|
How to conduct a survey.
This isn't gonna be very fleshed out, but here it goes.
I am writing this to hopefully influence those who post poll questions here on the TFP. People seem to use a little bias when filling out their poll options. True, not enough people actually vote on these polls to make them a basis for any scientific deduction, but for the sake of being anal, let's go over this.
Bias is impossible to escape. The very fact that you want to ask a question insinuates bias. The more obvious influence of bias can be seen in the questions asked.
Bad: "Do you like sex?" - (yes, it feels so good) (no, i'm a prude) (don't know, i'm a loser virgin)
Good: "How do you feel about sex?" - (positive) (neutral) (negative)
Yes, asking questions without bias can be very ambiguous, so polls aren't always great scientific means of gathering data. Sometimes an open-ended question will do the trick, but it needs to be very open ended. Asking a yes/no question is not open-ended.
Bad: "How do you like my new car?"
Good: "Please give your opinion on this car"
Argh, I know! Being unbiased means stripping the question of any semblance of personality and interesting terminology. So boring! However, if you want people to tell you the TRUTH, then you cant load your questions with anything that would influence them to answer in any particular direction. You see, people like to please the surveyor. If you go, "Do you like my new sunglasses?" chances are, the avg person isn't gonna go, "I think those sunglasses are a little too effeminate for you." instead they are gonna go, "Yeah, sure, I like them." I hope you are following me.
So anyways, the moral of the story is... if you want the TRUTH, you gotta sterilize your questions. If you want an ego boost, then you gotta go for the bias, big time, 'cause those sunglasses are WAY too effeminate for you.
|