Beyond Service Inventories: A Three-Dimensional Framework for Diagnosing Structural Barriers in Academic Library Research Dataset Management

Abstract

Academic libraries have assumed expansive research data management (RDM) responsibilities, yet persistent dataset underutilisation suggests systemic disconnects between services and researcher needs. This scoping review applied a three-dimensional diagnostic framework to examine why libraries struggle to advance beyond consultative roles despite sustained investment. Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) guidelines, this review analysed 34 empirical studies (2015–2025). Electronic databases, key journals, and grey literature sources were systematically reviewed, with 65% of studies originating from high-income (Global North) contexts. The analysis integrated the Institutional Readiness Index (IRI), Service Maturity Level (SML), and Information Flow Efficiency (IFE) to assess library engagement with research datasets. Three structural patterns constrain effectiveness. First, a capacity-complexity mismatch emerges as libraries manage increasingly diverse datasets without proportional infrastructure scaling, creating bottlenecks in discoverability, interoperability, and preservation. Second, structural progression barriers appear, where advancement requires simultaneous development across infrastructure, staffing, governance, and engagement rather than sequential improvement. Third, an implementation gap separates Findable, Accessible, Interoperable, Reusable (FAIR) policy awareness from operational capacity, as most institutions demonstrate standards knowledge without technical operationalisation ability. These patterns form interdependent constraints: infrastructure limitations correlate with restricted services, which are associated with persistent researcher skill gaps, reduced engagement, and constrained resource allocation, reinforcing the initial deficits. The review framework provides diagnostic specificity for identifying whether constraints stem from readiness, maturity, or implementation failures. This study advances RDM scholarship by explaining stagnation patterns rather than cataloguing services, offering an empirically grounded diagnostic tool. However, the findings reflect predominantly high-resource contexts and require validation across diverse institutional settings.

1. Introduction

Research datasets, entailing systematically collected and organised information, form the empirical basis for scholarly inquiry across disciplines [1]. They include numerical data, qualitative records, experimental results, and geospatial information, all essential for reproducibility, innovation, and evidence-based decision-making [2,3,4]. However, effective management and use of these datasets remain underexplored, despite their importance for advancing knowledge and addressing global challenges, including the United Nations’ Sustainable Development Goals (SDGs) [5,6,7].

Academic libraries have shifted from custodians of static collections to active participants within institutional research ecosystems, taking on responsibilities for access, preservation, and reuse of scholarly resources [8,9]. This expanded mandate requires examination of their readiness, particularly the adequacy of technical infrastructure, staff expertise, and strategic frameworks needed to support increasingly complex data services [10,11]. Libraries increasingly act as intermediaries between data producers and end-users, working to ensure research data are discoverable, accessible, and ethically reusable [12,13,14]. Their work is gradually aligning with international standards such as the Findable, Accessible, Interoperable, Reusable (FAIR) principles [15,16], with guidance from UNESCO and the Research Data Alliance (RDA) emphasising ethical sharing and sustainable governance [6,10,17].

Despite these developments, a substantial gap persists between the services offered and actual dataset utilisation. This disconnect reflects deeper systemic issues related to service design, technical capacity, and institutional priorities [14,18]. Barriers frequently stem from inconsistent cataloguing, insufficient metadata standards, limited repository awareness, and constrained data literacy skills [10,14,18]. Although many libraries invest heavily in data management services, they face challenges in evaluating usage patterns and tailoring services to the diverse research community needs [12,19,20]. This service-utilisation gap reduces the impact of infrastructure investments and constrains the transformative potential of data-driven scholarship [19,21].

The evidence base on research data services within academic libraries remains fragmented. Whilst a growing literature body addresses infrastructure, repository design, and curation practices, fewer studies examine how researchers interact with or derive value from datasets [22,23,24]. Much existing research focuses on isolated aspects, infrastructure development, policy formulation, or data-sharing behaviours, without offering comprehensive perspectives on usage patterns, disciplinary variations, or persistent barriers [23,24,25]. Furthermore, limited examination exists of how library practices align with international frameworks such as FAIR principles or contribute to global initiatives such as the SDGs [10,11,26].

This fragmentation highlights the necessity of a scoping review methodology, particularly suited to emerging and heterogeneous research domains [27,28,29,30]. Hence, this study employs a scoping review to map empirical research on academic libraries’ engagement with research data. To move beyond descriptive assessment, this review applies an analytical framework, operationalising three key dimensions:

-

Institutional Readiness Index (IRI) assesses foundational capacity: infrastructure, governance structures, and resource allocation required to support RDM services [8,9,10]. This evaluates what capacity exists to deliver services.

-

Service Maturity Level (SML) evaluates functional sophistication: progression from ad hoc awareness through structured development to transformational capabilities [12,14,18]. This assesses what services can be delivered at different maturity stages.

-

Information Flow Efficiency (IFE) measures operational effectiveness: how well institutional systems facilitate discoverability, accessibility, interoperability, and preservation of research datasets [15,16,17]. This assesses how effectively information flows in practice.

This framework enables systematic comparison across institutions and identification of critical factors influencing service effectiveness [31]. In this regard, the study objectives are:

• To examine the effectiveness of academic library services in supporting diverse datasets across disciplines, using institutional readiness indicators.
• To explore the challenges associated with managing, accessing, and using research data in academic libraries, mapped against service maturity levels.
• To establish patterns in how library services support data-driven research relative to international standards, assessed through information flow efficiency metrics.

The analysis identifies current practices, challenges, and priorities for future development, with particular attention to geographic and resource contexts shaping the evidence base.

2. Materials and Methods

To ensure integrity, methodological soundness, and accountability, this scoping review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) guidelines [32]. The guidelines facilitated a systematic and standardised approach to this review, encompassing all essential stages described in this section. To further enhance the robustness and credibility of the process, the study protocol was preregistered on the Open Science Framework (OSF).

2.1. Analytical Framework Development

To assess how academic libraries engage with research datasets, this review used a three-dimensional analytical framework. The framework comprises IRI, SML, and IFE, three interdependent dimensions where IRI assesses capacity (what infrastructure and resources exist), SML assesses capability (gaps in service delivery), and IFE assesses effectiveness (how well information flows in practice).

2.1.1. Institutional Readiness Index (IRI)

This dimension evaluated foundational capacity through three components:

-

Technical Infrastructure: repository systems, storage capacity, preservation tools, metadata systems, discovery tools.

-

Governance Structures: formalised policies, cross-unit coordination, decision-making authority, compliance frameworks, strategic planning integration.

-

Resource Allocation: dedicated staffing, operational budgets, training programmes, infrastructure investment.

Studies were assessed for the presence or absence of these indicators. Institutions were categorised as low readiness (minimal infrastructure, informal governance, limited staffing), moderate readiness (basic repository systems, developing policies, some staff with skill gaps), or high readiness (integrated infrastructure, formalised governance, specialised teams with adequate resources). Hence, the following were the coding criteria:

-

Coding criteria: Low = 0–3 indicators; Moderate = 4–6 indicators; High = 7–9 indicators across the three components.

2.1.2. Service Maturity Level (SML)

This dimension captured functional scope and sophistication across five stages:

-

Stage 1—Awareness: reactive, unstructured responses with no formalised services.

-

Stage 2—Initial: basic advisory services with designated contact persons and emerging FAIR awareness.

-

Stage 3—Developing: structured training programmes, repository deployment, emerging metadata support.

-

Stage 4—Managed: integrated workflows, formalised policies with enforcement, automated systems, operational cross-unit collaboration.

-

Stage 5—Transformational: advanced analytics services, leadership in standards development, innovative service models, full FAIR compliance with demonstrable impact.

