Sampling Terms

Elements
Population
Sampling
Sampling Frame
Enumeration Units
Sampling Units

Elements

Individual members of a population whose characteristics are to be measured

Population

The set of individuals to which study findings will be generalized

Sampling

Not necessary when all units in the population are identical or resources are unlimited

Sampling Frame

List of the study population

Enumeration Units

Units that contain one or more elements and that are listed in a sampling frame

Sampling Units

Units listed at each stage of a multistage sampling design.

Examples

Employees in a large factory are selected from the factory’s personnel list for a study of job satisfaction. The researchers discuss in their results the implications of their findings for "job satisfaction in contemporary factories." In sampling terminology, employees are the elements, the personnel list is the sample, and employees in contemporary factories are the target population.
In a study of DBU students’ attitudes, a group of classes was selected and then students within those classes were questioned.
DBU classes are the primary sampling units, and students are the secondary sampling units.

Sampling Error

Any difference between the characteristics of the sample and the population from which it is draw.
The larger the sampling error, the less representative the sample.

Representative and Unrepresentative Samples

Representative Sample

Sample that "looks like" the population from which it was drawn
The distribution of characteristics of elements in a representative sample is the same as the distribution of those characteristics among the total population of elements.

Census

Researchers study all of the elements of a population instead of taking a sample

Nonprobability or Probability Sampling

The most important distinction that needs to be made about samples is whether they are based on a nonprobability or a probability sampling method

Probability of selection

Likelihood that an element will be selected from the population for the sample

Probability sampling method

Sampling method where nothing but chance determines which elements are included (random sampling)
Allows one to know in advance how likelihood of selection
Eliminates systematic bias
Larger the sample, the more homogeneous the population, the more representative

Nonprobability sampling method

Sampling method where likelihood of selection is not known

Election Outcomes:Predicted & Actual

Chance = Random

Random sampling

The probability for selection is equal for every element
systematic
chance
probability
representative

Simple random sampling

Strictly on the basis of chance
In a class exercise, students in a research methods class were asked to wear numbered id tags and to collect sample data on the ages of members of the class. using a random numbers table, each student selected 10 numbers. Students then contacted those students matching the selected numbers and asked their ages.

Least sampling error

In the exercise described above, students’ sample means were compared to the population mean (computed by the teaching assistant). The sample mean closest to the population mean had the least sampling error.

Nonrespondent

In the exercise described above, a student who was not willing to give his or her age would be called, in sampling terms, a nonrespondent.

Replacement sampling

Sample in which elements are returned to sampling frame and may be sampled again

Systematic random sampling

After the first case is selected, every nth case is selected for the sample, where n is the sampling interval.

Cluster sampling

A probability sampling method
Random, multi-stage sampling procedure

Cluster Sampling

Rossi’s Homeless Cluster Sample

Stratified Random Sampling

Uses information known about the total population
gender, race, religion…

Proportionate Stratified Sampling

the proportion of each stratum that is included in the sample is proportionate to the general population
You must draw a sample of 1,000 from the population of a large state in order to estimate insurance coverage. A list of all state residents and their incomes is available from the state census bureau. Use a proportionate stratified random sample, using income for strata.

Sample Quality Questions

From what population were the cases selected?
What method was used to select cases?
Do the cases studied represent the population?

Partial Sampling Distribution:
Mean Family Income

Normal Sampling Distribution:
Mean Family Income

The Effect of Sample Size on Sampling Distributions

Availability sampling

Accidental or convenience sample

Quota sampling

Quotas are set to ensure that the sample represents certain characteristics.

Quota Sampling

Purposive sampling

Key informant survey
targets individuals particularly knowledgeable
Each sample element is selected for a purpose

Snowball sampling

Respondent-driven sampling

Respondent-Driven Sampling

Inferential Statistics

Estimate how likely it is that a statistical conclusion based on a random sample is representative of the population

The sample size must be larger:

Less sampling error (rather than more) is desired
An analysis involving many variables (rather than few) is planned
When a sample random sample (rather than a stratified random sample) is drawn
A more heterogeneous population (versus a more homogeneous population) is sampled