Any time you are using statistical sampling in order to estimate the qualities of a larger group, there is the opportunity to take multiple samples and present only the results that support a particular view. The samples will have random variation, so if MSNBC takes 10 polls and the average is a tie, there will be one which shows Obama ahead by 3 points. That is the one they will choose to make public, rather than the one which shows Romney ahead by 3 points. Sorry you don’t like FREEPERS, maybe you should post comments at a different site.
You are a bundle of joy.