I wonder if the study had a selection bias (eg. if it was a poll conducted on the parent website very.co.uk)