POFMA’ed v2020.05.08

Note: a more recent version of the dataset and its analysis is available (please check the Updates section of the website).


  • Inclusion of new variable (communication description).
    • The Codebook and site’s Methodology section have been updated to reflect this.
    • This new variable records a unique description for each distinct electronic communication using the following scheme: ACTOR MEDIUM CONTENT ID. Example: Actor A Facebook Post 2.
  • 2 prior cases have been amended, which results in the inclusion of 2 additional communications to the dataset:
    • As a result of the new coding procedure: Lawyers for Liberty Website (which was previously treated as the same communication as Lawyers for Liberty Website Article 1)
    • As a result of an error in coding: Mr Alex Tan Facebook Post 1 (which was previously treated as the same communication as States Times Review Facebook Post 3)
  • Inclusion of 1 new report, 1 new POFMA use and 2 new communications.
This dataset’s methodology and sources

A full description of the dataset’s methodology can be found here.

The full dataset can be downloaded here.

The sources used are government reports (POFMA Office press releases and Gov.sg Factually articles). Media reports and direct sources (the communications themselves) are also used to verify the accessibility of some communications.

How does the dataset define reports, uses and communications?

A report is a government press release issued to announce a use of the legislation.

Within a single report, one can find multiple distinct directions or orders under POFMA. Each of these are classified as a distinct use of POFMA.

Each direction or order can affect multiple distinct electronic communications, such as a Facebook post or website article.

For example:

  • According to Report 1, a Facebook post is the subject of a correction direction, issued to Individual A.
  • According to Report 2, this same Facebook post is also the subject of a targeted correction direction, issued to Facebook.

The POFMA’ed dataset would thus track that Facebook post as one communication, subject to two distinct POFMA uses, according to two distinct reports.

According to the Dataset:

  • One can find 45 distinct uses of POFMA affecting 41 electronic communications, recorded in 24 government reports.
  • You can find more details on the summary sheet of the Dataset’s excel file (which features some charts to visualise said data). Download the Dataset here.

Improving the Dataset

Are you thinking of using this dataset and do you find the current distinction between reports, uses and communications intuitive for your purposes? As a student, I’m ever eager to improve this dataset’s accuracy and useability to others, and am wondering if a different way of organising these cases might be better suited to what I’m measuring.

If you have thoughts, comments or suggestions, please do direct them towards POFMAed@gmail.com.