POLLS I can't sleep and my eyes are blurry
But I gotta get this outta my head someway impossibly shorr. The analytical minds can walk all over me I don't care.
For years I did analysis of defective production work and product failures.
The election polls fail accuracy because the samples are too small and area selective.
Unlike my analysis which had many possible categories of failure elections basically two comparison data points. The undecided are unkowns. The third parties can be tossed BUT you don't know which side they are from if they would vote at all without their candidate. You can ask them but the results could be meaningless. Are they be truthful or not.
SO i believe large samples would tell us more like 10,000 not 1000 and spread that sample farther out.
Statistically the sample could be biased depending on doing it but a true 50/50 sample would tell a truer story then we get now.
And for average I learned the hard way doing month to month trends tossing out the high and low number before calculation is a big mistake.
In production those two numbers were the beginning or end of trends that showed up the following period.
I got that off my chest and can go to sleep.
The analytical can tare it apart but production is much more difficult than these what should be simple polls.
Hillarys lose was proof enough.
For years I did analysis of defective production work and product failures.
The election polls fail accuracy because the samples are too small and area selective.
Unlike my analysis which had many possible categories of failure elections basically two comparison data points. The undecided are unkowns. The third parties can be tossed BUT you don't know which side they are from if they would vote at all without their candidate. You can ask them but the results could be meaningless. Are they be truthful or not.
SO i believe large samples would tell us more like 10,000 not 1000 and spread that sample farther out.
Statistically the sample could be biased depending on doing it but a true 50/50 sample would tell a truer story then we get now.
And for average I learned the hard way doing month to month trends tossing out the high and low number before calculation is a big mistake.
In production those two numbers were the beginning or end of trends that showed up the following period.
I got that off my chest and can go to sleep.
The analytical can tare it apart but production is much more difficult than these what should be simple polls.
Hillarys lose was proof enough.
