Data Sampling for Academic Research

From dissertationwiki
Revision as of 20:10, 27 February 2015 by Admin (Talk | contribs) (Created page with "A sample refers to the selection of a set of people out of an entire population to make inferences about a larger group. According to Rajan Kumar, usually the selection of the...")

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

A sample refers to the selection of a set of people out of an entire population to make inferences about a larger group. According to Rajan Kumar, usually the selection of the subset is identical or representative of the population since each person in population has equal chance of being selected for interview

The size of sample depends on time and resources. Try and take a larger sample since the larger your sample is, smaller will be the bias. There are different varieties of sampling which you can adopt as per your requirements.

Simple random sampling: Here the investigator prepares a complete list of population known as ‘Sampling Frame’ e.g. voter list or census data and then randomly selects a given number of people. For instance you want a sample of 50 students of 500 students, a computer can produce random numbers or a draw of lots can take place.

In systematic random sampling, assume your requirement is 100 students of 500 and you require one of every 50 students. Select any random number between 1 and 50, say 27. The second number would be 27 plus 50 and so on. For the stratified sampling method, sampling frame can include categories such as sex, age, income and religion. Response of variables differs across these subgroups. Draw samples from each of these subgroups. These subgroups comprise ‘strata’, hence the name stratified sampling.

For cluster sampling, say to check the instance of tuberculosis in India, divide the country into East, West, North & South. Secondly, select at random a few states from all regions. Third, divide a state into districts and select a few districts from various regions. Next, randomly select a few villages and cities and finally divide a village/city into households and randomly select a few households. While conducting a sample study, keep the following in mind. If the sample is small, the bias is likely to be large. A larger sample will lead to more precise results. However, both systematic and random errors are a possibility. It is also difficult to measure concepts through the use of sample.

The collective advantages of sampling are as follows: Highly representative if all subjects participate; Can ensure that specific groups are represented, even proportionally, in the sampleby selecting individuals from strata list, possible to randomly select , Inexpensive way of ensuring sufficient numbers of a study.

The disadvantages of sampling include higher administration expenses, a lot of advance planning and documentation, does not provide much concrete information about the product, certain ‘defective’ units may be included in the sample, and overall quality could be erratic. Generally though, the sample approach is effective, especially if you want people’s detailed opinions about a certain issue. This can help take your research forward, add nuance to it, and also suggest fresh areas of research which you might want to include in your overall study. However, the selection of sample is crucial. The manner in which you select your sample will depend on the nature of your research and its objectives.