Source Citation & Verification Standards

Version 1.0 · Effective 7/13/26

Contents

Introduction

Citing sources is not a formality. It is the mechanism by which DDC's readers — and DDC itself — can verify that what we publish is true. It is how we give credit to the researchers, institutions, and community members whose work we build on, and how we distinguish what the evidence says from what we think it means.

The Source Citation & Verification Standards, or SCVS, is a companion to DDC's Research Ethics Policy (REP). Where the REP establishes the ethical principles that govern DDC's research program, the SCVS establishes the specific standards for how sources are found, evaluated, cited, and maintained. Together they ensure that every factual claim DDC publishes can be traced back to its origin — and that that origin is worth tracing.

This policy is written for DDC's research team and for the public. Anyone who reads our work should be able to find our sources. Anyone on our team should know exactly what "properly cited" means before they submit anything for publication.

1. Purpose and Scope

1.1 Why This Policy Exists

DDC's research covers societal topics — the treatment of marginalized communities, the ethics of social institutions, the psychology of collective behavior, and the relationship between people and art and design. These topics carry real stakes. A badly sourced claim about a community can cause harm. A statistic pulled without verification can mislead readers and damage public discourse. A finding attributed to the wrong source misrepresents the record and undermines trust.

This policy exists so that every factual claim DDC publishes meets a consistent, documented standard — one the research team applies before publication and that readers can assess afterward.

1.2 Who It Applies To

The SCVS applies to everyone who produces, reviews, or approves research published under the Doodle Design Co. name — employees and contractors alike. It applies to all published formats: long-form pieces, short posts, internal reports that reference external sources, and any other format DDC research takes.

1.3 How This Policy Fits With the REP and Other DDC Policies

The SCVS gives procedural detail to obligations the REP establishes in principle:

  1. REP Section 4.1 (Attribution and Sourcing) states that every external source must be credited clearly enough that a reader could find it. The SCVS specifies exactly how.
  2. REP Section 4.5 (Review Before Anything Goes Public) includes verifying attribution as a pre-publication check. The SCVS defines what "verified" means.
  3. REP Section 5.3 (Verification Before Publication) requires that AI-sourced claims be checked against primary sources. The SCVS sets the standard for what that check looks like.

Where the REP and the SCVS appear to conflict in a research context, the more protective standard applies.

2. The DDC Citation Standard

DDC publishes societal research on its own website for a general public audience. The citation format used in published work must be clear enough for a general reader to understand and specific enough for a researcher to follow up.

DDC uses a hybrid citation standard: APA 7th Edition1 for academic and peer-reviewed sources and government data; inline hyperlinks for web-based sources; and a DDC Reference Format for its own published research and datasets. This approach mirrors the standard used by leading public-interest research organizations, which bridge academic rigor and accessible public publishing.

2.1 Format at a Glance

Source Type Format
Academic / peer-reviewed journals APA 7th Edition
Academic books APA 7th Edition
Government data and official reports APA 7th Edition with agency authorship
Web-based news, reports, and articles Inline hyperlink with source name in text
DDC's own published research DDC Reference Format (see 2.5)
DDC datasets DDC Dataset Format (see 2.5)
Interviews and direct sources Modified APA personal communication (see 2.6)

2.2 Academic and Peer-Reviewed Sources

Academic and peer-reviewed sources are cited using APA 7th Edition. This is the standard in psychology, sociology, and public policy — the social science fields that most closely align with DDC's research areas.

Journal article format:

Last, F. M., & Last, F. M. (Year). Title of article. Journal Name, Volume(Issue), page range. https://doi.org/xxxxx

Example:

Wallerstein, N., & Duran, B. (2010). Community-based participatory research contributions to intervention research. American Journal of Public Health, 100(S1), S40–S46. https://doi.org/10.2105/AJPH.2009.184036

Book format:

Last, F. M. (Year). Title of book: Subtitle if any. Publisher.

Example:

Freire, P. (1970). Pedagogy of the oppressed. Herder and Herder.

