CARIBBEAN (COMMENTARY – By Joel B. Liburd) - The quote above from Warren Buffet exemplifies exactly why pollsters are making a killing in recent times when an election is called. Polling is a methodology employed by social scientists to understand trends and behaviours. Scholars have spent decades perfecting it as a scientific tool to understand social phenomena. From Jurgen Habermas and his theory of the Public Sphere, to James Halloran and his research on multimodal communication, polling data has become an increasingly valuable commodity, ironically for a purpose for which it was not intended. Influence.
'Tis the season where politics meets politricks and where gimmickry melds with "gimme-gimme". Alas, elections are nigh. But apart from the regular expectations of campaigning, canvassing and complaining, the atmosphere is one that nurtures another facet of the process of democracy. The polls.
Kittitians especially have had it up to *here* with polls and surveys. In the beginning, many subscribed out of sheer curiosity, or were pleased to be the respectful centre of a stranger's attention for a few minutes. After a while, it just got silly. Many persons refuse to be polled, either out of fear of exposing their innermost feelings and opinions, or caution that there may be repercussions for their thoughts. It's the fundamental reason why democratic elections usually employ the "secret ballot" method, to protect the voters' opinion and decisions.
It means being able to interact comfortably with persons who hold opposing philosophical and political views, at least, until the next campaign season arrives.
Polling is not an easy task. To get a measurable snapshot of a population's views means designing questions that are easy, and not intimidating; being short, but concise; and most importantly, trying to ask key questions in different ways, to avoid bias or prejudice by the survey participant. Polls need to be tested rigorously internally, to ensure that credible and verifiable data can be extracted, and that the necessary adjustments can be made before the poll is finally distributed.
Distribution brings its own set of headaches. The biggest one is sampling. Since it is impossible to survey an entire population – especially multiple times – pollsters stick to a manageable number of responses (or representative sample) in order to extrapolate data to paint a picture of a community. But sampling has its inherent issues. Most people are familiar with "random sampling" as is used in the game show Family Feud. However, political polling has nothing to do with the types of questions that Steve Harvey asks. As a political tool, a randomly sampled poll is arguable quite flawed... some respondents might be party sycophants and slant their answers either all to the left or all to the right. Others may not intend to vote, and so add unnecessary data to the result. Some folks may not even be eligible to vote and will answer questions haphazardly, to simply give a response.
More appropriate methodologies tend to employ techniques that incorporate stratified or purposive sampling methods in order to hone in on the particular topic or effect being examined. This creates the foundation for more credible data... asking actual registered voters - away from party-centric geographical locations - their thoughts as they mull their decision.
Apart from the survey tool, there is the human element – that bias that determines who the interviewer approaches and who accepts the interviewer's approach. Even from that initial contact, bias plays a massive role, and can skew even the greatest error of margin as personal preferences and prejudices come into play for both parties involved in the exercise.
The poll agent will obviously have their personal views on the topic being examined, and because of mere human nature, these views can be projected subconsciously by body language and inflections of voice while administering the survey questions.
That's one of the key reasons why there has been a boom in online survey services – to remove the element of human bias. But this comes at the great cost of a trained pollster not being able to read the body language of the interviewee.
It is well known that unless you're in China, North Korea, Cameroon, Equatorial Guinea, Iran or the Republic of Congo (and the likes), leaders change. When a democratic nation goes to the polls (except in the cases of proportional representation), it's a first-past-the-post race. What is also very established is that most of these types of democracies are largely two-party systems.
So if there are two major parties, let's presume largely splitting the populations evenly, then how does the administration change? Simple. The swing vote. This is the Holy Grail for pollsters... identifying these voters and giving them every reason to turn them favourably on Election Day.
These voters are usually educated or experienced enough to disregard the electoral manna flowing from politicians' campaigning lips. They look, listen and think. Some voters even make up their minds while walking to the polling station. Some have a last look at the ballot sheet and then make the crucial decision. This points to fact that political parties can – and do – take their base supporters for granted, and that the focus is always on getting the swing vote your way, which begets the lofty promises, baby-kissing and the overwhelming amount of grassroots public appearances in the run-up to the big day.
Pollsters have realised that their data is another critical tool in peddling influence. No one likes to be on a losing team. It's probably why people groan when whenever the West Indies embark on a new tour, or why Lionel Messi doesn't smile while playing for the national team.
