| Surviving cohort at 45 years* | |||||
| Total† | Participant‡ | Not in 45-year sample | |||
| n = 17313 n (%) | n = 9377 n (%) | Refusal§ n = 2840 n (%) | Other¶ n = 5096 n (%) | % Bias** | |
| Occupational status | |||||
| Full-time work | 7489 (66.7) | 6115 (67.8) | 918 (60.9) | 456 (65.6) | 1.6 |
| Part-time work | 2100 (18.7) | 1724 (19.1) | 303 (20.1) | 73 (10.5) | 2.1 |
| Not working | 1634 (14.6) | 1181 (13.1) | 287 (19.0) | 166 (23.9) | −10.3 |
| Social class | |||||
| Non-manual | 6143 (64.3) | 5110 (65.4) | 744 (61.0) | 289 (54.4) | 1.7 |
| Manual | 3416 (35.7) | 2700 (34.6) | 475 (39.0) | 241 (45.5) | −3.1 |
| Marital status | |||||
| Married | 8010 (70.8) | 6530 (71.9) | 1089 (71.4) | 391 (55.3) | 1.6 |
| Divorced/widowed | 1882 (16.6) | 1453 (16.0) | 232 (15.2) | 197 (27.9) | −3.6 |
| Never married | 1426 (12.6) | 1102 (12.1) | 205 (13.4) | 119 (16.8) | −4.0 |
| Obese | |||||
| <30 kg/m2 | 9314 (84.5) | 7567 (85.3) | 1174 (80.3) | 573 (82.8) | 0.9 |
| ⩾30 kg/m2 | 1711 (15.5) | 1303 (14.7) | 289 (19.7) | 119 (17.2) | −5.2 |
| Smoking status | |||||
| Never smoked | 5283 (46.8) | 4334 (47.8) | 704 (46.2) | 245 (34.9) | 2.1 |
| Ex-smoker | 3110 (27.5) | 2541 (28.0) | 388 (25.5) | 181 (25.8) | 1.8 |
| Current smoker | 2902 (25.7) | 2194 (24.2) | 432 (28.3) | 277 (39.4) | −5.8 |
| Drinking frequency | |||||
| ⩾ Weekly | 8019 (70.9) | 6508 (71.7) | 1029 (67.5) | 482 (68.3) | 1.1 |
| ⩽ Monthly | 2694 (23.8) | 2111 (23.3) | 408 (26.8) | 175 (24.8) | −2.1 |
| Never | 598 (5.3) | 461 (5.1) | 88 (5.8) | 49 (6.9) | −3.8 |
*Excluding deaths by age 45 years. †The maximum with data at 42 years is n = 11 354 (11 419 responders at 42 years (table 1) of whom 65 had died by age 45 years. ‡At 42 years the maximum with data is n = 9087. §Permanent refusals (n = 1038) and refusals to 45-year survey (n = 1803); at 42 years the maximum with data is n = 1529. ¶Group comprises emigrants (n = 1300), no contact attempted (n = 3004), and eligible non-response not classified as refusal (ie 3633–2841 = 792) at 45 years; at 42 years the maximum with data is n = 738. **Percentage bias ((45-year sample% − total surviving cohort%)/total surviving cohort%); positive bias represents an overrepresentation of the characteristic in the sample compared with the total cohort, negative bias an underrepresentation.