Two independent reviewers coded each study using a decision matrix (available as Supplementary Material), with discrepancies resolved through consensus discussion. Inter-rater agreement for SML coding achieved κ = 0.79, indicating substantial agreement [33].

2.1.3. Information Flow Efficiency (IFE)

This dimension assessed how effectively institutional systems facilitate discoverability, accessibility, interoperability, and long-term preservation. Reported barriers were categorised as:

-

Discoverability: inadequate metadata, poor search functionality, and absence of persistent identifiers.

-

Accessibility: authentication restrictions, format incompatibilities, mediated deposit requirements.

-

Interoperability: heterogeneous metadata standards, proprietary formats, lack of API integration.

-

Preservation: insufficient backup systems, format obsolescence, and uncertain sustainability funding.

Institutions were classified as high IFE (0–1 bottlenecks), moderate IFE (2–3 bottlenecks), or low IFE (4+ systematic bottlenecks).

2.2. Methods for Identifying and Filtering Sources

A search strategy was employed to capture relevant academic and grey literature. Electronic databases included ERIC, JSTOR, PsycINFO, Web of Science, and Scopus, while key journals such as Journal of Academic Librarianship, College & Research Libraries, Data Science Journal, Information Science & Technology Abstracts, and Library Hi Tech were reviewed. Grey literature sources included OpenGrey, ProQuest Dissertations & Theses Global, EThOS, Google Scholar, and institutional repositories. This approach ensured a broad spectrum of insights on research data practices, trends, and challenges in academic libraries. The search strategy was structured using the PCC (Population, Concept, Context) framework:

-

Population: Academic libraries.

-

Concept: Research datasets, including management, curation, and utilisation.

-

Context: Academic and research environments (universities, colleges, and higher education institutions).

Building on the PCC framework, the multi-dimensional search protocol was structured around four distinct keyword categories to map the use of research datasets in academic libraries. The first category, “Academic Libraries,” addressed the institutional context with the search string: (“academic libraries” OR “research libraries” OR “university libraries” OR “college libraries”). The second category, “Research Data,” explored data types and management using: (“research data” OR “research datasets” OR “data management” OR “data curation” OR “data services”). The third category, “Researcher Engagement,” focused on practical data use and user challenges, employing: (“data literacy” OR “researcher support” OR “data use” OR “user behaviour” OR “data challenges”). The fourth category, “Library Service Models,” examined the role of library services in facilitating data-driven research with: (“open science” OR “FAIR principles” OR “data infrastructure” OR “library services” OR “data policy”). Boolean operators were applied to combine these categories, refine the search results, and exclude studies outside the scope, specifically literature reviews, systematic reviews, and meta-analyses, using: NOT (“literature review” OR “systematic review” OR “meta-analysis”). This structured keyword strategy ensured a comprehensive and focused capture of literature, encompassing the multifaceted nature of research data use in academic libraries, and providing a robust foundation for the scoping review. In addition to the fundamental PCC framework, the following eligibility criteria were established to refine the literature search for this study:

-

Language: English only.

-

Publication period: January 2015–September 2025.

-

Methodological and conceptual relevance: Studies with clear research data concepts, rigorous methodologies, and findings that inform library support for data-driven research.

The PRISMA-ScR flow diagram (Figure 1) illustrates the systematic process of searching literature, study screening, and selection for inclusion in this scoping review, as guided by Tricco et al. [32].

2.3. Study Selection Process

The study selection process, illustrated in Figure 1, was guided by the PRISMA-ScR framework to ensure methodological rigour and transparency. Initial searches across multiple databases and sources yielded 951 records, which were imported into Rayyan software (version 1.5.0) for deduplication [34]. Using Rayyan, 311 duplicates were removed, and 291 non-English records were excluded based on predefined language criteria, leaving 349 unique records for further screening. Screening was conducted using ASReview software (version 2.0), an open-source, machine learning-assisted tool that employs active learning to prioritise studies based on relevance predictions [35]. At this stage, 172 non-peer-reviewed studies were excluded, leaving 177 records for title and abstract evaluation. From these, 103 were excluded for misalignment with the research objectives. Accordingly, 74 studies were retained for detailed quality appraisal.

Quality assessment employed the Critical Appraisal Skills Programme (CASP) checklists, which provide structured, design-specific tools for evaluating study validity, reliability, and applicability [36]. During this appraisal, 28 studies were excluded due to inaccessible or incomplete full texts, and 12 were removed for methodological or conceptual limitations, including inadequate design, insufficient evidence, or lack of theoretical coherence. To ensure transparency, CASP appraisal focused on three domains: validity, results, and applicability, each addressing critical aspects of methodological quality and practical relevance [36].

Table 1. Application of CASP Domains in the Study Selection Process.

CASP Domain	Key Criteria	Example Questions	Exclusion Reasons	Studies Excluded (n)
Validity (Internal Rigour)	Clarity of aim; appropriateness of design; recruitment/sampling adequacy; transparency in data collection	- Was the research question clearly focused? - Was the methodology appropriate? - Was recruitment suitable?	Studies with unclear aims, inappropriate design, or poorly described data collection	5
Results (Reliability of Findings)	Consistency, credibility, transparency of findings; adequacy of analysis; bias handling	- Are the results clearly presented? - Was the analysis rigorous? - Were limitations acknowledged?	Studies with inconsistent results, inadequate analysis, or unaddressed bias	4
Applicability (Relevance and Contribution)	Relevance to objectives; transferability; implications for practice/policy	- Do the findings address review objectives? - Are the results transferable? - Are implications meaningful?	Studies with limited relevance, weak applicability, or unsupported findings	3

summarises the operationalisation of these domains, the guiding questions, and the studies excluded at each stage.

The selection process eventually resulted in the inclusion of 34 studies that demonstrated sufficient methodological quality and conceptual relevance, forming the final evidence base for this review. Given the heterogeneous global landscape of research data practices, Figure 2 visualises the regional distribution of the included studies to show where empirical evidence is concentrated.

2.4. Inter-Rater Reliability and Consistency

To ensure methodological rigour and minimise subjective bias, inter-rater reliability was maintained throughout the study selection and appraisal process. Two independent reviewers (authors) screened titles, abstracts, and full texts against the predefined inclusion and exclusion criteria. Cohen’s Kappa (κ = 0.82) was calculated based on independent screening of 100% of title/abstracts and full texts by both reviewers, indicative of strong agreement [33,36]. Discrepancies were resolved through dialogue, and where consensus could not be reached, a third reviewer provided a settlement [33]. Consistency was further reinforced during the application of the CASP appraisal checklists, with both reviewers independently evaluating validity, reliability of findings, and applicability. This iterative process of independent review, cross-checking, and consensus ensured that study selection was systematic, transparent, and reproducible, thereby enhancing the trustworthiness of the final evidence base [32,33,37].

2.5. Data Extraction and Charting

A structured data extraction framework guided by the analytical dimensions was used. Key extracted information included:

-

Bibliographic details: Author, year, journal/source, country.

-

Study context: Library type, disciplinary focus, institutional setting.

-

Dataset characteristics: Types, formats, size, metadata standards.

-

Library services: Access, curation, training, repository platforms, policy guidance.

-

Researcher engagement: Dataset usage, barriers, and FAIR awareness.

-

Standards alignment: FAIR principles, open science, SDGs.

-

Analytical framework indicators: IRI, SML, IFE.

-

Key findings: Trends, challenges, innovations.

The extraction framework was pilot-tested and refined for clarity and feasibility, particularly distinguishing reported versus measured researcher engagement. Data were charted in tabular and narrative formats to facilitate pattern identification across study contexts and service provision.

