Margins of gross error: Pollsters at work, but for who?

39
observer-image
Independent Observer

IO, Jakarta – Three weeks from the 17 April 2019 Elections, numerous survey agencies have released the results of their research. Several survey agencies favor Joko Widodo-Ma’ruf Amin, others favor Prabowo Subianto-Sandiaga Salahuddin Uno. Surveys that favor Presidential Candidate Pair Number 01 generally show a gap of votes with the second Presidential Candidate Pair of 15-20%. This amazing gap level begs the question, “Are these real results or based on an order from a specific Presidential Candidate Pair?” After all, we note that Presidential Candidate Pair Number 02’s campaigns are always flooded with people. Wherever Number 02 holds a campaign event, whether in Manado, Makassar, Lombok, Bali, Sragen, Lamongan, or Banyuwangi, there is always a flood of people attending. The people’s enthusiasm in attending Prabowo-Sandi’s campaign shows that they want a change of national leadership.

Map of Power
Firman Noor, Head of the Indonesian Institute of Sciences (Lembaga Ilmu Pengetahuan Indonesia – “LIPI”)’s Political Studies Center, concludes that when we correlate the power map of Presidential Candidate Pairs 01 and 02 with survey results, objectively Candidate Pair 01 is still leading, but not by what you would term “a safe and comfortable margin”. After all, many other surveys show that Candidate Pair 02 is surging ahead, such as the survey carried out by Rumah Demokrasi. Furthermore, the number of undecided voters remain large: surveys from Rumah Demokrasi noted 14%, Polmark said 18%, and the ultra-establishment Kompas newspaper admitted there were 13%. “So there is still an opportunity for either pair,” Firman said.

The flood of people to Pair Number 02’s campaign should be interpreted as the people’s aspiration, as citizens come to such events because they want change. The Campaign Team’s duty is to convince the voters. If a campaign is quiet, we need to ask why it is so. True, the flood of people to a specific candidate’s campaigns is not the same as election results, but we can consider that the candidate with the bigger crowd is the one whom the people pin their hopes and desires on. It can be inferred that the people like a certain candidate for giving them hope.

Ujang Komarudin, Al-Azhar University Indonesia (Universitas Al-Azhar Indonesia – “UAI”) Jakarta’s Political Observer, concludes that the power map between Candidate Pairs 01 and 02 is still balanced. The incumbent is strong, but so is the opposition. If we correlate this with the Kompas survey, Ujang notes that these survey results bestow an advantage on Presidential Candidate Pair 02. “Psychologically, this gives second wind to the Candidate Pair 02’s team. However, this is a double-edged sword for Candidate Pair 01’s supporters. This fact may bring their morale down, but it may also serve as a check to evaluate their work so far and allow them to make necessary corrections while there is time,” he said. “The true competition is still 3 weeks away; anything is possible. Either party may still win, depending on the strategy they deploy.”

Indonesia’s Educational University (Universitas Pendidikan Indonesia – “UPI”)’s Political Communica­tions Observer Karim Suryadi said that the power of each party can be seen from the amount of mass who attend the campaign events being held. The bigger the number of people who show up at a campaign, the bigger the penetrative value of the campaign itself, especially if it is greatly broadcasted. “So the amount of mass that attends a campaign would only generate power for the candidates if the public knows about it,” he said. “The great number of people turning out for a campaign event can signal a wish for change; what matters is how the image is managed or broadcast through a clear message that people can read, because facts only have power when the public comprehends them.”

Presidential Candidate Pair 01 perpetually claims that he has done real work, but Presidential Candidate Pair 02 also has experience. This is what Presidential Candidate Pair 02 must emphasize – that they have the experience to do something for the country outside of their capacity as President, such as their entrepreneurial experiences. The most important asset a candidate can have is political capital, which is their experience in managing public affairs relevant to the position they aspire to.

Jokowi’s Electability vs. Prabowo’s
Other than the power map of both sides, we also need to know about their electability rates. Political observer and founder of Lingkar Madani (LIMA) Indonesia Ray Rangkuti observes in the Kompas R & D Survey that some phenomena may impair the electability of Presidential Candidate Number 01.

