PAP's secret illegal election poll

[aurvandil]

PAP's secret illegal election poll

Round about cooling off day, we saw all sorts of rumors surfacing of mysterious bookies and mysterious data mining done by the PAP.

Now that the dust has settled, the truth is starting to emerge. During the elections, the PAP or a group related to the PAP secretly commissioned an illegal election poll. This was done by an Australian based polling company called UMR Research. You can see the topline findings of this poll here:

http://umrresearch.com.au/doc/Singapore_Pre-Election_Study_May11_Final.pdf

The data collection was from 3rd May to 5th May. The method of data collection is given as online interview. This is however almost certain to be incorrect since you cannot obtain a statistically random sample via online means in such a short period. As interviewers doing face to face would have been seen, the methodology was most likely via telephone interviews conducted from overseas (hence using an Australian polling company).

The total sample size of the poll was 522. This was stratified into sub samples of 52 each, meaning that they covered 10 GRCs/SMCs of interest. The Confidence Level is 90%. The n and p values of the sample yield a standard deviation of about 3% which was bandied about in this forum,

On interest is slide 6 where the respondent was directly asked whether they would vote for PAP or the Opposition. At the overall aggregate level, the sample size is 522. The Confidence Level is approximately 96%. This incidentally is illegal.

Also of interest is slide 10. This shows the perception Singaporeans have of the various parties. WP is off the chart and coming very close to matching the PAP. The concerns about SPP surfaced as this study showed that it was the most popular among the other Opposition parties less WP. NSP was the surprise package. Their poll results in Marine Parade and Tampines were almost as good as WP. Their image according to this poll however puts them on par with SDP. Clearly there is something else happening which is not being captured in the poll.

This illegal election poll shows just how far the PAP will go to win. Just like the unfair handling of election complaints (Opposition disqualified for being 30 seconds late Vs No action against Tin Pei Ling for campaigning on Cooling Off day), it is unlikely the police will take any action even if a police report is made. If the PAP is not prepared to act fairly and ethically, then what are the chances that they will keep all those promises they made to win the elections?

 

[Cestbon]

Poll is not illegal. In every country most of them have their own poll by newpaper,TV,Media.......
How accurate or it will help the party to win vote is another matter.
Next elextion 2016 will be deciding factor for opposition . Will they manage to grab 1/3 of the seat. If yes the 2021 will be the interesting election.
If Opposition still cannot manage to hold 1/3 of the seat. Then opposition still need to break that 1/3 barrier 1st.

 

[Sperminator]

PAP's secret illegal election poll

Round about cooling off day, we saw all sorts of rumors surfacing of mysterious bookies and mysterious data mining done by the PAP.

Now that the dust has settled, the truth is starting to emerge. During the elections, the PAP or a group related to the PAP secretly commissioned an illegal election poll. This was done by an Australian based polling company called UMR Research. You can see the topline findings of this poll here:

http://umrresearch.com.au/doc/Singapore_Pre-Election_Study_May11_Final.pdf

The data collection was from 3rd May to 5th May. The method of data collection is given as online interview. This is however almost certain to be incorrect since you cannot obtain a statistically random sample via online means in such a short period. As interviewers doing face to face would have been seen, the methodology was most likely via telephone interviews conducted from overseas (hence using an Australian polling company).

The total sample size of the poll was 522. This was stratified into sub samples of 52 each, meaning that they covered 10 GRCs/SMCs of interest. The Confidence Level is 90%. The n and p values of the sample yield a standard deviation of about 3% which was bandied about in this forum,

On interest is slide 6 where the respondent was directly asked whether they would vote for PAP or the Opposition. At the overall aggregate level, the sample size is 522. The Confidence Level is approximately 96%. This incidentally is illegal.

