By Ma. Lourdes Tiquia | The Manila Times
IF you want to revolutionize campaigning in the country, the Commission on Elections (Comelec) should make public the data set from the 2010 to 2019 elections by clusters so that people in the business of studying demographics and psychographics in relation to Filipino voters can make their correlation studies, establish models, contribute to the literature and bring data science to political campaigns. Since 2003, the Comelec has sold aggregates of voter data, not in soft copy form. One pays P1 per page and one cannot convert the data because it is in Word format. In 2010, another attempt was made to buy aggregates and still what the Comelec had was Total Registered Voters (TRV), Turn Outs (TO) and Votes Cast, until the leak that got everyone lining up to capture it because it was electronic. People tried to make sense of the data, but without time series, one cannot do trend analysis.
Actionable information is critical in political campaigns. Data has to be timely and accurate. We have been so used to reading data that is dated, alarming everyone because of a dip that was measured a month and a half ago. Mainstream media aggravates the situation because they just report the horse races without context and knowledge of sampling and margins of error. Electoral history analysis cannot be made because there is no effort on the part of some people at the Comelec to actually close the record every elections. Data may not be complete in the central office and one needs to go to the specific province to get the final breakdown. One shudders at the thought of how the Comelec handles and manages voter data.
Many kibitzers think that campaign management is simply getting a candidate to win. “Simply” is an understatement. That is the glitzy part but somewhere behind the engine of a campaign are data scientists, number crunchers, people studying scenarios and planning horizons. How can a candidate and his/her team perform well if the standard data of Comelec cannot be parsed by region, province, city, municipality across election cycles on voting trends? Imagine if the data set for voter precincts are made accessible across election cycles, and the sampling designs for polls are based on voter precincts, strategists can really put together models per candidate. The badly needed precision could be attained because the sample set is real voters, those who actually voted.
Take the case of three foreign campaigns which used data with razor-like precision: Clinton, Obama and Trump. All invested in data. All interplayed data. One was a comeback kid, drilling the fringes; another used social media in a data cave operation where analytics was central to every strategy adopted, while Trump actually built models, used historical data and layered it with actionable information that embodied the outliers in the reds and blue states. In our jurisdiction, very few actually build data year after year. Political candidates stop embracing data science after elections. For how can a senator win and grow his base after six years (single term) at a measly 3 percent?
Our country has been having automated elections since 2010 and the only thing that Comelec releases are the following data set: gender and age. No cross-correlation with results. And I am not asking for how Juan de la Cruz personally voted because that is crossing the bounds of privacy.
Political commentators and analysts insisting there is a Catholic, women’s, youth and labor vote do not know what they are talking about. We have none of those since the Comelec does not want to crunch data and do analytics. What is true is that we have a huge youth vote every election cycle, new blood that potentially can alter how things are. Get the Philippine Statistics Association data on birth rate, first-time voters, etc., and bump these with voter aggregates, and we get actionable information. The question there is, do they register? If they do, do they vote? How about senior citizens? And yet the data is there, somewhere in the labyrinth that is the Comelec. As someone suggested during the recent Joint Congressional Oversight Committee meeting, Comeelc would have to get out of the rut. Modernize. Embrace data science and technology as an organization. Help the public, candidates and campaigners understand the huge data set they have. Learn from each other by drilling the data. Future campaigns can be all about strategies formed, adopted and implemented in the campaign period because of precision in interpreting data sets thereby resulting in actionable information.
Then candidates do not need to overload the system pre-campaign by spending half a billion to reach universal awareness and adopting a machine-gun approach in TV ads where candidates compete only on prime-time slots, with high viewership as the only determining factor. When one station lords it over all others, queueing removes strategic advantages of candidates because cash is lord.
Just imagine if Comelec releases that data from 2010 to the Present? They have the aggregates and Comelec can just give everyone a guide, using their basic gender and age data by region, city, municipality. Try studying urban and rural Philippines with those data and we can come up with so many eureka moments. Do you know that there are 35 marginal provinces (M325) that have never been included in sorties? A candidate once had to be dragged to visit these island provinces. He was surprised that conversion was instantaneous because he was the first national candidate to have done so. M35 is 8 million votes. In a close election, it could be the kicker.
Again, data from Comelec would show that the old bellwethers of national candidates known as the Lingayen-Lucena corridor is no longer there because of the growth areas that have overtaken the same. In the 2019 elections, the North and South corridors are again alive. There is NCR34 when it comes to the vote-rich regions and there is the Top 20 vote-rich provinces or the highly urbanized city approach. Reality on the ground showed that in 2019, an incumbent president with a high 81 percent approval rating delivered the 23 percent percent of Mindanao. Duterte’s personal candidates — Go, dela Rosa and Tolentino — got in because of his endorsement. But dela Rosa stood out on his own during the CNNPh debate and from there he was his own man. Go stuck strictly to the game plan of less talk, no errors. Tolentino was the hardest considering he ran in 2016 and yet his conversion was slow burn. PRRD had to plead in his last TVC for Tolentino. Reelectionists “JV” Ejercito and “Bam” Aquino understand data and analytics. But Ejercito needs to settle the family issue first while Aquino will have to build his own bailiwick.
How about it, Comelec? Actionable information is the way to go.