Back in February of 2016, I told the world that all polls regarded the 2016 Presidential Election between Hillary Clinton and President Trump would be wrong (see: WHY EVERY PRESIDENTIAL ELECTION POLL FOR 2016 YOU SHALL SEE OR HEAR WILL BE WRONG). Further, I developed a highly accurate mathematical model that called the victory for president Trump within seven electoral college seats.
How could I do this? I could do this because I have professional experience in scientific public opinion polling, graduate school study of advertising and marketing research, as well as undergraduate study of statistics. Also, my gifted IQ intellect lets me see things that most can not.
The same design flaw that plagued the pollsters and their pseudo-scientific polls in 2016 in the USA, plagued Australian pollsters in 2019.
Australians use the Westminster system of anointing their PM from the party that wins the most seats in their House of Representatives.
Currently, there are 151 House seats. A simple of majority of House seats comes to 76.
To accurately poll, scientifically, requires polling enough of the 151 separate districts to discover which candidate has the majority and doing so all on the same day and about the same time so as to minimize the effects of new news.
To get a margin of error of +/- 5% requires 376 completed questionnaires taken at random to be sure that 95 out of 100 times, the results will be the same.
But to poll for the PM, requires enough separate surveys in enough districts to discover who will garner the majority in each of those districts. In the simple case, that would mean 76 separate surveys, or one for each of 76 districts if the same candidate won a majority in each of those districts.
Thus, one would need to complete an eye-popping 28,576 interviews (376 x 76) all on the same day and about the same time.
Yet, if you were to examine the polls published by Australian media, you would find paltry, inefficient sample sizes of 800 to 1,000, tops. And their method would be wrong, because the pollsters would fail to adhere to the design of the Australian government.
In short, in Australia, no one can survey the population at random, ignoring House districts, and then project those results to the universe of voters to predict who will become the next prime minister of Australia.
Polls are wrong because polling has been unscientific when it comes to politics. Before the Internet age, controlled media could sway public opinion toward fake polling results. And even when results were slightly off relative to the polls, such anomalies were overlooked.
No longer can that happen in the Age of the Internet. Would-be voters seek out intel and info from various sources that conform to their general personalities and predilections.
Real scientific polling requires consideration of population boundaries.
Gab It
