Small size sampling

Rakesh R. Pathak

Abstract


Based on the law of large numbers which is derived from probability theory, we tend to increase the sample size to the maximum. Central limit theorem is another inference from the same probability theory which approves largest possible number as sample size for better validity of measuring central tendencies like mean and median. Sometimes increase in sample-size turns only into negligible betterment or there is no increase at all in statistical relevance due to strong dependence or systematic error. If we can afford a little larger sample, statistically power of 0.90 being taken as acceptable with medium Cohen’s d (<0.5) and for that we can take a sample size of 175 very safely and considering problem of attrition 200 samples would suffice.


Keywords


Systematic errors, Distribution, Sample size, Probability

Full Text:

PDF

References


Kenny David A. The two group design. In: Kenny David A, eds. Statistics for the social and behavioral sciences. Boston: Little, Brown; 1987:215.

Sample size calculator. Available at http://www.surveysystem.com/sscalc.htm. Accessed 2 August 2012.

Sample size calculator. Available at http://www.raosoft.com/samplesize.html. Accessed 2 August 2012.

Sample size table. Available at http://research-advisors.com/tools/SampleSize.htm. Accessed 2 August 2012.

Glader BE. Glucose-6-phosphate dehydrogenase deficiency and related disorders of hexose monophosphate shunt and glutathione metabolism. In: Wintrobe's Clinical Hematology, 10th ed, Lee GR, Foerster J, Lukens J, et al. (Eds), Baltimore, Williams & Wilkins; 1999:1178.

Mead R. The design of experiments. Cambridge, New York: Cambridge University Press; 1988:620.