ABSTRACT
In this chapter we leave asymptotic theory and elaborate on some general inference princi-
ples initiated in Chapter 1, Sections 1.2 and 1.3 and Chapter 3, Sections 3.3 and 3.4. The
results will guide the construction of procedures with optimal asymptotic properties for
models much more general than those of this chapter. Recent work relating to the topics in
this chapter deals with optimal inference after data based model selection. See Fithian, Sun
and Taylor (2014).