Basic statistical concepts 21
involved this will never be the case and some level of approximation whe
n
constructing a model will have to be used. Therefore, there is some level
of subjectivity used when a model is assumed. Therefore, so-called “objec-
tive” a nalysis frequently lacks objectivity as well because the model is never
exact.
Subjectivity is unavoidable in decision making. We should not for get
t
hat the endpoint of any analysis in medical imaging or for that matter in
medicine is not the posterior or likelihood but rather the decision. Assum-
ing that statistical analysis is p erformed objectively the decision (disease
present or not, optimal course of therapy, etc.) still ha s to be ma de and this
decision will require a subjective judgment. ...