Why studying central tendency alone is insufficient and calculating dispersion is always needed in both descriptive and inferential statistics? What does a small value of standard deviation indicate?