Random Sampling Methods

Simple random, systematic, stratified, cluster sampling.

Notes

Random Sampling Methods

Class 11 Statistics — Simple Random Sampling, Restricted Random Sampling Methods

Introduction to Random Sampling

Random Sampling (Probability Sampling)
A method in which every item in the universe has a known chance of being chosen for the sample. The selection of sample items is independent of the person making the study.

Key Takeaways

  • Also known as 'Probability Sampling' because selection works on Law of Probability.
  • There is no room for discrimination in random sampling.
  • Random sample is a replica of the universe; measurements of such a sample estimate the measurements of the population.
  • Two methods: (i) Simple Random Sampling; (ii) Restricted Random Sampling.

Simple Random Sampling

Simple Random Sampling (Unrestricted Random Sampling)
A sample in which every item of the population has an equal chance of being selected. The method of selection shall not favour one item or another.

Merits and Demerits of Simple Random Sampling

Merits

  • No Personal Bias: Every item has equal chance — free from personal bias.
  • Based on Probability: Rules of probability are applicable.
  • Accuracy can be assessed: Magnitude of sampling errors can be estimated.
  • Representative as size increases: As size increases, sample becomes more representative.

Demerits

  • Unsuitable for small sampling: If sample is not large, it may not represent the population.
  • Difficult to prepare sampling frame: May be difficult for large or infinite population.
  • Time Consuming: Numbering population units and preparing slips is time consuming and uneconomical for large populations.

Restricted Random Sampling — The Four Methods

Summary Table — Types of Restricted Random Sampling

Types of Restricted Random Sampling
MethodDescriptionWhen to UseKey Limitation
Stratified Random SamplingPopulation divided into groups (strata); items taken from each group at randomHeterogeneous population; when representation from all subgroups is neededDifficult to determine size of strata; may be expensive
Systematic SamplingEvery nth item selected from complete listComplete list available; simple random sampling not feasibleBiased if periodic features exist in data
Cluster SamplingPopulation divided into clusters; some clusters selected at randomWhen list of all units is difficult to prepareRisk of unrepresentative clusters
Multistage SamplingCluster sampling carried out in multiple stagesVery large populations; when resources are limitedHigh subjectivity; portions of population cut out