Pollster Bill Johnson

Below is a statement by pollster Bill Johnson regarding the results of surveys his firm Johnson Survey Researchconducted on behalf of the St. Lucia Labour Party ahead of the general election held, last Monday.

Johnson Survey Research (JSR) conducted a series of four different polls for the election: November 18 and 19, November 25 and 26, December 5, 6 and 7.

It should be noted that on November 16 the Prime Minister announced the date of the election, December 11. The media and the on-the-ground campaigns for both parties started in earnest several days later.

Our first and second polls consisted of national samples of 888, and enhanced samples in six 'marginal' constituencies for a total sample of 1200 in each poll.

Both polls showed similar results with the ruling party, the St Lucia Labour Party, having a lead of 17 points, 49 per cent to 32 per cent, and leading in 14 of the 17 constituencies in the second poll. It should be emphasised that the second poll was conducted on November 25 and 26, more than two weeks before election day - December 11. This was the poll that was ultimately leaked to the press in St Lucia two days before the election.

Although this was the last national poll that we conducted before election day, we did conduct two other series of polls between November 26 and election day that indicated trouble ahead for the SLP.

On December 2 and 3 we interviewed 200 eligible voters in each of the six "marginal" constituencies, for a total sample of 1,200. A comparison of the cumulative results of these six polls to the cumulative results of the polling in these six constituencies in the two previous polls, showed that the SLP's margin had been cut in half in a one week period, which was clearly a foreboding sign.

We also conducted "tracking" polls on December 5, 6 and 7 with respective samples sizes of 400, 600 and 600. Each of these polls showed that the election could be considered a statistical dead heat.

All of these results were discussed in detail with our client.

On Saturday, December 9, for whatever reason, someone in the SLP campaign leaked the results of the two-week-old poll to the media (but not the results of the subsequent polls).

Retrospect

In retrospect, this probably hurt the SLP more than it helped it, as it appears to have energised the supporters of the opposition party, the United Workers Party, and resulted in complacency among some SLP supporters who incorrectly decided the election was over and didn't bother to come out and vote for one reason or another on election day.

The final election results showed the UWP receiving approximately 36,000 votes and the SLP 34,000, or a 51 per cent to 49 per cent popular vote majority for the UWP. However, because of very close outcomes in many constituencies, this resulted in an 11 to 6 UWP majority.

Let me emphasise that I stand by the results of our polls that were conducted on November 18 and 19 and 25 and 26 as being accurate representations (within accepted sampling errors) of the views of the electorate at the time they were conducted.

It has almost become a cliché that a poll is just a snapshot at a particular time, however, a series of properly conducted polls can become a series of snapshots, or a virtual moving picture, of what is transpiring in an electorate over time. The November 25 and 26 poll that was publicised was the second in a series of four snapshots, and was certainly not intended to be the final snapshot in the moving picture of the electorate's opinions.

In concluding, I would like to say that I have built up an enviable reputation for integrity and accuracy in a very challenging profession over the past 30 plus years.