Characteristics of the analytic sample categorised by potential confounder and mediator variable status
| Confounders | Mediators | ||
| % | % | ||
| Ethnicity | Birth weight | ||
| White | 89.65 | Heavy | 11.13 |
| Mixed | 0.95 | Normal | 82.02 |
| Indian | 1.85 | Low birth weight | 6.86 |
| Pakistani | 2.74 | Missing | <0.01 |
| Bangladeshi | 0.72 | Gestational age | |
| Black Caribbean | 1.08 | 37 weeks or over | 91.86 |
| Black African | 1.19 | Pre-37 weeks | 7.39 |
| Other (including Chinese) | 1.83 | Missing | 0.75 |
| Maternal education | Gestational diabetes | ||
| NVQ level 1 | 8.09 | No | 98.85 |
| NVQ level 2 | 29.83 | Yes | 1.15 |
| NVQ level 3 | 14.42 | Delivery | |
| NVQ level 4 | 29.9 | Vaginal | 68.28 |
| NVQ level 5 | 3.65 | Forceps | 3.93 |
| Overseas qualifications | 2.22 | Vacuum | 5.86 |
| None of these | 11.79 | Other | 0.72 |
| Missing | <0.01 | Planned caesarean | 9.26 |
| Mother’s age band | Emergency caesarean | 11.96 | |
| 14–24 | 23.65 | Labour complications | |
| 25–34 | 59.07 | Yes | 33.53 |
| 35–51 | 17.28 | No | 66.47 |
| Missing | <0.01 | ||
| Siblings in house | |||
| None | 42.63 | ||
| One | 36.21 | ||
| Two | 14.61 | ||
| Three or more | 6.55 | ||
| Step-siblings in house | |||
| None | 90.83 | ||
| At least one | 9.17 | ||
| Smoked during pregnancy | |||
| Did not smoke | 71.62 | ||
| Gave up in first month | 7.00 | ||
| Smoked after first month | 21.38 | ||
| Alcohol during pregnancy | |||
| Never | 67.03 | ||
| Less than once per month | 15.03 | ||
| 1–2 times per month | 7.86 | ||
| 1–2 times per week | 8.05 | ||
| 3–7 times per week | 2.01 | ||
| Missing | <0.01 | ||
Percentages relate to weighted number of cases (n=17 113). All estimates obtained using ‘svy’ commands in Stata.