This post was edited by abramicus at 2014-9-21 10:10|
Anomalies in the registration of voters have only now surfaced, after Alex Salmond spared no time to concede defeat, without inspecting the troops to see if they indeed had lost the battle, or have merely been sidetracked. The media attack on Dundee and Glasgow which are strongholds of the Scottish Patriots only shows the underhanded techniques of the pro-English factions in plotting to undermine any YES victory, if it should take place. But, it does raise the question that if Glasgow and Dundee could permit impostors to vote YES, then why cannot the rest of the country permit still more impostors to vote NO? To be fair, a recount is absolutely necessary to restore the public trust in the electoral system, which now is itself suspect of bias.
The solid and consistent results of 7 polls taken (just one day before the referendum for four, and 2 days before the referendum for three others) was held, declared on the average that the YES vote had 48% of the votes, and the NO vote had 52% of the vote.
Here were the polls before election day, on September 17, 2014, as follows:
Survation/Daily Mail 48% 52%
| POLL|| YES|| NO|
| YouGov/Sun|| 48%|| 52%|
| Survation/Daily Record|
| 47%|| 53%|
| Ipsos-MORI/Evening Standard|
| 47.4%|| 55.6%|
| Panelbase|| 48%||52%|
| 48%||52% |
| ICM/Scotsman |
| 48%|| 52% |
Ipsos-MORI/STV 49% 51%
The following is a simple statistical analysis of the results of the polls and of the referendum, and the simple question we are asking is "Are the polls and the referendum measuring the will of the same population?"
SCENARIO I: Assume that the percentage of YES voters is as claimed by the SEVEN major polls (four, one day before the referendum, and three, two days before the referendum) is 48%, and that the percentage of the NO voters is 52%.
QUESTION: What is the probability that after 3.6 million votes are cast, you will get a result of 44.7% YES and 55.3% NO votes?
ANSWER: Z = (p-P)/Sqrt(P*(1-P)/n)where
p = 44.7%
P = 48.0%
n = 3,623,344
Z = -125.7
Prob(Z<-125.7) = 0.00000000000000000...
CONCLUSION: If the real proportion of YES voters is 48% as all seven different polls claimed just before the referendum is true, then the PROBABILITY OF THE REFERENDUM REPRESENTING THE SAME POPULATION OF VOTERS IS VIRTUALLY ZERO.
SCENARIO IIa: Ipsos-MORI/EveningStandard asked 991 persons minus those who are undecided, or about 941 persons, and found the YES vote to be 47.4% and the NO vote to be 52.6%.
ASSUMPTION: That the true percentage of YES votes is as claimed by the reported referendum results is 44.7%, and the percentage of NO votes is 55.7%.
QUESTION: What is the probability that in such a pre-referendum survey, you will get a result of 48% YES and 52% NO votes?
ANSWER: Z = (p-P)/Sqrt(P*(1-P)/n)
p = 48.0%
P = 44.7%
n = 941
Z = 1.67
Prob(Z > 1.67) = 0.0498
A one-tailed test is used since we are testing whether the percentage YES in the referendum is really less than the percentage YES in the polls. And, therefore, we can state at a 0.95 confidence level that the two populations tested, i.e., the polls and the referendum, are NOT IDENTICAL.
Scenario IIb, IIc, IId, IIe, IIf, IIg from six other polls all come to the same conclusion (to be supplied in a separate table). The probability of all 7 poll YES percentages being different from the referendum by random chance is 0.000000000000019.
CONCLUSION: If the real proportion of YES voters is 44.7% as reported in by referendum is true, then the PROBABILITY OF THE IPSO-MORI POLL REPRESENTING THE SAME POPULATION OF VOTERS IS LESS THAN 5%. THEREFORE, AT A 95% CONFIDENCE LEVEL (ALPHA =0.05), ONE MUST CONCLUDE THAT THE POPULATION OF ALL REGISTERED VOTERS SAMPLED BY THE SEVEN POLLS IS NOT THE SAME AS THAT COUNTED IN THE REFERENDUM. SINCE THE POLLS ARE PERSON-FOR-PERSON TAKEN FROM THE POPULATION OF THE REGISTERED VOTERS, AND CAN BE VERIFIED BY THE POLLSTERS TO BE IN FACT RANDOMLY SELECTED FROM THE POPULATION OF REGISTERED VOTERS, THEREFORE THE REFERENDUM SAMPLE IS NOT REPRESENTATIVE OF THE POPULATION OF REGISTERED VOTERS. AS SUCH, THE VOTE COUNT FROM THE REFERENDUM IS NOT A VALID MEASURE OF THE WILL OF THE REGISTERED VOTERS OF SCOTLAND.
SUMMARY POLL VS REFERENDUM RESULTS:
|Date(s) of Survey|
Total Persons Surveyed minus the Undecided
Probability of YES in Referendum being less than in the Polls
|Total Y/N votes||YES votes||NO votes|
| 3,623,344 || 1,619,635 || 2,003,709 |
|16–17 Sep||Ipsos MORI/Evening Standard|
| 941 || 446 || 496 |
|16–17 Sep||Survation/Daily Record|
| 1,056 || 499 || 557 |
|15–17 Sep||YouGov/The Times/The Sun|
| 3,043 || 1,457 || 1,586 |
| 954 || 452 || 502 |
|15–16 Sep||Ipsos MORI/STV|
| 1,318 || 645 || 673 |
|12–16 Sep||ICM/The Scotsman|
| 1,011 || 482 || 529 |
|12–16 Sep||Survation/Daily Mail|
| 920 || 440 || 480 |