Explaining an Opinion Poll to project election outcome 

Feb 26, 2024 Letters

---

Dear Editor,

Kaieteur News – There appears to be a misunderstanding of the purpose of a poll and its findings as one dentist commentator suggests (KN Feb 19) his personal view is better than a survey. A poll is taken to determine a candidate’s (or a party’s) political fortunes (support) in elections or on some other issue. It is merely a guide or estimate of support at a given time. The support is not static. For example, Nikki Halley was trailing Trump four to one in South Carolina primary. On eve of Election Day, polls showed she trailed by 20 percent. Actual outcome, 60-40.

Errors are possible in polls and happen all the time. The closer to an election, the more reflective is the results of actual outcome.  That is why pollsters conduct tracking surveys to determine changes in support. Support is dynamic. American pollsters conduct polls for clients on a weekly basis to track support. In Guyana, polls are hardly conducted because of financing although a political party has employed the services of American pollsters. I am not contracted or commissioned to conduct polls by any party in Guyana.

Contrary to what the dentist commentator feels, a poll is not the opinion of one person. A poll collects information through a survey by asking many people the same questions and recording and tabulating their responses to each.

The opinion is not subject to verifiable proof unless several polls are conducted by different pollsters at the same. It is impossible to verify the view of each interviewee who is and supposed to be anonymous. It is a value judgment or preference for a party and changes over time. All polls taken around the same time should reflect similar findings as has been the case in polls conducted in USA in all elections including for 2024 to determine the Republican and Democratic Party expected candidate for President and the matchup between projected contenders — namely Biden V Trump. Support has vacillated back and forth between the duo. Trump was ahead two weeks ago. Now, it is a statical tie. The election is just over eight months and can go either way.

Polls are more or less accurate indicators of the electoral portrait of a nation at a particular time provided they are done objectively and scientifically. Accurate polls are based on scientific techniques which is a very complex process. No need to explain it here! Objectively conducted polls, not conditioned by biases of the pollster or his/her own party preference, should accurately estimate the support of candidates and outcome of elections. Respondents are interviewed randomly allowing every voter an equal opportunity to be interviewed to represent a subcategory of voters — like racial and religious composition, class, age, gender, etc.

Polls conducted on radio and television or on the Internet are straw polls and are not truly reflective of the views of the population because they are not based on any scientific method to reflect the demographics of the population. As such, these results should not be taken seriously. Similarly, the view of the dentist should not be accepted as that of the nation when he suggests Norton will defeat Irfaan. Opinion must be based on evidence and of a large number of people.

After opinions are collected of a large sample, say between 450 and 2500 people, pollsters usually analyse the results at a 95 percent significance level which means that in 19 out of 20 cases the results are accurate; 95% is standard in Sociology. The analysis produces a sample error that is obtained based on a standard formula statisticians use. The size of the sample determines the error. The larger the sample, the smaller the error. The sample error tells how much the actual findings could deviate — generally between 3 to 5 points in either direction. A larger sampling error is not accepted in social science research.

The dentist disagrees with recent findings of polls I conducted on support for Mr. Aubrey Norton and Irfaan Ali. He writes Norton “will go up and up” based on his personal view. The polls I conducted over last year among thousands found Norton going downwards in support.

There has been growing and widespread disgruntlement with the incumbent PPP administration but there is no effective opposition to exploit the dissatisfaction. Other contracted polls will reveal similar findings. The poll found Amanda Desir and Roysdale Forde have more support than Norton. Nigel Hughes also has significant traction in the APNU base and among some Indians. Indians have made it clear they will not vote APNU under Norton’s leadership. Some Africans voiced similar view, suggesting the coalition will not get the same percentage of support as in 2020 unless it has a likable leader. The coalition cannot win without Indian support. Right now, the poll projection from a January survey is Irfaan Ali will return to office with less Indian turnout than in 2020. But the PPP is projected to gain seats because of the disapproval of Norton who also faces lower turnout.

Yours truly,

Vishnu Bisram (PhD Pol Sci)

---

• dentist commentator, Letters, Opinion Poll, Vishnu Bisram

---