In practice: When a DOI2 is available, use it in place of a URL — DOIs are permanent identifiers that remain stable if a journal moves to a new website. If no DOI exists, use the journal homepage URL. If the article is behind a paywall, cite it in full and note "(Paywall)" at the end so readers know they may need institutional access to retrieve it.

2.3 Web-Based Sources

For sources accessed primarily online — news articles, organizational reports, think tank publications, and similar material — DDC uses inline hyperlinks. The source name and year appear naturally in the sentence, and the link takes the reader directly to the original.

Format:

...according to a [Year] [publication type] from [Source Name]([URL])...

Example:

...according to a 2023 report from the Pew Research Center on racial wealth disparities...

In practice: The hyperlinked text must identify the source specifically — not generic phrases like "this report," "click here," or "a recent study." The source name and document type should be clear from the sentence itself, so that a reader who cannot click the link still knows what is being cited.

2.4 Government Data and Official Reports

Government data and official reports are cited in APA 7th Edition format, with the issuing agency listed as the author.

Format:

Agency Name. (Year). Title of report or dataset. Department or Ministry. https://...

Example:

U.S. Census Bureau. (2023). American community survey: 1-year estimates. U.S. Department of Commerce. https://census.gov/programs-surveys/acs/

In practice: When citing a specific figure from a large dataset — for example, a single statistic from a census table — note the specific table or figure number in addition to the main citation, so the exact data point can be located independently and not just the report as a whole.

2.5 DDC's Own Research and Datasets

When a DDC publication references an earlier piece of DDC research or a DDC dataset, it cites it using the DDC Reference Format. DDC's own work is cited with the same rigor as any external source.

Published research format:

Doodle Design Co. (Year). Title of research. https://sharpsticker.store/research/[slug]

Dataset format:

Doodle Design Co. (Year). Title of dataset [Data set]. Available at sharpsticker.store or on request at hi@sharpsticker.store.

In practice: DDC research that is not yet publicly available — for example, raw data gathered during an ongoing study — is noted as "(DDC unpublished data, [Year])" rather than cited as a published source. Unpublished data citations must have a corresponding record in the project's Notion file confirming the data exists and how it was collected.

2.6 Interviews and Direct Sources

When DDC cites a direct source gathered during research — an interview, a focus group contribution, or a firsthand account — the format depends on whether the participant consented to being identified by name.

Named source (identified participant):

[Full Name], [Title or Role if relevant], personal communication, [Month Year].

Anonymous source:

DDC Research Interview, [Month Year] (anonymized per participant consent).

In practice: Named citations require that the participant explicitly agreed to be identified by name in the published piece — this agreement must be documented in their consent record (per REP Section 3.2). If consent was given only for anonymous use, the anonymous format is used regardless of how much identifying information DDC holds internally. No participant is identified by name without documented consent to that specific use.

3. Source Tiers

Not all sources carry equal weight. DDC uses a four-tier system to assess the quality and credibility of any source before citing it in published work. A source's tier directly affects the verification requirements in Section 4.

3.1 Tier 1 — Primary and Authoritative Sources

Tier 1 sources are the strongest DDC cites. They are either original data, rigorously reviewed scholarship, or DDC's own original research.

Tier 1 includes:

  1. Peer-reviewed academic journals3 in relevant fields — psychology, sociology, public policy, design studies, cultural studies, and related disciplines.
  2. Official government statistics and databases — U.S. Census Bureau, Bureau of Labor Statistics, CDC, NIH, Department of Justice, and equivalent agencies in other countries.
  3. International body data and reports — United Nations, World Health Organization, OECD, World Bank, and similar authoritative international organizations.
  4. Court documents, legal records, and official government transcripts.
  5. DDC's own original research and datasets, collected and documented in accordance with this policy and the REP.

In practice: When a Tier 1 source is available for a claim, it is used in preference to any lower-tier source making the same claim. A news article that cites a government report is not a substitute for the government report itself.

3.2 Tier 2 — Established Research and Institutional Sources

Tier 2 sources are credible and widely used in research contexts, but do not carry the same level of independent scrutiny as peer-reviewed or official government sources.

