RSM423H1 Lecture Notes - Lecture 3: Confidence Interval, Motivated Reasoning, Decision Aids
Document Summary
Equivalency of two decisions: mus if achieved p>mm, then rejected. Bayes approach (in terms of any risks): if prob of mm > planned risk, then reject. Example: assume population of 10,000, illustrates how aug 41 par 42 is applied. Confidence level of 95% desired so split distribution curve to 95% and tail is the 5% risk accepted. Confidence interval in our example is (0, 0. 063). Reasonable range that the auditor will consider acceptable. If estimate falls within the range, therefore not sure if there is mm. If esti(cid:373)ate does(cid:374)"t fall withi(cid:374) the ra(cid:374)ge we are 95% (cid:272)ertai(cid:374) of mm. Decreasing assurance from testing: pg 17 lecture notes (assurance vs. Inductive logic: being able to come to conclusion about those items which we do not sample. Randomization/selection process have to be able to predict the probability that something will be sampled or not i. e. requires predictable probability of selecting an item. Majority survey shows this is a true statement.
