Anyone remotely familiar with the scientific method understands that just like a ruler or a telescope, statistics is a tool. Scientists use the tool primarily for one purpose: To answer the question, "Is my data meaningful?" Properly used, statistics is one of science's most powerful tools. But used improperly, statistics can be highly misleading.
What is a scientific poll? First, it is a misnomer. There is nothing scientific about a poll. Second, it is conducted using sound statistical techniques. What's more, savvy politicos know that not just any poll will do.
Gallup's recent poll on race relations asked a loaded question and lacked a control group. This is a recipe for very bad social science.