Why Do All Election Forecasters, Political Scientists, Academics and Media Pundits Avoid the Systemic Fraud Factor? Was a big heaping bowl full of infuriating inability to understand modeling.
Alaska excluded, a general rule of thumb is that rural areas tend to be more Republican and urban areas tend to be more Democratic. Another general rule of thumb is that an exit pollster is worth more in an urban area, where he or she can collect more data, than they are in a rural area. A pollster at a precinct that gets 2000 voters in a day collects more data than a pollster in a precinct that gets 40 votes on election day. But the whole premise of the above mentioned piece is that all the raw votes collected from the small and biased sample ought to be counted as equally, and that any attempt to correct for the known bias in collection methods is really just an attempt to cover for fraud.
It just goes to show that even a degree in Math doesn’t mean that you understand numbers at all.