Also of interest is slide 10. This shows the perception Singaporeans have of the various parties. WP is off the chart and coming very close to matching the PAP. The concerns about SPP surfaced as this study showed that it was the most popular among the other Opposition parties less WP. NSP was the surprise package. Their poll results in Marine Parade and Tampines were almost as good as WP. Their image according to this poll however puts them on par with SDP. Clearly there is something else happening which is not being captured in the poll.

This illegal election poll shows just how far the PAP will go to win. Just like the unfair handling of election complaints (Opposition disqualified for being 30 seconds late Vs No action against Tin Pei Ling for campaigning on Cooling Off day), it is unlikely the police will take any action even if a police report is made. If the PAP is not prepared to act fairly and ethically, then what are the chances that they will keep all those promises they made to win the elections?

Great Article.

A sample size of 522 of random sampling will not even come close to represent the national average.

On 7th May 2011, I've started my experiment, truly random, based only on hear say information, and integrity, the sample size collected was 352, Random Sampling, Normal Distribution, Degree of Confidence 5%.

Results ; is that OPPOSTION have about 50% - 60% Support | PAP have about 38% - 48% Support | Void generally taken as 2%

And the above data is not even close... due to Gerrymandering of GRC... I would say that if there is no such thing as GRC, my data will be more accurate than the Australian Based Polling Company, as my data is based on the 7th May 2011, from people who already voted. (can only base on integrity on this)

On top of that, the poll information will always change day to day, hour to hour, based on what the politicians say in Television, Interviews, Rallies, Incidents... every sentence, even every word, even their facial expression will sway votes to their advantage or disadvantage...

Politics is emotional, it has got nothing to do with logic... if any politician who is charismatic enough, and can attend to the following segments;

21 - 35 (Youth Segment) (to be subdivided to Lower | Middle | Upper | Elite classes) (to be subdivided to races) (to be subdivided to issues they are happy with / not happy with) (to be subdivided to GRC in SGP case / SMC in SGP case) (to be subdivided to which political party they support and why) (to be subdivided to which Politician they support and why)

36 - 55 (Economic Class) with the above conditional statements

56 - Oldest (Senior Citizens) with the above conditional statements

Then I can say that the polling will be fairly accurate, with degree of confidence up to 1%.

However, even with that degree of confidence, the VOTES will swing on a hourly to hourly, day to day, basis, depending on what the politicians are saying... as noted clearly, Aljunied GRC was lost because of Lee Kuan Yew's statements of asking Aljunied GRC to repent and regret for the next 5 years if they choose Workers Party...

From my perspective, Statistics / Facts don't lie... but Polling MUST BE TAKEN within 1 day do not truly represent the peoples' sentiments...

From true information on the ground;

RP did not do a good job by doing house visitations, putting up more posters, giving brochures, giving more rallies, in the last 5 years, and they don't even have a good agenda...

Originally, if, provided if, RP did the same amount of effort of doing house visitation, putting up posters, giving brochures, giving rallies, in the last 9 days, and also did their rounds of house visits, or neighbourhood visits, in the last 5 years, I can be confident that RP on 7th May 2011 can have more than 50% votes...

So, what happened to the 19% vote? There are alot of people, prior to 7th May 2011 were talking about voting for opposition party... many thought that Workers Party were contending against Pinky, however, mostly were disappointed when they found out that it was RP...

On the actual day itself, many people switch... initially the morning, the mood was to support opposition, but suddenly in the afternoon... the mood was swing towards PAP... many people are still supportive to PAP.

The above information are taken from the ground, and polling for statistical analysis should not be done in Internet... another thing is that, Singaporeans would not be willing to share what they have voted for... expect for opposition supporters, however, even so, some Singaporeans say they voted for opposition, in actual fact they voted for PAP... making the information even not reliable...

 

[lockeliberal]

Dear Arun

Ahhh that explains Seren's high confidence :_)) in his numbers prediction. I am starting to have a clearer picture as to how his modelling simulation works. The Poll itself is random but represents a national average. The national average than has to be modeled against local precints in detail. What this paper leaves out is the details necessary for local detailing with high confidence which I suspect was done accordingly.

i.e Income Age and Sex and how they poll accordingly would have been important.




