Overview
This lecture covers additional sampling methods for collecting data in statistics, focusing on cluster, stratified, and systematic sampling techniques, their features, and potential biases.
Cluster Sampling
- Cluster sampling collects data from multiple groups (clusters) within the population rather than individuals one by one.
- Groups (clusters) must be randomly selected to minimize bias and ensure all population members have a chance of inclusion.
- An example is randomly selecting several classes at a college and surveying every student in those classes.
- If groups are not randomly chosen, such as only using classes taught by the researcher, the sample becomes biased.
- Using only one group (e.g., a single class) is not true cluster sampling and resembles convenience sampling.
Stratified Sampling
- Stratified sampling is used for comparison studies between different subgroups or populations.
- Divide the population into strata (groups) and take a simple random sample from each group.
- For example, compare average salaries by taking random samples from California and Arizona working adults.
- Stratified samples do not require equal sample sizes from each group; the focus is on random selection within each group.
- Choosing only friends or acquaintances as samples introduces bias rather than representing the groups.
Systematic Sampling
- Systematic sampling involves selecting subjects using a regular interval or system, not pure random selection.
- For example, choosing every 5th person who enters a store introduces potential bias if not all members of the population are equally likely to enter.
- A better approach is to select every nth person from a complete, ordered list of the population.
- To improve randomness, randomly choose a starting point, then select every nth person (this can produce a simple random sample).
- Systematic sampling can be biased if the system excludes segments of the population.
Key Terms & Definitions
- Cluster Sample — Sampling by randomly choosing groups (clusters) from the population and surveying all members in those groups.
- Stratified Sample — Sampling method where the population is divided into subgroups (strata) and random samples are taken separately from each.
- Systematic Sample — Sampling method that selects items from an ordered list at regular intervals (e.g., every 50th person).
Action Items / Next Steps
- Review examples of cluster, stratified, and systematic samples for better understanding.
- Prepare for next lecture on bias in sampling and how poor sample methods impact data quality.