We know that eye witness testimony can be notoriously unreliable for a whole host of reasons. Even though they may be a hundred % sincere, how we know that they're actually a hundred% accurate? So in the case of seeing a ghost, or even a crime for that matter. If you can't test the experience, the thing that's actually happened, how do you research this? How do you go about testing the alternative explanations of what somebodi's remembered and why they've remembered that, for example?