MANILA, Philippines – University of the Philippines (UP) statistics professor Peter Cayton on Thursday, February 8, questioned the survey of the National Police Commission (Napolcom) which named Caloocan cops as the most trusted in Metro Manila.
"Medyo malabo 'yung survey at results, maraming questions na kailangan pa sagutin to explain possible errors (The survey and its results are still unclear, there are a lot of questions that need to be answered to explain possible errors)," Cayton told Rappler in an interview.
The Napolcom survey was conducted from late October to early November in 2017, just months after the controversial teen killings which led to the sacking of the entire Caloocan City police force.
The deputizing commission polled 570 participants of different age groups and occupations from the National Capital Region (NCR).
Aside from learning that Caloocan City cops were the most trusted and respected police station in Metro Manila, they noted that Filipinos in the metro trust police less despite feeling safer.
"Errors," Cayton explained, meant a shortfall in gathering data, not a mistake, in statistical terms.
No margin of error?
He first pointed out the survey results did not give a margin of error, a staple in statistical studies, however big. The margin of error refers to the amount of variation of the results allowed to factor in for changing circumstances and possible miscalculations.
"It's vague why they didn't give a margin of error, especially for a survey of this form, because if you've seen surveys from Pulse Asia or SWS (Social Weather Stations), even surveys by the Philippine Statistics Authority, they would give you an idea of sampling error such as the margin of error," Cayton said.
"That measure should exist, at 'yun 'yung wala akong makita sa survey (and that's what I couldn't see in the survey)," he added. (READ: Understanding Duterte's decline in SWS, Pulse surveys)
The National Capital Region Police Office (NCRPO) requested for the survey. A copy of the results sent to them and obtained by Rappler does not show any figure representing a margin of error in their findings that hailed Metro Manilans feeling 11% safer but lamented a 3% drop in police trust.
Cayton then questioned how the Napolcom thought of assigning sample sizes per city in the region, comprised of 17 independent local government units (LGUs)
"I would need justification for them... There's also cost considerations and I respect that, but for me to just see 50, 50, 50 without any form of statistical accounting or justification it is quite lacking," Cayton said in a mix of English and Filipino.
Based on the sample sizes used disclosed by the Napolcom, they surveyed 50 people from Manila, Quezon City, and Caloocan, while they polled 30 people each from the 14 other LGUs.
With this sampling, Quezon City – which has around 3 million residents – has the same statistical weight in the results as Manila, which only has around 1.8 million people, and Caloocan, which has around 1.6 million locals.
Meanwhile, smaller LGUs, like Pateros with around 60,000 residents, have the same weight as much larger cities like Pasig and Taguig, which both have around 800,000 citizens each.
With these questions, is the survey credible? Cayton could not say just yet.
According to him, the Napolcom might have its own working methods, which they have been using for years, to gather data for cops. (READ: Should we trust surveys and opinion polls?)
"We're questioning the survey instruments, we are questioning them because we don't know whether they would be good enough, and the way of building up the results, you don't have to give us the data, but these are results. Each should be accounted, the margin of error and whatnot," Cayton said.
The Napolcom-NCR office, which officiated the survey has been reached for comment but has yet to respond. – Rappler.com