Hypothesis Testing and Statistical Inference: Study Guide for Statistics Students

Hypothesis Testing: Fundamentals

Introduction to Hypothesis Testing

Hypothesis testing is a statistical procedure used to evaluate claims about a population based on sample evidence and probability. It is a cornerstone of inferential statistics, allowing researchers to make decisions or inferences about population parameters.

• Null Hypothesis (H0): The statement to be tested; it usually represents no effect or no difference.

• Alternative Hypothesis (H1 or Ha): The statement the researcher is trying to find evidence to support; it represents an effect or difference.

Types of Hypothesis Tests

The direction of the test is determined by the form of the alternative hypothesis.

• Two-Tailed Test: Tests whether the parameter is different from a specific value (not equal). $H_0: \text{parameter} = \text{value}$ $H_1: \text{parameter} \neq \text{value}$

• Left-Tailed Test: Tests whether the parameter is less than a specific value. $H_0: \text{parameter} = \text{value}$ $H_1: \text{parameter} < \text{value}$

• Right-Tailed Test: Tests whether the parameter is greater than a specific value. $H_0: \text{parameter} = \text{value}$ $H_1: \text{parameter} > \text{value}$

Examples of Hypothesis Formulation

• Proportion Example: For students at a college, if the completion rate is believed to be higher than 0.399: $H_0: p = 0.399$ $H_1: p > 0.399$

• Mean Example: If home prices are believed to be lower than $245,700:$H_0: \mu = 245,700H_1: \mu < 245,700$

• Difference Example: If monthly revenue per cell phone is suspected to be different from $48.79:$H_0: \mu = 48.79H_1: \mu \neq 48.79$

Test Statistics and Decision Making

Test Statistic Calculation

The test statistic is a value calculated from sample data to make decisions about the null hypothesis.

• Proportion: $\hat{p} = \frac{x}{n}$, use z-test for proportions.

• Mean: Use z-test or t-test depending on sample size and population standard deviation knowledge.

• Standard Deviation: Use chi-square ($\chi^2$) test.

Level of Significance and p-value

• Level of Significance (\alpha): The probability threshold for rejecting the null hypothesis, commonly set at 0.05.

• p-value: The probability of obtaining a test statistic as extreme as the observed one, assuming the null hypothesis is true.

Decision Rules

• Reject H0: If p-value < \alpha, there is enough evidence to support the alternative hypothesis.

• Fail to Reject H0: If p-value > \alpha, there is not enough evidence to support the alternative hypothesis.

Errors in Hypothesis Testing

• Type I Error: Rejecting a true null hypothesis (false positive).

• Type II Error: Failing to reject a false null hypothesis (false negative).

Testing Claims about Proportions

p-value Approach

The p-value approach uses probability values to determine whether to reject the null hypothesis.

• State hypotheses ($H_0$, $H_1$).

• Set significance level ($\alpha$).

• Calculate p-value (e.g., 1-PropZTest).

• Reject $H_0$ if p-value < \alpha.

• State conclusion.

Critical Value Approach

This approach uses standard deviations to compare the sample statistic to the population parameter.

• State hypotheses ($H_0$, $H_1$).

• Set significance level ($\alpha$).

• Compute test statistic ($z_{stat}$).

• Determine critical value ($z_{crit}$).

• Reject $H_0$ if $z_{stat}$ is in the rejection region.

• State conclusion.

Testing Claims about Means

Testing Hypotheses Regarding the Population Mean

• Use random samples and normal distribution (or n > 30).

• State hypotheses ($H_0$, $H_1$).

• Set significance level ($\alpha$).

• Calculate t-statistic or p-value (T-Test).

• Reject $H_0$ if t-statistic is in the rejection region or p-value < \alpha.

• State conclusion.

Example Applications

• ACT math test scores: $H_0: \mu = 22$, $H_1: \mu \neq 22$; p-value < 0.05, reject $H_0$.

• Machine calibration: $H_0: \mu = 4.05$, $H_1: \mu \neq 4.05$; t-statistic in rejection region, reject $H_0$.

Testing Claims about Standard Deviation

Testing Hypotheses Regarding the Population Standard Deviation

• Use random samples and normal distribution.

• State hypotheses ($H_0$, $H_1$).

• Set significance level ($\alpha$).

• Calculate $\chi^2$ statistic: $\chi^2 = \frac{(n-1)s^2}{\sigma^2}$

• Identify critical values from $\chi^2$ tables.

• Reject $H_0$ if $\chi^2$ is in the rejection region.

• State conclusion.

Example Applications

• Chinese birth weights: $H_0: \sigma = 56.1$, $H_1: \sigma \neq 56.1$; fail to reject $H_0$.

• Supermodel heights: $H_0: \sigma = 2.6$, $H_1: \sigma < 2.6$; p-value < 0.01, reject $H_0$.

Summary Table: Hypothesis Testing Approaches

Test Type	Statistic	Distribution	Decision Rule
Proportion	z	Normal	p-value or critical value
Mean	t	t-distribution	p-value or critical value
Standard Deviation	chi-square ($\chi^2$)	Chi-square	p-value or critical value

Additional info: The notes cover hypothesis testing for proportions, means, and standard deviations, including both p-value and critical value approaches, and provide examples and visual aids for each method.