Designing and Evaluating a Digital Archive of 21st-Century English Poetry

Thesis Abstract

The rapid expansion of digitally accessible poetry in the early 21st century has outpaced conventional archiving practices, creating gaps in discoverability, authenticity, and scholarly reuse of contemporary English verse. This study addresses the problem of how to design, implement, and evaluate a scalable digital archive that captures the diversity, metadata richness, and intertextual relationships of 21st-century English poetry while ensuring long-term preservation and user-centered access. The aim is to develop a functional digital archive prototype and rigorously assess its usability, interoperability, and scholarly utility. Specific objectives are to (1) catalog a representative corpus of 1,200 poems composed between 2000 and 2020 from multiple genres and regions within the English-speaking world; (2) implement a modular archive architecture employing TEI-XML encoding, IIIF-compliant media presentation, and linked data principles to enhance cross-referencing with authorial, thematic, and publication data; (3) design metadata schemas and taxonomies for poets, themes, forms, and contextual links that support advanced search and semantic querying; (4) evaluate the archive using a mixed-methods approach that integrates usability testing, expert appraisal, and a comparative performance analysis against existing poetry repositories; and (5) provide an evaluative framework and set of guidelines for sustainable maintenance, including preservation strategies and roadmap for broader adoption. A multi-method research design combines design science and qualitative-quantitative evaluation to produce both a usable artifact and evidence of its value. The population includes poets active during 2000–2020, poetry editors, literary scholars, and graduate students as representative end-users. A purposive sample of 60 participants is drawn for the usability study, complemented by 12 literary scholars for expert appraisal. Data collection instruments include task-based usability tests (think-aloud protocol and System Usability Scale), a structured expert review rubric, semi-structured interviews, and a digital repository analytics log capturing search success rates, time on task, and feature usage. The archive prototype comprises a curated corpus of 1,200 poems, digitized and encoded in TEI-XML with linked metadata, IIIF-enabled image and manuscript views, and a RDF-based linked data layer using SKOS for subject terms and FOAF/vocabularies for authorial data. Validity and reliability are addressed through triangulation of qualitative interviews, expert reviews, and quantitative task performance metrics, as well as inter-rater reliability checks on metadata tagging and encoding consistency. Data analysis uses a mixed-methods sequence. Descriptive statistics summarize participant demographics, task performance, and feature utilization. Inferential analyses examine differences in usability outcomes across experience levels via ANOVA and regression analyses identifying predictors of task success and perceived usefulness. Thematic analysis of interview transcripts and think-aloud data uncovers user needs, friction points, and perceived scholarly value. A comparative benchmark against two established poetry repositories assesses interoperability, search precision, and retrieval effectiveness, employing precision, recall, and F-measure metrics. The conceptual framework integrates the Digital Humanities Archive Model and Kuhlthau Information Search Process, extended with semantic web principles to account for 21st-century poetics. A conceptual model maps relationships among poets, works, themes, and publication contexts, guiding evaluation criteria for depth of metadata, linkage richness, and reuse potential. Anticipated findings indicate that a TEI-based, IIIF-enabled, RDF-enhanced archive significantly improves discoverability and scholarly reach for 21st-century poetry, enabling nuanced searches by form, theme, and intertextual reference, while maintaining robust preservation metadata. Usability outcomes are expected to show high satisfaction (mean SUS score ? 78) among graduate users and positive expert appraisals regarding metadata completeness, interoperability, and archival fidelity. The study will reveal key design tensions between granular metadata capture and system performance, offering evidence-based recommendations for scalable encoding workflows and governance strategies. The contribution to knowledge lies in (i) a replicable design pattern for modern literary archives that balances scholarly utility with preservation imperatives, (ii) an empirically grounded evaluation framework for digital poetry repositories, and (iii) practical guidelines for metadata schemas, controlled vocabularies, and linked data practices tailored to 21st-century verse. In conclusion, the project demonstrates that a thoughtfully designed digital archive can transform access to contemporary English poetry, supporting rigorous scholarship, pedagogy, and public engagement. Recommendations include adopting modular architectures to accommodate evolving poetic forms, investing in metadata standardization and collaborative curation, expanding the sample to include multilingual traces of poetry, and developing community governance to sustain long-term preservation and ongoing enhancement.