This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Section 6.3 Using Significance Tests Section 6.3 1 Introduction Significance tests are widely used in reporting the results of research in many fields of applied science and in industry. Statistical significance is valued because it points to a result that is unlikely to occur simply by chance alone. Software makes the computation of test statistics and p-values straightforward. Applying tests appropriately can be challenging. Section 6.3 2 What is a convincing p-value? The purpose of a significance test is to describe the evidence that the sample provides against the null hypothesis. The p-value is a quantitative measure of this evidence. How small should the p-value be to be considered convincing evidence? One option is to use our fixed level of significance , but this is somewhat arbitrary. Section 6.3 3 Considerations for p-values Some considerations in assessing p-values: How plausible is H ? If H represents a long-standing belief, strong evidence (a small p-value ) will be needed to persuade the audience. What are the consequences of rejecting H ? If rejecting H in favor of H A requires expensive adjustments, strong evidence will be needed to justify the expense. Would even small amounts of evidence be useful? Often in pilot studies or multi-variable studies, modest p-values will give clues about variables or relationships that should be investigated more thoroughly. Section 6.3 4 Reporting p-values Recall that the choice of significance level is arbitrary. Traditional standard levels have included 1%, 5% and 10%....
View Full Document
- Spring '08