Of course data can be manipulated
This is pretty straight forward stuff though. You comment could be interpreted to mean all statistical analysis all the time are worthless, which is preposterous. In this particular case, unless the actual data that is input is skewed or falsified in some way, the results are accurate. And those results are also telling.
Another data point I read recently was that over 50% of self proclaimed Tea Party members who are on social security and medicare say they get no government assistance whatsoever.