Latest NACTA poll raises more questions
Vishnu Bisram of NACTA fame did another one of his infamous polls in mid-August 2011. He did one in July 2011. Bisram interviewed 780 voters in July. He questioned only 420 voters in August. A major difference of 360 voters between two polls. A whopping 46 percent smaller sample size from his July poll.
Put it another way, the closer we got to the election the smaller the sample size Bisram chose. Smaller sample sizes are more prone to larger sampling errors and mistakes particularly when many voters are sitting on the sidelines.
Now, Bisram did not tell us what accounted for this discrepancy in sample size between his polls of July and August. He did not confirm whether due to the smaller sample size he focused on voters in specific regions only in his August poll, which would raise concerns of bias.
A significant difference in sample size is a serious problem when polls are asking the same question. Bisram asked the same question in both polls, which is the biggest question of them all: popular support for each party in the election. He found the support of the PPP increased by one percent, APNU decreased by 1percent and AFC increased by 2 percent in the more statistically suspicious and weaker poll of August 2011.
Interestingly, Bisram did not disclose his sampling error in his July poll in his letter to the newspaper. He did so in his August poll (plus or minus 5).
Bisram’s August 2011 poll covered three areas. As usual, Bisram selectively releases information on two areas and withholds the findings on the popularity ratings for President Bharrat Jagdeo. This is why I have stated unequivocally that reputable news organisations must demand everything before they allows Bisram to publish anything.
Voters were asked who they preferred as President of Guyana. Bisram claims that 46 percent opted for Ramotar, 23 percent chose Granger, eight percent Ramjattan, one percent Sharma and less than one percent all other candidates, with 21 percent not making a choice.
For party support, Bisram claims the PPP is polling 47 percent, APNU 24 percent, AFC seven percent, JFAP one percent, other parties less than one percent, and 20 percent undecided or no response.
Bisram tells us of Ramotar’s national support for President in August (46 percent). He never told us of Ramotar’s national support in July, which was a significantly bigger poll. I extrapolated from his July poll that Ramotar’s national support in July was only 34 percent. Bisram now tells us Ramotar has the support of 98 percent of PPP voters.
Bisram is saying that after scandal following scandal and Ramotar’s failure on sugar and Guysuco seeking foreign management, Donald Ramotar rose from 73 percent of support from PPP voters to 98 percent.
This is either evidence that people will support something wrong or that Bisram is trying to sell us some statistical subterfuge by choosing a smaller poll sample in August as opposed to July.
Bisram has to tell the public whether he asked the same question in his July 2011 poll that he asked in his August 2011 poll about national support for presidential candidates.
Stabroek News and Kaieteur News should demand that Bisram provide proof in the form of his July questionnaires and names of his interviewers to verify whether this question was asked.
For this goes to the very heart of the contention I have made against Bisram that he selectively presents and withholds findings and statistics of his polls. In this last August poll, he deliberately withheld President Jagdeo’s ratings after admitting that his poll questioned voters on this issue.
A clear trend that has emerged in the 2006 election and is rearing its head again in the coming 2011 poll is non-voters, the people who stay home.
In 2006, it was 31 percent of the electorate. So far this year it is just over 20 percent. For some strange reason, Bisram has never really examined this problem in his polls despite its statistical significance and the long shadow it casts over polling since 2006.
He has not asked who is not voting and why. Yet this man calls himself a pollster. Newspapers with reputations to preserve should be wary that they are not pawns to statistical propaganda.