Tier 2 includes:

  1. Reports from established, non-partisan research institutions — Pew Research Center, Brookings Institution, Urban Institute, RAND Corporation, and similar organizations with documented methodology.
  2. Reports from recognized advocacy organizations, when cited for data those organizations collected themselves — such as ACLU, NAACP Legal Defense Fund, Human Rights Watch, and similar bodies. These are Tier 2 with required disclosure (see below).
  3. Academic books published by recognized university or academic presses.
  4. Industry research from organizations that publish their methodology and funding sources.

In practice: When citing a Tier 2 source from an organization with an advocacy position, the published piece notes that position openly. For example: "According to the ACLU — which advocates for civil liberties — ..." This is not a disqualifier; it is honest context that lets readers assess the source themselves. Failing to disclose an advocacy source's position is a form of misrepresentation.

3.3 Tier 3 — Journalism and Industry Sources

Tier 3 sources are useful for context, recent developments, and direct quotations, but are not sufficient on their own to support factual claims about trends, patterns, or communities.

Tier 3 includes:

  1. Reporting from established national and international news outlets — The New York Times, The Guardian, Reuters, Associated Press, BBC, NPR, ProPublica, and equivalent outlets with editorial standards and correction policies.
  2. Industry trade publications with documented editorial standards.
  3. Expert commentary and opinion in recognized publications.

In practice: Tier 3 sources may be cited to provide context, to quote a named expert, or to reference a recent event. They are not sufficient alone to support a statistical claim or a claim about a community's experience. When a Tier 3 source references a study, a dataset, or a report, DDC follows the citation back to the original and cites the Tier 1 or 2 source directly. Citing the news article instead of the underlying study is not acceptable.

3.4 Tier 4 — Sources to Avoid or Use With Extreme Caution

Tier 4 sources are generally not acceptable as citation sources in DDC research. If a Tier 4 source is the only available source for a claim, that claim is removed, reframed as unverified, or held until a higher-tier source can be found.

Tier 4 includes:

  1. Wikipedia and other crowd-sourced references.4
  2. Social media posts — including official organizational or government accounts.
  3. Press releases and marketing or promotional materials.
  4. Personal blogs and opinion websites without editorial oversight or accountability.
  5. AI-generated content — see Section 5 for how AI output is handled.
  6. Sources with undisclosed funding or conflicts of interest.5
  7. Anonymous sources without a documented basis for the claim.
  8. Outlets with a documented history of publishing misinformation.

In practice: Wikipedia may be used as a starting point for research — particularly to locate primary sources through its references — but the primary source itself is what gets cited, not the Wikipedia article. Wikipedia is a research tool, not a research source.

4. Verification Requirements

4.1 Verification by Claim Type

The minimum number of independent sources required to support a claim before publication depends on the nature of that claim.

Claim Type Minimum Sources Minimum Tier
Basic uncontested fact 1 Tier 1 or 2
Statistical figure or data point 1 Tier 1 strongly preferred
Trend or pattern claim 2 independent Tier 1 or 2
Causal claim ("X causes Y") 2 independent Tier 1
Claim about a community's experience 2 independent Tier 1 or 2, plus community source where available
Contentious or counter-intuitive claim 3 or more independent Tier 1 or 2
Historical claim 1 Tier 1 or 2

In practice: "Independent" means the sources did not draw from each other for the specific claim being made. Two news articles that both cite the same underlying study are one study cited twice — not two independent sources. Following all citations back to their original source is required before counting sources toward a verification threshold.

4.2 Contentious and High-Stakes Claims

A claim is considered contentious or high-stakes if it meets any of the following criteria:

  1. It challenges a widely held belief or contradicts prevailing consensus.
  2. It makes a specific assertion about a named individual, organization, or community.
  3. It involves a politically sensitive topic where the claim could be misused or weaponized out of context.
  4. It contradicts existing peer-reviewed research without a strong methodological explanation for the discrepancy.
  5. It relates to a marginalized community's experiences in a way that could cause harm to that community if wrong or if framed incorrectly.

