19385 calls, 504 responses, margin of error +/-4.6% Axne 43% Young 41% Undecided 11%.
It took 19,000 calls to get 504 responses? This poll is meaningless, it is self-selecting. The sample is unreliable and even the margin of error is meaningless. For even a semblance of accuracy they need to have some form of control, otherwise the sample is biased/tainted. For example who is likelier to answer a call? An unemployed or employed? A retiree or a student? Man or woman? White or non-white? And of these groups which is more likely than others to vote? When you call 19,000 times and get only 500 responses, you are basically getting demographics skewed.
Many polls adjust for such issues, based on historical data, and so they randomly strike out some responses in order to balance according to their demographic projections. Which obviously further taints the data because they cannot account for what may be motivating voters this year vs their historical models.
It took 19,000 calls to get 504 responses? This poll is meaningless, it is self-selecting. The sample is unreliable and even the margin of error is meaningless. For even a semblance of accuracy they need to have some form of control, otherwise the sample is biased/tainted. For example who is likelier to answer a call? An unemployed or employed? A retiree or a student? Man or woman? White or non-white? And of these groups which is more likely than others to vote? When you call 19,000 times and get only 500 responses, you are basically getting demographics skewed.
Many polls adjust for such issues, based on historical data, and so they randomly strike out some responses in order to balance according to their demographic projections. Which obviously further taints the data because they cannot account for what may be motivating voters this year vs their historical models.