Design and analysis are two parts of an Observational study.
There must be a design. The design must come first. A design is inherently subjective. The analysis should respect the design.
Statistics is an epistemological tool.
Has it been succssful? “Mehh.”
Will it be successful? My paper [here/here] tries to keep the “positivism” alive.
Causal inference from observational studies.
Can inferences from observational study prove a causal effect? Logically, never. But don’t be disappointed, statistical inference was never meant to be used to prove any statement.
Objectivity of probability.
Remains a sort of puzzling question for over a century. Interesting read in this article forthcoming in BJPS.