Partap does not understand polling

THE EDITOR: I am responding to the spurious allegations levelled against the most recent NACTA poll by Harry Partap, MP for Nariva, as published in your paper. Partap attacks the integrity of NACTA, imputes far-fetched motives for NACTA’s polling, questions the methodology of the sample and disputes the reliability of the findings. The facts and the techniques used by NACTA do not support his baseless allegations. His conclusions are totally inaccurate and his critique of the sampling methodology is without merit.

I would advise Partap to pick up a simple basic book on statistical sampling and polling to get a better understanding on how polls are conducted. Then he would be in a better position to question experts on polling. Readers should note that Partap has not disputed the actual findings showing that UNC supporters are disenchanted with the representation provided by the MPs. But he is seeking to discredit NACTA and the methodology of the poll. Partap argues that a sample of 288 among a voting population of 250,000 (the approximate supporters of the UNC) is too small and cannot represent the opinion of party’s supporters. Compare that with a sample of about 1,200 as used by pollsters in the US to represent a voting population of 210 million voters (one for every 175,000 voters). Using such an infinitesimal sample, American pollsters predict election results that are on target. NACTA uses a much larger sample (one for every 1,000 voters) in Trinidad than that used in the US by American pollsters.

Since the findings of all of NACTA’s polls have been on target, then the methodology cannot be flawed. And the recent NACTA findings must therefore be reliable contrary to Partap’s assertion. It should also be noted that the methodology used by NACTA and the size of the sample is not very different from those used by other pollsters in Trinidad. And the findings on many of the issues do not deviate significantly from those obtained by other pollsters. Partap did not attack those polls but chose to train his anger at NACTA. Also, several UNC MPs and local government councillors have praised NACTA’s work and indicated that the findings of NACTA are in agreement with the findings on the ground in their own unscientific polling. Talk radio, newspaper internet, and TV people meter polls also substantiate the findings of the NACTA polls. Should we conclude that all of them are unreliable? It should be noted that NACTA uses a smaller sample size in Guyana and all of the polls’ findings there have been accurate since 1992.

In Trinidad society where race determines voting behaviour and the party obtains their support almost exclusively by race, it is not necessary to interview voters in every street, village or town to obtain a representative opinion of the total population (called the “universe” in polling jargon). The NACTA poll indicated that it randomly interviewed voters who claim they are or were supporters of the UNC and the sample represents the demographic distribution (age, sex, geography, occupation, religion, etc) of the population. The findings of the entire sample taken as a whole revealed certain trends and these were published; these findings are reliable when assessed against statistical sampling techniques used in the social sciences. Also, my work has been recognised by several political parties and governments throughout the Caribbean, the USA, Fiji and India. I have been invited to conduct polls in other islands.

I am not an agent of any political party. What is found on the ground (or in the fields) is what is actually published and people respect me for my honesty and forthrightness (calling a spade a spade and not burying my head in the sand). There is no truth that “NACTA showcases a poll whenever former Attorney General Ramesh Maharaj has a meeting that attracts one hundred or two hundred supporters.” Partap claims that “Maharaj attracted 200 supporters at a meeting in Naparima Bowl two Sundays ago and NACTA was ready with a poll on Tuesday denouncing the UNC.” Let the record show that the poll was conducted in early July and the findings released before Maharaj’s meeting. NACTA did not denounce the UNC but released the survey’s findings some of which are positive towards the UNC. I also want to be on record that I recently attended several UNC meetings which were broadcast live on radio; the largest of these meetings was 54 individuals (inclusive of the speakers) held at the Trinzuela College.

Of the 54, I was reliably informed that only five were actually from the San Fernando West constituency and the rest were a rent-a-crowd. Compare the size of recent meetings with the huge crowds of four years ago and one would conclude that the UNC has lost massive support.There is also no truth that the NACTA poll “denounces the UNC whenever the PNM’s back is against the wall or deflect public attention away from the PNM.” The recently published findings indicated that the PNM is weak on crime and that 93 percent of UNC voters feel the UNC government outperformed the PNM government in combatting crime and also performed better in managing the economy. There is no truth in Partap’s claim that “NACTA is playing on the minds of UNC supporters and has not published a poll reflecting people’s thoughts on the PNM’s performance.”

Partap has not been honest here as the facts show otherwise. For the record, the findings of NACTA’s poll, as published in virtually all of the papers, have been severely critical of the PNM on the important issues of health care, education, national security, crime, CEPEP, cocaine in the diplomatic pouch, corruption, moving the parliament, commission of inquiries, Unit Trust, etc. Partap should go back to the newspaper’s archives in early June 2004, November 2003, September 2003, July 2003, June 2003, April 2003, November 2002, October 2002, August 2002, July 2002, etc.

Partap and some of the MPs, who ill-advise the leader of the UNC, need a reality check. There is only one way to test the veracity of NACTA’s findings — through an actual election. And I am confident if one were to be held now, NACTA will be proven right again and vindicated. The UNC, as currently constituted with tainted MPs, will be wiped out from the face of the Trinidad’s political map. Finally, throughout the world, every serious political party and politician conducts polls or wants to know what the polls say about their support. Partap and the UNC should appreciate polls and use them to reconstruct the party rather than attack the poll.

VISHNU BISRAM
NACTA