(1) The stock market has closed up for four days in a row; therefore, it will close up on the fifth day.
(2) It has rained for two weeks straight; therefore, it will rain tomorrow.
(3) The sun has risen for X days; therefore, it will rise tomorrow.
(4) War has occured for three thousand years; therefore, it will occur for the next three centuries.
(5) People who earn a higher income are happier.
(6) People who go to university are smarter than people who go to college.
(7) People who drink more are smarter.
(8) People who drink wine are healthier.
In each case, nothing is "proven" by the statistic, only evidence is given drawing our attention to a correlation. In regards 5-8, one can at best say that these are averages based on a limited sample size and there are countless other factors that unsophisticated stastical models do not take into account that can taint the results. Actually, it turns out that marketers love statistics because it is very easy to design surveys in a way to get a desired result, which can then "intellectually legitimize" the product/service being sold/offered. It therefore takes people like myself and others to point out that this method is not bulletproof, contrary to what you and others feel.