The Double-Edged Sword Called Opinion Polls

[Lite Beer]

EDITORIAL
The double-edged sword called opinion polls
The Nation March 10, 2013 1:00 am

BANGKOK: -- It seems that pre-election survey predictions altered voters' behaviour and had an effect on the result of last weekend's city gubernatorial contest

Opinion polls have always been suspected of influencing election outcomes one way or another. One theory about the March 3 Bangkok gubernatorial election is that Sukhumbhand Paribatra won largely because most pre-election polls had predicted he would lose, thus causing "disillusioned" Democrat supporters to stick with the party instead of going for an independent candidate.

Not that the authorities and others involved have not been aware of this phenomenon. In fact, pre-election surveys have always been a contentious issue. There were times when the news media were strictly barred from naming names when it came to reporting the results of opinion polls. In the not-so-distant past, newspapers only said "the leader" of a given political party was leading his closest rival by a given per cent. It was frustrating for the media, but did it make the game fairer? It most likely did.

Survey results can have big impacts. They can discourage voters who intended to vote for the "loser" from exercising their right if they hear their candidate has little chance of winning. On the other hand, outright leaders in opinion polls can suffer from complacency on the part of their supporters. In a neck-and-neck race, popularity surveys might help democracy by contributing to a large turnout, but the truth is that close contests are few and far between. And it's not uncommon for politicians who trail in the ratings to resort to nasty tactics that undermine democracy.

For last weekend's election, opinion polls played their controversial role against a highly complicated political backdrop. They predicted a comfortable win for Pheu Thai candidate Pongsapat Pongcharoen, causing the Democrats to launch fierce attacks that the government might have described as blows below the belt. One of the Democrats' strategies was to drum up the "fear factor" by telling voters that a Pongsapat victory would benefit Thaksin Shinawatra more than flood- and traffic-weary Bangkokians.

The outcome of the gubernatorial election refuted all the opinion polls. Quite a few analysts believe that "fears" caused by those surveys, or by the Democrats exploiting those surveys, contributed significantly to Sukhumbhand's record victory. Despite being dubbed the most colourless governor in recent times, he won a second term with more than 1.2 million votes. Pongsapat also became the first gubernatorial election loser to receive more than a million votes.

Looking back, Sukhumbhand and his party might have to thank the pollsters. Pongsapat and Pheu Thai, on the other hand, must be cursing them. However, in Thailand, the problem is that a problem is not always a problem. Pheu Thai never complains when rating surveys benefit them. The Democrats, meanwhile, lashed out at pro-Pongsapat survey results, but they know deep down that they owe Sukhumbhand's amazing triumph to those predicting he would lose.

Opinion polls, whether they are impeccably scientific or sloppily executed, can affect voters' actions. The question that everyone needs to revisit is whether the pros outweigh the cons. The polls surely provide jobs and make democracy more attractive, engaging and exciting, but they can also be manipulated or exploited, or by themselves give a false impression about a political situation.

The media certainly love opinion polls, which provide attractive headlines. But jurists in court cases are quarantined from the outside world for good reasons. The question we have to ask is whether voters equal jurists when it comes to deciding a country or city's future. If so, should voters be treated the same way as their courtroom counterparts? If not, things can be allowed to go on as usual, but we will have to live with the fact that a lot of voters will not make their choice based on candidates' merits alone, but also on pre-election numbers that might be right or wrong.

-- The Nation 2013-03-10

[Bellweather]

A lot has to do with how questions are framed and what range of answers are available, especially if tied to yes or no. Unless the pollsters are truly interdependent they can get the results, ore or less, favourable to which party etc they represent. In any case does everyone tel the truth when polled ?

[chainarong]

Opinions are like arso!!es everyone has one, some people will think , whats it got to do with them and tell polsters anything, some will be honest and some just don't give a rats arse.The chicken has arrived, time for lunch.

[MacChine]

There was a time in the USA when opinion and exit polls were excellent indicators of who was going to win an election.. Then a companies named Diebold ans ESS started selling ballot counting software.

Has Thailand turned to electronic balloting , too?

[jamhar]

"You never count your money when you're sittin' at the table."

polls that you take are tactical information. you use it to guide your plans. If you're doing really well or really poorly, you may not want to even share it with your own staff for fear it may affect their morale. This is really valuable information. Why the hell would you share it with everyone?

All i can do is shake my head in puzzlement.

[Bluespunk]

Polls have one edge and it is a dull and bloody pointless one. They tell you what you want to hear or at least what the pollsters think those paying for the poll want to hear.

Edited March 10, 2013 by Bluespunk

[chiangmaikelly]

A lot has to do with how questions are framed and what range of answers are available, especially if tied to yes or no. Unless the pollsters are truly interdependent they can get the results, ore or less, favourable to which party etc they represent. In any case does everyone tel the truth when polled ?

