Understanding the Distribution of Sample Means

Chapters 7 and 8 serve as the central bridge to the remainder of the material in the statistics course.
Chapter 7 focuses on the distribution of sample means, which is the fundamental concept that enables the practice of hypothesis testing.
Chapter 8 introduces the specific methodologies and steps of hypothesis testing.
Once the basics of hypothesis testing are understood in Chapter 8, every subsequent chapter in the class will involve a different form of hypothesis testing:
- Chapter 9: Hypothesis tests for a specific kind of study design.
- Chapter 10: New hypothesis tests for study structures where Chapter 9 methodologies are not applicable.
- Chapters 11 and 12: Continued exploration of various hypothesis tests until the end of the course.
- Correlation and Regression: Even these topics at the very end of the course will incorporate hypothesis testing elements.

The course is currently moving out of descriptive statistics (Chapters 1–6) and into inferential statistics.
Inferential Statistics Definition: The process of studying a sample, calculating values regarding that sample, and then using those findings to make conclusions or inferences about a larger population.
In research, studying every member of a population is rarely possible. Therefore, research involves studying samples to make inferences about populations.
Polling Example (Fall 2020 context):
- A pollster might sample $100\,\text{people}$ to ask how they plan to vote.
- Inferential statistics is the step of using that sample of $100\,\text{people}$ to draw conclusions about the voting behavior of the entire country, which has approximately $300,000,000\,\text{people}$ (total population) or roughly $150,000,000\,\text{voters}$ .
Clinical Drug Trial Example:
- A researcher gives a drug to a sample of $30\,\text{people}$ with depression.
- The goal is to draw a conclusion about the drug's effectiveness for every person in the population who suffers from depression.

The Distribution of Sample Means is defined as the collection of sample means for all the possible random samples of a particular size $n$ that can be obtained from a population.
This concept can be highly abstract, but it is best understood by visualizing the process step-by-step.
Key Components of the Definition:
- Population: You start with a population of scores (e.g., ages of students in a class or number of siblings).
- Sample Size ( $n$ ): You must select a specific size for the samples you are drawing (e.g., picking $3\,\text{students}$ out of a class of $30$ ).
- All Possible Random Samples: You must account for every possible combination of students/scores of that fixed size $n$ .
- Collection of Means: You don't just collect the samples; you calculate the mean for every single one of those possible samples. That collection of means forms the distribution.

Consider a very small population of scores ( $N = 3$ ):
- Individuals measured on the variable "number of siblings."
- Population scores: $1, 2, 3$ .
Consider a sample size of $n = 2$ .
Sampling with Replacement: Based on the requirements for random sampling (introduced in Chapter 6), we must sample with replacement. This means a single score can be selected twice for the same sample.
Generating All Possible Samples of $n = 2$ from the Population (1, 2, 3):
1. $(1, 1)$
2. $(1, 2)$
3. $(1, 3)$
4. $(2, 1)$
5. $(2, 2)$
6. $(2, 3)$
7. $(3, 1)$
8. $(3, 2)$
9. $(3, 3)$
There are $9\,\text{possible samples}$ in this specific scenario.

To create the distribution of sample means, we calculate the arithmetic average for each of the samples generated above:
- Sample (1, 1): $M = \frac{1 + 1}{2} = 1.0$
- Sample (1, 2): $M = \frac{1 + 2}{2} = 1.5$
- Sample (1, 3): $M = \frac{1 + 3}{2} = 2.0$
- Sample (2, 1): $M = \frac{2 + 1}{2} = 1.5$
- Sample (2, 2): $M = \frac{2 + 2}{2} = 2.0$
- Sample (2, 3): $M = \frac{2 + 3}{2} = 2.5$
- Sample (3, 1): $M = \frac{3 + 1}{2} = 2.0$
- Sample (3, 2): $M = \frac{3 + 2}{2} = 2.5$
- Sample (3, 3): $M = \frac{3 + 3}{2} = 3.0$
This resulting set of means ( $1.0, 1.5, 2.0, 1.5, 2.0, 2.5, 2.0, 2.5, 3.0$ ) is the distribution of sample means for $n = 2$ .

A distribution of sample means should always be identified by its sample size ( $n$ ).
A sample mean of $1.0$ does not reveal how many scores were in the sample; it could represent $2\,\text{scores}$ , $15\,\text{scores}$ , or $1,000\,\text{scores}$ .
Large-scale populations make manual calculation impossible:
- In the example above, a population of $3$ and a sample of $2$ resulted in $9\,\text{means}$ .
- If the population was $20,000,000\,\text{voters}$ (e.g., a state like California) and the sample was $100\,\text{voters}$ , the number of possible samples is a number so large it does not have a common name.
Because we cannot manually calculate every possible mean for large populations, we rely on the Central Limit Theorem.
The Central Limit Theorem provides information about the distribution of sample means without requiring the calculation of every possible sample.

The Parent Population refers to the original distribution of scores from which the samples were drawn.
In the logic of Chapter 7, the distribution of scores "gives birth" to the distribution of sample means.
This term is specifically used in the context where a distribution of scores is being used to derive all possible samples of size $n$ .
Samples are drawn from the parent population, and the means calculated from those samples form the new distribution.