With mere hours to go before the polls open, this country's pollsters are projecting varying results, cementing the fact that the 2015 election is too close to call.
Citizens are no better off than they were at the start of the campaign, since they are still left wondering just which party has the edge in tomorrow's general election–the ruling People's Partnership (PP) or the People's National Movement (PNM).
However, H.H.B & Associates pollster Louis Bertand believes that even though there are slight variations in the predictions of the local polls, citizens can make an assessment of who will win tomorrow since local polls were proven accurate in the past.
"Polls like any other things you are trying to forecast, by and large, have been proven to be more accurate than inaccurate. It is like the weatherman who tells you it is going to rain and it does not, but it rains the next day," he explained.
Over the last week, pollsters have announced varying projections for the outcome of the highly anticipated September 7 polls.
H.H.B & Associates, which conducted the Guardian Media Ltd (GML) poll, is projecting a PNM victory, with the opposition party taking 22 seats and the PP capturing 19. This finding contrasted with the North American Caribbean Teachers Association (Nacta) poll by Dr Vishnu Bisram, which predicted yesterday that the PP would win 21 seats, the PNM 18 seats, with two marginals up in the air. The UWI Constitutional Affairs and Parliamentary Studies Unit (Capsu) poll released last week also found that the PP is leading in five marginal seats in the election race. Solution by Simulation tracking poll by Nigel Henry projected yesterday that the PP was leading by 38 per cent, the PNM had 36 per cent, and a whopping 25 per cent were still undecided.
Dr Vishnu Bisram, in his Nacta poll, said the percentage of undecided voters had shrunk considerably from a month ago, from "a high of 31 per cent to just six per cent now," and they could very well decide the outcome of the election.
Political analyst Dr Maukesh Basdeo agreed that the undecided voters could swing the election either to the PP or to the PNM.
He agreed that with the varying results, one might be left wondering if polls were accurate, especially given the situation with the UK elections in May.
However, he said, it was "very difficult" to predict human behaviour.
"Polls are a possible outcome, it is not to say 100 per cent accurate, because you are speaking about human behaviour. The poll is a snapshot in time, a person's view."
He explained that someone may give an opinion on a political issue like whom he is voting for and 24 hours later, they could be exposed to new information which could change their opinion and their behaviour would change.
In the UK, pollsters came under fire after they predicted a hung parliament with the Conservatives, led by Prime Minister David Cameron, needing a coalition partner to remain in government. However, the opposite occurred, and Cameron's party received a landslide victory securing 331 seats, five more than needed for a House of Commons majority.
Dr Hamid Ghany explained that the UK situation was different from T&T's because pollsters there used methodologies such as online polling, telephone polling, and limited face-to-face intercept.
"We have had, in T&T, telephone polling and face-to-face intercept," he said.
The key to polling, he said, was the methodology used for gathering the data and the size of the margin of error.
"If your margin of error is more than +/- five per cent then the poll is unreliable, if is it less than +/- five per cent then the poll is reliable. In order to get a margin of +/- five per cent, that is determined by the number of questionnaires," he said.
Henry, managing director of Solution by Simulation, said polling had a long history of successes, but recently it had been facing challenges for a number of reasons.
However, he said, it was very important because "we do not just need to know the numbers at the end of the day, but we need to know why people are doing what they are doing in order to improve governance."
He said the question should be: how can we make the polls more accurate? �2See Pages A6 & A8