WHAT can we conclude from the polls we have seen in the run-up to Sunday’s presidential elections?
It is hard to find many points of agreement among the myriad of polls published, but there are a few sound conclusions that we can deduce.
First, DIKO candidate President Tassos Papadopoulos is in the lead in almost every poll.
This lead is not failsafe, for it is almost systematically less than the allowed margin of error in a given poll.
Margins of error range between 1.8 per cent and 3.5 per cent.
Second, the difference between the other two candidates is also less than the margin of error, thus adding uncertainty to the final result.
Predictions for the second round are even more tentative.
This is because they do not take into account what will happen in the interim week and how much this may or may not affect voters’ behaviour.
With this in mind, two further trends can be delineated.
In the case of a run-off on the second Sunday between Papadopoulos and DISY candidate Ioannis Kasoulides, Papadopoulos seems to have a lead of at least five per cent.
But, there is a high percentage of those polled who are undecided or say they will not vote for either of the two remaining candidates in the second round.
A total of 515,994 voters will cast their ballots in the compulsory voting, including 390 Turkish Cypriots.
One week before Election Day, polls came to an end on Sunday, in an effort to allow the public to make its decision in peace.
BUT ARE polls reliable?
Constantinos Lordos, in an opinion piece, has expressed his doubts.
The overall argument is that “opinion polls are more an art rather than a science. That is to say, every opinion poll is yet another opinion”.
His suggestion is sure to raise the hackles of pollsters, who see their surveys (heavily reliant on statistics) falling under the remit of social sciences, while most would also question the implication that arts are mere opinions without concomitant scientific rigour.
Lordos goes on to argue that the process of telephone polls affects the result because “the interviewer’s introductory speech, the timing, the manner, the tone, the sequence in which the questions are asked all have an important psychological effect on the interviewee, which affects the content of his answers given.”
This is called the ‘Observer’s Paradox’, the phenomenon where the observation of an event or experiment is influenced by the presence of the observer/investigator.
Christina Kokkalou of Insights Market Research which conducts rolling polls for Simerini admitted “he does have a point: it’s true, people are less affected when they vote inside a polling booth as opposed to when they answer questions over the phone.
“However, you can deal with this by employing professional interviewers who we train for the job.”
A spokesman from RAI which conducts the Phileleftheros polls, agreed “that there is some truth in Lordos’ argument.
“However, opinion polls are a practical affair, and if you do not have any experience of them, theorising can only go so far.
“What matters are the trends which emerge, rather than one-off polls; trends are better insulated from those who may not be honest in their answers.”
Lordos goes on to argue that size matters: “One in three or four people are directly or indirectly related to the state.
“In larger countries, countries which have provided us with the methods of opinion polls, this does not happen.
“When polled, won’t these people think that their answer might affect their own position in the state in a negative or positive manner?”
Kokkalou said the point was neither here nor there.
“Why should ties to the state affect answers in opinion polls, any more than they would actual votes?” she wondered.
Another argument Lordos makes is that the statistical error which all polls refer to is nothing but “a resort to subterfuge, because statistics deal with numbers, and numbers, according to science and tradition, cannot be wrong.”
Polls survey a certain number of people, known as the ‘population sample’. The numbers are correct for that sample.
But the margin of error is there to accommodate the extrapolation when the figures are applied to the whole of the electorate.
Pollsters will argue that, unless one samples the whole of the population (as will happen on polling day), a margin of error is necessary so as to define the reliability of the results.
A polling expert (who wished to remain anonymous) said the argument Lordos makes “doubts the whole enterprise of Western science”.
“Even successful medical research by which advances are made is conducted within what are called ‘acceptance and/or confidence levels’, which are analogous to the margin of error in polls.”
What the latest polls say
THE presidency is up for grabs, according to results published in the last polls before the elections.
Three polls were published in the Sunday newspapers.
Simerini hosted the last poll in a chain of rolling polls conducted by Insights Market Research (IMR).
