Behavioral data contradict Pulse Asia’s polling of PRRD’s 91% approval rating.
Behavioral data contradict Pulse Asia’s polling of PRRD’s 91% approval rating.
A year ago, when the Philippines economy was vibrant, before the double blasts of Ta ‘al and COVID, Pulse Asia surveyed and found President Duterte had a 78% approval rating. A year later, after the COVID pandemic triggered both the PhilHealth implosion and the economic seizure, the same firm found the President’s approval rating skyrocketed to 91%.
AutoPolitic disagree. Based on social data, the President’s social approval is around 49% — 55% at end of October.
The Pulse Asia poll was anomalous because it opposes the existing, polling trends and continuous behavior index, and without casual causal factors preceding their poll.
To determine the general movement of the public’s attitude, we use episodic polls by other firms and our continuous data. In 2019 Q4, Social Weather Station had PRRD at 72%. Publicus Asia had the president’s approval at 69% by April and 64% by August. The diagram shows all polling and our continuous calculation. PRRD’s approval has been declining during the COVID epidemic.
If a reasonable person believes PRRD’s popularity has suffered during COVID, then the Pulse Asia 91% approval polling is problematic. We diagram 30 days of social attitude to search for factors that account for Pulse Asia’s result.
The 9/7 anger spike is caused by President Duterte’s absolute pardoning of US Marine Joseph Scott Pemberton. And the smaller spike on 9/13 is when Pemberton is freed.
While the sole Love spike on 8/31 was due to President Duterte’s kissing of the Jolo, Sulu blast ground. ‘From a rational view, these were the factors on the Philippine people’s mind before the Pulse Asai survey, and none of the events or the magnitude calls for a 91% social approval.
We delayed this publication to check with the October behavioral data to decide if our analysis is warranted.
Anger Spikes:
9/22 PRR’s criticism of VP Robredo and his suggestion to spray the entire country with pesticide to kill the coronavirus
9/30 PRRD’s plan to infuse the PhilHeath with cash to prevent bankruptcy.
10/5 91% approval rating reported by major media
10/7 PRRD refusal to prosecute Secretary Duque & NPA burning
10/8 PRRD addresses the nation 10/15 PRRD has full trust and confidence in Mark Villar 10/19 PRRD SALN
Love Spikes
9/23 PRRD UN Speech
10/5 Pulse Asia 91% survey reporting
10/8 Pulse Asia 91% survey reporting reposted
10/28 PRRD visit parent’s grave
If we do live in a world where 91% of the Philippines people approve of PRRD, we’d see a much higher reaction to the positive news and much less negative reaction. Instead, the positivity is short-lived while the negativity is lasting.
While the net approval rating for October is 80%, the numbers show it is a strange month. Due to its low post about the PRRD, lowest sharing ratio, and low reaction ratio, we infer this either as a lack of genuine emotional news against the everyday COVID fatigue, or a general disinterest in news related to the president. We decompose the social attitude from 9/21 to 10/31 to understand any causal factors.
Based on the overall emotional context, the narrowing of the social approval/disapproval ratio, and the lack of corresponding responses to events, we believed the Philippine’s social approval of the president is around 49% — 55%.
AutoPolitic deploys a continuous behavior index, social intelligence to advise our government and political clients across Asia.
Disclaimer: This is an independent analysis provided by AutoPolitic and has not been paid for, requested or solicited by other parties or individuals.
Data Description: Data from 1/1/2020 to 10/31/2020. A total of 88,619,300 data points from publicly available Facebook data were geolocated to the Republic of the Philippines. No demographic or personally identifiable data is acquired or used. The data is acquired through observation and not solicitation, thus have minimal framing bias. Because the data source is only from those Philippines with social media accounts, the data inherits the demographic bias exhibited by the major social media platforms. Other factors such as access to the internet due to location, cost, and media consumption habits contributes to the passive selection bias of the model.
Model Description: The AutoPolitic behavior index models public attitude as a net binary attitude, which removes ambiguous or undecided responses. Also, the model does not give weight to the gradation of support or opposition. Furthermore, the model does not determine if the user is a genuine individual or sock puppet but does remove data points that do not have corroborating social resonance to reduce the possibility of intentional biasing the sample selection.
Copyright Assertion. No part of this blog may be quoted or reprinted without attribution.
Comments