Thanks


Locke

PAP's secret illegal election poll

Round about cooling off day, we saw all sorts of rumors surfacing of mysterious bookies and mysterious data mining done by the PAP.

Now that the dust has settled, the truth is starting to emerge. During the elections, the PAP or a group related to the PAP secretly commissioned an illegal election poll. This was done by an Australian based polling company called UMR Research. You can see the topline findings of this poll here:

http://umrresearch.com.au/doc/Singapore_Pre-Election_Study_May11_Final.pdf

The data collection was from 3rd May to 5th May. The method of data collection is given as online interview. This is however almost certain to be incorrect since you cannot obtain a statistically random sample via online means in such a short period. As interviewers doing face to face would have been seen, the methodology was most likely via telephone interviews conducted from overseas (hence using an Australian polling company).

The total sample size of the poll was 522. This was stratified into sub samples of 52 each, meaning that they covered 10 GRCs/SMCs of interest. The Confidence Level is 90%. The n and p values of the sample yield a standard deviation of about 3% which was bandied about in this forum,

On interest is slide 6 where the respondent was directly asked whether they would vote for PAP or the Opposition. At the overall aggregate level, the sample size is 522. The Confidence Level is approximately 96%. This incidentally is illegal.

Also of interest is slide 10. This shows the perception Singaporeans have of the various parties. WP is off the chart and coming very close to matching the PAP. The concerns about SPP surfaced as this study showed that it was the most popular among the other Opposition parties less WP. NSP was the surprise package. Their poll results in Marine Parade and Tampines were almost as good as WP. Their image according to this poll however puts them on par with SDP. Clearly there is something else happening which is not being captured in the poll.

This illegal election poll shows just how far the PAP will go to win. Just like the unfair handling of election complaints (Opposition disqualified for being 30 seconds late Vs No action against Tin Pei Ling for campaigning on Cooling Off day), it is unlikely the police will take any action even if a police report is made. If the PAP is not prepared to act fairly and ethically, then what are the chances that they will keep all those promises they made to win the elections?

 

[Sideswipe]

So, what happened to the 19% vote? There are alot of people, prior to 7th May 2011 were talking about voting for opposition party... many thought that Workers Party were contending against Pinky, however, mostly were disappointed when they found out that it was RP...

On the actual day itself, many people switch... initially the morning, the mood was to support opposition, but suddenly in the afternoon... the mood was swing towards PAP... many people are still supportive to PAP.

you are suggesting 19% of electorate who are favorable to voting for oppositions eventually decided to cast their votes for PAP on 6th or 7th May.

19% of electorate is a lot. imo it's unlikely that such a large number of voters would still be undecided or switch support from oppositions to PAP 1 day before or on polling day unless very dramatic events happened to influence the election.

 

[scroobal]

The polls were indicated on polling by Reuters

http://in.reuters.com/article/2011/05/07/idINIndia-56846820110507

SINGAPORE | Sat May 7, 2011 7:06pm IST

SINGAPORE (Reuters) - Singaporeans voted on Saturday in the city-state's most hotly contested parliamentary election since independence, although Prime Minister Lee Hsien Loong's long-ruling People's Action Party (PAP) is almost certain to retain power.

Voting ended at 8 p.m. (1200 GMT) without any major incidents and results are expected to start coming in within a few hours.

Analysts say the PAP should easily win, albeit with a smaller share of the popular vote. At the last election in 2006, it won about 67 percent of the vote and 82 of 84 seats.

This year the opposition is contesting 82 of 87 seats in parliament, the most ever, with the PAP returned unopposed from one five-seat constituency. In 2006, just over half the seats were contested.

"Given the strength of the opposition slate, we expect an electoral swing of 7-12 percentage points from the PAP," Prasenjit Basu, Daiwa Capital Markets regional chief economist, said in a note to clients.

