The criticisms leveled against control charts in the wiki again elucidate a common problem with the state of statistics today. If experts, who have thought long and hard about the philosophical implications of likelihood versus significance, cannot reach any consensus, then how do we teach experimenters or engineers to correctly apply practical statistical tests or draw the right kind of inferences?
Comments
I'm wondering if there is added value in replacing the proposed simple alarm function with control charts: https://en.wikipedia.org/wiki/Control_chart
The criticisms leveled against control charts in the wiki again elucidate a common problem with the state of statistics today. If experts, who have thought long and hard about the philosophical implications of likelihood versus significance, cannot reach any consensus, then how do we teach experimenters or engineers to correctly apply practical statistical tests or draw the right kind of inferences?