A friend and I are thinking of a talk about the books by Nassim N. Taleb. He is a trader and probabilist. Our thinking broadened to include "How to Lie With Statistics" by Darrell Huff (1954). The copyright date plus the fact that the book is still in print and is also available in Kindle format indicates that book's popularity and usefulness.
Many students dread statistics. Most of the mathematical expressions in the subject look scary, especially for those used to arithmetic. But in Huff's book, there are no equations. You don't have to look at
[Sigma of Xi, i=1, N]/N
You can read "Add the numbers and divide by how many there are."
Fifth graders often forget the division and find the average to be much greater than any figure in the set.
Finding the average is a good example of sampling. We can't remember all the numbers. We want one figure to represent all those in the group, the sample, the population, the community, the nation, the world, whatever. So, we try to choose an single number that is "close to them all."
If you want to get technical, and most likely you don't, get hold of the book "75 measures of central tendency". How to Lie with Statistics makes clear that the arithmetic mean (not the geometric mean or the harmonic mean), the median and the mode are all often called the "average", after choosing whichever one seems to serve the purpose of the communication or analysis best.
But sampling is deeper. Look around. You can only take in a sample of what is to be seen. Use a microscope for a different set of views. Travel to the moon and look (you will only see a dot). Or, read a book and tell us what you read. Ok, just a sample, just what lingers in memory. Sure, it is a personal sample, not a random one, but it is a sample.
What happened yesterday? We only want a sample. Just what you can remember of what happened to you, not counting when you were sleeping or what happened on another continent. Or planet. It's all samples!
