ELECTIONS 2022: Marcos keeps lead in Pulse Asia survey with ‘high error margin’

 

ELECTIONS 2022:

Marcos keeps lead in Pulse Asia survey with ‘high error margin’

The voter preference for dictator’s son Ferdinand “Bongbong” Marcos Jr. remains the same at 53% in Pulse Asia’s April 2022 nationwide survey, making him the frontrunner so far with only days away from the elections.

However, some social media users are pointing out something odd about the survey: its representation of socio-economic classes.

In Pulse Asia’s most recent survey, only six people responded for Class AB combined, 180 people responded for Class C, 1,839 people for Class D, and 375 people for Class E – totaling 2,400 people.

The April survey’s error margins report also shows that AB has a 40% error margin, C has 7.3%, D has 2.3% and E has 5.1%. Class AB’s margin is anomalously high relative to surveys in the previous months.

According to Statistics How To, a 4% margin of error in a survey with a 95% confidence interval means that 95% of the time, your statistic may be four percentage points of the real population.

Applying this to the Pulse Asia survey, a 40% error margin for Classes AB means that the statistic for AB could be 40 points higher than it actually is.

But Pulse Asia only has ABC presented. Classes ABC has a voter preference of 29%. Using Pulse Asia’s margin of error for ABC (7.2%), this means that Marcos could actually be at least 49.8% preferred in the class, while the competing Leni Robredo could actually be at 36.2% at most.

This is just for ABC. Classes D and E have margins of error of 2.3 and 5.1, respectively, with values of 56% and 57%, respectively.

Because March had no AB and C representation at all, Pulse Asia did not include a margin of error for those classes at the time. The January and December surveys, meanwhile, had no comprehensive error margin reports available.

The last time that Pulse Asia included Classes AB before April was during the December 2021 survey.

Pulse Asia mentioned before that their group uses multi-stage random sampling, a method of selecting respondents that would be randomly chosen within a subgroup. For the survey group, they choose within Metro Manila, the rest of Luzon, Visayas, and Mindanao.