First, the 01 Team neglected to maintain the “spirit of Reform”. They did not offer change – instead, they are too focused on boasting about the great achievements made by Presidential Candidate 01. On the other hand, 02 Team gives hope of change. They maintain the spirit of Reform and continue to focus on providing solutions for people’s problems, such as job opportunities, prices of basic necessities, etc.

Second, several activists and scholars felt so uneasy about the 01 Team that they migrated to the 02 Team. Such discomfort is caused, among others, by unsolved cases (e.g. the attack against Corruption Eradication Commission (Komisi Pemberantasan Korupsi – “KPK”) investigator Novel Baswedan, conspiracy criminals who are not yet brought to court) or wrongful arrests (e.g. Robertus Robert’s arrest for making and publishing songs that insult the military).

The third thing Ray noticed as a cause of the decline in Jokowi’s electability is a lack of coordination in his National Campaign Team (Tim Kampanye Nasional – “TKN”)’s in carrying out their duties, especially the fact that it tends not to be involved when many supporting groups make their declaration of support.

Meanwhile, Para Syndicate Executive Director Ari Nurcahyo states that the feeble militancy of 01 Team supporters can erode the Candidate’s electoral power, especially if they do not vote in the Voting Points (Tempat Pemungutan Suara – “TPS”) on Voting day.

We all know that the 02 Team supporters’ militancy is higher. This is something that 01 supporters should imitate, such as by arriving early at the TPS. With 300 votes available for each 4-cubicle TPS, each voting cubicle need only house 75 votes. With voting duration set at 6 hours, there is about 4 minutes available for each citizen to vote. However, each citizen has to vote for 5 offices using 5 individual forms. The average amount of time required for this recorded in several simulations is 8-11 minutes. Therefore, we also need to tell the voters that they need to vote quickly and save everybody a lot of time.

Ari further noted that the 01 Team tends to forget to mitigate the inevitable negative impact generated by Jokowi’s achievements, especially in terms of infrastructure construction, when they campaign. “They forget that many people may be victimized by toll road construction – the economies of many small towns, such as those in the Northern Coast of Java (Pantai Utara Jawa – “Pantura”), have become depressed and died off because they are no longer passed through by drivers. They should have addressed the mitigation of this negative impact too,” he said.

Vote Migration
Firman stated that an advantage in a survey should not be used as a benchmark. This is because people tend to migrate their votes along various issues and it is hard for survey agencies to capture this tendency, such as what took place with the decision by the British to exit the European Union (“Brexit”). “When I was in London to watch the election process, which ended at 12.00 noon, the trend until then seemed to incline toward victory for those who approved of remaining in the European Union. Yet it later turned out that the Brexiters won the day. This proves that there is something that the survey agencies fail to catch, and that there are many other things that we need to review because the phenomenon is real,” he said.

This phenomenon is also occurring in Indonesia: to be exact, the results of the Regional Head Elections in DKI, West Java, and Central Java. We know that survey results do not determine things significantly, so let’s just discuss the other factors that we think may affect the issue of migrating votes:

First, the public suffers an accumulative dissatisfaction that neither campaign teams or survey agencies catch.

Second, there is a change in how the Presidential Candidate Pairs present themselves. Prabowo, who had the image of a harsh warrior on horseback, now dances relaxed in public. Even during Presidential Candidate Debates, he seems reluctant to name anyone in order to attack them. Prabowo now tends to moderate himself. This is a good strategy: if he had stayed with his old image, there might be few advantages for him. After all, the recent trend is to humanize leaders, so that the people will like them. On the other hand, Jokowi used to be calm and had a low profile. However, during the Debate, he was tough and attacked a lot. This is dangerous for him, because to repeat: the current trend among the people is that a leader must be down-to-earth, humanistic, and work properly. Appearing tough is no longer necessary.