Did you mean to write interdependent?

[JetsetBkk]

There was a time in the USA when opinion and exit polls were excellent indicators of who was going to win an election.. Then a companies named Diebold ans ESS started selling ballot counting software.

Has Thailand turned to electronic balloting , too?

Does Thailand have a 'hanging chad' problem?

[chiangmaikelly]

There was a time in the USA when opinion and exit polls were excellent indicators of who was going to win an election.. Then a companies named Diebold ans ESS started selling ballot counting software.

Has Thailand turned to electronic balloting , too?

Does Thailand have a 'hanging chad' problem?

No and the thread is about opinion polls not elections.

[chiangmaikelly]

There was a time in the USA when opinion and exit polls were excellent indicators of who was going to win an election.. Then a companies named Diebold ans ESS started selling ballot counting software.

Has Thailand turned to electronic balloting , too?

So what elections were called wrong?

[chiangmaikelly]

Opinions are like arso!!es everyone has one, some people will think , whats it got to do with them and tell polsters anything, some will be honest and some just don't give a rats arse.The chicken has arrived, time for lunch.

Your opinion is not supported by fact.

[waza]

Like I have said before the pre-election polls are flawed because Bangkok has 20 million residents but only around 5 million can vote. However, that doesn't explain the exit polls which indicated a large win to the PTP candidate. But in the end the only number that matters is the election results

[Locationthailand]

There in nothing 'swordlike' about any opinion poll in Thailand, more like a soiled nappy.

[surangw]

a lot of it depends in who they choose to ask

[Bellweather]

A lot has to do with how questions are framed and what range of answers are available, especially if tied to yes or no. Unless the pollsters are truly interdependent they can get the results, ore or less, favourable to which party etc they represent. In any case does everyone tel the truth when polled ?

Did you mean to write interdependent?

Sorry but my keyboard can't spell but it's Thai ! Best I can think of

[otherstuff1957]

All of this "thoughtful discussion" about the polls is just a waste of time.

Either they are grossly incompetent or else they were producing the results that they were paid to produce.

[Thai at Heart]

I can't believe this amount of hand wringing over an opinion poll. Its a prediction.

What do they expect? That it is going to be accurate to the second decimal place. I really wouldn't be surprised if half of the people they ask, tell the surveyor to buzz off. In the political climate in Bangkok, that is hardly surprising. Half of the people in the city can't even vote anyway, so if you stand on a street corner, half of the people you ask, aren't even eligible to vote anyway before you start. So, unless you are going to ask whether they are even eligible to start with, the numbers can become completely skewed.

They should just ask a consumer survey company to add the questions into their research for advertising. At least that may become "representative" of the entire population of Bangkok. It's as if, getting an opionion poll wrong is some horrendous affront to the face of Thai academia. Well if that's the extent of it, strewth, they should start working out why Thailand is at the bottom of so many academic rankings before they start gnashing about getting a survey wrong.

What do they intend to do? Ban polling? Looking at the quality of half of theopinion polls we see quoted in the news here, that might be a start.

I did find it interesting that watching the news during the count, they interestingly had some correlation of votes in comparison with stated earnings, i.e. 10 to 20k per month, 20 to 30k per month etc. Who the hell answers questions like that correctly?

[A_Traveller]

Hm.. if you don't like the data flow change the parameters. By the by a Prime Minister [PM] is not Head of State but is the Head of Government: in most cases elected by the members of the house not by direct election [AKA The Westminster Model].

As to poll accuracy there have been many over time which have proven to be inaccurate, just because a poster is not aware of them does not make it so. The US Presidential election polls run by the media have become more accurate over time by a combination of factors, including but not limited to combining multiple poll into a single dataset, more detailed and comprehensive statistical analytics to reduce errors and finally it is by definition a somewhat simpler model, as opposed to multi constituency elections.

Edited March 11, 2013 by A_Traveller

[metisdead]

Off topic posts and replies have been removed.

[MacChine]

No and the thread is about opinion polls not elections.

There was a time in the USA when opinion and exit polls were excellent indicators of who was going to win an election.. Then a companies named Diebold ans ESS started selling ballot counting software.

Has Thailand turned to electronic balloting , too?

So what elections were called wrong?

OT, as you pointed out

[cycloneJ]

a lot of it depends in who they choose to ask

years ago, in a land far away, I worked in the electoral system for 4 years; polls were conducted by telephone - landline to be precise, thereby immediately narrowing the field of participants - ruling out all the people who had moved on from a fixed landline at home to mobile/cellphone only. 18-25s being the most fickle, and least 'fixed' on a party, they were also the group most likely not to be contacted via a fixed phone line.

Skews any poll somewhat.