SAMPLING METHODS IN RESEARCH
WHY SAMPLING IS IMPERATIVE
In most research procedures, it is not possible to investigate 100% of the components of the entire units that need to be researched. The cost, the lack of time and the sheer amount of work involved would make it almost impossible. Researchers have, therefore, found a way around this problem by selecting what could be regarded as representative units of the population or universe. The alternative will be to carry out a complete census, which means studying every component unit of the population. This could be prohibitively expensive.
Population (Universe ): This refers to any group of people or objects which are similar in one or more ways, and which form the subject of research study in a particular survey (Chisnall, 1975:54). Population units could be made up of individuals, households, companies, factories, transactions, ethnic groups, category of employees, etc.
The practice of sampling presumes that, if carefully selected, a sample drawn from a population could reflect the general tendencies of the population from which it is drawn.
Sample: This refers to a subset of a larger population. This occurs when a number of sampling units (fewer than the aggregate) is drawn from a population and examined in some detail (Sudman and Blair, 1998:723) Chisnall (1973:55).
Advantages of Sampling
• Reduced costs
• Greater speed
• Greater accuracy
• Greater depth of information
• Preservation of units, especially in the testing of products such as ammunition.
Sampling Theory: This is concerned with the study of relationships existing between the population and units or samples drawn from it. Using statistical inference and employing probability theory, researchers hope that certain conclusions can be drawn about a population from a study of samples drawn from it.
Sampling Frame: Before a study, using sampling procedure, is undertaken, it is important to define the population that is to be surveyed very carefully. Yates (1953) has provided a general framework that researchers have to note in designing a sampling frame as follows:
- Adequacy - A sampling frame should cover the population to be surveyed and it should do so adequately as related to the purposes of the survey.
- Completeness - Sampling frame should include all units that should be included; otherwise, the missing units will not have the opportunity to be selected, thus, rendering the sample biased to that extent.
- No Duplication - No unit should be counted or entered more than once. Doing so can result into bias.
- Accuracy - Non-existent units should not be included.
- Convenience - the sample list or units should be accessible and suitable for the arrangement and purposes of the sampling.
Types of sampling
- Simple Random Sampling
- Systematic or Quasi-random Sampling (proportionate, disproportionate)
- Cluster Sampling
- Area Sampling
- Multistage Sampling
- Replicated Sampling
- Master Sampling
- Multiphase Sampling
Although the attempt to list so many types of sampling methods may impress you, the list is not exhaustive. There are other types.
Obtaining Primary Data on Attitudes
• By Use of Numbers
• Nominal Scale
• Ordinal Scale
• Internal Scale
• Ratio Scale
• By Measuring Attitudes Using:
• Semantic differential
• Projective techniques
• Thurston differential
• Likert Attitude Scale
• Q sort
• By Measuring Salience Using:
• Comparative weights
• Paired comparisons
• Trade-off analysis (conjoint analysis)
• Multi-dimensional scaling |
