1.Bayesian can be used when there is no prior statistics derived from traditional methods. i.e. it can be used for one-time events or subjective issues. For example, what is the probability that the universe results from a Big Bang (non-repeatable event)? What is the probability that God exists (subjective)?
2. if X happens, what is the probability of Y? If Y, what about Z? If Z, how about W...etc. One can form an inference network like this. In Educational Testing Service (ETS), psychometricians use Bayesian network to develop tests.
3.. What should we do when we have the full population? Can we close the case and declare a definite answer? Long story in short: even if we have the number (e.g. test score) from every body, but each person still has within-subject variations. For example, if you retake a test (e.g. SAT, ACT, GRE), you might not get the same score. I took GRE three times and each time I got a different score. Do we REALLY have the FULL information? This is a very deep question and it is beyond the scope of this course.
4. In inferential statistics we make estimations about a larger set based on the sample data. What is the name of that "larger data set"?
5. Who is winning the war on terror? According to Richard Wackrow, in some sense terrorists are winning. The objective of terrorism is to provoke fear and hatred. Since the 911 attack, Americans have been overreacting and living under threat. Thus, terrorists have accomplished their objectives.
Are we really overreacting? Wackrow uses statistical data to point out that the probability of being killed by terrorist attacks is much lower than that of other causes of death.