Philstar.com's investigations are based on a set of standards that prioritize robustness of proof
This page sets out the standards Philstar.com applies when investigating influence operations and foreign information manipulation. However, our reports specify which is potentially an influence operation, what evidence we require, how we collect and count it, what we are prepared to conclude from it, and what we are not.
In most cases, each of our investigations is published with the dataset it is based on. This page explains how those datasets are assembled and what they can and cannot support.
Most of our investigations originate from our own monitoring. Philstar.com's reporters, editors and social media producers, most of whom have received training in identifying coordinated behaviour, monitor platforms, keyword sets, pages, groups and public discussions as part of routine work.
Credible tips from readers and concerned individuals sent to our reporters' and editors' emails or through our social media accounts.
Published findings by researchers and other institutions, where the finding is newsworthy and where we are able to examine the underlying activity ourselves.
Detecting suspicious activity does not by itself produce a story. Before committing resources, and again before publication, we assess:
Is the activity widespread enough to warrant reporting, or does it show signs of growing?
Is it directed at Filipino audiences or at audiences abroad? Is there a potential for cross-seeding of narratives?
Does it have anything to do with elections, public safety, health, foreign policy, or the safety of identifiable people? Does it involve corruption of moral values and shared understanding of democracy and due process?
Is anything outright false? If the content is factually accurate but arranged to mislead, what truth is being obscured, and who benefits?
Where the main audience is abroad, we assess whether there is an attempt to muddle global understanding of Philippine news and current affairs, shape opinion of a prominent Filipino individual or group, and / or whether there is value in recording that a narrative was made prominent overseas first.
We do not report on campaigns that are not spreading where doing so would expose identifiable people to harm.
Philstar.com drops potential stories and investigations at this stage regularly.
We do not treat content alone as evidence of foreign or partisan involvement. A narrative may match a foreign government's or party's messaging, serve its interests, or repeat its talking points. However, we recognize that an individual or group can hold a range of sincere views on the issues of the day. Some common topics include the West Philippine Sea, foreign policy and domestic politics. Philstar.com does not believe that holding a view similar to, or identical to, some foreign government or party, is by itself evidence of coordination or sponsorship.
We do not designate any country as a default source of manipulation. Where evidence points to alignment with a particular state's messaging, we communicate that in our reports, and we state what the evidence establishes and what it does not.
Readers will see that many narratives Philstar.com has reported on, align with Chinese state messaging. That reflects the issues currently dominating Philippine public debate, particularly the West Philippine Sea. This, however, is not a starting assumption.
We use two public frameworks alongside our own editorial judgement.
The ABCDE Framework separates a potential IO into the following: Actor, Behaviour, Content, Degree (of spread), and Effect. In our reports, we identify each of these elements, except the Effect, which refers to evaluating a campaign's overall impact and who it affects—a separate action from the act of journalistic reporting.
DISARM is an open framework cataloguing techniques used in influence operations. Several OSINT researchers use it as a periodic table of bad behavior. When needed, we use it to identify and name the malign tactic we encounter.
is the activity widespread enough to warrant reporting, or does it show signs of growing?
is it directed at Filipino audiences or at audiences abroad? Is there a potential for cross-seeding of narratives?
We do not publish an investigation without a quantitative basis. Reports state how many posts, accounts, pages or videos were examined, over what period, on which platforms, and what characteristics they shared.
We rely on samples. Collecting every post on a topic is not possible and we do not claim to have done so.
We treat around 30 content pieces as a working threshold for a body of evidence. This is a rule of thumb as in many cases, the number may be less than 30 but bear other kinds of hallmarks of IOs. We communicate all of this in our reports.
Data collection at Philstar.com is primarily manual wherein reporters collect by hand, assisted where useful by our newsroom's in-house scraper tool or by extracting elements from public web pages, and examine account characteristics individually using open-source investigation methods.
Material is preserved using SingleFile, screenshots and the Internet Archive's Wayback Machine. We archive before approaching any subject for comment.
Our influence operation reports carry no rating label, unlike our fact checks.
High up in each report, we state what was observed, its scale and where it spread. We use qualified language, including "suspected," "bears the hallmarks of" and "bears the markings of," where that is what the evidence supports. We do not use terms such as "foreign-linked" or "state-backed" unless we have evidence relating to the actor rather than the content.
We do not publish confidence ratings. In their place we publish the number of evidence collected, how much, over what period, and what it had in common. The linked dataset allows readers to check this.
Where a question cannot be answered with the evidence available, our reports say so.
Material posted publicly by accounts that chose to publish publicly is generally reported without redaction. This is assessed case by case. We consider the person's prominence, whether they hold or seek public office, whether they appear to be operating professionally, the potential harm of publication, and whether identification is necessary to the finding. Ordinary users are treated differently from those who appear to be operating the network.
We do not report that any individual is a foreign agent or paid operator on the basis of narrative alignment. This applies to public figures as well.
Where we document a narrative, we do not assert that everyone repeating it is part of the campaign. Our investigations rest on behavior. Where we identify accounts that amplified a narrative, including cases where timing is significant, we describe what they did and when, without characterising motive beyond what the evidence shows.
As much as possible, we ensure our investigations are published with their underlying dataset, hosted as a Google Sheet on Philstar.com's Google Drive and linked from the report. Datasets typically include the archived posts or accounts examined, collection dates, and the attributes recorded.
Where publishing a field would cause disproportionate harm to a person who is not a subject of the finding, that field is redacted and the redaction noted in the sheet.
Readers may raise corrections or disputes through email or our social media accounts. Editors will review, and where a correction is warranted the report is corrected and an Editor's Note appended at the end.
Philstar.com's reporting on influence operations and foreign interference is not funded by any external party. It is produced by our own staff on our own initiative.
Members of the team have previously received training and operational support from outside organizations. Where a specific report was produced with such support, this is disclosed on that report.
No funder or partner has been allowed to select a subject, shape an angle, review evidence or see a story before publication. All angles, collection decisions and conclusions are those of Philstar.com reporters and editors.
We treat allegations that our work is paid for or directed by others seriously and will respond to them specifically.