2.6. Data Synthesis and Analysis

Data were synthesised using descriptive and thematic approaches consistent with scoping review methodology.

• Descriptive analysis: summarised publication year, country, library type, and disciplinary focus.
• Thematic analysis: identified recurring concepts through iterative coding.
• Framework coding: positioned studies along IRI, SML, and IFE dimensions.

Triangulation of quantitative and qualitative findings highlighted consistencies, discrepancies, and gaps across studies. All procedures were documented to ensure transparency, reproducibility, and compliance with PRISMA-ScR guidelines. A summary of the extracted data is presented in tabular form in Appendix A.

3. Findings

This section presents findings structured according to the three-dimensional analytical framework aligned with the research objectives: Institutional Readiness Index (IRI), Service Maturity Level (SML), and Information Flow Efficiency (IFE). Before examining these dimensions in detail, it is important to contextualise the evidence base. Given the heterogeneous global landscape of research data practices, Figure 2 visualises the regional distribution of the included studies, illustrating where empirical evidence is most heavily concentrated.

The distribution reveals a marked concentration of studies in the Global North, particularly the United States, the United Kingdom, Australia, and Europe broadly, indicating that much of the empirical knowledge on research data services originates from high-income research environments. In contrast, studies from the Global South are fewer and unevenly represented, with clusters emerging only in Sub-Saharan Africa and parts of South and Southeast Asia. This imbalance suggests that global understandings of research data practices are shaped predominantly by contexts with comparatively advanced infrastructure and resources, underscoring the need for more geographically diverse empirical contributions. Therefore, an important caveat is that findings presented below reflect this geographic imbalance. Patterns identified may not generalise uniformly across all institutional contexts, particularly those with substantially different resource profiles or governance structures.

3.1. Effectiveness of Academic Library Services in Supporting Diverse Research Datasets: An Institutional Readiness Analysis

To examine the effectiveness of academic library services in supporting diverse datasets across disciplines (Objective 1), the IRI framework evaluated foundational capacity through technical infrastructure, governance structures, and resource allocation.

Table 2. Dataset Diversity and Institutional Readiness Index (IRI) Classification.

Dataset Category	Specific Types	Key Management Practices	Studies (n)	References
Textual and Documentary	Theses, dissertations, reports, manuscripts, survey transcripts	Appraisal, metadata creation, preservation workflows, and institutional repository curation	2	[38,39]
Numerical and Statistical	Experimental results, survey data, spreadsheets, quantitative research outputs	Data management plans (DMPs), statistical analysis training, and secure storage solutions	2	[18,40]
Multimedia and Geospatial	Geographic Information Systems (GIS), images, video recordings, audio files	Digital asset management, format migration, and limited preservation infrastructure	1	[41]
Scientific and Agricultural	Laboratory data, chemical/biological datasets, and agricultural research outputs	FAIR-compliant management, domain-specific metadata, specialised repositories	1	[41]
Open and Government	Public sector datasets, NGO data, government statistics	Reuse training, ethical compliance guidance, licensing support	1	[40]
Student-Generated	Undergraduate/postgraduate research outputs, datasets, posters, and capstone projects	Citation training, data cleaning instruction, and visualisation support	2	[18,42]
Metadata and Documentation	Data dictionaries, codebooks, catalogues, annotations, schema	FAIR evaluation tools, metadata standards implementation, quality assessment	2	[38,43]
Sensitive and Ethical	Human subjects’ data, health records, personal information	Anonymisation protocols, licensing frameworks, and ethical compliance oversight	2	[38,39]
Educational and Institutional	Teaching datasets, budgeting data, institutional planning records, and staffing information	Internal use management, decision support, and benchmarking applications	2	[44,45]
Behavioural and Engagement	Library usage data, student engagement measures, psychological assessments, and alumni loyalty data	Analytics support, survey data management, engagement tracking	1	[45]
Faculty Research	Researcher-generated data across lifecycle stages (creation, analysis, storage, archiving)	Full lifecycle support, research output integration, and long-term preservation	1	[46]
Repository-Hosted	Collections in institutional/disciplinary repositories (DSpace, Digital Commons, Zenodo, Dataverse, Dryad)	Platform management, deposit workflows, discovery optimisation	4	[47,48,49,50]
Administrative and Statistical	Library operations data, staffing records, budget information, service statistics	Benchmarking, advocacy support, and national database contributions	1	[51]
Policy-Governed	Datasets with eligibility restrictions, file type requirements, and sensitive data protocols	Embargo management, Creative Commons licensing, and DOI assignment	2	[47,52]
Medical and Health Research	Biomedical datasets, clinical research outputs, and health research linked to publications	Specialised biomedical repositories, ethical oversight, restricted access protocols	2	[53,54]
Impact and Indigenous	Bibliometric data, altmetrics, Māori data sovereignty (taonga) datasets	Impact assessment, culturally appropriate stewardship, and sovereignty protocols	1	[55]
Cross-Disciplinary	Multi-domain datasets spanning the humanities, sciences, and social sciences	Subject guides, interdisciplinary repositories, and cross-domain training programs	6	[48,49,56,57,58,59]
Digital and Open Access	Born-digital datasets, open platforms content	Discoverability enhancement, reuse facilitation, and LIS professional curation	1	[60]
Big Data and Open Science	Large-scale datasets, open data collections, open access resources	Advanced analytics integration, computational infrastructure, scalable repositories	2	[59,61]

Note: n = number of studies; FAIR = Findable, Accessible, Interoperable, Reusable; LIS = Library and Information Science; DMPs = Data Management Plans; DOI = Digital Object Identifier; NGO = Non-Governmental Organisation; GIS = Geographic Information Systems.

documents dataset diversity and associated management practices categorised by institutional readiness indicators.

Table 2 shows that academic libraries support a wide spectrum of dataset categories, each requiring distinct management practices, though coverage across studies is uneven. Textual, numerical, and metadata-related datasets receive moderate attention, typically involving appraisal, metadata creation, data management planning, and secure storage. More specialised categories, such as multimedia, geospatial, scientific, agricultural, medical, and sensitive datasets, appear less frequently and require domain-specific workflows, specialised repositories, and strict ethical or technical protocols. Libraries also provide targeted support for student-generated, educational, administrative, behavioural, and faculty research data, ranging from training and analytics to full lifecycle assistance. Repository-hosted and cross-disciplinary datasets are the most widely addressed, reflecting a strong emphasis on platform management, deposit workflows, discovery optimisation, and interdisciplinary training. Emerging areas such as Indigenous data sovereignty, open access data, and big data analytics highlight evolving expectations for culturally attuned stewardship, enhanced discoverability, and scalable computational infrastructures.

Table 3. Library Services Effectiveness Mapped to IRI Components.

