Women who stop smoking during pregnancy have a greater chance of giving birth to an easy-going child, a study has shown.

Smoking is known to be linked to smaller babies, but the link to problem behaviour is less well-established. Some studies have shown a link, but it has never been clear if this is a result of smoking, or if it is the personality of mothers who smoke which affects the personality of the child.

Researchers led by Kate Pickett of the University of York used data from the Millennium Cohort Study, which followed 18,000 children born in 2000-02. They compared the temperaments of children born to non-smoking mothers, mothers who quit, and mothers who continued to smoke during pregnancy.

Their conclusions, in Journal of Epidemiology and Community Health, is that quitters have babies with the easiest temperaments, while smokers who continue have the most difficult babies. The conclusion is somewhat undermined by the fact that lifelong non-smokers have more troublesome babies than quitters.

The authors suggest that quitting marks out mothers with an urge to protect the baby, rather than any intention to quit in the long term. Relapse rates are high after the birth, they say. Quitting even for a short time indicates the capacity to adapt to different circumstances and the ability to plan and to delay gratification, characteristics which seem to be missing in those who carry on smoking, they say.

The differences are more marked when the psychological traits of the children are put into categories, such as mood, how distressd they are by new experiences, and irregularity – which includes regular sleeping. The biggest differences were found in irregularity, which was low in the babies of quitters, higher in those of mothers who were light smokers, and higher still in heavy smokers (defined as ten or more cigarettes a day). But even these differences were not huge.

The authors say this is the first study to go beyond the obvious effects of smoking, such as low birthweight, and to examine more nuanced effects. Their conclusion is that both smoking and giving up have effects on unborn babies, and that quitters have babies that are easier to deal with.

They admit that they relied on mothers to report the temperament of their babies, which may have introduced a bias.