Accuracy of Quick Count Guards Quality of Election

KOMPAS/HERU SRI KUMORO

The atmosphere in a quick counting room for the West Java, Central Java and East Java regional elections by Kompas research and development at the Kompas daily office in Jakarta on Wednesday (27/6/2018).

JAKARTA, KOMPAS — The accuracy and transparency of the quick count is a priority of Kompas research and development (R&D) to ensure credibility of the general election’s quick-count results. Ensuring the accuracy and transparency of the quick-count process is part of an effort to guard the quality of the election results.

As in a number of previous general and regional elections, Kompas R&D will again hold a quick count and exit poll on voting day, April 17. The quick count will cover the results of legislative and presidential elections.

Deputy Executive Director of Kompas Budiman Tanuredjo said transparency and accuracy, which have been the hallmarks of the Kompas R&D quick count, will be a top priority. "All processes will be carried out transparently. This is the contribution of Kompas to provide an assessment and measure the quality of the election itself," he said at the launch of the Kompas R&D quick count program on Tuesday (9/4/2019), in Jakarta.

Present at the event were Deputy Executive Directors of Kompas Liliek Oetama and Rikard Bagun, Chief Editor of Kompas Ninuk Mardiana Pambudy and General Manager of Kompas R&D Harianto Santoso.

The Kompas R&D quick count will use a sample of 2,000 polling stations (TPS), out of a total of 810,329 polling stations. As many as 2,000 volunteers will be deployed to the polling stations to collect voting data as raw data to be processed for the exit polls and quick count.

KOMPAS/HENDRA A SETYAWAN

A total of 400 freelance workers and 70 field coordinators of Kompas’ quick count who are assigned as interviewers at polling stations participate in a simulation for the quick count procedure at the Kompas daily office in Jakarta on Sunday (12/2/2017). The Kompas research and development division held a quick count for the Jakarta regional elections on Feb. 15, 2017.

Determination of the samples

Harianto Santoso explained that quick-count preparations had been carried out by Kompas R&D since February 2019. In determining the quick-count samples, Kompas R&D uses a final voter list (DPT)-based system issued by the General Elections Commission (KPU) and not based simply on TPS data alone. The selection of samples based on the DPT is more complex and detailed than if it was based on TPS data.

The use of polling station samples based on the DPT was first carried out in the quick count of the Central Java, West Java and East Java regional elections in June 2018.

Kompas R&D compresses the DPT data from the total 190,770,329 voters (according to the second change of the DPT) to 95,385 voters according to the total samples from 2,000 polling stations. The next random program is to select one voter data as a benchmark for determining the location of the first polling station sample. With the same pattern, the process will continue to select the next polling station samples up to a total of 2,000 samples.

Such a determination system is believed to be able to strengthen the accuracy of quick count results. "From the beginning we have emphasized that there must be guarantees and confidence in accuracy. This is because the results of the quick count will eventually be \'judged\' by the actual results within a short time span," said Harianto.

In this quick count, Kompas uses Open Data Kit support technology for its data input system. The volunteers will send a report from the polling stations in the form of a photo of C1 form via the program.

The incoming data will be processed by a team in the situation room, Jakarta. The quick count process in Jakarta will be carried out openly by inviting a number of observers and political consultants from various top universities in the world.

To ensure data accuracy, the team in Jakarta will confirm the truth of the volunteers\' reports by contacting the KPPS officers at a number of polling stations. After that, there is a team of validators who will reconfirm the data.

Kompas/Wawan H Prabowo

Quick count Kompas Members of the Kompas research and development team make a final check for the hardware to be used for the quick counting of results of the Jakarta regional election at the Kompas Gramedia Building, Palmerah, West Jakarta, on Tuesday (14/2/2017).

Track records

Kompas R&D has thus far conducted 14 quick counts. The first quick count was conducted in the 2007 regional election and the last one was in 2018 regional elections in West Java, Central Java and East Java.

So far, the results of the quick counts by the Kompas R&D have never been at odds with the results of the KPU recapitulation. In the 2014 election, the difference between the quick count results of the Kompas R&D and the results of the KPU recapitulation was 0.81 percent for the presidential election and 0.18 percent for the legislative elections.

The closest difference that has ever been achieved was counting the 2017 second round of the Jakarta gubernatorial election. At that time, the difference in results between the Kompas quick count and the final result of the KPU recapitulation was only 0.04 percent.

Ninuk Mardiana Pambudy said that the activities of the Kompas R&D have so far received recognition from various parties. This shows the ability to maintain the credibility and integrity of the Kompas R&D.

With regard to independence, Harianto guaranteed the Kompas R&D is an independent institution that is not related to any political element. "The cost of implementing this quick count is fully borne by Kompas," he said. (AGE/FRD)