Best Sample Size for Dissertations
Survey is now a basic tool in social sciences and in some fields of specialization which reliance is from the sampling procedures. Little or unacceptable knowledge will be gained if the sample size is poorly designed and executed: no matter how good the questions are and no matter how impressive the analysis is (Kalton, 1987, p. 4).
The important question of how to determine the size of the sample is vital for estimating the parameters (Singh and Chaudhary, 1986, p. 38). Our common sense would suggest that a large sample size will be better than a small one since an increase in sample size will decrease the sampling error.
This is not always the case because, for example, having a large sample size with a sampling frame with very similar types of respondents will just be a waste of time. Determining the best sample size will thus depend on the compromise of practical constraints (i. e. money, time, ethical issues) and theoretical considerations. For the simplicity of analysis, given that there are no practical constraints, the following are the rule of thumb: 1. ) If the population is about 50 or less, it is best to sample the whole population; 2. If you have to sample a population of 50 or more, then try to sample around 30 and 3. ) If you will be using stratified sampling, aim to have at least five sampling units from each stratum or category (White, 2002, p. 65-66). With the load of work dissertation requires researchers can not help but to consider the practical constraints. The credibility of the study can thus be sacrificed so it must be the goal of the researcher to find ways to maintain credibility while considering constraints.
Researcher bias is another issue that may hinder the determination and use of sample size because unless it is a purposive sampling, samples should represent the whole population as much as possible. For example, a bias researcher will decide a sample size that is easy for him to gather and that is fit for his/her desired results. There can never be a perfect or right sample size but there is always a best sample size depending on the nature of your study and on the given population.