Significance Testing for Proportions: Exhaustive Study Guide

Defining Hypotheses for Significance Tests

• Parameter of Interest: The variable under investigation is denoted by $p$, representing the true proportion of a specific characteristic within a context or population.
• The Null Hypothesis ($H_0$): This is the statement of the status quo or no effect. It is formally expressed as:     * $H_0: p = \text{null value}$
• The Alternative Hypothesis ($H_a$): This is the claim for which evidence is being sought. It can take three forms depending on the direction of the test:     * One-Sided (Lower Tail): $H_a: p < \text{null value}$     * One-Sided (Upper Tail): $H_a: p > \text{null value}$     * Two-Sided: $H_a: p \neq \text{null value}$

The P-Value and Statistical Conclusions

• Interpretation of the P-Value: The p-value is a conditional probability. The exhaustive interpretation template is as follows:     * "Assuming $H_0$ is true (where $p = \text{null value}$), there is a [actual p-value] probability of getting a sample proportion ($\hat{p}$) of [observed value] or more extreme purely by chance."     * The direction of "more extreme" must match the direction specified in the alternative hypothesis ($H_a$).
• Reaching a Conclusion: The decision to reject or fail to reject the null hypothesis is based on a comparison between the p-value and the significance level ($\alpha$).     * When the Outcome is Not Significant: If the $p\text{-value} > \alpha$, the correct decision is to fail to reject $H_0$. Consequently, there is not convincing evidence for the claim made in $H_a$.     * When the Outcome is Significant: If the $p\text{-value} < \alpha$, the correct decision is to reject $H_0$. In this case, there is convincing evidence for the claim made in $H_a$.
• Conclusion Template: "Because [p-value] [> or <] [$\alpha$], we [reject/fail to reject] $H_0$ and we [do/do not] have convincing evidence for [context of $H_a$]."

Practical Application: Investigation into Teenagers' Perspectives on Mothers

• Study Background:     * Claim Source: Factinate.com.     * Original Claim: $84\%$ (or $0.84$) of teenagers think highly of their mother.     * Research Objective: Investigate if the true proportion of teens who think highly of their mother is actually greater than $0.84$.
• Sample Data Collection and Statistics:     * Researcher: A school psychologist.     * Sample Size ($n$): $150$ teenagers.     * Count of Successes: $132$ teenagers think highly of their mother.     * Calculated Sample Proportion ($\hat{p}$):         * $\hat{p} = \frac{132}{150} = 0.88$
• Hypothesis Formulation:     * The parameter $p$ is defined as the true proportion of all teens who think highly of their mother.     * Null Hypothesis ($H_0$): $p = 0.84$     * Alternative Hypothesis ($H_a$): $p > 0.84$
• Evidence Evaluation:     * Does the result provide evidence for $H_a$?: Yes, because the sample proportion ($\hat{p} = 0.88$) is numerically greater than the null value of $0.84$.
• Significance Testing and Interpretation:     * Provided P-Value: $0.091$     * Verbatim Interpretation: "Assuming $H_0$ is true ($p = 0.84$), there is a $0.091$ probability of getting a $\hat{p}$ of $0.88$ or greater purely by chance."     * Note: The term "greater" is used in the interpretation to match the "greater than" direction of the alternative hypothesis ($H_a$).
• Final Statistical Decision:     * Testing Level (Alpha): $\alpha = 0.05$     * Comparison: Since $0.091 > 0.05$, the result is not statistically significant at the $5\%$ level.     * Final Conclusion: "Because $0.091 > 0.05$, we fail to reject $H_0$ and we do not have convincing evidence that the true proportion of teens who think highly of their mother is greater than $0.84$."