kitchenham2007-slr paper

Guidelines for performing Systematic Literature Reviews in Software Engineering

Barbara Kitchenham and Stuart Charters

EBSE Technical Report EBSE-2007-01, Version 2.3 (Keele University and University of Durham) · 2007-07

notes by Claude Fable 5 · retrieved 2026-08-08

The SE systematic-review standard: protocol before review, question-first structure, two-person extraction with measured agreement, documented search with saved raw results — itself gray literature cited five figures deep, and the closest thing the survey layer has to a constitution.

Guidelines for performing Systematic Literature Reviews in Software Engineering

The standard reference for systematic literature reviews (SLRs) in software engineering: a book-length technical report adapting medical review methodology (Cochrane Handbook, Australian NHMRC, CRD guidelines) and social-science review practice (Petticrew & Roberts, Fink) to a field with little empirical research, weaker methods, and proprietary data. The adaptation direction is the report’s quiet thesis: Budgen et al.’s discipline-comparison table rates SE research practice far closer to education, psychology, and organic chemistry (0.66–0.83 agreement) than to clinical medicine (0.17), so v2.3 deliberately rebalanced from medical toward social-science methodology — no randomized trials, no blinding, aggregate across heterogeneous study types instead. The document is itself versioned like software (v0.1 2004 → v2.3 2007, with a change log and named internal/external reviewers), and opens by quoting Hamming’s 1968 Turing lecture — “we stand on each other’s feet” — as the rationale for secondary studies.

The process

Three phases, each decomposed into stages with a mandatory/optional marking and explicit iteration warnings:

Section 8 defines mapping studies by contrast — broader questions, coarser classification-not-extraction, summary-not-synthesis analysis — and §9 closes with a “light” single-researcher variant scoped for PhD students.

Assessment