Notes on Population, Sample, and Interpreting Proportions

Population and Sample: Key Terms

Population: the entire group of interest for a study (all individuals or units that could be studied).
Sample: a subset selected from the population to observe and analyze.
Parameter: a numeric characteristic of the population (unknown in most cases).
Statistic: a numeric characteristic calculated from the sample (observed data).
Relationship: a statistic is used to estimate or infer the corresponding parameter of the population.
Notation examples:
- Population proportion: $p = P(X=1)$ or more generally, a population parameter such as the mean $\mu\ and \sigma^2$ .
- Sample proportion: $\hat{p} = \frac{k}{n}$ where $k$ is the number of successes in the sample of size $n$ .
- Sample mean: $\bar{x} = \frac{1}{n}\sum{i=1}^n xi$ .

Parameter vs Statistic; core idea

The parameter describes the entire population (often unknown): $\text{parameter} = p\, (\text{or } \mu, \sigma, \ldots)$
The statistic describes the sample (observed value): $\text{statistic} = \hat{p}, \bar{x}, s, \ldots$
Inference goal: use statistics to estimate parameters and quantify uncertainty.
Common sampling theme: ensure the sample is representative of the population to justify generalizations.

Population, Sample, and their typical questions

When you read a study, you should identify:
- What is the population of interest? (Who or what could vary in the entire group?)
- What is the sample actually studied? (Who or what was measured?)
- What statistic is reported from the sample? (e.g., proportion, mean, median, etc.)
- What parameter would this statistic estimate in the population?
Language cues to watch for:
- “X% of people…” may refer to a sample proportion unless the population is explicitly defined.
- “Compared to Y, X%…” requires clarity on the group being referred to and whether a sample or population is involved.

Interpreting Proportions: the 9/10 example and its issues

Example statement discussed: "Nine out of 10 people frequently recycle waste." and related lines about the UK.
Key idea: Distinguish between statements about a sample vs statements about the population.
Correct interpretation if based on a sample:
- If a sample of size $n$ yields $k$ successes (e.g., people who frequently recycle), then the sample proportion is $\hat{p} = \frac{k}{n}$ (e.g., $\hat{p}=0.90$ if $k=90$ and $n=100$ ).
- This does not automatically imply that the entire population has proportion $p=0.90$ unless the sample is representative and we account for sampling error.
A proper phrasing when referring to a sample: "In this sample, $\hat{p}=0.90$ of respondents frequently recycle."
If attempting to compare to the UK population, you would need data for the UK population or a clearly defined population corresponding to the UK, with appropriate confidence statements.
Common misinterpretations to avoid:
- Treating a sample proportion as if it were a population proportion without caveats.
- Confusing the unit of analysis (the sample) with the unit of the population (the population).
- Overgeneralizing from a non-representative sample.
The calculation example: if the sample shows $k$ successes out of $n$ trials, then the point estimate is $\hat{p}=\frac{k}{n}$ and a standard error for a proportion is often approximated by $SE\approx\sqrt{\frac{\hat{p}(1-\hat{p})}{n}}$ for large enough $n$ .
Confidence interval for a population proportion (as a typical next step):
- $CI = \hat{p} \pm z{\alpha/2} \sqrt{\frac{\hat{p}(1-\hat{p})}{n}}$ where $z{\alpha/2}$ is the critical value from the standard normal distribution (e.g., $z_{0.025}=1.96$ for 95% CI).

The three statements exercise: what’s wrong (as seen in the transcript)

