SOCC23H3 Lecture Notes - Lecture 6: Statistical Significance, Test Statistic, Sampling Distribution
Document Summary
When we test a hypothesis, we"re testing a statistic that we obtain from one or more random samples against some value that is stated to characterize a population or populations. Today we"ll start with the simplest version: a claim about the value of a population parameter (mean or proportion) that we evaluate with sample data. This is known as for a one-sample hypothesis test. Data from the early 1990s found that 80% of people thought that infidelity was always wrong. we think that now a lower proportion of people think it"s wrong. We can test this expectation by drawing a random sample and asking people about infidelity. Suppose that fewer than 80% of people in our sample said that infidelity was always wrong. An alternative (or research) hypothesis (signified by h1 or ha: the alternative hypothesis can be one tailed vs. two-tailed.