Editorial: A love-hate relationship with polls

0
754
- Advertisement -

If you want to create a stir in the world of politics, then host a mock election in the midst of an anxious mood leading up to what could possibly be an early election.  That is exactly what the Free and Fair Election League did recently and wow, did they ever throw ‘the cat among the pigeons’ when they declared that
the results were in the United Progressive Party’s favour.
You all know our feelings on polls and statistics.  We adhere to the assessment of American writer E.B. White who eloquently said, “The so-called science of poll-taking is not a science at all but mere necromancy. People are unpredictable by nature, and although you can take a nation’s pulse, you can’t be sure that the nation hasn’t just run up a flight of stairs.”  Plus, the amateur statistician in us knows that statistics can be manipulated to show anything.  So, when we hear about polls, we do not get too hyped on the results.  
In the case of the mock election recently completed, we have been inundated with people who want to know what we think.  Is the sample too small?  What about the way the poll was executed?  Etc. Etc. Etc.  To be honest, these are question that are better asked of the Free and Fair Election League than us.  That said, we will give you our thoughts on what we have seen with some upfront caveats and disclosures.  We are not statisticians.  We have no knowledge of the mock election design.  And we only know what we have been told by the Free and Fair Election League.
Now to the results (as presented).  There are many people who have criticised the poll for the size of the sample but the sample size is not the problem, the selection of the sample probably is (but more on that later).  In the last election, there were 47,720 registered voters.  So, let’s say that today there are 50,000 (a nice round conservative numbers).  The sample size required for a margin of error +/- 5 percent and a confidence level of 95% is only about 400.  The League’s sample was just short of 1,000 so they could boast better figures than the standard just cited.   
All well and good, if this was a perfectly random sample.  And this is where we believe the mock election may start to lose its scientific strength.  For any polling to be considered “accurate” the sampling must be as random as possible.  We do not know how the mock election was designed nor executed so we cannot make any proclamation on how the organisers were able to ensure randomness.  Add to that, the claim by the Antigua and Barbuda Labour Party (ABLP) that it dissuaded its members from participating and the randomness would be affected (if that request materialised in lower ABLP supporters’ participation).
If the randomness of the mock election can withstand scrutiny, then focus would then fall to the question asked.  In a real election, voters are asked to place their “X” near the name of the candidate that they feel will best represent their constituency.  In this case, the voter was asked to choose the party that they feel should run the country.  An entirely different question but considering that our politics break down along party lines rather than individuals, many will concede that there is little difference between the two questions.  That might be the case, but the question also asks who should run the country rather than who will you vote for.  We may be splitting hairs but if that is how the question was asked, then it does not mimic the question faced by voters in the polling booth.  
All said and done, the mock election exercise is a good one and we give credit to the Free and Fair Election League for its effort.  Regardless of the results and the criticism, these mock elections are good for our democracy.  They raise awareness and encourage debate on political issues.  And to give ‘Jack his jacket’ they were not too far off in 2014 when they used a sample size of 801 to predict an ABLP win.
Polls are polls.  The ‘winners’ love them and the ‘losers’ detest them.  Had the poll revealed a win for the ABLP, we are fairly certain that the mock elections would not have been dismissed as “irrelevant” by the party’s leaders and its supporters.  We would have probably heard the ‘tek dat’ chorus, while the UPP would have been the ones offering harsh criticism. In fact, even before the dust had settled on the Free and Fair mock election results, the Caribbean Development Research Services (CADRES), a polling and research company headed by Peter Wickham, released the results of a poll that suggested that the ABLP would win the election if it were called today. So there! With a little less than 48 hours after Free and Fair, CADRES says, “not so fast, “ and the ABLP supporters get their chance to (as predicted) say, “tek dat!”
Who can forget the early 2014 mock election conducted by the League?  The results showed the ABLP winning with 495 votes as compared to the UPP’s 221 votes.  Chairman of the UPP Leon ‘Chaku’ Symister dismissed the simulated election as “an exercise in futility,” and said his party placed no credence on the results of the poll.“  Never mind that, in the hours following the Free and Fair exercise this year, the jubilant political leader of the opposition UPP welcomed the results and proclaimed that they were consistent with the party’s polling.  Of course, in September 2013, the mock election results showed a 16-seat election victory for the ABLP. At that time, the results were gladly embraced by the then- leader of the opposition, Gaston Browne, who said, “The mock election has really borne out what our polls suggested a few months ago.”  Today, Gaston Browne says, “We base nothing we do off of the Free and Fair Election League. The League’s mockery poll bears no relation to reality.”
All so familiar – same sounds from different mouths.  Oh, how times have changed (or, at least the poll results).  

- Advertisement -