Day 18
Math 216: Statistical Thinking
Sampling Distributions using Proportions
- Objective: Understand how the sample proportion \(\hat{p}\) estimates the population proportion \(p\).
- Context: Whether estimating voter preferences, customer behaviors, or biological occurrences, knowing how to infer population characteristics from sample data is crucial.
Motivation and Real-World Examples
- Voter Preference: Estimate the proportion of voters in favor of a new bill.
- Customer Behavior: Understand the percentage of customers using store credit cards.
- Conservation Efforts: Track the fraction of endangered species born in captivity.
Sampling distribution for \(\hat{p}\)
Properties of Sampling Distribution of \(\hat{p}\)
- Mean of \(\hat{p}\)
- The mean equals the population proportion \(p\).
- \(\hat{p}\) is an unbiased estimator of \(p\). \[ E(\hat{p}) = \mu_{\hat{p}} = p \]
- Standard Deviation of \(\hat{p}\)
- The variability of \(\hat{p}\) decreases as sample size \(n\) increases. \[ \sigma_{\hat{p}} = \sqrt{\frac{p(1-p)}{n}} \]
Central Limit Theorem for \(\hat{p}\)
- Approximately Normal: For large samples, \(\hat{p}\)’s distribution is nearly normal.
- Rule of Thumb:
- Consider large if \(np \geq 15\) and \(n(1-p) \geq 15\). \[ n \cdot \hat{p} \geq 15 \quad \text{and} \quad n(1-\hat{p}) \geq 15 \]
Normal Approx. for Binomial
- Binomial Properties: For \(X \sim \text{Binomial}(n, p)\):
- Mean: \(\mu = np\)
- Variance: \(\sigma^2 = np(1-p)\)
- Normal approximation: Valid if \(np \geq 15\) and \(n(1-p) \geq 15\).
- Sampling Distribution of Sample Proportion:
For proportions \(\hat{p} = X/n\), the sampling distribution is approximately normal:
\[
\hat{p} \approx N\left(\mu_{\hat{p}} = p,\ \sigma_{\hat{p}} = \sqrt{\frac{p(1-p)}{n}}\right)
\]