Hypothesis Tests Regarding a Parameter: Structured Study Notes

Chapter 10: Hypothesis Tests Regarding a Parameter

The Nature of Hypothesis Testing

Hypothesis testing is a fundamental aspect of inferential statistics, allowing us to make decisions about population parameters using sample data. It is widely used in fields such as medicine, manufacturing, and social sciences to assess claims or test the effectiveness of interventions.

• Hypothesis: A statement about a population parameter (e.g., mean, proportion).

• Null Hypothesis (H0): The hypothesis to be tested, usually stating no effect or no difference.

• Alternative Hypothesis (Ha): The hypothesis against which H0 is tested, indicating an effect or difference.

• Example: Testing if a new packaging process meets a quality standard (e.g., mean weight of pretzel bags is 454g).

Classifying Decisions in Hypothesis Testing

When conducting a hypothesis test, there are four possible outcomes based on the reality and our decision:

Decision	H0 is True	Ha is True
Do not reject H0	Correct Decision	Type II Error (Probability = β)
Reject H0	Type I Error (Probability = α)	Correct Decision (Probability = 1-β)

• Type I Error (α): Rejecting H0 when it is true.

• Type II Error (β): Failing to reject H0 when Ha is true.

• Power of a Test: $\text{Power} = P(\text{Reject } H_0 \text{ when } H_a \text{ is true}) = 1 - \beta$

Significance Level and Statistical Significance

The significance level (α) is the probability of making a Type I error. Common values are 0.05 and 0.01. A result is statistically significant if the P-value is less than α.

• Statistically significant: P-value < α

• Not statistically significant: P-value ≥ α

Types of Hypothesis Tests

Depending on the research question, hypothesis tests can be:

• Two-tailed test: Tests if the population mean is different from a specified value. $H_0: \mu = \mu_0$ $H_a: \mu \neq \mu_0$

• Left-tailed test: Tests if the population mean is less than a specified value. $H_0: \mu = \mu_0$ $H_a: \mu < \mu_0$

• Right-tailed test: Tests if the population mean is greater than a specified value. $H_0: \mu = \mu_0$ $H_a: \mu > \mu_0$

Errors and Diagnostic Test Concepts

• Sensitivity: Probability of a test correctly giving a positive result when the condition is present. $= 1 - \beta$

• Specificity: Probability of a test correctly giving a negative result when the condition is absent. $= 1 - \alpha$

P-Value Approach to Hypothesis Testing

The P-value approach is the most common method for hypothesis testing. It quantifies the probability of observing a test statistic as extreme as the one obtained, assuming H0 is true.

• P-value: Probability of obtaining a result at least as extreme as the observed, under H0.

• Decision Rule: If P-value ≤ α, reject H0. Otherwise, do not reject H0.

Steps in the P-value approach:

1. State the null and alternative hypotheses.

2. Decide on the significance level (α).

3. Compute the value of the test statistic.

4. Determine the P-value.

5. Compare P-value to α and make a decision.

6. Interpret the result in context.

Hypothesis Testing for One Population Mean When σ is Unknown

When the population standard deviation (σ) is unknown, we use the sample standard deviation (s) and the t-distribution.

• Test Statistic: $t = \frac{(\bar{x} - \mu)}{(s / \sqrt{n})}$

• Degrees of Freedom: $df = n - 1$

• Use t-tables or statistical software to find P-values.

Example: For a left-tailed test with n = 12 and t = -1.938, use the t-table or R software to estimate the P-value.

Relationship Between Confidence Intervals and Hypothesis Tests

Confidence intervals provide a range of plausible values for the population parameter. If the confidence interval does not contain the value specified in H0, we reject H0.

• Example: 95% confidence interval for mean spending does not contain $1,736, so we reject H0.

Classical (Critical Value) Approach to Hypothesis Testing

The classical approach uses critical values to define rejection and nonrejection regions for the test statistic.

• Rejection Region: Values of the test statistic that lead to rejection of H0.

• Nonrejection Region: Values that do not lead to rejection.

• Critical Value: The boundary value(s) separating the regions.

Test Type	Critical Value(s)
Two-Tailed	±tα/2
Left-Tailed	−tα
Right-Tailed	+tα

If the test statistic is more extreme than the critical value, we reject H0.

Statistical Significance versus Practical Significance

A result can be statistically significant due to a large sample size, even if the actual difference is trivial. Always consider the practical importance of findings.

• Statistical significance: Driven by sample size and standard error.

• Practical significance: Relates to the real-world impact of the result.

Controlling the False Discovery Rate

When conducting multiple hypothesis tests, the probability of false positives increases. The false discovery rate (FDR) is the proportion of rejected null hypotheses that are actually true.

• False Discovery Rate Formula: $\text{FDR} = \frac{\text{Number of false positive results}}{\text{Number of total positive test results}}$

• Benjamini-Hochberg Procedure: Controls FDR by ranking P-values and selecting a threshold (q-value).

Example: If 1,000 tests are run at α = 0.05, the FDR can be much higher than 5% due to multiple comparisons.

Worked Examples

• Consumer Spending: $t = \frac{1923 - 1736}{(355 / \sqrt{25})} = 2.664$ Compare t to critical values and P-value to α to make a decision.

• Engine Miles per Gallon: $t = \frac{27.5 - 26}{(3.5 / \sqrt{12})} = 1.485$ If P > α, do not reject H0.

Summary Table: Types of Errors and Test Outcomes

Outcome	Probability	Description
Type I Error	α	Reject H0 when H0 is true
Type II Error	β	Fail to reject H0 when Ha is true
Power	1 − β	Correctly reject H0 when Ha is true

Additional info: These notes expand on the original file by providing definitions, formulas, and structured examples for clarity and completeness. The Benjamini-Hochberg procedure and the distinction between statistical and practical significance are included for academic context.