Modelling Outcome‐Related Risk in Choice Experiments