Hypothesis Testing for the Population Proportion of a Binomial Distribution

• Population Proportion ($p$): This is the parameter being tested, representing the probability of 'success' in a single trial of a binomial experiment.

• Null Hypothesis ($H_0$): The default assumption that the population proportion is equal to a specific value ($H_0: p = p_0$).

• Alternative Hypothesis ($H_1$): The statement that contradicts the null hypothesis, suggesting the proportion has increased ($p > p_0$), decreased ($p < p_0$), or simply changed ($p \neq p_0$).

• Test Statistic ($X$): In this context, the test statistic is the observed number of successes in a sample of size $n$.

• Discrete Probability: Unlike normal distribution tests, the binomial distribution is discrete, meaning probabilities are calculated for specific integer values of $X$.

• The p-value Approach: This calculates the probability of obtaining a result as extreme as, or more extreme than, the observed value, assuming $H_0$ is true. If this probability is less than the significance level ($\alpha$), $H_0$ is rejected.

• Actual Significance Level: Because the distribution is discrete, the probability of falling into the critical region may not exactly equal the nominal significance level (e.g., 5%). The actual significance level is the true probability of a Type I error (rejecting $H_0$ when it is true).

Step-by-Step Procedure

• Step 1: Define the parameter $p$ in the context of the problem and state $H_0$ and $H_1$.
• Step 2: Identify the sample size $n$ and the significance level $\alpha$.
• Step 3: Define the test statistic $X \sim B(n, p)$ under the assumption that $H_0$ is true.
• Step 4: Calculate the p-value based on the observed number of successes ($x$). For $H_1: p > p_0$, calculate $P(X \geq x)$. For $H_1: p < p_0$, calculate $P(X \leq x)$.
• Step 5: Compare the p-value to $\alpha$. If $p\text{-value} \leq \alpha$, reject $H_0$. Otherwise, fail to reject $H_0$.
• Step 6: State the conclusion in the context of the original claim.

• Check the Inequality: When calculating $P(X \geq x)$, remember that calculators often provide cumulative probabilities $P(X \leq k)$. Use the complement rule: $P(X \geq x) = 1 - P(X \leq x-1)$.

• Two-Tailed Halving: In a two-tailed test, always remember to halve the significance level for each tail when finding critical values. If $\alpha = 0.05$, you look for a probability $\leq 0.025$ in each tail.

• Contextual Conclusion: Never just say 'Reject $H_0$'. Always follow up with 'There is sufficient evidence at the $x\%$ level to suggest that [restate the claim in context].'

• Critical Value Definition: The critical value is the first value that falls within the critical region. For a 'greater than' test, it is the smallest integer $c$ such that $P(X \geq c) \leq \alpha$.

• Incorrect Tail Direction: Students often calculate $P(X \leq x)$ for a 'greater than' claim. Always align the probability calculation with the direction of $H_1$.

• Discrete Boundaries: Forgetting that $P(X \geq 10)$ is $1 - P(X \leq 9)$, not $1 - P(X \leq 10)$. In discrete distributions, the boundary value matters significantly.

• Sample Proportion vs. Population Proportion: Confusing the observed sample proportion ($\hat{p} = x/n$) with the hypothesized population parameter ($p$). The test is always about the population parameter.