For any contentious or high-stakes claim, three or more independent sources at Tier 1 or 2 are required before publication. If that threshold cannot be met, the claim is either reframed as provisional — "Some evidence suggests..." or "Limited data indicates..." — or removed entirely.

In practice: The reviewer in REP Section 4.5 is specifically responsible for flagging contentious or high-stakes claims that have not met the three-source threshold. A piece does not go live until every such claim either meets the threshold or has been revised.

4.3 Conflicting Sources

When credible sources disagree — for example, when two peer-reviewed studies reach different conclusions about the same question — DDC does not resolve the conflict by selecting the source that supports the preferred conclusion.

What happens instead:

  1. The conflict is reported honestly: "Studies are divided on this point. [Source A] found X, while [Source B] found Y."
  2. Higher-tier sources carry more weight when assessing which finding is more methodologically credible — but the existence of a conflicting source is always disclosed.
  3. Where the conflict reflects genuine scientific uncertainty, it is presented as uncertainty, not resolved artificially.

In practice: The lead researcher flags any source conflict in the project's Notion file so the reviewer can confirm the published piece addresses it honestly rather than appearing to have settled a genuinely unsettled question.

4.4 When a Claim Cannot Be Verified

If a factual claim cannot be verified against the minimum required sources at the minimum required tier, one of three outcomes applies:

  1. Remove the claim. If the claim is not essential to the piece's argument, it is cut.
  2. Reframe the claim. If the claim matters but the evidence is limited, it is reframed to accurately reflect what is and is not known: "Some reporting suggests..." or "Based on limited available data..."
  3. Delay publication. If the claim is central and verification is realistically achievable with more time, publication waits.

The one outcome that never applies: publishing an unverified claim and hoping no one checks.

5. Citing AI-Assisted Research

AI tools may assist with research — searching for sources, summarizing findings, suggesting connections — but AI-generated content is not itself a citable source. The following rules apply, consistent with REP Section 5:

  1. AI output is a lead, not a source. If an AI tool produces a statistic, a study name, or a finding, that output must be traced back to the original primary source before it can be cited. The citation is the primary source — not the AI tool that pointed toward it.
  2. Unverified AI-generated claims are not published. If a claim produced by an AI tool cannot be traced to a verifiable primary source, it is removed.
  3. AI hallucinations6 are treated as unverified claims. If an AI tool generates a citation that does not exist, misattributes a finding, or produces a plausible-sounding statistic with no real source, that error is corrected or removed — it is never published.
  4. AI tools are not listed as sources in the bibliography or source list. The original primary source is what gets cited. If AI assistance was used to find it, that is disclosed in the AI disclosure note per REP Section 5.2, not in the citation itself.

In practice: Before submitting a draft for review, the lead researcher confirms in the project's Notion file that every claim originating from AI assistance has been traced back to and verified against its actual primary source.

6. Source Archiving and Link Maintenance

6.1 Archiving at Time of Publication

Online sources can disappear. URLs change, reports get deleted, websites shut down. To ensure DDC's published research remains verifiable long after publication, every online source cited in a published piece is archived at the time of publication.

What happens before a piece goes live:

  1. The lead researcher saves a copy or screenshot of every cited online source in the project's Google Drive folder alongside the project file.
  2. For Tier 1 and Tier 2 sources that anchor the piece's central claims, a Wayback Machine7 link (web.archive.org) is also saved in the project file. If the original link breaks after publication, this archive link is added to the live piece.

6.2 Broken Links After Publication

When a source link in a published piece stops working after publication, DDC does not simply delete the link or leave it pointing nowhere.

What happens:

  1. If an archived version exists via the Wayback Machine or the project folder, the broken link is updated to point to the archive, and a note is added: "(Original link no longer active. Archive accessed [date].)"
  2. If no archive is available and the source was Tier 1 or 2, DDC attempts to locate the source through the issuing organization directly before treating it as lost.
  3. If the source genuinely cannot be recovered and it was material to the piece's claims, the piece is reviewed under REP Section 4.6 (Correcting Mistakes Publicly) to determine whether a correction or explanatory note is warranted.

In practice: Broken links discovered after publication are logged in the project's Notion file with the date discovered and the resolution taken, so there is a complete record of every change made to published source citations.

