Summary
Highlights
The video begins by introducing the concept of the level of significance, denoted by alpha (α), which represents the degree of significance used to accept or reject a null hypothesis. It highlights that 100% accuracy in hypothesis testing is not possible, and the significance level is the probability of making a Type I error (rejecting a true null hypothesis). Common alpha levels are 0.01 (public health), 0.05 (social science), and 0.10 (other studies).
When the alternative hypothesis is 'not equal' (≠), indicating a two-tailed test, the alpha level is divided by two. For instance, an alpha of 0.01 becomes 0.005, 0.05 becomes 0.025, and 0.10 becomes 0.05. This division accounts for the critical regions in both tails of the distribution.
Several examples are provided to demonstrate how to determine the appropriate alpha level. For a problem stating a 5% significance level and an alternative hypothesis that 'the current percentage of unmarried couples is different from 34 percent,' the alpha of 0.05 is divided by 2, resulting in 0.025 because 'different from' implies a two-tailed test.
The core distinction between one-tailed and two-tailed tests is elaborated. A two-tailed test is used when the alternative hypothesis contains 'not equal to' (≠), or phrases like 'different from,' 'changes from,' or 'not the same as.' A one-tailed test is used when the alternative hypothesis involves 'less than' (<) or 'greater than' (>).
For one-tailed tests, if the alternative hypothesis states 'less than,' it's a left-tailed test where the rejection region is on the left side of the normal curve. If it states 'greater than,' it's a right-tailed test, with the rejection region on the right side. The video uses examples like 'less than' and 'greater than' in enrollments to illustrate these concepts.
A self-test section challenges viewers to identify if a given alternative hypothesis leads to a one-tailed or two-tailed test. Phrases like 'less than' and 'lower than' indicate one-tailed tests, while 'not equal to' and 'differ' indicate two-tailed tests.
The video explains how to find the alpha level when a confidence interval is given. If a 90% confidence interval is mentioned, the alpha level is calculated as 100% - 90% = 10%, or 0.10. Similarly, for a 93% confidence interval, the alpha is 0.07 (100% - 93%).