Each person of a "type" will still have unique differences The biggest problem with it is that there’s no validity to these “types”. Taking a unimodal normal d

Each person of a "type" will still have unique differences The biggest problem with it is that there’s no validity to these “types”. Taking a unimodal normal distribution and just calling everyone below the mode one type and everyone above the mode another type, is not only completely arbitrary but just plain wrong from a statistical perspective. That is, it won’t yield predictions with any accuracy. In order to show that there actually are real types they need to show a bimodal (or multimodal) distribution. As far as I’m aware, no one has done that.