Do the items of the AQ 50 equally measure autistic traits across groups? So is it the same for women and older people as it is for younger men? It seems that some items are not that great. When groups are compared it seems better to use the AQ-28 as in the AQ 28 the majority of items are good enough. This is all described in the paper by Joost Agelink van Rentergem (Negatively phrased items of the AQ function differently for groups with and without autism)