sampling

"There are three kinds of lies: lies, damn lies and statistics"

This web page is not providing an introduction in statistics, but in sampling and testing. Statistics is a science dealing with uncertainties. And probably because of that statistics is often associated with mending or bending the truth. One of the most often used quotes on statistics is "There are three kinds of lies: lies, damn lies and statistics", attributed to Benjamin Disraeli. Googling for this quote yields roughly a million references, but many do not point so much to statistics, but to the other parts of Disraeli's statement: lies.
Statistics does not have a particular good press - see again the quote above - and this may be related to its nature: dealing with uncertainties. So, where I know how I created this text (on a qwerty keyboard linked to a pc), you, the reader, can only guess. You may be able to work out that there is over 80 % chance that it was written on a pc using a keyboard, and that the keyboard was probably a qwerty keyboard. But than we're getting into statistics, dealing with uncertainties.

On sampling and seed testing

Obvious as it may seem, it is not a simple task to make a statement on the quality of a product like seeds. Testing of seeds typically is a destructive process. To achieve 100 % certainty all seeds should be tested, but this goes with the inevitable problem that after testing nothing is left to sell. Therefore only samples are tested and the results of such tests predict quality with a certain level of uncertainty.