Juggling figures to indicate to the swing voters that they may be on the losing team does two things; either influences them to change their vote, or encourages them to abstain from voting, which in both cases is a win for the pollster and the party that is paying him. This is a highly unethical practice.
As stated before, polls were designed to be a measuring tool, and not a weapon against a person's own opinion in his/her independent thought.
This leads to the recent string of rather desperate utterances from pollster Peter Wickham, who is crunching the figures for Prime Minister Timothy Harris' Team Unity arrangement. It is rather disappointing that the vastly experienced Wickham has stooped so low as to wring every potential drop of political promise from his dated dataset, to appease the coalition and its supporters.
That's the very reason why former Barbados DLP Member of Parliament Robert Morris accused Wickham of unscientific methods and sensational analysis of what were decidedly unsexy figures, in the hindsight of Mia Mottley's historic win at the 2018 polls. A frustrated Wickham, on that occasion, seemed unable to respond to Morris academically, and chose instead to use insulting language to defend the accusation, calling the MP a "common yard fowl".
On the other hand, renowned pollster Bill Johnson is crunching the numbers for the St Kitts-Nevis Labour Party (SKNLP), and apart from a single major press release last year stating that Labour was comfortably poised, there's been nothing else from that quarter. It's very close to a proper strategy – polls should advise the campaign machinery on its area of focus, and keep the other side wondering about the current status and the next move. According to former US Secretary of State Henry Kissinger, "Leaders are responsible not for running public opinion polls, but for the consequences of their actions." Using that thought as a yardstick, SKNLP Political Leader Dr Denzil Douglas certainly seems to be in the driver's seat.
But Barbados was not the only faux pas for Wickham's Caribbean Development Research Services (CADRES) outfit. Although he called the winner for Dominica, the myriad of reports of inconsistencies and subsequent protests in that November 2019 election, further illustrates the fact that his figures were pointing the wrong way. He had projected that Prime Minister Roosevelt Skerritt would lose support, but hold on to the Prime Ministership. In reality, all Wickham did was play all the numbers on the roulette table... you can't really win, but you definitely don't lose.
Last year as well, Wickham foolishly misread the Jamaican vibes in the East Portland bye-election, and confidently pronounced that People's National Party candidate Damion Crawford was miles ahead of Jamaica Labour Party's Ann-Marie Vaz. When the votes were counted, the constituency had their first JLP MP in more than two decades.
But Wickham's memory of his inconsistent, unscientific methods seems to have faded. In 2006, his polls favoured St. Lucia Labour Party for the win, but it was the United Workers Party that won a comfortable victory.
Then in 2008, he went to Grenadian election would be won by the New National Party. Instead, the National Democratic Congress took 11 of the 15 seats.
Again, in 2013, Wickham tried predicting the election in his home country of Barbados. He said that the political pulse indicated, that the Barbados Labour Party would form the government. In the end, his "scientific" techniques and pulse-reading were proven wrong as the incumbent Democratic Labour Party won another term in office.
Most shamefully was Wickham's abject failure in Trinidad and Tobago in 2007. CADRES was hired by Winston Dookeran's "third force" Congress of the People (COP) party at the end of September, to poll for the General Election that took place on November 5 that year. His projection was 30% for COP, with Patrick Manning's People's National Movement (PNM) and Basdeo Panday's United National Congress trailing with 23% and 10% respectively. Although the COP managed a massive 148,000 votes, the party did not win a single one of the 41 seats up for grabs, and the PNM romped into office with 26 seats, leaving the UNC Opposition with 15. Wickham fled Trinidad with a very firm tail between his legs.
One probable cause for Wickham's difficulty in accurately predicting outcomes could be his stated preference for a form of polling known as "swing analysis", which was popular worldwide in the 80s and 90s. However, this technique yielded highly inaccurate results – most notably in the United Kingdom in 1992 - and has been largely abandoned.
Wickham – who recently married his openly gay long-time partner, Italian Giancarlo Cardinale – is now showing exactly how much he really has in common with Harris and PAM Leader Shawn Richards. The most glaring is the lack of leadership and an increasing lack of diligence, ethics and responsibility. Wickham, as a consultant contracted to the Team Unity client, is now behaving as if he is a part of the Federal government, filling either a role for financial enrichment, or filling a hole of disenchantment within the tenuous coalition arrangement.
Wickham of all people should know that when the election dust is cleared, he has to search for the next engagement, and simply saying (incorrectly) "I've never been wrong" is not as comforting as actually doing it right!
Joel B. Liburd
Communications Consultant, Basseterre/Quebec