GVC
Well-Known Member
This is beside the point. Concluding that there is no clutch because you have no good evidence that clutch exists is not any more reasonable than concluding that there is clutch because you have no good evidence that clutch doesn't exist. That there exists a very poorly conducted "study" shouldn't have any bearing on anything.Choosing a null is not about what makes more sense, it's about determining whether an effect is statistically significant. If "clutchness" exists, but can be separated from random variation, it may as well not exist. You always choose the null as saying there is not noticeable effect.
Last edited: