Appropriate sampling method 2 / 2

Sampling method in a specific situation

  1. Random Sampling

    A simple random sample is defined as a sample taken in such a way that every member of the population has an equal chance of being selected. The normal way of achieving this is by numbering each item in the population.

    If a sample of, say 20, items is required then 20 numbers from a table of random numbers are taken and the corresponding items are extracted from the population to form the sample (sampling frame) 

    e.g. in selecting a sample of invoices for an audit. Since the invoices are already numbered, this method can be applied with the minimum of difficulty.

    This method has obvious limitations when either the population is extremely large or, in fact not known. The following methods are more applicable in these cases.

  2. Systematic Sampling

    If the population is known to contain 50,000 items and a sample of size 500 is required, then 1 in every 100 items is selected. 

    The first item is determined by choosing randomly a number between 1 and 100 e.g. 67, then the second item will be the 167th, the third will be the 267th... up to the 49,967th item.

    Strictly speaking, systematic sampling (also called quasi-random) is not truly random as only the first item is selected randomly. However, it gives a very close approximation to random sampling and it is very widely used.

    There is danger of bias if the population has a repetitive structure. For example, if a street has five types of house arranged in the order, A B C D E A B C D E... etc, an interviewer visiting every fifth home would only visit one type of house. 

    Systematic sampling should not be used if the population follows a repetitive pattern.

  3. Stratified Sampling

    If the population under consideration contains several well defined groups (called strata or layers), 

    e.g. men and women, smokers and non-smokers, etc, then a random sample is taken from each group. 

    The number in each sample is proportional to the size of that group in the population and is known as sampling with probability proportional to size.

    For example, in selecting a sample of people in order to discover their leisure habits, age could be an important factor. 

    So if 20% of the population are over 60 years of age 65% between 18 and 60 and 15% are under 18, then a sample of 200 people should contain 40 (20% x 200) who are over 60 years old, 130 (20% x 65) people between 18 and 60 and 30 (20% x 15) under 18 years of age.

    This method ensures that a representative cross-section of the strata in the population is obtained, which may not be the case with a simple random sample of the whole population.

    The method is often used by auditors to choose a sample to confirm receivables’ balances. In this case a greater proportion of larger balances will be selected.

  4. Multi-Stage Sampling

    This method is often applied if the population is particularly large, for example all TV viewers in Malta. The process involved here would be as follows:

    Step 1 The country is divided into areas (towns and villages) and a random sample of areas is taken. 

    Step 2 Each area chosen in Step 1 is then subdivided into smaller areas and a random sample of this is taken.

    Step 3 Each area chosen in Step 2 is further divided into roads and a random sample of roads is then taken.

    Step 4 From each road chosen in Step 3 a random sample of houses is taken and the occupiers interviewed.

    This method is used, for example, in selecting a sample for a national opinion poll.  Fewer investigators are needed and hence it is less costly.  

    However, there is the possibility of bias if a small number of occupiers are interviewed.

  5. Cluster Sampling

    This method is similar to the previous one in that the country is split into areas and a random sample taken. Further sub-divisions can be made until the required number of small areas have been determined. 

    Then every house in each area will be visited instead of just a random sample of houses. In many ways this is a simpler and less costly procedure as no time is wasted finding particular houses and the amount of travelling by interviewers is much reduced.

  6. Quota Sampling

    Quota sampling is a non-probability sampling method in which the chance of each member of the population appearing in the sample is not known.  

    With quota sampling, the interviewer will be given a list compromising the different types of people to be questioned and the number of quota of each type 

    e.g. 20 males, aged 20 to 30 years, manual workers; 15 females, 25 to 35, not working; 10 males, 55 to 60, professionals, etc. The interviewer can use any method to obtain such people until the various quotas are filled. 

    This is very similar to stratified sampling, but no attempt is made to select respondents by a proper random method, consequently the sample may be very biased.

We use cookies to help make our website better. We'll assume you're OK with this if you continue. You can change your Cookie Settings any time.

Cookie SettingsAccept