Just stepping away from the context and looking at the methodology here, it’s infuriating that they try to apply statistical (sorry, data science) techniques which only really make sense with continuous data, to bullshit measures which are flags or at best discrete values. That bollocks CDM could be improved, from a certain point of view…
Just stepping away from the context and looking at the methodology here, it’s infuriating that they try to apply statistical (sorry, data science) techniques which only really make sense with continuous data, to bullshit measures which are flags or at best discrete values. That bollocks CDM could be improved, from a certain point of view, by using a single question “do you have breasts?” And taking the values zero and one. This would give 100% improvement after surgery with a completely sound statistical basis, but telling us nothing of any use.
I could understand this if the audience for the articles was the general public this is published in a scientific journal. Are those people really so statistically illiterate that they just let this go without a hint of worry?
Just stepping away from the context and looking at the methodology here, it’s infuriating that they try to apply statistical (sorry, data science) techniques which only really make sense with continuous data, to bullshit measures which are flags or at best discrete values. That bollocks CDM could be improved, from a certain point of view, by using a single question “do you have breasts?” And taking the values zero and one. This would give 100% improvement after surgery with a completely sound statistical basis, but telling us nothing of any use.
I could understand this if the audience for the articles was the general public this is published in a scientific journal. Are those people really so statistically illiterate that they just let this go without a hint of worry?