Sampling & Inference: The Big Picture
Sampling is taking a small, manageable piece of a larger group (population) to study. Inference is using that sample to draw conclusions about the whole population. It's the foundation of surveys, political polls, and scientific studies. We do this because measuring everyone is often impossible or too expensive.
Key Steps for Inference
- Define the Population & Parameter: Identify the entire group and the number you want to know (e.g., the true mean GPA of all 11th graders).
- Collect a Sample & Find a Statistic: Randomly select a sample and calculate a value from it (e.g., the mean GPA of 50 students).
- Check Conditions: Ensure the sample is random, and for means, that the sample size is large enough (n ≥ 30) for the Central Limit Theorem to apply.
- Calculate a Confidence Interval: Create a range of values that you're confident contains the true population parameter.
- Interpret the Results: State your conclusion in context.
Worked Examples
Example 1: Confidence Interval for a Mean
A sample of 35 eleventh graders has a mean study time of 1.5 hours/night (s=0.5 hrs). Find a 95% CI for the true mean.
- Parameter: μ = true mean study time for all 11th graders.
- Statistic: x̄ = 1.5 hrs, n=35, s=0.5 hrs.
- Conditions: Random sample? (Assume yes). n=35 ≥ 30, so CLT applies.
- Calculation: CI = x̄ ± (t*)(s/√n). For ~95% CI, t* ≈ 2.
Margin of Error = 2 * (0.5/√35) ≈ 0.17 hrs.
95% CI: 1.5 ± 0.17, or (1.33, 1.67) hours. - Interpretation: We are 95% confident the true mean study time for all 11th graders is between 1.33 and 1.67 hours per night.
Example 2: Inference for a Proportion
In a poll of 200 random students, 45% support a new dress code. Find a 90% CI for the true proportion.
- Parameter: p = true proportion of all students who support the code.
- Statistic: p̂ = 0.45, n=200.
- Conditions: Random sample, n*p̂=90 & n*(1-p̂)=110 are both ≥10.
- Calculation: CI = p̂ ± (z*)(√[p̂(1-p̂)/n]). For 90% CI, z*=1.645.
Standard Error = √[(0.45*0.55)/200] ≈ 0.035
Margin of Error = 1.645 * 0.035 ≈ 0.058
90% CI: 0.45 ± 0.058, or (0.392, 0.508). - Interpretation: We are 90% confident the true proportion of all students supporting the dress code is between 39.2% and 50.8%.
Common Mistakes to Avoid
🚫 Claiming Certainty: Never say "The true mean is in the interval." Confidence level (e.g., 95%) describes the *method*, not a single interval.
🚫 Misinterpreting the Interval: A 95% CI means that if we took many samples, 95% of the intervals we build would contain the true parameter. It does NOT mean there's a 95% chance the parameter is in your specific interval.
🚫 Forgetting Conditions: Always check for randomness and sample size. Inference from a biased sample is worthless.
🚫 Confusing Population & Sample: Parameters (μ, p) describe populations. Statistics (x̄, p̂) describe samples.
Tips & Tricks
💡 Memory Aid: "We are [Confidence Level]% confident that the [Population Parameter] is between [Lower Bound] and [Upper Bound]." This is the perfect interpretation template.
💡 Bigger Sample = Narrower Interval: To get a more precise estimate (smaller margin of error), increase your sample size (n).
💡 Parameter vs. Statistic: P's go together: Population Parameter. S's go together: Sample Statistic.
How to Practice
- Find real-world polls in the news and identify the sample, population, statistic, and parameter.
- Practice calculating confidence intervals for both means and proportions until the steps are automatic.
- Focus on writing the interpretation in a full, contextual sentence. This is often where points are lost on exams.
- Work with a study partner to explain the logic behind each step. If you can teach it, you know it.