Probabilistic sampling

PROBABILISTIC SAMPLING

  • In probabilistic sampling, the law of probability determines which unit of the population is to be included in the sample. But, non-probabilistic sampling is not based on the theory of probability. Hence, it does not provide a chance of selection to each unit.
  • The types of probabilistic sampling are given as follows:
  • Simple random sampling
    • This sampling technique gives each element (unit) an equal and independent chance of being included in the sample.
    • Ex: In a population of 100, each element theoretically has 1/100th chance of being selected.
      • Enumerate all elements in the population.
      • Prepare a list of all elements, giving them numbers in a serial order.
      • Draw sample numbers by any one of the following methods:
        • Lottery method
        • Table of random numbers (Tipett) or
        • Computer/calculator
      • This sampling is suitable only for a relatively small and homogenous population.
  • Stratified random sampling
    • It is used when the researcher wants to ensure that there should be representation of certain characteristics in the sample. For example when we want to study the profile of livestock farmers, to ensure that there is representation from different livestock farmers (viz landless, marginal, small and large size farmers) in our sample, then the livestock farmers are categorized into groups (strata) based on land size and then from each group (stratum) of farmers draw random sample to get required sample size.
      • Population is subdivided into homogenous groups (strata with the desired features).
      • A random sample is selected from within each group (stratum).
      • Pool all the samples, which are selected from various strata to get required sample size.
    • This sampling is suitable, when the population is large, heterogeneous and overlapping in nature.
    • Note: Drawing a sample from each stratum in proportion to the stratum’s share in the total population, is known as Proportionate Stratified sampling. Contrary to this, if a sample gives over or under representation of some strata, it is known as Disproportionate Stratified sampling.
  • Systematic Random Sampling
    • It is similar to random sampling. A random starting point is selected and every Kth unit (for example every 10th unit) is chosen from systematically arranged observations. This selected sample unit determines other sample units.
      • Arrange observations in a systematic order (ascending or descending or alphabetical)
      • Select a sample from first row or stratum by simple random sampling technique.
      • Pick up other samples by identifying observations in constant interval from the first selected sample unit.
    • Ex: To select a sample of 20 farmers from a list of 200 farmers, a number (farmer) between one and 10 is chosen by random method from the systematically arranged list of farmers. Suppose the selected number is 7 then the farmers numbered, 7, 17(10+7), 27(17+10)….197 are selected as sample. Major advantage in this method is that, it is easier to carry out than other random sampling methods.
  • Cluster sampling
    • It is a type of random sampling in which the samples are selected in the same way as it is done in simple random sampling technique. But in this case, sampling units are not individual elements of the population, but group of elements (i.e, each unit is a cluster of population elements). From each selected sampling unit, a sample of population element is drawn by random method.
  • Multistage sampling
    • In this method, items are selected in different stages at random. After the population is initially sampled the resulting sub population is again sampled. This procedure can be repeated as many times as desired to get required characteristics.
    • For example: To know average milk yield of crossbred cows in Tamilnadu, five districts in the first instance are selected at random. Then, of these five districts, 10 villages per district will be selected in the same manner. Now in final stage again by random selection five cows from the list of crossbred cows from each selected village. Then, milk yield of 250 cows all over Tamilnadu is ascertained and average yield is calculated.
      • Domain or Universe – Tamilnadu
      • Districts – I stage -- 5 Districts
      • Villages – II stage -10 Villages
      • Cows – III stage --200 Cows

Last modified: Friday, 1 October 2010, 9:40 AM