Third, the lousy workings of the party and volunteer mechanism. In both the recent Regional Head and Presidential Elections, it was not the party mechanisms that moved the campaign. Most of the parties have deflated; they are poorly motivated and it is difficult for them to generate quality cadres. This is why they depend so heavily on volunteers. When volunteer networks move by their own volition, they are much more militant than parties. It is the militancy of this network that generates the magic of change in several regional elections that survey agencies fail to detect, such as during the DKI Gubernatorial Elections.

Fourth, program offers are more realistic. One of the weak points of the current Government is its management of the economy, especially in terms of providing job opportunities and maintaining the people’s buying power. Presidential Candidate Pair 02 capitalized on this fact, as we have seen how they always turn everything back to the economic issue in the First, Second, and Third Presidential Candidate Debate. This is quite relevant to the current condition of the people, as many of the incumbent’s programs have not been on target and may thus be criticized. For example, the results of the Indonesian Sciences Institute (Lembaga Ilmu Pengetahuan Indonesia – “LIPI”) survey show that the public considers that there are things that are more important than infrastructure.

Fifth, “X factor”. For example, in the last American election campaigns, Hillary Clinton was shown by various surveys to have the advantage. Later, information from the FBI stating that Hillary’s personal e-mails were mixed up with Government-related e-mails tarnished her image. Furthermore, Donald Trump’s primordial campaigns drew many votes towards his victory. These factors are more significant and reliable than survey results when attempting to recognize possible voter migration.

Similarly with Firman Noor, LIPI researcher Aisyah Putri Budiatri concludes that vote migration might occur during the 2019 events, and that change of electability levels would probably continue all the way until Election Day on 17 April 2019. Such migration might occur because of the following causes:

First, the high number of unidentified voters. “Unidentified voters” are voters who have not made up their minds, who are still doubtful about who to vote for in the 2019 Presidential Elections, and voters whose opinions are not recorded by surveys because they decided to keep them a secret or refused to answer when being surveyed. The majority of survey results show that the level of unidentified voters is still in the 10%-17% range. This is a significant number, and it is more than able to change the voting composition of Presidential Candidate Pairs’ electability rate.

Second, Presidential Candidate Pairs’ campaigns and debates would definitely affect the results of the election. They can sway undecided voters, who would only suddenly discover for themselves which candidates and proposed programs appeal to them. They might simply suddenly change or make up their minds after reviewing campaigns and debates.

People tend to be surprised when survey results, such as the one published by Kompas daily news­paper, inform them that the gap of preference between Presidential Candidate Pair 01 and 02 was only 11%. This is laughable, because this level is entirely possible in any election. What should the candidates and their teams do then? Undecided voters are highly rational and doubtful voters; you cannot easily sway or persuade them to choose either Presidential Candidate Pair through emotional means. Therefore, campaign teams need to campaign more massively and reach out more widely in their campaign programs.

Campaign teams must also ensure that their voters actually go to the TPS and vote, as from experience, there was a high level of non-voters in both the 2009 and 2014 Elections, at 26% and 29% respectively. This is a huge gap from the statement made by many survey agencies at the time, that only 10% were “unidentified voters”. That means that more than a quarter of voters fail to arrive at TPS, while even 10% is more than enough to significantly determine victory. “The last American election also gave us a lesson in why Hillary lost: that’s because her loyal supporters did not go to their TPS because they were sure that Hillary was going to win,” Aisyah said.

The DKI gubernatorial Election was different from the condition during the Presidential Election, because there is the blasphemy issue surrounding Ahok as a candidate. This caused survey results to fail to predict the actual outcome. Similarly, the Central Java and West Java Regional Head Election contained unmeasured factors, i.e. the high number of unidentified voters and the militancy that loyal supporters ensure that they sway their compatriots who were still of two minds, and supporters’ militancy in ensuring that they themselves went to TPS’s to vote. Within the remaining campaign period from now until 13 April 2019, Campaign Teams must ensure the sterling performance of their Presidential Candidate Pair in the remaining two debates. Political parties must also continue to map out the regions that they can take over and fight for. Afterwards, each Presidential Candidate Pair must ensure that their supports actually go to the TPS’s to vote. “Campaign Teams must really work on making clear-cut, beneficial programs, because this is a significant way to make swing voters feel safe and certain of their choice,” Aisyah said.

