By Dr. D.K. Duncan, Contributor 
D.K. Duncan
A PHOTO finish in the October 16, 2002 General Election is once again a real possibility. The People's National Party (PNP) won the last general elections by a margin of 17 points over the Jamaica Labour Party (JLP). The most recent Observer/Stone poll, which was conducted between October 8 and 10, 2002, shows the PNP with a one point lead over the JLP. Therefore, at this time there is a 16-percentage point swing in favour of the JLP since the December 1997 elections. (See table 1)
Table 1 also shows that the JLP has increased its support amongst registered voters by 10 points when the poll findings are again compared to the 1997 General Election results. The 16-point swing includes these 10 points as well as a 6-point decrease in voter support for the PNP over the corresponding period.
When this 16-POINT SWING is used as a PREDICTOR, the likely results of a General Election held on October 10 would be 33 seats for the JLP and 27 for the PNP. (See table 2) This is just six days before the actual elections are due to be held on Wednesday, October 16, 2002.
Using only the 10-POINT INCREASE in JLP support as a PREDICTOR, the likely result of an election held on October 10 would have been JLP 25 seats, PNP 35 seats. This is without factoring in any decrease in support for the PNP. (See table 3) Up to this point, I have been using this predictor - the most conservative approach, as the onus is usually on the opposing party to gain support.
With voting day scheduled for only six days after this last poll was conducted, the most appropriate predictor is the overall swing (16-points).
CONSTITUENCY PROJECTIONS - METHODOLOGY
The seat-count and constituency projections are arrived at by comparing the results of the 1997 General Elections with the October 8/10 Stone findings. (See Table 1)
This comparison as shown in Table I demonstrates that:
The PNP won the 1997 General Elections by a margin of 17 points (56 per cent to 39 per cent) over the JLP.
The October 8/10 Obser-ver/Stone findings show the PNP with a margin of one point (50 per cent to 49 per cent) over the JLP.
There is a 16-point decrease in the margin of victory by PNP over the JLP since the 1997 General Elections (from 17 to 1).
The PNP with 50 per cent support is 6 points behind the 56 per cent support it received in the 1997 elections.
The JLP has increased its 1997 support by 10 points, from 39 per cent to 49 per cent.
Tables 2 and 3 lists the 60 constituencies (numbered 1 - 60) in descending order of the magnitude of the margins of victory by the PNP over the JLP in the 1997 elections.
As stated in my article of October 4, 2002, twenty-three seats have definitely been decided and are not likely to change barring a major disaster, 16 for the PNP and seven for the JLP. These are highlighted as SAFE seats in both tables and charts. (Constituencies nos. 116 for the PNP and nos. 5460 for the JLP).
A PREDICTOR
The number of BATTLEGROUND seats (too close to call) differs according to which predictor is used. Using the 10-point increase in JLP support as the predictor, there are eight battleground seats, constituencies nos. 31-38 as shown in Table 3. (The number of battleground seats is arrived at by including all those seats with a PNP margin of victory in the 1997 elections ranging from 7 to 13 points, that is 10 plus or minus 3.) This would leave 14 seats (constituencies nos. 17-30) where the PNP would have the advantage and 15 for the JLP (constituencies nos. 39-53). The overall count (outside of the 8 battleground seats) would be 30 PNP and 22 JLP as shown in table 3 and chart 3.
THE PREDICTION
As indicated earlier, the most appropriate predictor at this time is the overall swing (16 points as at October 10). Using this predictor, there are nine BATTLEGROUND seats (too close to call), constituencies nos. 24-32 as shown in table 2. (In this case, the number of battleground seats is arrived at by including all those seats with a PNP margin of victory in the 1997 elections ranging from 13 to 19 points, that is 16 plus or minus 3.) This would leave 7 seats (constituencies nos. 17 23) where the PNP would have the advantage and 21 for the JLP (constituencies nos. 33-53). The overall count (outside of the 9 battleground seats) would be 23 PNP and 28 JLP as shown in table 2 and chart 2. Without applying a three per cent margin of error either way, all constituencies with a margin of victory for the PNP of less than 16 points in the last election would go to the JLP. In the converse, all constituencies above the 16-point swing would be retained by the PNP. The result would be 33 for the JLP (constituencies nos. 28-60) and 27 for the PNP (constituencies nos. 1-27).
My own predictions are based on the findings of the Stone Team, which I have interpreted as indicated above and in the tables and charts. Any change in the party standings in subsequent poll findings will continue to influence these projections. The pollsters returned to the field over the last few days. They include the Stone team for The Observer, Don Anderson team for The Gleaner. Their findings should be
known by Tuesday, after this article has been written.
VOLATILITY AND REVERSALS
The trends in the poll findings as indicated in Table 4 and Chart 4 suggest that the Golding Factor is real. It has been the dominant issue since late September, the period between the last two polls.
Separately, in previous articles we noted that a significant trend in the Stone Poll findings since 1976 had been reversed. This reversal was first manifested in the August 17/18, 2002 poll when the PNP took the lead over the JLP after some 21 months. Up to that time the JLP had an unbroken lead. (Since November 2000) Prior to this any party that won a general election had maintained an unbroken lead in the public opinion polls for several months.
During September, the relative party standings seem to have been holding. The October 8-10 Observer/Stone Poll findings represent another dramatic reversal in the opinions of the registered voters. This volatility emerged since the findings of the poll conducted at the end of March 2002. When the findings of the period December 1997 to March 2002 is compared to those of March 2002 to present (see table 4 and Chart 4), it could be argued that the Jamaican electorate might be becoming much more sophisticated.
ELECTION NIGHT
Table 2 can be useful to readers who wish to make the predictions based on the next polls and also track the results after the polls are closed on Election Day. The predictor that is recommended to be used is the latest reported overall swing. Any constituency with a PNP margin of victory over the latest published swing is likely to go the PNP. In the converse any constituency with a PNP margin of victory less than the latest published swing is likely to go to the JLP. Up to this time that number is 16 (calculated using table 1).
There are ten columns in table 2. The third column shows the PNP margin of victory over the JLP in the last elections. Columns 9 and 10 are blank and for you, the reader to fill in. After applying the latest number representing the overall swing to column 3, put in the party you expect to win that seat in column 9. On election night, put in the actual results in column 10 and see how good an analyst you are.
Also remember, that the JLP won 10 seats in the last election, listed from nos. 51 to 60 in this table. If there is going to be change of Government, they will need to add another 21 seats. In the main, these 21 should come from 27 constituencies, nos. 24 to 50.
However, as I said in the article of October 4 where I had done a similar analysis, "the old adage reminds us, 'there is many a slip between the cup and the lip'. Unforeseen obstacles haunt the parties. We could be in for more electoral drama, especially if the polls continue to show volatility. The prospect of a photo finish is still on the horizon."
One Love, One Heart.
Former PNP General Secretary and Government Minister in the PNP Administration of the 1970's. Dr. Duncan, a dental surgeon, recently established the D.K. Duncan Political Institute. Email : dktruth@hotmail.com
