It is convenient to measure or estimate variation in samples or distributions in units of standard deviations. There are alternative methods of estimation of standard deviation aside from the conventional and well-known definition. Estimation of standard deviation for very small samples (as small as two), whereas not always ideal, might be useful in certain practical circumstances. A simulation study of five standard deviation estimation methods and the effects of small sample sizes 2 through 6 is made through description of sampling variations and biases. Consideration of the unbiased mean, small median biases, and algorithmic simplicity, show logically that Gini mean difference estimation is a competitive method.