"The enhanced opposition representation in the new parliament will give a substantive new voice to dissent, providing for greater debate on key policy issues."

The Elections Department said an opposition party had made several complaints, including an allegation that a PAP candidate had updated her Facebook page on Friday, contravening a 24-hour "cooling off" period before the start of polls when parties were not allowed to campaign.

"We had responded to National Solidarity Party that they may wish to lodge a police report," a spokeswoman for the department said.

As at 5 p.m. (0900 GMT), 1.85 million Singaporeans, or 83.5 percent of eligible voters, had cast their votes, the department said in a separate statement.

Voting is compulsory in Singapore, a former British colony which in 1965 became a republic after breaking up with neighbouring Malaysia.

Opinion polls are not published in Singapore and exit polls are banned, but an online poll conducted by Australian group UMR Research earlier this week indicated that the PAP's share of the vote may fall to 61 percent.

(Reporting by Kevin Lim; Editing by Raju Gopalakrishnan)

 

[aurvandil]

Dear Arun

Ahhh that explains Seren's high confidence :_)) in his numbers prediction. I am starting to have a clearer picture as to how his modelling simulation works. The Poll itself is random but represents a national average. The national average than has to be modeled against local precints in detail. What this paper leaves out is the details necessary for local detailing with high confidence which I suspect was done accordingly.

i.e Income Age and Sex and how they poll accordingly would have been important.




Thanks


Locke

Given that the poll asks the illegal "who you will vote for" directly, there is no need to use data mining or any other kind of black box technique. All you need to do is weight the responses by the demographics in the GRC / SMC of interest.

Here, a cardinal law of stats is that you cannot improve the power of any sample by weighting or other technique. At the end of the day, you are still limited by 52 observations per GRC/SMC giving you a 90% CI for whatever weighted number you come up with.

 

[Sperminator]

you are suggesting 19% of electorate who are favorable to voting for oppositions eventually decided to cast their votes for PAP on 6th or 7th May.

19% of electorate is a lot. imo it's unlikely that such a large number of voters would still be undecided or switch support from oppositions to PAP 1 day before or on polling day unless very dramatic events happened to influence the election.

What I am suggesting is that 19% of these people are Pro-PAP, but just talk about voting for Opposition.

And when they re-think about the fact that RP did almost nothing to win the votes, these people decided to vote for PAP no matter how angry they were about the policies. Besides, AMK GRC have the most PAP supporters in the whole of Singapore, and that is why Pinky choose to put his camp here.

For the 31%, they are generally committed, and they belong to the 21 - 35 group of people and 36 - 55 group of people... in AMK GRC, there are many 56 - Oldest (this group is die hard fans of PAP).

The 31% didn't vote for RP, they were voting for no confidence for PAP.

 

[Sideswipe]

What I am suggesting is that 19% of these people are Pro-PAP, but just talk about voting for Opposition.

And when they re-think about the fact that RP did almost nothing to win the votes, these people decided to vote for PAP no matter how angry they were about the policies. Besides, AMK GRC have the most PAP supporters in the whole of Singapore, and that is why Pinky choose to put his camp here.

point taken.

I also think a significant portion of middle ground or light PAP electorate always talk about voting oppositions but eventually vote for PAP.

 

[Rogue Trader]

Such accurate results based on a sample size 522 for 2.35 million voters? I am very amazed!

 

[Char_Azn]

you are suggesting 19% of electorate who are favorable to voting for oppositions eventually decided to cast their votes for PAP on 6th or 7th May.

19% of electorate is a lot. imo it's unlikely that such a large number of voters would still be undecided or switch support from oppositions to PAP 1 day before or on polling day unless very dramatic events happened to influence the election.

Actually for all his BSing, he may be right to a certain extent. 2 very dramatic thing took place in the 2nd half of the campaign.