Service Category	Specific Services Provided	IRI Component	Implementation Characteristics	Effectiveness Indicators	Studies (n)	References
Advisory and Consultation	RDM guidance, data curation consultation, FAIR principles advice, research mentoring, DMP review	Resource Allocation (staffing); Governance (advisory frameworks)	Offered through scheduled consultations; librarians serve as strategic advisors; consultative role well-established across institutions	Strong in institutions with moderate-high IRI; limited effectiveness in low IRI contexts due to infrastructure gaps	9	[38,40,41,44,47,49,56,57,62]
Repository and Technical Infrastructure	Institutional repository implementation, DMP tools (Version: v5.47) deployment, metadata support systems, storage solutions, preservation infrastructure	Technical Infrastructure (all components)	Technical services remain underdeveloped relative to demand; infrastructure gaps persist in long-term preservation, interoperability, and advanced analytics	Critical bottleneck limiting service effectiveness; requires substantial capital investment for improvement	9	[39,40,42,47,48,54,59,63]
Training and Capacity Building	Staff-led workshops, cohort-based training programs, data literacy curriculum integration, structured learning initiatives, embedded instruction	Resource Allocation (training programs); Governance (educational policies)	Active delivery through formal programs; emphasis on metadata, file management, and compliance; integration into academic curricula increasing	Effective when sustained and integrated into curricula; limited impact when delivered as isolated workshops	8	[38,41,45,46,50,62,64,65]
Policy Development and Cross-Unit Collaboration	RDM policy formulation, institutional guideline development, collaboration with IT and research offices, advocacy for best practices, and strategic planning	Governance (policy frameworks); Resource Allocation (coordination mechanisms)	Essential for institutional RDM growth; collaboration facilitates integrated service delivery; policy harmonisation remains a challenge	Effectiveness depends on enforcement mechanisms and cross-unit buy-in; weak in institutions with siloed structures	9	[47,48,55,56,61,66,67,68,69]
Open Data and Sharing Support	Open dataset promotion, data discovery facilitation, persistent identifier assignment, metadata standards implementation, and reuse guidance	Technical Infrastructure (discovery tools); Governance (open science policies)	Critical for supporting the Open Science movement; metadata quality and persistent identifiers are central to success; increasingly aligned with funder requirements	Effectiveness varies by metadata quality and repository sophistication; high effectiveness in high IRI institutions	5	[18,42,43,68,70]

Note: n = number of studies; RDM = Research Data Management; FAIR = Findable, Accessible, Interoperable, Reusable; DMP = Data Management Plan; IT = Information Technology.

analyses services through IRI components, examining how infrastructure, governance, and resources influence effectiveness.

Effectiveness Synthesis by IRI Level:

-

High IRI: All five service categories operational; integrated workflows; services complement infrastructure effectively [43,55,56,67]

-

Moderate IRI: Advisory, training, and policy services well-developed; infrastructure services lagging; effectiveness constrained by technical gaps [18,40,42,44,47,49,54,60,62,64,65,66,70]

-

Low IRI: Limited to advisory and basic training; infrastructure and policy services largely absent; services ineffective due to readiness deficits [38,39,41,46,48,50,52,53,58,63,69]

IRI analysis reveals asymmetry between strong consultative services and weak technical infrastructure. Technical infrastructure constitutes the primary constraint, being most resource-intensive. Governance readiness manifests in policy development, though implementation remains inconsistent. Resource allocation to training shows relative consistency, yet sustainability concerns persist. Libraries demonstrate moderate institutional readiness: possessing foundational governance and consultative expertise but constrained by infrastructure limitations preventing full-lifecycle RDM operationalisation.

Table 4. Data Literacy Competencies and Development Initiatives.

Competency Domain	Specific Skills	IRI Resource Analysis	Development Approaches	Studies (n)	References
Library Staff Technical Competencies	Technical RDM, curation, FAIR implementation	Critical constraint: Staff skill gaps limit readiness	Strong consultative skills; gaps in advanced technical competencies	10	[38,39,40,41,56,59,62,63,69,70]
Researcher Data Management Skills	Metadata creation, DMP development, and version control	Indirect readiness indicator: Competency gaps constrain utilisation	Structured training; cohort instruction; learning communities	4	[41,64,65,68]
Curriculum Integration	Programme embedding, learning communities, collaborative teaching	Strategic investment: Long-term capacity building	Systematic integration through librarian-faculty partnerships	2	[64,65]
FAIR and Open Science Literacy	FAIR principles, Open Science practices, data sharing ethics	Governance-resource intersection: Reflects staff expertise and policy alignment	Collaborative initiatives; targeted guidance; framework alignment	3	[61,67,70]

Note: n = number of studies; RDM = Research Data Management; FAIR = Findable, Accessible, Interoperable, Reusable; DMP = Data Management Plan.

examines data literacy, a key resource component, emphasising competencies required for service operationalisation.

Data literacy analysis reveals reciprocal deficits: staff lack technical competencies while researchers demonstrate management gaps. This creates interdependencies wherein effective service delivery requires competent librarians and meaningful utilisation demands literate researchers. From a resource allocation perspective, training investment faces sustainability challenges. FAIR literacy occupies the governance-resource intersection, requiring coordinated investment across multiple IRI components rather than isolated initiatives.

Table 5. Researcher Support Services and Interventions.

Support Type	Specific Activities	IRI Output Analysis	Service Delivery Model	Studies (n)	References
Direct RDM Guidance	Consultations, project support, curation guidance	Governance output: Translates policy into practice	One-on-one consultations; embedded librarians; lifecycle engagement	8	[38,39,40,41,42,56,66,67]
Training and Capacity Building	Workshops, cohort training, embedded instruction	Resource output: Reflects instructional infrastructure investment	Formal sessions; learning communities; discipline-specific workshops	4	[41,50,64,65]
Infrastructure and Tools Provision	Repository access, DMP tools, technical platforms	Infrastructure output: Direct manifestation of infrastructure readiness	Technology-mediated support; varying technical sophistication	5	[42,48,54,59,63]
Policy and Strategic Navigation	Policy interpretation, funder guidance, compliance support	Governance output: Demonstrates governance maturity	Advisory role in policy landscape; strategic positioning	4	[47,55,60,69]
Open Data and Reuse Facilitation	Open dataset guidance, persistent identifiers, and metadata consultation	Infrastructure-governance integration	Active sharing facilitation; Open Science alignment	3	[18,43,68]

Note: n = number of studies; RDM = Research Data Management; DMP = Data Management Plan.

examines researcher support services representing institutional readiness outputs.

Support services demonstrate how readiness translates into operational capacity. Governance outputs (guidance, policy navigation) are well-established regardless of readiness level, leveraging existing expertise without significant infrastructure. Infrastructure outputs reveal stark readiness disparities: only well-resourced institutions provide sophisticated platforms. The emphasis on guidance over technical services positions libraries as navigators and educators rather than providers of advanced analytical tools, reflecting readiness constraints wherein governance and capacity-building operate within existing allocations while infrastructure development requires substantial capital investment.

Integrated IRI assessment shows that service effectiveness is limited by uneven institutional readiness. Most libraries exhibit moderate governance readiness through consultative roles and developing policies, and moderate resource readiness with training programmes, but persistent staffing gaps. Technical infrastructure readiness is generally low to moderate, with basic repositories but limited preservation, interoperability, and analytics. This profile indicates operation primarily at consultative/stewardship levels rather than full-lifecycle RDM support, with effectiveness determined by the weakest component, typically technical infrastructure.

3.2. Challenges in Research Data Management: A Service Maturity Lens

To address Objective 2, this section maps challenges against the Service Maturity Level (SML) framework, which categorises capabilities from awareness through transformational stages.

Table 6. Researcher Challenges Across Service Maturity Levels.

