The Underappreciation of Small Sample Sizes
• John Vandivier
This article is a discussion of statistical methods and it argues that statistical analysis can and should be utilized more often for small sample sizes.
The Unremarkable Difference Between n = 3 and n = 30
There is a large difference between 0 and 3 observations. There is not much difference, however, between a sample size of 3 and a sample size of 30, other than the power of the test including the confidence in resulting estimates.- 1 sample means we can form rational estimates.
- 2 samples helps us estimate in which ways a population may be heterogeneous and homogenous. 2 samples is also the minimum number of points needed to create a line.
- 3 samples allows us to begin to asses dominance, or a usual case even in the face heterogeneity, which is a main point of statistical analysis. 3 samples is also the minimum number of points needed to create a curve.
- More than 3 samples allows increased confidence.