Research Populations

External Validity

Ability to generalize the results of a research study to people other than those in the study

What is the goal of sampling

maximize external validity

Stratum

Subsets of the population with common characteristic

Sample

goal of population elements used in a study

Target population

general population the study seeks to understand

Source population

specific individuals that represent the sample

Sample population

individuals asked to participate

Study population

eligible participants

Sampling error

difference in the distribution of characteristics between a sample and population

Probability Sampling

Every element of sampled population has a known probability of being included in a sample

Simple random sampling

all population elements have the same chance at being selected

Stratified random sampling

source population is divided into subgroups and simple random samples

Cluster Sampling

random sample of pre-existing clusters/groups are selected from a larger list of clusters within a population

Nonprobability Sampling

no way to estimate the probability each element has of being included in the sample

Population size is unknown

Convenience sampling

sampling method in which the most easily/readily available individuals are selected

Quota sampling

Subgroups of interest are identified, then quotas are set for the number of individuals to be collected within each subgroup

Snowball sampling

members of the initial sample form the target population enlist other members of the target population who meet inclusion criteria

Probability sampling is better because it is more likely to produce a

representative sample of the population

Non-probability is often more efficient, requires

fewer resources

Law of Large Numbers

the larger a sample, the more likely values form the sample will be similar values of the population

Importance of sample size if it’s too big or too small

waste resources

Larger sample sizes gives you

narrower ( more precise) confidence intervals for statistical measures

More statistically significant results

Power

ability of a statistical test to detect significant differences between subgroups of a population when differences really do exist