PM Lee's - I'm sorry
and George Yeo's - Let me be your voice for change

I know more then a few of may friends who changed their minds after hearing those speeches. LHL was kena screwed by the 2 old men in his party for the majority of the campaign. He came out and saved his ass at the lunch rally. GY's last min appeal gave his own team a fighting chance. If he had done it earlier, I believe the % of votes would be even closer. GY's final speech which was a replica of his online video was on most front page on most major papers on Friday. That had a dramatic effect. A relative of mine posted the article on Facebook and said she decided who to vote when she had indicated more then once before cooling off day that she would spoil her vote.

 

[Sperminator]

From my opinion, I think being a Singaporean, we do have the responsibility to do our own interviews regarding about Singapore's national issues, Singaporean Electorates Concerns, Cost of Living, Healthcare System, HDB pricing / affordability issues, FT Infux policies, etc... etc... and we have to keep our discussion organised.

Another thing I realize is that we have to maintain a polling station, to keep track of the sentiments of Singaporeans. I wonder if there are any brothers out there who have very good statistics background, and is able to create accurate statistical analysis models, based on specific Issues, and Electorate Segments...

If we pump in bullshit into the Statistical Model, we'll definitely get bullshit... however, if we supply true actual information into the Statistical Model, we can actually keep track of the results on a monthly to monthly basis / yearly to yearly basis... hopefully, these information can be supplied to our Opposition Party members for them to address the true issues in parliament.

Seeking forummers' views on this, do you think this will work?

 

[aurvandil]

Seeking forummers' views on this, do you think this will work?

Of course it will work. The only problem is that if you are with the Opposition, you will be quickly arrested for doing political polling. Only the PAP are above the law.

 

[aurvandil]

The polls were indicated on polling by Reuters

My pet theory for why they released this secret poll to the foreign news wires was because there were concerns that the financial markets would be spooked by the huge rallies and negative PAP sentiment sweeping Singapore. Hence the need to tell the foreign audience not too worry and that the PAP was still firmly in control.

 

[Sperminator]

Of course it will work. The only problem is that if you are with the Opposition, you will be quickly arrested for doing political polling. Only the PAP are above the law.

What if these people were your friends & relatives? would you also be arrested for talking with your family and friends about the national issues?

 

[aurvandil]

What if these people were your friends & relatives? would you also be arrested for talking with your family and friends about the national issues?

How can the poll be respresentative of Singapore if all of those taking part are your friends and family?

 

[Sperminator]

My pet theory for why they released this secret poll to the foreign news wires was because there were concerns that the financial markets would be spooked by the huge rallies and negative PAP sentiment sweeping Singapore. Hence the need to tell the foreign audience not too worry and that the PAP was still firmly in control.

This sounds reasonable. Afterall, there are many Foreign Investments in Singapore.

 

[Sperminator]

How can the poll be respresentative of Singapore if all of those taking part are your friends and family?

Each and everyone of the Forum, each account have Singaporean Electorate, family and friends.

Each forummer can interview the national issues, and they will also know the Electorate Segments (21 - 35 | 36 - 55 | 56 - oldest)

They can be their proxy to supply information for polling through this website, however, there must be a statistical model to work on, and analyzed.

Singaporeans are not daft... in this forum, I am sure there are many intelligent people, graduates, masters, Phd, doctors, lawers, financial whiz, people from all walks of life.

thus, the sampling will definitely be Random. The people being interviewed, don't even know they are being interviewed.

 

[aurvandil]

thus, the sampling will definitely be Random. The people being interviewed, don't even know they are being interviewed.

i appreciate the spirit and the sentiment.

Unfortunately there is no such thing as a "friends and family" random sampling methodology.

The results would be biased and wildy inaccurate.

 

[Sperminator]

i appreciate the spirit and the sentiment.

Unfortunately there is no such thing as a "friends and family" random sampling methodology.

The results would be biased and wildy inaccurate.

thanks for the inputs. I appreciate the honesty. how do we get random sampling? any ideas?