PP217 Lecture Notes - Lecture 20: Null Hypothesis, Confidence Interval, Fiduciary
Document Summary
State of neutral info: means that the clinician is genuinely uncertain about which treatment option sis best, she is operating under the null hypothesis the null hypothesis is the assumed starting position of all (minimally acceptable) clinical trials. P-values: the probability that you would get a sample that deviated from the norm if the null hypothesis were true. Ie. if the deviation were truly random: this is usually put at 5% or 1% to make it fairly easy for the null hypothesis to win , the 0. 05 or 0. 01 threshold is called a (alpha) In the case of 0. 05 or lower for p (assuming = 0. 05), then we have evidence against the null hypothesis: anything above 0. 05 means we default to the null hypothesis. Null hypothesis testing: what should i take away from this discussion, there is a mathematical way of cashing out statistical significance , this matters a great deal to our discussion of equipoise.
