Step 01
Discover
Find candidate cases through existing archives, scholarship, journalism, community contributions, targeted searches, playthrough evidence, and official material. Discovery identifies a lead; it does not confirm it.
Research Methodology
Press Q combines queer game studies, structured archive design, and constrained AI querying through Quiu to make representation patterns legible without flattening them.
Method
Structured data
Press Q keeps characters, game systems, and queer readings in separate datasets so interpretation is not mistaken for identity confirmation.
Method
Queer game studies
The project treats representation as cultural context, not just category counting, so the archive can support interpretation.
Method
AI-assisted querying
Quiu's responses are constrained to information registered in the Press Q dataset to reduce unsupported claims and keep answers traceable.
Method
Visual analytics
The interface surfaces patterns across identities and systems while giving contested or creator-refuted readings their own clearly qualified view.
Data boundaries
The archive is built for cautious interpretation.
Press Q uses three distinct research units rather than guessing missing identity information. Queer readings preserve criticism, audience interpretation, and creator responses, but never enter confirmed character-identity percentages. When details are absent, Quiu should surface that absence instead of filling the gap with an unsupported inference.
Research workflow
From research lead to revisable archive record
Step 02
Choose the unit
Decide whether the evidence concerns an identifiable character, a game-level affordance, or a documented queer reading. When more than one unit applies, create linked but separate records.
Step 03
Gather evidence
Record the source, evidence type, relevant quotation or scene context, language, platform or version, image provenance when applicable, and known counterevidence.
Step 04
Code conservatively
Use specific supported terms, preserve player-dependent outcomes, avoid inferring protected or personal identities, and leave a field unknown when the evidence cannot support a classification.
Step 05
Qualify the claim
Assign research status and evidence confidence, document limitations, and distinguish confirmation from implication, interpretation, rumor, or creator refutation.
Step 06
Review and revise
Check required fields and internal consistency before publication. Reviewed records remain versioned and correctable when new sources, releases, translations, or community knowledge emerge.
Coding protocol
Rules that prevent a label from outrunning its evidence
No default identities
Blank or unknown gender and sexuality fields remain visible as documentation gaps. They are never recoded as cisgender or heterosexual defaults.
Specific terms are not opposites
Trans man remains a man and trans woman remains a woman. Specific trans labels are stored for research visibility, not to place trans people outside their gender.
Identity is not behavior
A relationship, animation, costume, pronoun, or mannerism can be relevant evidence but does not automatically establish gender identity or sexual orientation.
Player choice stays conditional
Mutually exclusive routes are documented as player-defined outcomes. They are not split into several simultaneous canonical identities.
Systems do not assign NPC identities
Gender-independent romance documents a mechanic. It does not by itself make every compatible NPC canonically bisexual or pansexual.
Interpretation remains interpretation
Queer readings preserve reception history, including contested and creator-refuted cases, but never enter confirmed character-identity totals.
Intersectionality requires evidence
Race, ethnicity, religion, disability, class, nationality, and migration context are coded only when supported. Appearance and names are not sufficient.
The researched version matters
Platform, patch, expansion, localization, and release context can change content. Records should identify the version actually supported by the evidence.
Evidence confidence
Strength of support, not value
High
The specific claim is clearly supported by strong in-game, official, creator, or well-corroborated evidence.
Caution: Still revisable; it is not a score for representation quality.
Medium
The claim is supported but has a meaningful gap, indirect source, version restriction, translation issue, or interpretive qualification.
Caution: Not a numerical probability and not ‘half true.’
Low
The case is research-relevant but relies on limited, indirect, or substantially contested evidence.
Caution: Must be presented as uncertain and used as a lead for further research.
Analytics protocol
Every percentage needs a denominator and a boundary
Character pages use all documented character records as their denominator; system pages use queer-system records; and queer-reading pages use queer-reading records. Research coverage pages combine the three units only to examine workflow metadata.
Multi-value fields can place one record in more than one bar, so category assignments and percentages may add to more than the record total or 100%. Game-scale charts are weighted by character records rather than unique games.
“Unknown,” “Not recorded,” and “None documented” are different limits. They are never evidence of a presumed cisgender, heterosexual, white, able-bodied, or otherwise default identity.
Review, correction, and provenance
“Reviewed” describes a stage, not permanent truth.
Discovery source, evidence type, evidence source, platform/version, last-reviewed date, and confidence make the conditions of each record inspectable.
Conflicting evidence should be preserved in notes or counterevidence rather than silently removed. Creator-refuted and contested readings remain visible with their qualification.
Represent Me and the LGBTQ Video Game Archive are discovery sources with documented contributor overlap. Matching entries are deduplicated by subject, game, and version, retain both provenances, and are not counted as independent corroboration without separate underlying evidence.
When sources disagree, Press Q preserves each claim and its date, flags the record for human review, and avoids silently choosing or averaging labels. Corrections can change labels, status, evidence, or inclusion while retaining a change history.
Sources and framework
Methodological and terminological foundations
These sources inform the project’s definitions and safeguards. They do not make Press Q exhaustive, peer reviewed, or free from curatorial judgment.
Shaw, Adrienne, et al. “About (Please Read First!).” LGBTQ Video Game Archive.
Supports the distinction between explicitly coded content, creator statements, queer readings, and incomplete research coverage.
Queerly Represent Me Ltd. “About the Database” and “About the Site.” Represent Me.
Provides a near-comprehensive historical discovery source for queer content in games from 1974–2020. Because Represent Me identifies collaboration with the LGBTQ Video Game Archive, overlapping records are reconciled as potentially related evidence rather than assumed independent confirmation.
GLAAD. “Glossary of Terms.” Where We Are on TV 2023–2024.
Provides working definitions for terms including bisexual, nonbinary, queer, and transgender while emphasizing respectful, specific language.
D’Ignazio, Catherine, and Lauren F. Klein. Data Feminism. MIT Press, 2020.
Grounds the project’s attention to power, context, plural perspectives, and the limitations of binaries and apparently neutral classifications.
Crenshaw, Kimberlé. “Demarginalizing the Intersection of Race and Sex.” University of Chicago Legal Forum, 1989.
Provides the foundational account of why axes of power and identity cannot always be analyzed as isolated, mutually exclusive categories.
Wilkinson, Mark D., et al. “The FAIR Guiding Principles for Scientific Data Management and Stewardship.” Scientific Data 3, 2016.
Supports rich metadata, traceability, interoperability, and reuse while making the conditions and provenance of data visible.