mzadell
Mouseketeer
- Joined
- Apr 18, 2014
- Messages
- 173
Sorry , like I said not only do I not agree with this, I will go with the actual numbers I have seen on polls instead of a theory that is unproven.![]()
The fact that polls not offered to a random and large sample size are statistically inaccurate is not an unproven theory. I'm not saying the results of the poll are wrong, I'm just saying it's a poorly executed survey that can't guarantee accurate results.
Let me try to explain the concept of self-selection bias better:
I put up a booth in Times Square that offers a poll on the effectiveness of subways in New York City. Let's say 100 people walk by the booth and that of those 100 people, 70 people like the subway, 10 people dislike the subway, and 20 people HATE the subway because they had some terrible experience with it. Odds are very good that the 20 people that absolutely hated their experience will go out of their way to take that survey and express their opinions. Let's assume then that half of the rest of the people volunteer to take the survey. This gives us 20 negative responses (the haters), 5 more negative responses, and 35 positive responses for a total of 30 to 35, negative to positive. So by this survey, 46% of people do not like the subway, when in reality, only 30% of people don't like the subway.
Again, I'm not saying it's a fact that these polls are wrong or that MDE/FP+/MB's are perfect and universally loved (nor that they should be), I'm just saying that to say 50% of people who use the service are unhappy with it is not statistically useful given the method of polling.