The Logic of Hypothesis Testing
| dc.date.accessioned | 2025-04-25T12:00:04Z | |
| dc.date.available | 2025-04-25T12:00:04Z | |
| dc.date.issued | 0000 | |
| dc.description | This chapter discusses hypothesis testing, comparing it to estimation in statistics. The main question in estimation is determining the value of a population parameter, while hypothesis testing asks whether it’s reasonable to believe a parameter equals a specific value (like 100). Using an example involving IQ scores of fourth-graders, the process of hypothesis testing is explained: starting with a null hypothesis (e.g., the average IQ is 100) and testing it through a sample. If evidence from the sample contradicts the null hypothesis, the alternative hypothesis (e.g., the average IQ is not 100) is accepted. It also introduces the concept of errors in hypothesis testing—Type I errors (incorrectly rejecting the null hypothesis) and Type II errors (failing to reject the null hypothesis when it’s false). The chapter outlines the general procedure for hypothesis testing and the importance of setting a level of significance to control the risk of errors. | |
| dc.description.abstract | If you understood the logic presented in the chapter on estimation, you've gone a long way toward an understanding of hypothesis testing as well. Not that the two forms of inference are identical. Indeed, they seek answers to fundamentally different questions. However, they both make use of concepts like statistics, parameters, samples, and populations. And distributions of sample statistics (such as the normal distribution of sample averages you encountered in the last chapter) play a crucial role in both procedures. In estimation, the fundamental question we try to answer is, "What is the value of a population parameter?" In hypothesis testing, the Fundamental question is, "Is it reasonable to believe that the value of a population parameter is x?" The x in that question is just a number, like 0, or 100, or some other value that makes sense in the context of a particular problem. Hypothesis testing might be easier to understand if we con-sider a specific example. You're probably tired of it by now, but the example involving the average IQ-test score of all fourth-graders in a school system is as good as any, and at least it's familiar. To simplify the language of the presentation we'll assume that you're the superintendent of schools. | |
| dc.identifier.uri | http://192.9.200.215:4000/handle/123456789/515 | |
| dc.language.iso | en | |
| dc.subject | Null Hypothesis (H₀) | |
| dc.subject | Hypothesis Testing | |
| dc.subject | Level of Significance (α) | |
| dc.subject | Sample Statistic | |
| dc.subject | Alternative Hypothesis (H₁) | |
| dc.title | The Logic of Hypothesis Testing | |
| dc.type | Working Paper |