Skip to content
Sahithyan's S3
1
Sahithyan's S3 — Applied Statistics

Sampling Techniques

Sampling is the process of selecting a subset of individuals or items from a population to make statistical inferences about the whole. It helps researchers study large populations efficiently without examining every individual.

  • Feasibility – Studying an entire population is often impossible due to size or accessibility.
  • Economy – Sampling reduces cost by limiting the number of observations.
  • Speed – Results can be obtained faster than a census.
  • Accuracy – Quick collection helps maintain data relevance and reliability.
  • Conservation of Resources – Prevents exhausting data sources during measurement.
  • If sampling is biased, or not representative, or too small, the conclusion may not be valid and reliable.
  • In research, the respondents to a study must have a common characteristics which is the basis of the study.
  • If the population is very large and there are many sections and subsections, the sampling procedure becomes very complicated.
  • If the researcher does not possess the necessary skill and technical knowhow in sampling procedure.

A good sample must ensure both:

  • accuracy: no bias
  • precision: sample represents the population

A population with diverse characteristics or attributes among its members.

For groups difficult to access:

  • Use mixed sampling methods (snowball + purposive).
  • Maintain confidentiality and anonymity.
  • Gain community trust before data collection.

Every unit in the population has a known, non-zero chance of being selected.

Equal chance for all elements. With or without replacement.

From a population of NN and a sample size of nn, select an item randomly in the first nn items, and then every kkth item after it.

Here k=Nnk=\frac{N}{n} is the sampling interval.

Simple to execute. Evenly spread samples.

Can be biased if hidden periodicity exists.

Population is divided into strata (groups) based on shared characteristics. Inside each stratum, simple random sampling is performed and the results are combined.

Ensures representation of all groups. Provides greater accuracy when population is heterogeneous.

Population divided into clusters. Each represnets a mini-population. Randomly select some clusters, and include all members within selected clusters.

Cost-effective for large geographic areas.

Less precise; may require larger sample to maintain accuracy.

Units are chosen based on subjective judgment, convenience, or accessibility.

Population divided into categories. Fixed number of elements of each category is surveyed.

May introduce bias.

Aka. Purposive Sampling. Researcher selects sample based on knowledge or purpose. Used when expert choice is justified.

Sample selected from easily available respondents. Quick but highly biased.

Existing study subjects recruit future subjects from among their acquaintances. Useful for hidden populations.

Individuals voluntarily participate. Not diverse. Biased.

Drawn to test specific hypotheses or theoretical ideas. Common in grounded theory research.

Entire sampling process repeated at intervals (e.g., periodic surveys). Different samples at each iteration. Allows observation of changes over time but requires large samples.

Aka. Cohort Surveys. Same group studied repeatedly over long periods (longitudinal studies). Issues: fatigue, attrition, and order effects.

Sample is split into rotation groups. In each iteration, one rotation group is replaced with another. Mix of repeat and panel survey.

Below components must be defined before sampling:

  • Target Population
  • Parameter of Interest
  • Sampling Frame: list or source of all population elements.
  • Sampling Method
  • Sample Size

Nonresponse occurs when selected subjects do not provide data.

Can be caused by:

  • Refusal to respond
  • Ineligibility
  • Inability to locate respondent or contact