Challenge Category	Specific Barriers	Impact on Research Practice	SML Stage Correlation	Studies (n)	References
Discovery and Awareness	Lack of awareness of available datasets, unclear dataset locations, inadequate discovery tools, insufficient promotional efforts	Researchers are unable to locate relevant datasets; missed opportunities for data reuse; reduced research efficiency and reproducibility	Most prevalent in Stage 1–2 (Awareness/Initial); persists in Stage 3 (Developing); minimal in Stage 4–5	4	[50,53,58,65]
Technical Skills and Competencies	Insufficient data management skills, limited understanding of FAIR principles, inadequate metadata creation abilities, weak repository proficiency	Inability to effectively engage with datasets; poor data documentation; non-compliant data practices; barriers to contributing data	Critical barrier across all SML stages; mitigated by training in Stage 3–4 but not eliminated	5	[48,59,64,65,68]
Data Quality and Documentation	Incomplete metadata, inconsistent formatting, inadequate documentation, missing persistent identifiers, and poor data descriptions	Data reuse severely hindered; interoperability challenges; reduced trust in dataset validity; increased effort required for data interpretation	Reflects institutional SML level; severe in Stage 1–2; improving in Stage 3; addressed systematically in Stage 4–5	4	[47,49,57,68]
Access and Preservation Barriers	Technical access restrictions, mediated deposit requirements, paywall limitations, inadequate long-term preservation, and format obsolescence concerns	Delayed or prevented access to datasets; uncertainty about long-term availability; concerns about data loss; restricted data sharing	Infrastructure-related; severe in Stage 1–3; partially resolved in Stage 4; comprehensively addressed in Stage 5	4	[47,48,53,59]
Institutional Support Gaps	Limited guidance and mentorship, insufficient embedded support, lack of dedicated RDM staff, inadequate technical assistance for data-intensive projects	Researchers lack the necessary support for complex data management tasks; an increased burden on researchers; and suboptimal data practices	Directly correlates with SML level; severe in Stage 1–2; moderate in Stage 3; minimal in Stage 4–5, where embedded support exists	3	[50,56,66]
Policy and Compliance Confusion	Ambiguous data-sharing policies, unclear copyright and licensing requirements, complex funder compliance mandates, and conflicting institutional guidelines	Hesitancy to share data; compliance failures; risk aversion; delayed dataset publication; ethical concerns	Policy clarity improves with SML progression; severe in Stage 1–2; improving in Stage 3–4; clear frameworks in Stage 5	4	[48,55,60,61]

Note: n = number of studies; RDM = Research Data Management; FAIR = Findable, Accessible, Interoperable, Reusable; SML stages: 1 = Awareness, 2 = Initial, 3 = Developing, 4 = Managed, 5 = Transformational.

examines researcher challenges within this maturity context (see decision matrix: https://www.mdpi.com/article/10.3390/info16121046/s1, accessed on 28 July 2025).

Challenge-SML Correlation Summary:

-

Stage 1 (Awareness): All six challenge categories severe; no systematic interventions [52,58,69]

-

Stage 2 (Initial): All challenges present; limited mitigation through basic advisory services [38,39,41,46,48,50,53,63]

-

Stage 3 (Developing): Discovery, skills, and support gaps improving through training; quality, access, and policy challenges persist [18,40,42,44,47,49,54,60,62,64,65,66,70]

-

Stage 4 (Managed): Skills and support gaps largely resolved; quality and access improving; policy frameworks established [43,55,56,61]

-

Stage 5 (Transformational): Minimal challenges; systematic solutions across all categories [67]

Researcher challenges mapped against SML reveal systematic maturity deficits. Discovery barriers indicate Stage 1–2 stagnation, lacking the structured mechanisms characteristic of Stage 3. Technical gaps reflect Stage 2–3 deficits: training remains sporadic rather than systematic. Quality issues signal failure to progress beyond Stage 3: standards exist but lack enforcement, validation remains manual, hallmarks of Stage 4 maturity. Access barriers similarly reveal Stage 3–4 gaps: preservation infrastructure remains reactive, automated workflows undeveloped. Support gaps and policy confusion both reflect Stage 2–3 stagnation: services remain ad-hoc, staffing is insufficient, and policy frameworks are incomplete. Most institutions operate between Initial and Developing stages (2–3), with progression to Managed services (Stage 4) constrained by infrastructure, resources, and governance fragmentation.

Table 7. Library Institutional Challenges Across Service Maturity Levels.

Challenge Category	Specific Issues	Organizational Impact	SML Progression Pattern	Studies (n)	References
Technical Infrastructure Limitations	Inadequate repository platforms, insufficient storage capacity, unreliable backup systems, limited technical tools, and scalability constraints	Inability to support full RDM lifecycle; service delivery gaps; poor user experience; preservation risks; competitive disadvantage	Fundamental barrier preventing SML progression beyond Stage 2–3; requires capital investment to advance	6	[38,47,48,54,59,69]
Staff Skills and Competencies	Strong consultative skills but weak technical RDM competencies; limited data curation expertise; insufficient FAIR implementation knowledge; gaps in specialised domain knowledge	Service quality limitations; inability to provide advanced technical support; dependence on external expertise; staff frustration; restricted service expansion	Limits progression to Stage 4–5; professional development enables advancement	4	[57,58,62,66]
Resource and Funding Constraints	Limited budgets, insufficient staffing levels, competing institutional priorities, unsustainable service models, and the inability to invest in infrastructure	Restricted training programs; delayed repository development; inability to scale services; staff burnout; service discontinuation risks	Constrains all SML stages; particularly limits Stage 3–4 transition	6	[55,56,60,65,69,70]
Policy and Governance Gaps	Inconsistent institutional policies, unclear service mandates, poor coordination with IT and research offices, and the absence of strategic frameworks	Operational inefficiencies; duplicated efforts; confused service responsibilities; poor integration with research workflows; reduced institutional impact	Prevents systematic SML advancement; formalisation enables Stage 3–4 progression	4	[47,48,56,67]
Researcher Engagement Challenges	Low researcher participation, reluctance to deposit data, limited awareness of library services, and disciplinary cultural barriers	Underutilised services; limited RDM adoption; reduced library relevance; difficulty demonstrating value; perpetuation of poor data practices	Present across all SML stages; diminishes as services mature and demonstrate value	4	[48,50,53,54]
Data Heterogeneity and Standards	Diverse disciplinary practices, inconsistent data formats, multiple metadata standards, complex interoperability requirements, varying quality expectations	Management complexity; preservation challenges; discovery difficulties; resource-intensive customisation requirements; barriers to cross-disciplinary use	Complexity increases with SML advancement; it requires sophisticated solutions at Stage 4–5	4	[47,49,57,68]

Note: n = number of studies; RDM = Research Data Management; FAIR = Findable, Accessible, Interoperable, Reusable; IT = Information Technology; SML stages: 1 = Awareness, 2 = Initial, 3 = Developing, 4 = Managed, 5 = Transformational.

examines library challenges through SML, revealing how institutional constraints perpetuate researcher barriers.

Challenge-SML Relationship:

-

Stage 1–2 Transition: Overcoming awareness; establishing basic infrastructure; securing minimal resources.

-

Stage 2–3 Transition: Building technical infrastructure; developing training programs; formalising initial policies.

-

Stage 3–4 Transition: Resolving infrastructure gaps; professionalising staff competencies; establishing governance frameworks; achieving cross-unit collaboration.

-

Stage 4–5 Transition: Advanced analytics capabilities; innovative service models; comprehensive FAIR implementation; full researcher engagement.

Library challenges reveal associations with systematic maturity patterns. Infrastructure limitations constitute the primary Stage 3–4 barrier. Staff gaps represent Stage 2–3 limitations: consultative excellence without technical expertise for systematic service delivery. Combined with resource constraints, these function as cross-stage barriers. Governance fragmentation is associated with Stage 2–3 stagnation. Engagement difficulties both reflect and may reinforce early-stage maturity.

The researcher barriers (Table 6) and library limitations (Table 7) exhibit striking symmetry, suggesting interdependence rather than coincidence. This pattern indicates that library maturity constraints may contribute to researcher barriers, creating patterns in which reduced service utilisation, limited evidence of value, and constrained resource allocation appear to co-occur, potentially perpetuating maturity constraints.

