Interviews to prove a point, especially with random people, are generally meaningless. Unless video footage is provided of the entire interview (approaching someone, asking the question, and getting the whole response), they don’t say much. With editing, the viewer has no idea what background was given to the interviewee, or what question was asked.
Let me provide an example. If I wanted to show that most people are against animal cruelty, I could precede my question with a detailed explanation of how chickens are raised just to be killed. I would throw in incendiary language such as “slaughtered” or “murdered,” and give a negative impression of the chicken farming process. As a result, I would get a lot of responses agreeing with me and be able to edit together a video showing how the majority of people interviewed support the rights of chickens.
This same idea applies to polls. By phrasing the question differently, I can easily pull out the results I want.
Whenever you see a video or poll like this, be skeptical and understand what might be going on that you can’t see.
Chocolate Frosted Sugar Bombs 