Is randomisation necessary for measuring defocus curves in pre-presbyopes?