3.3. Alignment with International Standards: Information Flow Efficiency Assessment

Addressing Objective 3, this section assesses library services’ facilitation of data-driven research according to international standards using the Information Flow Efficiency (IFE) framework, which evaluates discoverability, accessibility, interoperability, and preservation, core FAIR dimensions.

Table 8. Library Service Alignment with International Standards: IFE-Based Assessment.

Service Area	Implementation Status	Alignment with Standards	IFE Classification	FAIR Dimension Supported	Studies (n)	References
RDM Advisory and Consultation Services	Advisory services, consultations, and DMP support are widely implemented; consistency varies across institutions	Aligns with OECD RDM guidelines; DMP guidance meets major funder requirements (NIH, NSF, ERC); consultative model consistent with international best practices	Moderate-High IFE: Supports Findability (guidance on metadata) and Accessibility (DMP advisory); limited impact on Interoperability and Reusability due to infrastructure gaps	Findable (advisory on metadata); Accessible (DMP compliance guidance)	6	[38,47,48,54,59,69]
Data Literacy Training Programs	Structured programs and learning communities enhance competencies in metadata, file management, FAIR principles, and open data	Supports FAIR compliance; strengthens capacity for data sharing and reuse; aligns with UNESCO Open Science recommendations and RDA guidelines	Moderate IFE: Improves Findability (metadata skills) and Reusability (documentation practices); coverage uneven; curriculum integration emerging	Findable (metadata creation); Reusable (documentation standards)	4	[56,62,64,65]
Repository and Technical Infrastructure	Institutional repositories implemented; mediated deposits common; metadata support systems variable; long-term preservation gaps evident	Supports FAIR principles (Findable, Accessible); contributes to global RDM stewardship frameworks; partial compliance with CoreTrustSeal requirements	Low-Moderate IFE: Basic Findability and Accessibility achieved; Interoperability limited by proprietary systems; Reusability constrained by preservation gaps	Findable (basic discovery); Accessible (mediated access); limited Interoperable and Reusable	6	[49,57,59,68,69,70]
Policy Development and Governance	RDM policies developed; institutional guidelines established; advocacy for responsible data management; alignment efforts ongoing	Complies with OECD principles, UNESCO recommendations, and major funder standards; ensures legal, ethical, and preservation compliance; contributes to national RDM frameworks	Moderate IFE: Establishes governance for all FAIR dimensions; implementation inconsistency creates IFE bottlenecks; enforcement mechanisms are weak	All FAIR dimensions (governance frameworks)	4	[48,56,61,67]
Cross-Institutional Collaboration	Collaboration with IT departments, research offices, and academic faculties; integrated service models; shared governance structures	Supports FAIR and Open Science frameworks promoting interoperable and collaborative stewardship; aligns with consortium models (LIBER, CNI)	Moderate IFE: Improves Interoperability through cross-unit coordination; organisational silos persist; strategic partnerships are developing	Interoperable (cross-system coordination); Accessible (integrated access)	4	[47,56,66,67]
Open Science and Open Data Support	Promotion of open access; persistent identifier assignment; metadata standards implementation; discovery tools; data reuse support	Directly supports FAIR principles, Open Science practices, and discipline-specific reproducibility standards; aligns with Plan S, UNESCO, and Berlin Declaration	Moderate-High IFE: Strong Findability (persistent identifiers); improving Accessibility (open access); Interoperability advancing (standards); Reusability enhanced (licensing clarity)	All FAIR dimensions (comprehensive Open Science support)	3	[49,61,68]

Note: n = number of studies; RDM = Research Data Management; FAIR = Findable, Accessible, Interoperable, Reusable; IFE = Information Flow Efficiency; OECD = Organisation for Economic Co-operation and Development; UNESCO = United Nations Educational, Scientific and Cultural Organisation; RDA = Research Data Alliance; DMP = Data Management Plan; LIBER = Association of European Research Libraries; CNI = Coalition for Networked Information.

summarises alignment incorporating IFE metrics to assess operational effectiveness beyond policy commitments.

The overall IFE-Standards Alignment Summary:

-

High IFE: Comprehensive FAIR implementation across all dimensions; minimal bottlenecks; transformational services [43,56,67].

-

Moderate IFE: Selective FAIR compliance; Findability and Accessibility prioritized; Interoperability and Reusability gaps [18,39,40,42,44,47,49,54,55,60,61,62,64,65,66,70].

-

Low IFE: Limited FAIR compliance; systemic bottlenecks across all dimensions; aspirational policies without operational capacity [38,41,46,48,50,52,53,58,63,69].

Across the service areas, the relationship between Standards Alignment and IFE shows that compliance with international frameworks improves information flow, but only when supported by adequate technical and organisational capacity. Services with strong alignment, such as RDM advisory, data literacy training, and Open Science support, demonstrate moderate to high IFE because standards-driven practices (e.g., FAIR metadata, DMP requirements, persistent identifiers) directly strengthen Discoverability, Interoperability, and Accessibility. However, areas with only partial standards alignment, most notably repository and technical infrastructure, show low to moderate IFE, as adherence to FAIR and CoreTrustSeal (The Hague, Netherlands) principles is undermined by weak metadata, limited search capability, and inadequate preservation systems. Policy development and cross-institutional collaboration align well with OECD, UNESCO, and Open Science frameworks, yet inconsistent implementation and siloed structures inhibit their ability to enhance IFE effectively. Therefore, the table demonstrates that standards alignment is a necessary but insufficient condition for high IFE: strong information flow depends not only on adopting recognised standards but also on the depth of technical infrastructure, enforcement, and organisational integration that make those standards operational.

Triangulation across the framework (IRI, SML, IFE) reveals interconnected constraints: low institutional readiness constrains service maturity advancement, perpetuating low information flow efficiency, reducing researcher engagement, limiting evidence of value, constraining investment in readiness, and creating reinforcing negative feedback loops. The results of this review indicate that breaking these cycles requires coordinated interventions across all three dimensions simultaneously: infrastructure investment (IRI), service systematisation (SML), and bottleneck elimination (IFE), supported by governance enabling cross-unit coordination and strategic resource allocation. The persistent gap between policy alignment and operational effectiveness indicates standards compliance cannot be achieved through policy adoption alone but demands addressing the infrastructure, maturity, and efficiency gaps separating aspirational commitments from operational capabilities.

4. Discussion

This scoping review synthesised evidence from 34 empirical studies examining academic libraries’ engagement with research datasets, analysing service provision, capacity constraints, and standards alignment through a three-dimensional framework. The analytical approach, integrating IRI, SML, and IFE, reveals that libraries face systematic structural patterns rather than isolated operational challenges. This section interprets key findings and their implications for research data ecosystems.

Geographic context caveat: The evidence base derives predominantly from high-income contexts (approximately 65% Global North), reflecting broader knowledge-production asymmetries in library and information science research [8,9]. Patterns identified below may not transfer uniformly to resource-constrained settings and require empirical validation across diverse institutional contexts before generalisation.

4.1. The Capacity-Complexity Paradox: Why Dataset Diversity Outpaces Institutional Capacity

The diversity of datasets managed by libraries (Table 2) represents more than incremental expansion; it signifies fundamental mission redefinition from collection stewardship to active data governance [12,14]. However, IRI analysis exposes a critical pattern: as libraries assume responsibility for increasingly complex datasets, progressing from textual materials to multimedia, geospatial, and big data, institutional capacity appears to grow incrementally whilst complexity demands grow exponentially. This creates what we term a capacity-complexity mismatch.

