Statistics is how observations are manipulated to describe a population and to draw conclusions about that population.   

Both descriptive and inferential statistics may be used to express truth or to simply make a point.   In either case an impression is made on the people who consider those statistics.   

The population sample from which observations are obtained is critical for describing the population accurately.   

A person who wants to accurately describe a population must poll every member or take a truly representative sample of that population.   

A person who wants to make a point arbitrarily selects the sample in order to ensure the statistics bear out the point they want to make.   

An example of how statistics may be manipulated and generate incorrect conclusion is this:
Is the forest in front of me hardwood or conifer?   
→ If I count y type every tree as to its type I obtain an indisputable conclusion
→ If I count by type only the tall trees my conclusion is disputable
→ If I count by type only the visible trees from outside the forest my conclusion is disputable
→ If I count by type every 5th tree my conclusion has a confidence level associated with it because my scheme assumes even distribution of hardwoods and conifers throughout the forest