Tweet
Personally, I’m surprised that the daily Gleaner continues to use the services of this man Bill Johnson.
This American pollster (Johnson) is, in my opinion, a joker of a pollster. Read the following extract from the newspaper article below:
The article below has been copied from The Trinidad Guardian, Sunday December 17, 2006
(http://legacy.guardian.co.tt/archives/2006-12-17/HamidGhany.html )
The victory by Sir John Compton and the United Workers Party (UWP) in St Lucia has undermined the credibility of two opinion pollsters in the Caribbean, namely Bill Johnson, who does polls for the Jamaica Gleaner and the People’s National Party (PNP) and Peter Wickham of the Barbados-based Caribbean Development Research Services (CADRES).
Both pollsters had predicted the return to power of Dr Kenny Anthony and the St Lucia Labour Party (SLP).
Perhaps, Johnson’s predictions were far more extreme than Wickham’s. Johnson predicted that the SLP would win 14 of the 17 seats, while Wickham’s worst-case outcome was predicted to be ten–seven in favour of the SLP.
The reality is that the final outcome was 11–six the other way, in favour of the UWP. The initial reaction to this is that there seemed to be a late swing in the campaign, or the methodology used by both pollsters was flawed.
In the case of Bill Johnson, his poll was conducted on November 25 and 26, but only published on December 10.
In the world of opinion polling, that kind of time delay can be misleading, because it fails to capture shifts in public opinion that have taken place in the interim.
The Johnson poll was conducted among 1,200 registered voters, and had a margin of error of plus or minus three per cent.
The poll indicated that the lead of the SLP over the UWP was 49 per cent to 32 per cent, with nine per cent undecided and ten per cent saying they would not vote.
When handling the trickier aspect of conducting polls in a first-past-the-post system, Johnson translated this into a 14–three projected majority for the SLP. He was so wrong.
The people of St Lucia voted resoundingly the other way. Preliminary results indicate that 74,476 of the 134,443 people registered to vote cast their ballots.
This would give a turnout of about 55.39 per cent. The split between the main parties was 38,427 (51.59 per cent) for the UWP and 36,049 (48.40 per cent) for the SLP.
If one is to believe the Bill Johnson poll, then the UWP reversed an adverse 49 per cent to 32 per cent finding on November 25–26 into a 51 per cent to 48 per cent turnaround by election day.
Did the people of St Lucia shift their opinions so massively in the space of two weeks, or was this a case of a faulty methodology?
This American pollster (Johnson) is, in my opinion, a joker of a pollster. Read the following extract from the newspaper article below:
Originally posted by Dr. Hamid Ghany
Originally posted by Dr. Hamid Ghany
(http://legacy.guardian.co.tt/archives/2006-12-17/HamidGhany.html )
Opinion polls vs the people in St Lucia
The victory by Sir John Compton and the United Workers Party (UWP) in St Lucia has undermined the credibility of two opinion pollsters in the Caribbean, namely Bill Johnson, who does polls for the Jamaica Gleaner and the People’s National Party (PNP) and Peter Wickham of the Barbados-based Caribbean Development Research Services (CADRES).
Both pollsters had predicted the return to power of Dr Kenny Anthony and the St Lucia Labour Party (SLP).
Perhaps, Johnson’s predictions were far more extreme than Wickham’s. Johnson predicted that the SLP would win 14 of the 17 seats, while Wickham’s worst-case outcome was predicted to be ten–seven in favour of the SLP.
The reality is that the final outcome was 11–six the other way, in favour of the UWP. The initial reaction to this is that there seemed to be a late swing in the campaign, or the methodology used by both pollsters was flawed.
In the case of Bill Johnson, his poll was conducted on November 25 and 26, but only published on December 10.
In the world of opinion polling, that kind of time delay can be misleading, because it fails to capture shifts in public opinion that have taken place in the interim.
The Johnson poll was conducted among 1,200 registered voters, and had a margin of error of plus or minus three per cent.
The poll indicated that the lead of the SLP over the UWP was 49 per cent to 32 per cent, with nine per cent undecided and ten per cent saying they would not vote.
When handling the trickier aspect of conducting polls in a first-past-the-post system, Johnson translated this into a 14–three projected majority for the SLP. He was so wrong.
The people of St Lucia voted resoundingly the other way. Preliminary results indicate that 74,476 of the 134,443 people registered to vote cast their ballots.
This would give a turnout of about 55.39 per cent. The split between the main parties was 38,427 (51.59 per cent) for the UWP and 36,049 (48.40 per cent) for the SLP.
If one is to believe the Bill Johnson poll, then the UWP reversed an adverse 49 per cent to 32 per cent finding on November 25–26 into a 51 per cent to 48 per cent turnaround by election day.
Did the people of St Lucia shift their opinions so massively in the space of two weeks, or was this a case of a faulty methodology?
.
Comment