The Stone Polling Organisation came within two percentage points of accurately predicting the percentage share of the popular vote for the two major parties in last Wednesday’s general elections, despite the unusual volatility among the electorate and Wednesday’s thunderstorms and floods that caused many voters not to cast their ballots.

In fact, when the margin of error of plus or minus 3.5 per cent on the final poll Stone conducted for the election season is taken into account, the poll, statistically, is right on target.

The final poll, conducted two days before the election and published on the eve of the vote, gave the ruling People’s National Party (PNP) a 38.4 per cent support among all Jamaicans 18 years and over. The Jamaica Labour Party (JLP) support among all voting-age Jamaicans was 29 per cent. The PNP, among adult Jamaicans, had a lead of nine percentage points over the JLP.

But Stone had pointed out that among the 70 per cent of the registered voters who had indicated they would cast their ballots, 54.9 per cent would vote for the PNP and 44.8 per cent for the JLP.

When the ballots were counted on Wednesday night, it turned out that 53 per cent had voted for the PNP and 47 per cent for the JLP.

“The Stone Organisation has polled consistently for the Observer since 1995 and with each outing has reinforced a well-earned reputation,” said the Observer’s editor-in-chief Paget de Freitas. “They got it right again, even in an environment as the one that characterised the election just past. This is a cherished partnership.”

In the past, fewer persons actually turn out to vote than the numbers who, during the campaign, indicate that they would cast their ballots.

“October 16 held to this trend, but was exacerbated by rain that made the turn-out the lowest in the last 20 years,” Stone noted in their initial comments on the election.

There has been some controversy over the accuracy of the polls because of the electoral volatility that was captured by Stone in the election series and the inaccurate reading of the findings by some commentators.

One common misconception among commentators was that the nine-point lead which the PNP had over the JLP among all voting-age Jamaicans – including the uncommitted and those who did not intend to vote – would translate into a landslide of seats for the ruling party.

With its 56 per cent of the popular vote, the PNP won 35 seats in the 60-seat House of Representatives and the JLP, 25. The seat counts were in close proximity to the expectations of informed analysts who relied on the Observer/Stone poll data.

From late 2000 until the middle of this year, the JLP had led the PNP in popular support in the Observer/Stone Polls, but the volatility among voters became increasingly apparent as the election campaign heated up and the date for the vote drew nearer.

For instance, in an August survey for the Observer, Stone found that the PNP (37.8 per cent) had established a 3.4 percentage point lead over the JLP (34.4 per cent).

Weeks later, in a mid-September survey, the PNP’s support had dropped by just under a point, while the JLP had declined by 2.7 points, to give the PNP a 5.3 percentage point advantage.

When Stone polled again over the last two days of September and the first three days of October, the PNP (40.6 per cent) had increased its lead over the JLP (31.7 per cent) to 8.9 percentage points. Those findings were published on October 11.

Stone, however, had gone back to the field for the Observer between October 8 and 10, and found that the JLP (33.3 per cent) had picked up 1.6 percentage points in support, while the PNP (34.5 per cent) had declined by 6.1 percentage points. That translated in popular vote to 50.6 per cent and the JLP 48.8 per cent – a 1.8 percentage point advantage for the PNP.

Given the volatility that was being displayed by the electorate, the Observer and Stone decided to check if the latest position would hold. So Stone went back to the field last week Monday, two days before the election. The Observer published the results in an afternoon extra on Tuesday.

In this survey, the PNP lead among the overall electorate was back to nine per cent.

Stone explained the volatility in the electorate, in part on the fact of the unusual circumstances of the elections – Jamaica had never previously had a three-term government and no major opposition party had previously lost three consecutive general elections. In nearly all the polls before the election, the party leaders were less popular than their parties and the return of Bruce Golding to the JLP also had the potential to change the dynamics of the election.

“We suspect that when an analysis of the voter behaviour in this election is done, many electoral truths will be shattered,” Stone said. “For example, weather vane seats that would normally be used as predictors of elections might have to be discarded.”

Stone Comment

There have been some questions raised about the Observer/Stone Polls published on Sunday, October 13 and Tuesday, October 15.

Even though these polls were published days apart, the Sunday, October 13 poll was done a week earlier than published. The field work for the October 15 poll was done on October 14.

In past elections (1993 and 1997), we did not poll as often as we did in this election. There are many reasons for this, not least among them being the volatility that we picked up months leading to the election of October 16. We spoke about the volatility in the electoral responses and we believe that it did exist.

The unusual circumstances of this election were very evident:

a. Jamaica had never had a three-term government;

b. we also had never had an opposition party which had lost three consecutive general elections;

c. in nearly all the polls leading up to the October 16 election, both of the party leaders were less popular than the party they led; and

d. the injection of Bruce Golding into the ranks of the JLP, with the potential for changing the dynamics of the election.

These factors contributed to the volatile nature of the election and both the pollsters and the electorate were going into uncharted waters.

We stand by all the Observer/Stone Poll that we published because all the indicators that we used as validity checks held true for all the polls.

The commentary on our final poll on October 15 quoted a nine per cent lead in our findings for the People’s National Party that included categories such as:

. don’t know if voting;

. not voting;

. won’t say who voting for; and

. the uncommitted.

A nine per cent lead in this context is not a landslide and the Stone team did not suggest that. In fact, it translated in our findings in the popular vote (use those who indicated that they would vote) as 54.9 per cent for the PNP and 44.8 per cent for the JLP. In this respect, our poll was correct, within the margin of error of plus or minus 3.5 per cent.

Traditionally, the voter turn-out reported by the poll has been higher than the turn-out on election day. October 16 held to this trend, but was exacerbated by rain that made the turn-out the lowest in the last 20 years.

We suspect that when an analysis of the voter behaviour in this election is done, many electoral truths will be shattered. For example, weather vane seats that would normally be used as predictors of elections might have to be discarded.