LN: Hevner et al. (2004) — Design Science in Information Systems Research

Bibliographic Reference

Citation: Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems research. MIS Quarterly, 28(1), 75–105. https://doi.org/10.2307/25148625


Pass 1 — Bird’s Eye View (5 Cs)

CAssessment
CategoryResearch methodology framework paper
ContextPublished in MIS Quarterly — the premier IS research journal. Establishes design science as a rigorous academic research paradigm distinct from (but complementary to) behavioural science
CorrectnessHighly cited (20,000+ citations); based on foundational work by Simon (1996) The Sciences of the Artificial
Contributions(1) Seven DSR guidelines; (2) Three-cycle model (Relevance, Design, Rigour); (3) Artefact taxonomy (Constructs, Models, Methods, Instantiations); (4) Evaluation criteria for design artefacts
ClarityExcellent — precise, canonical definition of what counts as DSR contribution

Relevance: ⭐⭐⭐⭐⭐

PUMA’s research paradigm. The PUMA benchmark framework is a DSR Instantiation artefact — a working system built to demonstrate feasibility and evaluate utility. Hevner et al. provides the methodological justification for this in PUMA Ch.3.


Pass 2 — Key Concepts

The Seven DSR Guidelines

#GuidelinePUMA Application
1Design as an artefact: DSR must produce a viable artefactPUMA benchmark = Instantiation
2Problem relevance: Research addresses important business problemsManual triage cost at scale
3Design evaluation: Artefact utility, quality, and efficacy must be rigorously evaluatedWilcoxon, F1-macro, MAE
4Research contributions: Must contribute to foundations of IS designLLM benchmark methodology
5Research rigour: Apply rigorous methods in construction and evaluationDSR + SLR + statistical validation
6Design as a search process: Explore the space of possible artefacts4 models × 4 strategies = 16 conditions
7Communication: Present results to both research and practice audiencesAcademic paper + GitHub repository

The Three Research Cycles

Environment ──→ Relevance Cycle ──→ Design Science Research
Knowledge Base ─→ Rigour Cycle ──→ Design Science Research
                                       ↓
                               Design Cycle (Build, Evaluate, Feedback)
  • Relevance Cycle: Grounds research in real business problems (PM inefficiency, manual triage cost)
  • Design Cycle: Iterative build-evaluate loop for the artefact (PUMA benchmark iterations)
  • Rigour Cycle: Connects to existing knowledge base (SLR of 42+ papers)

Artefact Types

TypeDescriptionPUMA Example
ConstructsVocabulary and symbols”triage accuracy”, “SPR”, “F1-macro”
ModelsAbstractions representing realityPUMA 5-stage pipeline model
MethodsAlgorithms and practicesKeshav Three-Pass, PRISMA-trAIce
InstantiationsWorking implementationsPUMA benchmark platform (Docker + Ollama + Python)

PUMA is primarily an Instantiation — a concrete, working system that demonstrates that LLM agents can perform PM tasks reproducibly.


PUMA Integration

  • PUMA Ch.3 Methods: DSR as the overarching research paradigm; cite Hevner et al. (2004) for methodological justification
  • Artefact claim: PUMA contributes a novel Instantiation artefact — a reproducible benchmark — satisfying Guidelines 1, 3, and 4
  • Evaluation design: The statistical validation pipeline (Wilcoxon, effect size) satisfies Guideline 3

MOCs