Top Highlights

• The probability histogram of the sample statistic. This is also called the samplin,g> distribution. Unlike a regular (or empirical) histogram that describes data, the probability histogram represents chance.
• understand that sample statistics are random variables with their own probability histograms.
• Hypothesis tests contain one more element. In addition to the claimed value for a parameter called the null hypothesis, there is a second hypothesis, called the alternative hypothesis, which contradicts the null. Alternative hypotheses come in two varieties: the first says that null is false, andthe second says in which direction it is false.
• Statistical inference, as exemplified by hypothesis testing, boils down to determining what the claim is about the data generating process, deciding on a box model that describes the data generating process, figuring out the shape of the sampling distribution, obtaining a measure of the likely size of chance error, deciding on a critical level of s...
• You have to use the right distribution for the particular sample statistic you are working with in order to get the hypothesis test right.