0.162 |
| Educational attainment category in years, mean % (SD) | |||
| 0–7 | 12 (32) | 7 (25) | 0.003 |
| 8–9 | 41 (49) | 43 (50) | |
| 10–11 | 21 (40) | 18 (39) | |
| > 12 | 27 (44) | 32 (47) | |
| Married, mean % (SD) | 90 (30) | 94 (23) | 0.003 |
| Smokers, mean % (SD) | 47 (50) | 46 (50) | 0.508 |
| Self-reporting poor health, mean % (SD) | 17 (38) | 07 (26) | < 0.001 |
| Mean body mass index, kg/m (SD) | 22 (3) | 22 (3) | 0.098 |
| Mean age, years (SD) | 68 (5) | 65 (5) | < 0.001 |
| Longest-held employment type, mean % (SD) | |||
| Professional or administrative | 26 (44) | 11 (31) | < 0.001 |
| Service or clerical | 17 (38) | 6 (23) | |
| Agriculture, forestry or fisheries | 6 (23) | 3 (17) | |
| Manual labour | 45 (50) | 13 (34) | |
| Self-employment | 7 (25) | 67 (47) |
SD: standard deviation.
a P-values of differences in variables between those not employed and those in employment. There is only one P-value for each of the combined education and employment categories because we only investigated whether a proportion of some category was different between the two groups.
b Welch t test for continuous variables and χ2 test for categorical variables.
Note: Inconsistencies arise in some values due to rounding.