7. Policy Administration

  1. Effective Date and Review — This policy takes effect on [EFFECTIVE DATE] and will be reviewed at least once a year alongside the REP, or sooner if DDC's source practices or the landscape of credible sources changes significantly.
  2. Amendments — Changes to the SCVS are decided by DDC leadership, recorded in the Changelog below, and communicated to the research team before they take effect.
  3. Questions — Questions about this policy can be directed to hi@sharpsticker.store.

Changelog

Version 1.0 — [EFFECTIVE DATE]

Initial release of the Source Citation & Verification Standards (SCVS). Establishes DDC's hybrid citation format (APA 7th Edition for academic and government sources, inline hyperlinks for web-based sources, DDC Reference Format for internal research and datasets), a four-tier source quality system, verification thresholds by claim type including a three-source minimum for contentious or high-stakes claims, rules for handling AI-sourced claims, and source archiving and broken link procedures. Companion policy to the Research Ethics Policy (REP).

Footnotes

  1. APA 7th Edition refers to the Publication Manual of the American Psychological Association, 7th edition (2020). APA format is the citation standard used across psychology, sociology, public health, education, and most other social and behavioral science disciplines — the fields that most closely align with DDC's research areas. It uses an author-date system for in-text citations (e.g., Smith, 2021) paired with a full reference list at the end of a document. It is widely recognized by both academic and general audiences and is well-adapted to citing online sources, including DOI-based journal articles. ↩
  2. A DOI (Digital Object Identifier) is a permanent, unique identifier assigned to a published academic article or dataset. Unlike a standard URL, a DOI does not change if the article is moved to a new web address or if the journal changes publishers. When a DOI is available, it is more reliable than a URL for citation purposes because it continues to resolve to the correct source regardless of where the article is hosted. DOIs in APA 7th Edition are formatted as full URLs: https://doi.org/[identifier]. ↩
  3. A peer-reviewed journal publishes research that has been independently evaluated by qualified experts in the relevant field before publication. This review process — commonly called double-blind peer review, meaning reviewers do not know who wrote the paper and the author does not know who reviewed it — is designed to catch methodological errors, unsupported conclusions, and logical gaps before findings reach the public. Peer review is not infallible, but it represents a substantially higher standard of scrutiny than most other publication formats and is the closest thing research has to a quality assurance system. ↩
  4. Wikipedia is a crowd-sourced encyclopedia that can be edited by anyone with a registered account. While it is often accurate and useful for general orientation on a topic, it is not a citable source in research because its contents can change at any time, contributors are largely anonymous, and it does not undergo editorial or peer review. It is, however, often a useful starting point for locating primary sources, which can then be cited directly. The standard practice in research is: use Wikipedia to find sources, not as a source. ↩
  5. A source with undisclosed funding or conflicts of interest is one whose financial backers, institutional affiliations, or personal interests in the topic are not disclosed to readers — making it impossible to assess whether those interests may have shaped the findings or conclusions. This does not mean that funded research is automatically unreliable. Undisclosed funding is, however, a significant credibility concern in research ethics because it prevents independent assessment of potential bias. DDC avoids citing sources that do not disclose their funding, methodology, or institutional affiliations. ↩
  6. AI hallucinations is the term used to describe instances in which an AI tool generates a confident, plausible-sounding response that is factually incorrect — for example, fabricating a study that does not exist, attributing a quote to someone who never said it, or producing a statistic with no basis in reality. AI hallucinations are a well-documented characteristic of large language models and are one of the central reasons AI output must be independently verified against primary sources before it is published. The fact that AI-generated content sounds authoritative is not evidence that it is accurate. ↩
  7. The Wayback Machine (web.archive.org) is a free service run by the Internet Archive, a non-profit organization, that automatically saves and stores archived copies of web pages over time. It allows anyone to access a past version of a webpage even if the original has been deleted, moved, or changed. DDC uses the Wayback Machine to create stable, permanent archive links for online sources at the time of publication, so that readers can still access those sources even if the original URL stops working in the future. ↩