Table 3.
Mean differences (95 confidence intervals) in IQ within families between singletons and twins, at ages 7 and 9
| Adjustment for confounders within families | At age 7 | At age 9 |
|---|---|---|
| None | 5.3 (1.5 to 9.1) | 6.0 (1.7 to 10.2) |
| Sex | 5.3 (1.6 to 9.1) | 6.0 (1.7 to 10.2) |
| Maternal age at delivery | 5.5 (1.8 to 9.3) | 6.2 (1.9 to 10.5) |
| No of older siblings | 5.1 (1.3 to 8.9) | 5.3 (1.0 to 9.6) |
| Birth weight (continuous *) | 2.8 (−1.3 to 6.9) | 4.3 (−0.3 to 8.9) |
| Gestational age (continuous †) | 4.7 (0.8 to 8.6) | 5.7 (1.3 to 10.1) |
| Birth weight z score (continuous ‡) | 4.0 (0.1 to 7.9) | 5.0 (0.6 to 9.4) |
| Sex, maternal age at delivery and number of older siblings | 5.4 (1.6 to 9.3) | 5.7 (1.4 to 10.0) |
| Birth weight and gestational age (continuous *†) | 2.6 (−1.5 to 6.7) | 4.1 (−0.5 to 8.8) |
| Birth weight z score and gestational age (continuous *†‡) | 2.9 (−1.2 to 7.0) | 4.3 (−0.3 to 8.9) |
See statistical methods for details of transformation notation.
Birth weight was modelled by using a transformation to the (−2,−2) power, where the second transformation differs from the first by multiplication by log (birth weight), for the analysis of IQ at age 7 and (1,2) power for the analysis of IQ at age 9.
Gestational age was modelled by using a transformation to the (3,3) power, where the second transformation differs from the first by multiplication by log (gestational age), for the analysis of IQ at age 7 and at age 9.
Birth weight z score was modelled by using a transformation to the (0.5,3) power for the analysis of IQ at age 7 and (1,3) power for the analysis of IQ at age 9.