I assume it was simply down to having a pre-existing, conveniently-accessible large set of potential participants to analyse, in the form of those involved in the UK Million Women Study (sponsored again by Oxford), who had been asked about their phone use all the way back in 2001.