Assessing self-reported use of new psychoactive substances: The impact of gate questions