Ramdansyah, Rumah Demokrasi Founder, said that within the next 21 days, both Presidential Candidate Pair 01 and 02 should fight to win as many swing voters’ hearts as possible to increase their voting percentage…and they must do so fairly. The victory of the upcoming 2019 Presidential Election winner is absolutely determined by the migration of these unidentified voters, after all.

Several survey agencies have taken note of the potential. Rumah Demokrasi’s survey found that there remain 14.25% voters with no identifiable preference as of yet, a mere one month before the Election Day. This is nearly the same level as discovered by survey results during the 2016 American Presidential Elections. This percentage is much higher than those of the two previous Presidential Elections, which was only about 8% in 2012 and 6% in 2008.

The untouched and unidentified voters either feel that they are insufficiently informed to make an intelligent choice, or that they do not feel the need to vote. In the 2016 US Elections, these two groups of swing voters added their ranks to the group of depressed partisan voters, independent voters who hate the establishment being represented by Hillary Clinton, and uncaring voters who were affected by negative campaign framing against her, causing Donald Trump to win. The experience we have from the 2016 US Presidential Election is that identity politics affect the public’s emotions so strongly, that they cause the mass movement of unidentified voters’ votes.

The Kompas R & D release on 20 March 2019 and the Rumah Demokrasi survey on 15 March 2019 show that the gap between the two Presidential Candidate Pairs have shrunk, with 10%-14% voters still unidentified. This reflects the presence of the Bradley effect, a field bias caused by respondents’ sense of religiosity and/or fear that makes them more wary in giving information. They are reluctant to answer, which means that the results of the election may become very different from what is expected.

Presidential Candidate Pair 01 is trapped in formal means to reach the unidentified voters’ group. The incumbent tends to approach crowds more. On the other hand, Presidential Candidate Pair 02 went with business as usual, by traveling to all corners of Indonesia and reach out to untouched groups. Vice-Presidential Candidate 02 Sandiaga Uno has even claimed that he has visited 1,500 points all over Indonesia within 7 months. Sandi has even visited regions known as the opponent’s voter bases, such as Central Java, East Java, and even Bali. The target is to erode Presidential Candidate Pair 01’s electability. In order to generate extreme exodus from the untouched group, an emotional and confronting attitude must be built among them during the open campaign from 24 March 2019 to 13 April 2019. This is the only way.

Currently, swing voters’ level is 13%. This is an extremely strong determinant. “So, we still cannot predict who the winner might be. Therefore, both sides must work on swing voters, who are generally made up of millennials. As the people are currently polarized into 2 parties, this group will determine who wins the Election. Whoever get their votes will be at an advantage,” Ujang said.

The best way to win over millennials is to provide them with solutions for their needs. “Millennials are restless and anxious. Originally, they were apolitical. They generally aspire to be entrepreneurs with digital-based business. This is what millennials want nowadays – they need a figure with measurable programs to walk along their path,” Ujang said.

Karim Suryadi stated, “Even now I am sure that the actual number of swing voters is still larger than the level stated by several survey agencies. These swing voters generally dislike Jokowi, but they cannot accept Prabowo fully. They continue to “wait and see”. I believe that actually about 30% voters are still swing voters, higher than the electability levels of either Presidential Candidate Pair 01 or 02.”

Survey Agencies’ (Lack of) Independence
Ramdansyah stated that even though many survey results have been released, the people should not put full trust in survey agencies. “If survey agencies want the public to trust them and remain in existence, they must deliver true survey results,” he said.

The tendency of survey results being too similar with each other indicates that they are likely paid off by a certain Presidential Candidate Pair. Therefore, Aisyah requests that the public look at survey results with a critical eye, among others by checking whether these survey agencies are members of a Survey Association or not. Membership in such an association would place additional limitation on the agencies, because they normally abide by a Code of Ethics. Furthermore, the Association’s members must have similar population throughout Indonesia, both in terms of demographic types and population comparison per region. We must not look from raw electability levels only.

