Investigative Journalism: Why Bernie may have actually won New York
from reddit
https://www.reddit.com/r/SandersForPresident/comments/4g3x5h/investigative_journalism_why_bernie_may_have/
An election exit poll is a poll of voters taken immediately after they have exited the polling stations. Unlike an opinion poll, which asks for whom the voter plans to vote, or some similar formulation, an exit poll asks for whom the voter actually voted. Pollsters – usually private companies working for newspapers or broadcasters – conduct exit polls to gain an early indication as to how an election has turned out, as in many elections the actual result may take hours or even days to count. Exit polls have historically and throughout the world been used as a check against, and rough indicator of, the degree of election fraud.
After all votes are tabulated, exit polls are “adjusted” to match recorded results. According to NPR, for this election cycle, a firm called Edison Research conducts the polling used by major networks. Exit polling has not been conducted for every contest thus far. Here are the unadjusted exit polls against the final results (significant discrepancy | state flip; data source):
see chart:
https://www.reddit.com/r/SandersForPresident/comments/4g3x5h/investigative_journalism_why_bernie_may_have/
Side note: although Edison Research did not conduct exit polling in Arizona, a local newspaper called the Daily Courier did – but only for Yavapai County. Official results have Clinton winning the county 52.9-44.7; however, the Courier’s exit polling had Sanders crushing her 62-37. Possible explanation: heavy early voting advantaged Clinton; nonetheless, Arizona was a quagmire.
Excluding Arizona (because only one county was polled), Sanders has suffered an average 6.92% deviation among all contests with exit polling. In particular, assuming that New York exit polling was conducted correctly, the statistical likelihood of a 12% deviation from exit polling is 1/126,000. Theoretically, the results would be equally likely to deviate in either direction; the probability that the 18 of the 19 exit polls above swung to Hillary’s advantage is 0.000038 (that is, fewer than four in one hundred thousand elections would roll this way due to chance).
= new reply since forum marked as read
