Aligning Tri-Agency RPP reporting with CRediT and Dictionary

Profile

Composite institution profile

The challenge

What problem were they trying to solve?

The Tri-Agency progress-reporting workflow had become a known pain point for PIs at the university — different agencies wanted different things, the institution's grants office did the consolidation by hand, and the result was that PIs filled in the same contribution information three different ways for three different agencies. CIHR's preference for structured contribution metadata, NSERC's acceptance of structured-or-narrative, and SSHRC's long-standing humanities tradition of narrative-only progress reporting meant that any institutional template had to handle all three formats without forcing humanities PIs into structured-only forms or biomedical PIs into long-form narrative. The grants office wanted a single source of truth — the institutional CRIS — that could emit each of the three agency's preferred formats on demand. A complicating factor was that the Tri-Agency Open Access Policy (2015, updated 2023) was about to be extended to data-management plans, and the institution wanted the new workflow to handle DMP reporting in the same pipeline.

The approach

How they implemented it

The grants office and the library co-led the project, with the research-systems team as the technical implementer. The architectural decision early on was that the CRIS would hold contribution data in a normalised CASRAI-Dictionary-aligned form, and three agency-specific exporters would translate from that source of truth into each agency's preferred format. The CRediT taxonomy filled the contribution-role layer of the Dictionary, with the eighteen CIHR contribution categories cross-walked to CRediT roles wherever a mapping existed and flagged as CIHR-specific where it did not. SSHRC narrative reporting was handled by surfacing the CRediT data as a prompt-and-context structure inside the narrative form: the PI saw their structured contribution data on the left of the screen and wrote their narrative on the right, with the option to leave the narrative as the only output if they preferred (preserving SSHRC's humanities-friendly convention). NSERC reporting used a hybrid format that emitted both the structured data and a short narrative auto-drafted from the CRediT roles, which the PI could edit before submission. The DMP integration was bolted on in year two by adding the FAIR-principles-aligned Dictionary terms to the same source of truth, so a single DMP could emit Tri-Agency-compliant variants on demand.

Timeline

Rollout phases

1. Months 1–4

   Cross-walk: 18 CIHR categories → CRediT

   Working group of three grants officers, two librarians, one research-systems lead. Mapped CIHR's contribution categories to CRediT; documented the four CIHR-specific categories with no clean CRediT mapping (preserved as CIHR-specific extensions in the Dictionary-aligned schema).

2. Months 5–7

   CRIS source-of-truth refactor

   DSpace-CRIS contribution-data model refactored to align with CASRAI Dictionary terms. Existing data migrated in two batches over six weeks; QA pass caught ~3% of legacy records needing manual cleanup.

3. Months 8–10

   Three agency exporters built

   CIHR (structured), NSERC (hybrid), SSHRC (narrative-with-context) exporters built. PI-facing UI deliberately identical across exporters; format selection is automatic from the grant's funding source.

4. Months 11–12

   Pilot with 24 PIs across 3 agencies

   Eight PIs from each agency volunteered for the pilot. Reported a 40–60% reduction in time spent on the progress-reporting step; one SSHRC PI initially preferred the legacy narrative-only form but converted after seeing the auto-prompt structure.

5. Year 2

   Institution-wide rollout + DMP integration

   Workflow rolled out to all 3,200 active grants. Tri-Agency DMP integration added, surfacing FAIR-principles Dictionary terms in the same workflow.

Outcomes

Illustrative outcomes

Every metric below is illustrative — synthesised from observed patterns across multiple adoption journeys, not attributed to a single real institution.

Lessons learned

What they would tell the next institution

• 01Modelling contribution data in a CASRAI-Dictionary-aligned source-of-truth and translating outward is much easier than maintaining three parallel agency-specific data models.
• 02SSHRC humanities PIs were not the obstacle people feared. When given a narrative form with structured context on the left of the screen, almost all preferred it over the legacy narrative-only form.
• 03The four CIHR categories with no clean CRediT mapping were flagged honestly rather than forced. Pretending they mapped would have produced wrong data; the explicit "CIHR-specific extension" approach made the system trustable.
• 04The pilot scale (24 PIs) was big enough to surface UX issues and small enough to be intimate. Anything smaller would have missed format-specific edge cases; anything larger would have lost the conversation quality.
• 05Year-2 DMP integration was almost free because the source-of-truth refactor in year 1 had already aligned data to the Dictionary. That investment paid off twice.

What's next

Planned next steps

The grants office is now extending the workflow to handle CIHR's "Strategic Plan 2021-2031" requirement that all funded researchers maintain an active ORCID record with CRediT-mapped contribution history, and is working with the CRediT Standing Committee at NISO to propose two extensions to the taxonomy that would cover the four CIHR-specific categories cleanly. A peer U15 university has expressed interest in adopting the same architecture.

Q&A with the composite project lead

Composite project-lead Q&A

The questions and answers below are composite — synthesised from interview patterns across multiple real project leads. They are not attributed to a specific real person.

How did you handle the four CIHR categories that did not map to CRediT?

We kept them in the schema as CIHR-specific extensions and documented the mismatch publicly. CIHR's "knowledge mobilisation" category, for example, partially overlaps with CRediT's Writing — review and editing and partially with Visualisation, but neither role captures the policy-impact dimension cleanly. We chose to preserve the CIHR concept rather than force-fit it; the alternative would have produced data that looked clean but lied about what the PI had actually done.

Why DSpace-CRIS rather than a commercial CRIS?

It was the existing institutional choice; we did not have a remit to replace it. The DSpace-CRIS contribution data model needed refactoring to align with the CASRAI Dictionary, but that was a contained engineering project. A commercial CRIS migration would have been a multi-year procurement and we would not have shipped the Tri-Agency work for three years.

How did the SSHRC PIs react to seeing CRediT prompts inside a narrative form?

Better than we expected. The original brief specifically protected SSHRC narrative tradition, so we built the exporter so that PIs could ignore the structured prompts entirely if they wanted to. In the pilot, only one of the eight SSHRC PIs took that option, and even that one switched back to using the prompts after seeing how they preserved her time when she came back to amend the report two months later.

What does the year-2 DMP integration look like?

The CASRAI Dictionary terms for data-management planning (data stewardship, data lifecycle, FAIR principles assessment, data-protection impact assessment) sit alongside the contribution data in the source of truth. When a PI writes a DMP for a CIHR grant, the workflow auto-populates the Tri-Agency DMP template; when the same PI later submits a progress report, the DMP commitments are surfaced in the report context.

Cited CASRAI resources

Internal CASRAI resources referenced

• ›Dictionary: Data Management Plan
• ›Dictionary: FAIR principles
• ›Dictionary: CRIS
• ›CRediT role: Writing — review and editing
• ›CRediT role: Visualization
• ›CRediT role: Funding acquisition
• ›CIHR CRediT statement guidance
• ›NSERC CRediT statement guidance
• ›SSHRC CRediT statement guidance
• ›Implement: DSpace-CRIS
• ›CRIS integration notes
• ›Implementation checklist: research data