“Rolling polls do more than take a snapshot of the electorate: they show trends,” said Christina Kokkalou of IMR.
All three main candidates show an increase in their percentages: DIKO candidate Tassos Papadopoulos is ahead with 32.6 per cent, followed by AKEL candidate Demetris Christofias with 32.1 per cent and DISY candidate Ioannis Kasoulides at 30.5 per cent.
Most notably, Kasoulides shows an impressive increase of 0.4 per cent.
“We are in for an election thriller,” Kokkalou said.
“This is because the Kasoulides candidacy is gaining momentum, and if it continues, it will be a very, very close affair.”
Independent candidates MEP Marios Matsakis and Costas Themistocleous, in unison, only manage 1.1 per cent of the vote (0.9 and 0.2 per cent respectively).
There remains a 3.3 per cent who are undecided or said that they will cast a blank vote.
The three main parties show a voter mobilisation of more than 82 per cent, while EDEK and the European party are around 74 per cent.
In the second round, Papadopoulos would beat Kasoulides, but he would lose to Christofias by a hairline margin of less than one per cent.
Christofias would also beat Kasoulides, this time by a margin of 2.2 per cent.
Under this scenario, 58.4 per cent of Papadopoulos voters remain undecided or say they would cast a blank.
Politis housed the Noverna poll, which continued its firebrand results showing Kasoulides ahead of Christofias (30.4 per cent and 30.3 per cent respectively), but Papadopoulos still first at 32.1 per cent.
The undecided and blank votes are at 4.2 per cent, while the independents Matsakis and Themistocleous fare better than they did in the IMR poll.
Matsakis is at 2.3 per cent and Themistocleous at 0.7 per cent.
In the second round, Papadopoulos would beat both the main candidates, with more than 10 per cent in Kasoulides’ case but less than 1 per cent if he were to face Christofias.
Christofias would beat Kasoulides, but only by 0.5 per cent.
The poll also shows the preferences of Cyprus’ districts.
Nicosia is split almost evenly amongst the three candidates, Limassol-Paphos is won by Papadopoulos and Larnaca-Famagusta goes to Kasoulides.
Papadopoulos has the vote of people over 60 and of those who think that the most important criterion is the Cyprus problem.
Kasoulides wins the youth vote (18-29) while also that percentage of the electorate which will vote based on the policy and opinions of a candidate.
Christofias wins those who cast their vote based on the candidate’s personality, as well as those who think that the most important criteria are social and economic policy.
Only 24 per cent of refugees will vote for Papadopoulos, the rest evenly split between Christofias and Kasoulides.
The reverse occurs in the non-refugee vote, Papadopoulos winning the lion’s share (37 per cent), the rest again being evenly split between Christofias and Kasoulides.
Phileleftheros published a poll conducted by RAI, which shows Papadopoulos in the lead with 32.7 per cent, Christofias at 30.3 per cent and Kasoulides at 30 per cent.
Matsakis and Themistocleous are at 0.5 per cent and 0.1 per cent respectively.
The blank votes are at two per cent and those who would not answer at 3.9 per cent.
RAI’s poll shows the three main parties backing the three candidates with more than 80 per cent mobilisation, while Papadopoulos-backing EDEK and European party are lower than 74 per cent.
RAI broke down the results according to gender, town and age.
Women are shown to prefer Papadopoulos by a margin greater than five per cent.
The male vote is still up for grabs, with Kasoulides sporting a slender lead.
Nicosia, Limassol and Paphos all prefer Papadopoulos, while Larnaca and Famagusta prefer Kasoulides.
Christofias wins the votes of ages between 25 and 54, Kasoulides leads in the under 25s, and Papadopoulos has the over 55s on his side.
Interestingly, RAI’s three second round scenarios hand victory to every candidate.
Papadopoulos would beat Kasoulides by an 8.6 per cent margin, Christofias would beat Papadopoulos by 0.5 per cent and Kasoulides would beat Christofias by 0.9 per cent.
