Smoking is associated with poor educational outcomes, but whether this is due to confounding is hard to tease out. This study aims to investigate smoking during adolescence, and whether this is associated with educational and IQ outcomes (in essence replicating Mokrysz et al's 2016 ALSPAC study on cannabis and educational outcomes but focussing on smoking as the exposure).

As well as this, smoking during pregnancy will also be investigated, to ascertain whether this predicts offspring performance on a neurocognitive task at age 18. In order to attempt to rule out confounding, a negative control design will be used, whereby associations between maternal smoking and offspring neurocognition will be compared to those between paternal smoking during pregnancy and offspring neurocognition, where an intra-uterine effect cannot have occurred.