Libraries managing traditional datasets demonstrate moderate infrastructure readiness [18,38,39,40], possessing adequate systems for established workflows. However, those handling advanced formats [41,59,61] reveal infrastructure gaps, not necessarily from lesser commitment but because these formats demand fundamentally different capabilities: computational resources for big data analysis, specialised preservation for multimedia formats, and domain expertise for scientific datasets. Dataset diversity is not simply additive; each new category introduces unique technical, ethical, and governance requirements that compound rather than accumulate [2,4].

This mismatch helps explain why libraries with strong consultative services [38,40,41,44,47,49,56,57,62] simultaneously exhibit weak technical infrastructure [39,40,42,47,48,54,59,63]. Consultative excellence leverages existing information literacy expertise, requiring minimal additional infrastructure. Advanced data curation demands specialised technical skills, computational infrastructure, and domain knowledge, resources requiring substantial capital investment and sustained institutional commitment [55,56,60,65,69,70]. Libraries thus become positioned as advisory intermediaries rather than comprehensive service providers, a positioning potentially reflecting resource constraints rather than strategic choice.

These findings imply that rhetorical adoption of data stewardship roles does not automatically translate into operational capacity [8,11]. Meaningful transformation requires acknowledging that certain RDM functions demand capabilities beyond traditional library expertise and infrastructure. This suggests libraries may need to pursue selective specialisation rather than comprehensive coverage, focusing on areas where existing capabilities provide a comparative advantage, consultative services, metadata expertise, policy development, whilst collaborating with IT services and research computing for infrastructure-intensive functions [47,56,66,67].

These findings extend Cox et al.’s [57] maturity model by identifying the capacity-complexity mismatch as a specific barrier to progression. Whilst Cox et al. [57] documented service stages, the review framework reveals why libraries struggle to advance despite effort. This complements rather than replaces existing models, providing diagnostic specificity about constraint sources.

4.2. Structural Barriers to Service Progression

SML analysis reveals that most libraries in the sample operate at Stages 2–3 (Initial–Developing), struggling to progress toward Stage 4 (Managed) despite sustained effort [48,57,59]. This pattern reflects not inadequate commitment but what this review terms structural progression barriers that may prevent incremental advancement. A critical barrier appears at the Stage 3–4 transition, where progression appears to demand simultaneous advancement across multiple dimensions, infrastructure, staffing, governance, and researcher engagement, creating interdependencies that isolated interventions cannot easily resolve. Researcher challenges (Table 6) and library challenges (Table 7) exhibit striking symmetry: discovery barriers [50,53,58,65] correlate with infrastructure limitations [38,47,48,54,59,69]; technical skill gaps [48,59,64,65,68] correspond to staff competency deficits [57,58,62,66]; policy confusion [48,55,60,61] reflects governance fragmentation [47,48,56,67].

This symmetry suggests potential interdependence: library maturity constraints may contribute to researcher barriers, which reduce service utilisation, limiting evidence of value, constraining resource allocation for advancement, and perpetuating maturity constraints, a pattern consistent with reinforcing feedback loops. However, causal directionality could not be established from cross-sectional data. The progression barrier appears to operate through what this review terms threshold effects: advancing from Stage 3 to Stage 4 may require crossing simultaneous thresholds in infrastructure sophistication, staff expertise, policy integration, and user engagement. Crossing one threshold without others may produce minimal benefit, improved infrastructure without staff expertise to leverage it, enhanced training without infrastructure to practise on, and clear policies without enforcement mechanisms. This could explain why incremental investments in individual dimensions appear not to produce maturity advancement in the studies examined.

This review, therefore, suggests that sequential development strategies, beginning with infrastructure, followed by skills development, and concluding with policy implementation, are insufficient for achieving Stage 4 maturity. Instead, libraries might require bundled interventions that simultaneously address infrastructure, expertise, governance, and engagement. This demands different resource allocation models: rather than distributing limited resources across dimensions incrementally, concentration on achieving threshold-crossing in a coordinated fashion may prove more effective, even if this means delaying advancement in some areas to enable breakthrough in others. However, this hypothesis requires empirical testing through intervention studies.

In terms of methodological contribution, the review framework differs from linear maturity models by revealing interdependencies that create progression barriers. The DCC Curation Lifecycle Model [54] provides process guidance but does not explain stagnation. This review framework identifies where constraints originate (IRI), what capabilities exist (SML), and whether systems function effectively (IFE), enabling targeted diagnosis rather than generic capacity-building.

4.3. The Implementation Gap

IFE analysis exposes systematic discrepancies between policy awareness of international standards and operational effectiveness in enabling information flows. Libraries demonstrate policy awareness of FAIR principles [49,61,68], OECD guidelines [38,47,48,54,59,69], and UNESCO recommendations [56,62,64,65], reflected in documented policies and training content. However, operational reality reveals persistent bottlenecks across IFE dimensions, discoverability, accessibility, interoperability, and preservation, creating what this review terms the implementation gap.

This gap manifests distinctly across information flow dimensions. Discoverability suffers from inadequate metadata and search systems [47,49,57,68], despite widespread recognition of metadata importance. Accessibility faces technical barriers [47,48,53,59] that contradict open access policy commitments. Interoperability remains constrained by standards heterogeneity [47,49,57,68] despite FAIR principle endorsement. Preservation shows greatest deficits [38,47,48,54,59,69], with infrastructure inadequacy potentially threatening long-term data survival despite policy acknowledgement of preservation importance.

The implementation gap appears to occur because standards compliance is not a linear technical problem but a complex socio-technical challenge requiring coordinated transformation across institutional systems, workflows, and cultures [15,16,17]. Libraries can articulate FAIR principles without possessing infrastructure to operationalise them, adopt open data policies without technical capacity to implement open repositories, and endorse preservation standards without resources for sustained preservation activities. The gap thus appears to reflect resource and capacity constraints rather than knowledge deficits.

A critical insight emerging from service-specific patterns is that advisory and training functions achieve moderate-to-high IFE [38,47,48,54,56,59,62,64,65,69] because they operate through human-mediated channels that circumvent infrastructure bottlenecks. In contrast, repository and technical services demonstrate low-to-moderate IFE [49,57,59,68,69,70], reflecting their fundamental dependence on underlying infrastructure. This suggests differential pathways to effectiveness: services leveraging human expertise can achieve reasonable effectiveness despite infrastructure limitations, whilst services requiring technical systems remain constrained until infrastructure adequacy is achieved.

The strategic implication is that library positioning should account for these differential service pathways. In resource-constrained contexts, infrastructure-dependent services should be prioritised only when sustainable investment is assured, while consultative and educational services—being less infrastructure-dependent offer a pragmatic means to maximise impact in the interim. This recommendation is most applicable to contexts comparable to those represented in the review sample and may not generalise universally. The underlying strategic imperative derives from this review’s core analytical contribution, summarised in Figure 3: three interdependent structural barriers that constrain academic library RDM effectiveness through reinforcing feedback loops, requiring coordinated, multidimensional intervention.

4.4. Ecosystem Dependencies and Geographic Considerations

The triangulated analysis across IRI, SML, and IFE frameworks reveals that effective RDM requires ecosystem-level coordination extending beyond individual library control. Three critical dependencies constrain institutional effectiveness. First, standards heterogeneity [47,49,57,68] creates interoperability challenges that individual libraries cannot resolve unilaterally. Disciplinary metadata standards, repository platforms, and preservation approaches vary across institutions and domains, requiring coordinated standards adoption and system integration at consortium or national levels [17,26]. Libraries adopting FAIR principles individually may achieve limited impact if surrounding ecosystem partners use incompatible systems.

