Data Saturation and Sample Size and its Limitation on Qualitative Research
Education research is primarily qualitative. This is due to the fact that the data to be studied is already available from secondary sources of information. In qualitative research, a small size is normally chosen compared to quantitative research which factors in large sample sizes from the data is collected. Another primary problem that arises in qualitative research is that of data saturation. The issue of data saturation comes up rampantly in PhD theses. Data saturation refers to a situation where the researcher can no longer hear or see new information from the research that he or she is carrying on. There are various factors that have been put forward to date that lead to clues which may explain the limitations due to the occurrences of data saturation and sample sizes. This paper will analyse and discuss data saturation and sample size as well as the limitations in qualitative research with respect to the PhD research paper on Education.
This paper aims to develop a thesis on the ability of training children who are born with autism and are living with it; using virtual reality technology. Data has to be collected first and foremost on the number of children that live autism from the sample population. Then from the children population, the study will consider a sample that will be for the assessment training. Qualitative research is normally labour intensive and time consuming hence the need for sampling. According to the recommendations made on sampling while using the qualitative research methodology, the smallest acceptable sample should be fifteen. The sample size to be considered has to be small. This is due to the fact that in choosing a very large data sample it can be quite repetitive and superfluous hence leading to research results that are highly inaccurate and unreliable. However, there are various disadvantages or limitations that normally arise during sampling with respect to the size of the sample. Besides size, the sample relies on other factors namely homogeneity and convenience.
There are a number of limitations associated with a small sample size. From a given sample size, tests that are based on the statistics should yield traits that exist truly in the population of the statistic. This is what is referred to as the statistical power. There exists a correlation between the sample size and the statistical power. A decline in the size of the sample leads to a decline in the statistical power. This may yield unreliable results contrary to what is sought after by the researchers. Another limitation is referred to as Type II Error which represents a “false negative”, that is, the result of a given statistic is not true. There is a correlation between sample size and type II error. When the sample sizes it too small, the type II error increases commensurately. The limitation of significance also emerges. Significance is whereby the difference is relatively large for it to matter. Larger samples can detect significant differences between statistical values compared to smaller samples hence they are preferred. The sample population must at times be subdivided and tested in scenarios or conditions that are different which may lead to a severe problem in the distribution of data.
Several factors that lead to data saturation can be advanced. They include the scope of study, the nature of the topic under research, the quality of data that has been collected, the and the study design. Furthermore, the problem may be exacerbated by the use of data that is shadowed. The heterogeneity of the population and the availability of a budget and resources are other factors that may take centre stage. Data saturation is when data begins to be repetitive during data collection and no new information is presented; and responses begin to overlap (Bogdan &Biklen, 2007). Only five teachers participated in the study limiting the opportunity to reach data saturation. Although data saturation was not met, the study sample was sufficient because the collected data presented a sufficient amount of information displaying the benefits of VR as an intervention tool.
Bogdan, R. C., & Biklen, S. K. (2007). Qualitative research for education: An introduction to theories and methods (5th ed.). Boston, MA: Allyn & Bacon
Giuffrida, M. A. (2014). Type II error and statistical power in reports of small animal clinical
trials. Journal of the American Veterinary Medical Association, 244(9), 1075-1080.
Horvath, S. (2011).Association Measures and Statistical Significance Measures. In Weighted
Network Analysis (pp. 249-277). Springer New York.
Mason, M. (2010, August). Sample size and saturation in PhD studies using qualitative
interviews. In Forum Qualitative Sozialforschung/Forum: Qualitative Social
Research (Vol. 11, No. 3).
Murphy, K. R., Myors, B., &Wolach, A. (2014).Statistical power analysis: A simple and
general model for traditional and modern hypothesis tests. Routledge.
Neuman, W. L. (2003). Social research methods (5th ed.). Upper Saddle River, NJ: Prentice Hall.
O’Reilly, M., & Parker, N. (2012). ‘Unsatisfactory Saturation’: a critical exploration of the
notion of saturated sample sizes in qualitative research. Qualitative Research, 1468794112446106.
Suri, H. (2011). Purposeful sampling in qualitative research synthesis.Qualitative Research
Journal, 11(2), 63-75.