Firman stated that we should not put blind trust in surveys, because there might be something behind all that. He further stated that our suspicions should be raised whenever survey agencies show large voting margins, even up to two digits. “I think that there is another agenda operating here,” he said. Surveys are actually merely supplemental, and they are also frequently used to direct the public’s opinions. Survey results during the DKI, West Java, and Central Java Regional Head Elections do not match reality. This means that they are not that reliable as a reference.”

Ujang believes that one of the purposes of surveys is to build up public opinion. However, he admits that while some surveys are true, others are presented based on a specific agenda. “It’s all useless if survey agencies play around with their survey results,” he said.

If we correlate survey results during the 2017 Regional Head Elections in DKI Jakarta, West Java, and Central Java, where the survey turned out markedly different from the results of Regional Elections Commission Count, Ujang said, it would not be too far-fetched if we think survey agencies were in fact working as part of the campaign team. This caused survey results to be less than objective. “This is why Ahok ended up losing even though DKI survey agencies favored him. I saw a large gap between the survey results and actual vote count. The people had a massive movement to support change; that’s the same result as the elections in West Java and Central Java. So the people want change. Therefore, surveys nowadays tend to apply only on that day, i.e. when they were held, they do not apply for the next day. They only validly measure electability on that day. Now what we worry about is if survey agencies are no longer neutral and become sold off to either Candidate’s campaign team,” Ujang said.

Political observer Rocky Gerung also questions the independence of survey agencies nowadays. However, he admits that nowadays the public can differentiate between agencies that are true “surveyors” from agencies that are actually “consultants”. “Political consultants masquerading as survey agencies will make statements according to the agenda of specific parties,” he said.

Karim Suryadi said that in terms of survey agencies’ independence, the difference in surveys result from different conditions between voters where the survey was created and Indonesia. In other words, electability surveys in Indonesia and in America are not the same, even though both surveys are performed honestly and independently, or there is no data manipulation involved, the results would naturally be different. This is because electability survey results in Indonesia are nearly never a good predictor of victory. For example, during the latest Regional Head Election in West Java, a survey held -3 Days before Elections generated an exact reverse of the actual voting. This is because Indonesian voters are different from those of other countries. In Indonesia, people generally decide on their choice at the very last minute. Surveys may be performed mere days before actual election day, but voters actually may have not made up their minds on who to vote for when they are surveyed!

“Therefore, I always say that we should never trust surveys at face value, especially if there is no explanation about how a survey arrives at its conclusion. Furthermore, even though survey agencies have already explained the methodology that they use, candidates must still remain careful, because it is quite possible that the respondent have not determined their choice when they are asked. This is because of cultural differences among voters: our citizens do not determine their political preference for a long time beforehand. This is the exact opposite with how things are in America, where families would have said proudly that “We are Democrats”, “We are Republicans” long before election day, and also show their loyalty by putting up stickers on their homes and cars. In Indonesia, instead of being predicting tools, most survey results are used as bandwagons for directing voters and working on their certainty,” he said.

Karim notes something strange about Indonesian survey results: when a survey agency shows a different result, the others would gang up and pass judgment on it. They would call the results different, unreliable and unfair. For example, the survey agencies that predict the victory of Presidential Candidate Pair 02, or that report a thinning gap between the electability of Presidential Candidate Pair 01 and 02 tend to be judged. “I think these survey agencies confirm each other. The more random the differences are, the more it’s obvious that these survey agencies are killing each other. This is because even if they perform their surveys properly, the result should not be too different from each other. With different sampling, but the same percentage of the population and the same technique, the results should not be too different. Therefore, any survey agency with significantly different results would weaken the others. This weakness is actually generated by survey agencies themselves, because survey agencies tend to confirm or correct one another,” he said.

Agencies’ survey results in Indonesia are nothing but a tool to sway public opinion, to generate electoral effects. They do not reflect actual conditions, as even the construction of these surveys tend to be on the hypothetical side.