Sampling Concepts from Ryzin Ch 5 (Video)

Vocabulary flashcards covering key concepts from sampling, generalizability, sampling methods, biases, and related statistics.

Generalizability (External validity)

The extent to which findings can be generalized to other populations, contexts, and times beyond the study.

Population of interest

The complete group the study aims to learn about—the true population characterized by the study’s goals.

Sampling frame

A list or map representing the population from which the sample is drawn.

Random sample

A subset selected so every member of the population has an equal chance of inclusion, increasing representativeness.

Census

Data collection from the entire population, rather than a sample.

Sample vs Population

A sample is a subset of the population; the population is the entire group of interest.

Sampling variability

Natural differences that occur between different random samples drawn from the same population.

Standard error

The standard deviation of the sampling distribution; measures how much a sample statistic will vary from the population parameter; decreases with larger samples.

Confidence interval

A range around a sample statistic within which the population parameter is expected to lie with a specified probability (e.g., 95%).

Margin of error

Half the width of a confidence interval; roughly the amount the estimate could be off due to sampling variability (often linked to $2 \cdot SE$ for 95% CI).

Simple random sampling

A method where each population member has an equal probability of selection, often via random numbers or drawing names.

Systematic sampling

A sampling method that selects every $k^{th}$ unit from a frame after a random start.

Stratified sampling

The population is divided into strata (groups), and random samples are drawn from each stratum to ensure representation.

Disproportionate sampling / oversampling

Sampling more from certain groups than their share in the population to improve analysis; requires weighting in analysis.

Multistage sampling

Sampling in stages (e.g., clusters, then households) to study a group efficiently over time.

Cluster sampling

Randomly select clusters (groups) and then sample within those clusters.

Snowball sampling / respondent-driven sampling

Start with a few participants who refer others; useful for hard-to-reach populations.

Quota sampling

Nonprobability method that fills predefined quotas for subgroups; may bias results due to nonrandom selection.

Purposive sampling

Deliberately selecting participants with specific characteristics or roles to gain insight into particular issues.

Nonprobability sampling

Sampling methods that do not give all units a known or equal chance of selection (e.g., voluntary, convenience, snowball, quota).

Coverage bias

Bias when the sampling frame systematically excludes or overrepresents certain groups relative to the population.

Nonresponse bias

Bias arising when individuals who do not respond differ in relation to the study’s variables.

Propensity to respond

The likelihood that a person will participate in a survey; related to potential bias if linked to what is being studied.

Exit polls

Polls conducted at polling places as voters leave to gauge election results; rely on a specific sampling frame of voters.

Parameter vs. Statistic

A parameter is a population characteristic; a statistic is a value calculated from a sample to estimate the parameter.

Replication

Repeating a study with different samples, in different places, times, or designs to test the stability of results.

Meta-analysis

A method of combining results from multiple smaller studies to produce a larger overall estimate.