Second, researcher behaviour and expectations [48,50,53,54] reflect disciplinary cultures and funder requirements shaped by factors beyond library influence. Low researcher engagement [48,50,53,54] stems partly from library service limitations but also from disciplinary norms, research workflow incompatibilities, and competing demands. Libraries can improve services, but cannot unilaterally transform research cultures [14,19,23]. Third, resource allocation decisions [55,56,60,65,69,70] occur at institutional levels where libraries compete with other priorities. Infrastructure investment, staffing levels, and strategic positioning reflect institutional decision-making influenced by multiple stakeholders. Libraries advocating for RDM investment face competing demands from teaching support, collections, physical spaces, and emerging technologies.

These ecosystem dependencies help explain regional disparities observed: well-resourced institutions in developed contexts achieve higher maturity and IFE [38,47,49,56] not solely through superior library management but through advantaged positioning within resource-rich ecosystems featuring coordinated standards adoption, supportive research cultures, and sustained institutional investment. Resource-constrained institutions [50,53,58,63] face compounded disadvantages: inadequate infrastructure, fragmented standards environments, weak institutional support, and contexts where data-sharing norms remain underdeveloped.

A critical limitation relates to this review’s sample’s geographic imbalance (65% Global North), which means these ecosystem patterns reflect predominantly high-resource contexts. Whether similar dependencies operate in different ways across resource-constrained settings, or whether entirely different factors dominate, remains empirically underexplored. Addressing these disparities requires ecosystem-level interventions, national repository infrastructures, coordinated standards adoption, policy harmonisation, and capacity-building initiatives, which transcend individual institutional capabilities [6,10,17].

4.5. Framework Contribution and Diagnostic Utility

This study’s analytical framework contributes methodologically by demonstrating that single-dimensional assessments may misrepresent RDM capacity. Evaluating libraries solely on service presence (SML), infrastructure availability (IRI), or policy alignment can produce an incomplete understanding. The three-dimensional approach reveals:

-

Libraries with strong IRI may exhibit low SML if unable to operationalise infrastructure into effective services.

-

High SML without adequate IRI may prove unsustainable as service demands exceed capacity.

-

Policy alignment (apparent high maturity) may mask operational ineffectiveness (low IFE) when implementation gaps persist.

-

IFE assessment exposes whether stated capabilities translate into actual information flows.

The framework thus provides diagnostic specificity: identifying whether effectiveness constraints stem from readiness deficits (IRI), maturity limitations (SML), or implementation failures (IFE). This specificity enables targeted interventions, infrastructure investment addresses IRI constraints, service development addresses SML gaps, workflow redesign addresses IFE bottlenecks, rather than undifferentiated capacity-building that may miss critical limiting factors. The review framework complements rather than replaces existing tools. Cox et al.’s [57] maturity model provides stage descriptions; this review framework explains why progression stagnates. The DCC Curation Lifecycle Model [54] guides processes; this review framework diagnoses capacity gaps preventing process implementation. FAIR assessment tools (version 1.0) [43,71] measure data compliance; this review framework identifies institutional barriers to achieving compliance. This positions this review contribution as diagnostic rather than prescriptive, identifying constraint sources to inform strategic decisions.

The limitations of the framework arise from its use of categorical indicators (low/moderate/high; Stages 1–5), which necessarily simplify complex institutional realities. Variation exists within each category, and the staged model implies a linear progression that may not reflect actual development trajectories. Although interdependent elements are separated for analytical clarity, this segmentation and the broader simplification, while useful for cross-study comparison, risk overlooking contextual nuance. Consequently, the framework requires empirical validation through prospective application across diverse institutional settings before wider adoption.

5. Conclusions and Recommendations

Academic libraries occupy critical yet precarious positions within research data ecosystems. They have taken on stewardship responsibilities that exceed their traditional capabilities, creating expectations that many lack the resources or infrastructure to meet. Evidence, drawn largely from high-resource settings, indicates that effectiveness constraints arise less from insufficient effort than from structural misalignments between service expectations and institutional capacity, reinforced by ecosystem dependencies beyond library control. Three key patterns emerge: a capacity–complexity mismatch, where dataset diversity grows exponentially while infrastructure scales only linearly; structural progression barriers, as advancement from developing to managed services appears to require concurrent progress across multiple dimensions rather than sequential development; and an implementation gap, in which standards awareness does not translate into operational effectiveness without adequate infrastructure and resources. These patterns are interconnected: low institutional readiness aligns with limited-service maturity and low information flow efficiency, which in turn relate to reduced researcher engagement and constrained resource allocation. Nonetheless, the causal mechanisms underlying these relationships require confirmation through longitudinal research.

In terms of recommendations, library practitioners should focus on areas of comparative advantage, such as consultative support, metadata expertise, and policy navigation, while collaborating with IT for infrastructure-heavy functions. They should address infrastructure, staffing, governance, and engagement together, and use the framework to identify readiness, maturity, and implementation constraints. Institutional leaders should recognise ecosystem dependencies, align expectations with capacity, and sustain investment in technical infrastructure. Policymakers and funders should support ecosystem-level initiatives, account for resource-context diversity, and fund longitudinal research to clarify causal mechanisms.

6. Limitations and Future Research Directions

This analysis has several acknowledged limitations that constrain generalisability and interpretation. First, it is limited to English-language studies from 2015–2025, potentially excluding relevant research from other linguistic contexts or historical periods that might reveal different developmental trajectories. Second, approximately 65% of studies originate from Global North contexts, limiting transferability to resource-constrained settings and institutions with differing resource profiles or governance structures. Thus, this geographic distribution of included studies reflects scholarly publication infrastructure rather than RDM practice distribution. The predominance of research from high-income contexts indicates where LIS research is conducted and published, not necessarily where RDM activities or challenges are most prevalent. Context-specific factors (institutional funding models, national policies, disciplinary cultures) likely influence how findings transfer across settings, but systematic investigation of these contextual moderators remains a priority for future research. Third, predefined databases and keywords, while systematic, may not have captured all relevant literature, and the deliberate exclusion of systematic reviews and meta-analyses omitted consolidated quantitative evidence. Fourth, the application of the framework involves subjectivity: categorical indicators (low/moderate/high IRI; Stages 1–5 SML; IFE classifications) simplify complex realities, within-category variation exists, and reviewer interpretations may introduce bias despite high inter-rater reliability (κ = 0.82 for study selection; κ = 0.79 for SML coding). Fifth, the predominance of cross-sectional data prevents causal claims regarding reinforcing feedback loops or barriers to progression; longitudinal studies are required to confirm causal mechanisms. Sixth, publication bias may skew findings toward successful implementations, underrepresenting failures. In addition, the IRI–SML–IFE framework itself requires empirical validation through prospective application across diverse contexts and has not yet been tested with practising librarians to assess its diagnostic utility in operational settings.

Building on the limitations of this scoping review, several areas for further research emerge. Geographic expansion beyond English-language publications and the current timeframe would capture insights from diverse linguistic, cultural, and historical contexts, enhancing global applicability. Comparative research examining how institutional context moderates RDM patterns would clarify which findings reflect universal challenges versus context-specific phenomena. Also, longitudinal studies following institutions as they develop RDM capabilities could test whether observed correlations reflect causal mechanisms and reveal trajectories of successful development. Intervention studies comparing bundled, simultaneous approaches to sequential strategies would empirically evaluate strategic recommendations. Framework refinement is needed, including the development of quantitative indicators for IRI, SML, and IFE, and exploration of alternative, non-linear maturity pathways suited to varied contexts. Comparative cross-institutional research could illuminate how differences in practices, policies, and infrastructure affect effectiveness and scalability. Finally, practitioner validation through workshops or pilot implementations would assess the framework’s diagnostic utility and identify refinements for operational use.