Hypothesis Testing: The Language and Logic of Statistical Inference

Chapter 10: Hypothesis Tests Regarding a Parameter

Section 10.1: The Language of Hypothesis Testing

This section introduces the foundational concepts and terminology of hypothesis testing, a core procedure in inferential statistics. Hypothesis testing allows statisticians to make decisions about population parameters based on sample data.

Learning Objectives

• Determine the null and alternative hypotheses

• Explain Type I and Type II errors

• State conclusions to hypothesis tests

Understanding Hypothesis Testing

Hypothesis testing is a formal process used to assess claims about population parameters using sample data and probability theory.

• Hypothesis: A statement regarding a characteristic of one or more populations. In this context, hypotheses concern a single population parameter.

• Hypothesis Testing: A procedure based on sample evidence and probability, used to test statements about population characteristics.

Steps in Hypothesis Testing

1. Make a statement regarding the nature of the population (formulate hypotheses).

2. Collect evidence (sample data) to test the statement.

3. Analyze the data to assess the plausibility of the statement.

Null and Alternative Hypotheses

Hypothesis testing involves two competing statements:

• Null Hypothesis ($H_0$): The statement to be tested, representing no change, no effect, or no difference. It is assumed true until evidence suggests otherwise.

• Alternative Hypothesis ($H_1$): The statement for which we seek supporting evidence, representing a change, effect, or difference.

Types of Hypothesis Tests

• Two-tailed test: $H_0$: parameter = some value; $H_1$: parameter ≠ some value

• Left-tailed test: $H_0$: parameter = some value; $H_1$: parameter < some value

• Right-tailed test: $H_0$: parameter = some value; $H_1$: parameter > some value

In all cases, the null hypothesis contains a statement of equality.

One-tailed vs. Two-tailed Tests

• One-tailed tests (left- or right-tailed) specify a direction of the effect (less than or greater than).

• Two-tailed tests look for any difference, regardless of direction.

Examples: Forming Hypotheses

Scenario	Null Hypothesis ($H_0$)	Alternative Hypothesis ($H_1$)	Test Type
Proportion of children experiencing headaches with a new drug vs. 2% (competing drugs)	$H_0$: p = 0.02	$H_1$: p ≠ 0.02	Two-tailed
Mean time to complete an exam (claimed 60 minutes)	$H_0$: μ = 60	$H_1$: μ > 60	Right-tailed
Standard deviation of detergent bottle contents (old machine: 0.23 oz)	$H_0$: σ = 0.23	$H_1$: σ < 0.23	Left-tailed

Type I and Type II Errors

When conducting hypothesis tests, two types of errors can occur:

• Type I Error: Rejecting the null hypothesis when it is actually true.

• Type II Error: Not rejecting the null hypothesis when the alternative hypothesis is true.

Outcomes of Hypothesis Testing

Conclusion	Reality: $H_0$ True	Reality: $H_1$ True
Do Not Reject $H_0$	Correct Conclusion	Type II Error
Reject $H_0$	Type I Error	Correct Conclusion

Probability of Errors

• Probability of Type I Error ($\alpha$): $\alpha = P(\text{Type I error}) = P(\text{rejecting } H_0 \text{ when } H_0 \text{ is true})$

• Probability of Type II Error ($\beta$): $\beta = P(\text{Type II error}) = P(\text{not rejecting } H_0 \text{ when } H_1 \text{ is true})$

Level of significance ($\alpha$): The probability of making a Type I error. Common values are 0.05 or 0.01.

There is a trade-off between $\alpha$ and $\beta$: as the probability of a Type I error increases, the probability of a Type II error decreases, and vice versa.

Examples: Type I and Type II Errors

• Type I Error: Concluding that the proportion of children experiencing headaches is different from 0.02 when it is actually 0.02.

• Type II Error: Failing to conclude that the proportion is different from 0.02 when it actually is different.

Stating Conclusions to Hypothesis Tests

When interpreting the results of a hypothesis test, it is important to use precise language:

• We never "accept" the null hypothesis. Instead, we say we "do not reject" the null hypothesis, similar to a court declaring a defendant "not guilty" rather than "innocent."

• If the null hypothesis is rejected, there is sufficient evidence to support the alternative hypothesis.

• If the null hypothesis is not rejected, there is not sufficient evidence to support the alternative hypothesis.

Example: Stating the Conclusion

• If $H_0$ is rejected: "There is sufficient evidence to conclude that the proportion of children who experience a headache as a side effect is different from 0.02."

• If $H_0$ is not rejected: "There is not sufficient evidence to say that the proportion of children who experience a headache as a side effect is different from 0.02."

Additional info: These notes cover the essential language and logic of hypothesis testing, including the formation of hypotheses, types of errors, and proper interpretation of results. This foundation is critical for understanding subsequent chapters on hypothesis testing for means, proportions, and other parameters.