Young drivers' elevated crash risk is linked to inexperience, low risk awareness, and underdeveloped cognitive control. To improve driving testing, several countries use hazard perception tests, and Sweden is piloting a simulator-based screening test as a potential complement to the on-road driving test. These screening tests would have to be developed in several versions and would have to be randomized. The aim of this study was to examine whether introducing controlled visual variations into traffic situations would alter the overall difficulty structure of the simulator-based screening test, that is, whether the relative difficulty of the situations remained stable across different test versions. An experimental between-group design was used, with 127 Swedish automotive high school students randomly assigned to one of three versions of a simulator-based screening test. All versions included the same 14 traffic situations across urban, rural, and highway environments, but differed in minor design features. The screening test evaluated participants' performance in each situation (pass/fail) and the outcome from the three versions was analyzed with a Rasch analysis. The findings suggest that small variations did not compromise overall test consistency, although added pedestrians and roadside objects increased difficulty in some traffic situations. These insights support the development of fair and reliable simulator-based screening test, with potential benefits for driver testing and traffic safety.