Statement 1 (roughly): "We compared, like, nine out of 10 people."
- Issue: The subject of the claim is unclear; it mixes a comparison with a proportion without specifying the population, sample, or time frame.
- Improvement: Specify the population and whether the statement refers to a sample statistic or to the population proportion. Example reformulation: "In this sample of size $n$ , $\hat{p}=0.90$ of respondents frequently recycle."
Statement 2 (roughly): "When compared to the UK, we're nine out of 10 people. Nine out of 10 people frequently recycle waste."
- Issue: Grammar and scope; unclear what is meant by "we're nine out of 10 people"; also conflates a UK comparison with a local statement about recycling.
- Improvement: Break into clear pieces:
- Define the comparison group clearly: population or sample from the UK (or another specified population).
- State the statistic for the group of interest: e.g., "In our sample, $\hat{p}=0.90$ of respondents frequently recycle," and, if possible, compare to a population value from UK data with appropriate caveats.
Statement 3 (roughly): "the subjects are the 3,900"
- Issue: This sounds like mislabeling the population when the number (3,900) is more likely a sample size. The population is not identified; the count 3,900 likely refers to the sample, not the entire population.
- Improvement: Distinguish population vs sample size: If the study has 3,900 participants, report: "The sample consists of 3,900 participants; the population is [defined group]." Then proceed with the statistic and its inference to the population.
Takeaway about the errors:
- Do not define the population by a sample size; specify the population clearly.
- Do not generalize a sample proportion to the population without appropriate sampling methodology and uncertainty quantification.
- Use clear, consistent phrasing when describing comparisons (which group is being described, and whether you refer to a population or a sample).

Corrected and clarified phrasing: examples you can use

Example A: Population vs sample
- Population: all adults in country X.
- Sample: a randomly selected group of 1,000 adults from country X.
- Reported statistic: $\hat{p} = 0.62$ (58–66% in 95% CI) for frequently engaging in activity Y.
- Proper interpretation: "In this sample of 1,000 adults, 62% reported engaging in activity Y. This provides an estimate of the population proportion $p$ with a confidence interval (e.g., 95% CI)."
Example B: Comparing to another population
- If you want to compare to the UK population proportion, ensure you have data for that population and report both estimates with CIs, or state clearly if the comparison is qualitative.
- Correct phrasing: "The sample proportion in our study is $\hat{p}=0.60\pm SE$ . The UK population estimate is $p_{UK}=0.65$ (with its own CI). Direct comparison should consider sampling error and context."
Example C: Misinterpreted statement corrected
- Incorrect: "9 out of 10 people in the UK recycle." (unverified universal claim)
- Correct: "In the sample, 9 out of 10 respondents ( $\hat{p}=0.90)) frequently recycle."; If you want to generalize to the population, you need proper sampling and a reported confidence interval.</li></ul></li></ul><h3 collapsed="false" seolevelmigrated="true">Practical guidelines for reading and writing these statements</h3><ul><li>Always identify:<ul><li>Population: who is the target group?</li><li>Sample: who was actually observed?</li><li>Statistic: what numeric value is reported from the sample?</li><li>Parameter: what population value is the target of estimation?</li></ul></li><li>Use consistent language: write clearly whether you are describing a sample statistic or a population parameter.</li><li>Report uncertainty: include standard error or confidence intervals when inferring from a sample to a population.</li><li>Be careful with comparisons: when using phrases like "compared to X," specify the comparator's population and ensure the comparison is valid given sampling design.</li></ul><h3 collapsed="false" seolevelmigrated="true">Quick practice prompts</h3><ul><li>Given a study with a sample size$ n=500 $and 220 report frequent recycling, compute the sample proportion:$ \hat{p}= \frac{220}{500} = 0.44 $.</li><li>Compute the 95% confidence interval for the population proportion using the normal approximation:<ul><li>$ CI = \hat{p} \pm z_{0.025} \sqrt{\frac{\hat{p}(1-\hat{p})}{n}} $with$ \hat{p}=0.44 $and$ n=500$$.
Interpret the result in terms of population inference, noting any assumptions and potential limitations.

Summary: key points to remember

Population vs Sample definitions are essential for correct interpretation.
Parameters describe population; statistics describe samples.
A proportion like "9 out of 10" is a fraction that must be tied to a defined sample or population and must be reported with context and uncertainty.
The exact population size (e.g., 3,900) does not automatically define the population; clarify whether that number is a sample size or population size.
Always phrase results with clear reference to whether you are discussing a sample statistic or a population parameter, and report uncertainty where appropriate.