Editorial Policy

Source Hierarchy

Together | side by side uses a five-tier hierarchy to weigh evidence. The goal is not to exclude non-government perspectives, but to ensure that the strongest, most auditable sources anchor the most important claims.

Tier 1: Government (primary)

Government sources are the default foundation for factual claims. This includes official program documentation, legislation, regulations, official datasets, departmental reports, and published statistical tables. These sources are weighted highest because they are authoritative, traceable, and directly tied to policy design and administration.

Tier 2: Parliament (primary or secondary)

Parliamentary sources include House of Commons and Senate publications, committee reports, evidence transcripts, written responses, and parliamentary debates. These sources can be primary (for what was said, proposed, or recorded) and secondary (for interpretation). They are weighted highly because they capture the legislative intent, oversight, and public accountability mechanisms around policy.

Tier 3: Academic (secondary)

Academic sources include peer-reviewed research, credible working papers, and well-established research institutes that provide method transparency. Academic evidence is weighted strongly for causal claims, economic modeling, program evaluation, and international comparisons. When academic findings diverge, we summarize the range and label confidence accordingly.

Tier 4: NGO (secondary or supplementary)

NGO sources include policy briefs, monitoring reports, and analyses from reputable organizations. These sources can add context, field observations, and synthesis, but they often reflect specific mandates or advocacy positions. We treat them as secondary or supplementary, and we avoid basing core claims solely on NGO reporting when primary sources are available.

Tier 5: Media (supplementary)

Media includes reputable journalism, investigative reporting, and explainers. Media is useful for timelines, public statements, and identifying issues that merit verification. However, media is generally treated as supplementary because it may summarize data without full methodology and can be sensitive to framing. We use it to guide questions, not to substitute for primary evidence.

Confidence Labels

Together | side by side assigns a confidence level to major claims to signal how strong the supporting evidence is. Confidence is about evidence quality and stability, not about how persuasive a claim feels.

• Established: Multiple primary sources, government data, and no credible dispute. Use:

  Established

• Probable: Strong secondary evidence that is consistent across sources, with minor gaps or limited primary confirmation. Use:

  Probable

• Preliminary: Limited data, emerging evidence, or claims that rely on a single credible source. Use:

  Preliminary

• Under Review: Contradictory evidence, active policy changes, or data that is outdated, ambiguous, or being updated. Use:

  Under Review

When we label a claim "under review," we treat it as a prompt for more verification, not as a conclusion. If a claim is updated later, the change is logged with dates and a brief explanation.

Corrections and Retractions

Together | side by side takes corrections seriously. We distinguish between corrections (a claim remains broadly true but needs adjustment) and retractions (a claim is factually incorrect and must be withdrawn).

Corrections process

1. An error is discovered internally or reported by a reader.
2. We verify the issue against the highest available sources.
3. A correction note is added with the correction date and a short description of what changed.
4. The original claim is updated to reflect accurate information.
5. A changelog entry is created, including the affected page and the nature of the correction.

Corrections are not hidden. The goal is to improve accuracy while keeping the revision trail visible.

Retractions process

If a claim is determined to be factually incorrect, we retract it. Retractions include:

• A clear statement that the claim was incorrect
• The reason it was incorrect (for example, misread dataset, outdated rule, incorrect inference)
• The date of retraction
• The corrected framing or replacement claim, if available
• A changelog entry that links to the affected page revision

Changelog

Every content change on Together | side by side is logged with:

• Date of change
• What changed (short description)
• Which pages were affected
• Why the change was made (new data, correction, clarified wording, policy update)

Revision history is accessible from each page, so readers can see what changed over time. This supports auditability and reduces confusion when policies or datasets evolve.

Responsibilities

Together | side by side keeps role ownership simple and explicit:

• Policy analysis: Payman Khortalab
• Data verification: Payman Khortalab
• French translation: Verified by native speakers

When external verification is used (for example, translation checks), we record the verification status in the changelog or revision notes.

Update Cadence

Core content is reviewed quarterly to ensure it stays aligned with current rules, data releases, and public records. Data is refreshed whenever government sources publish new datasets or revisions that materially affect a claim.

For breaking policy changes, Together | side by side aims to reflect updates within 48 hours. If the situation is evolving or contradictory, we mark relevant claims as:

Under Review

and document what is changing and what is still uncertain.