A review organises a field rather than listing it

The previous unit answered that a protocol makes the search, the screening and the appraisal auditable. This one asks what makes a review last. This unit argues that the reviews that last carry an organising idea a later reader can test. That makes a review a theoretical contribution in its own right. The typologies of review types show that choosing a genre is choosing the kind of claim the review can make. A review can be method-transparent and still be a list: a protocol earns trust in the counts and does not supply the idea. The reader is part of that test: a claim carried only in a diagram cannot be checked by a reader who uses a screen reader or a reader in later life who needs larger text, since the map is not restated in the text.

The typologies: a genre is a promise about what can be claimed

Grant and Booth examined fourteen review types against a four-part framework they call SALSA, for Search, AppraisaL, Synthesis and Analysis. Their conclusion is this unit's starting point: few review types possess prescribed and explicit methodologies, and many fall short of being mutually exclusive (Grant and Booth, 2009).

Pare, Trudel, Jaana and Kitsiou built a typology of nine review types on seven core dimensions and applied it to 139 reviews in top information systems journals. The majority were theoretical, then narrative, and the sample included work labelled hybrid, a category outside the nine types (Pare et al., 2015). Templier and Pare distinguish four categories of review and key their guidelines to the type, so the criteria change with what the review is trying to do (Templier and Pare, 2015).

What a review contributes, and what is not a contribution

Webster and Watson state the claim most directly. An effective review, they write, creates a firm foundation for advancing knowledge and facilitates theory development. They place a field's theories on a hierarchy, from classifications and taxonomies to frameworks with propositions (Webster and Watson, 2002).

Whetten supplies the test that a list fails. His two questions for a conceptual paper are what is new and so what, the second asking whether the work is likely to change research practice. Relationships, not lists, are the domain of theory (Whetten, 1989). Sutton and Staw add the negative case: references, data, variables, diagrams and hypotheses are routinely offered in place of theory. Theory is the explanation of why a relation holds (Sutton and Staw, 1995). A review that arranges variables into a diagram has assembled the raw material of an explanation, not the explanation.

An organising idea is a typed claim, and its type fixes the judgement

Gregor's taxonomy answers the objection that an organising idea is too vague to be used. She distinguishes five interrelated types of theory, for analysing, for explaining, for predicting, for explaining and predicting, and for design and action. Disagreement about what counts as theory often turns on which type the disputants have in mind (Gregor, 2006). A review should say which type of claim it makes, because a map and a predictive framework are judged differently.

Shepherd and Suddaby show that such claims can be built. Their systematic review of the theory-building literature organises the tools around the five elements of a good story, conflict, character, setting, sequence, and plot and arc. They also propose pragmatic empirical theorising, which uses quantitative findings to stimulate theory rather than to close it (Shepherd and Suddaby, 2017).

Some questions need a narrative review, and it may end in no conclusion

Baumeister and Leary defend the position-taking review on grounds a protocol cannot match. Narrative reviews can tackle broader and more abstract questions than empirical reports, and theorise after reading without capitalising on chance. They can also reach any of four conclusions. The hypothesis is correct, or it is not conclusively established but is the currently best guess. It is false, or the evidence permits no conclusion (Baumeister and Leary, 1997).

Snyder adds the matching rule and a warning. Three broad approaches, systematic, semi-systematic and integrative, answer different questions, and the question should choose among them. An integrative review exists so that a new framework can emerge. Her caution matters here: reviews labelled integrative are frequently summaries and not truly integrative (Snyder, 2019).

Turning a corpus into a position, and the test a reader applies

Wolfswinkel, Furtmueller and Wilderom supply a procedure for that demand. They adapt grounded theory to the literature review as a five-stage process, from the protocol through the coding of the studies to a conceptual model. They argue that the method produces theory from a corpus rather than a summary of it (Wolfswinkel et al., 2013).

Boote and Beile supply the instrument for judging the result. Their twelve-item rubric groups into five categories, coverage, synthesis, methodology, significance and rhetoric. They applied it to twelve dissertations, four drawn from each of three colleges of education. Mean review quality was 2.09 on a three-point scale, and scores for individual items ran as low as 1.08. The authors record low average scores on criteria such as synthesising the literature and gaining a new perspective (Boote and Beile, 2005).

An HCI case: a field's theory as a live problem

Rogers's HCI Theory: Classical, Modern and Contemporary charts the theoretical development of human-computer interaction in three waves. It sets the rhetoric against the reality: how each theory was conceptualised, what was promised for it, how it was used, and which made the most impact (Rogers, 2012). It is the HCI instance of the genre this unit argues for, and its three-wave account is the organising idea the unit asks a review to provide. The field's theoretical history is a claim to be tested, not an inventory to be recited.

Contested question: does the protocol or the idea make a review last?

The literature agrees that rigour matters and divides over what rigour can supply. The method camp, Grant and Booth (2009), Pare and colleagues (2015) and Templier and Pare (2015), holds that reviews differ by type and should be judged against the standard of their type. It locates the field's weakness in a shortage of type-specific method. Brereton and colleagues applied the medical systematic review process to software engineering and recorded what it cost. They found the basic process appropriate and protocol validation especially valuable. The infrastructure worked against it. The indexing databases were inadequate for a reliable search, and abstracts were too poor to judge relevance from.

The idea camp, Webster and Watson (2002), Whetten (1989) and Sutton and Staw (1995), holds that a review earns its place by the explanation it adds. On their criteria, a method-transparent review of references, variables and a diagram passes neither of Whetten's two questions.

Boote and Beile's audit is claimed by both sides, which makes it the sharpest evidence in the dispute. Its synthesis criteria scored low. The method camp reads that as a standard not yet written down and taught. The idea camp reads it as proof that the organising idea is the part no standard can supply.

Three pieces of evidence would settle it. The first is a reliability study of synthetic judgement: whether independent reviewers agree when they decide that a review gained a new perspective. A criterion that cannot be agreed on cannot become a standard. The second is a follow-up study of reviews that state their organising idea as propositions in Whetten's or Gregor's terms. It would test whether later work cites the framework as a claim to test or only as a source of references. The third is a comparison of matched systematic and integrative reviews of one literature, asking which changed a decision.

The strongest counter-argument to this unit's thesis is the second of those studies run the other way. Boote and Beile found synthesis and method failing together, so their audit cannot tell a reader which constraint binds. If the method bar is the one that can be checked reliably, then the protocol of the previous unit is logically prior. A demand for an organising idea may then only produce confident reviews with no better grounding. The reply is that a review must pass both, which is how Boote and Beile treat the categories. The very thing this unit asks for describes the best reviews in their sample, the critical examinations that set the stage.

The evidence is unevenly distributed. The method camp rests on three typologies of review types and on one reported attempt to transfer the medical process into a computing field. The idea camp rests on editorial arguments and on one conceptual paper, which state the test rather than apply it. The only measurement either camp owns is a rubric applied to twelve dissertations, and both camps read it their own way. While the dispute is open, a designer should ask what organising claim a review makes and whether that claim can be tested, rather than counting its references.

Where the argument is heading

The direction is towards reviews that declare their genre and state their idea. Templier and Pare key guidelines to the type. Snyder matches the approach to the question and warns that integrative labels are often unearned. For HCI the next step is a review of a frontier area that does both. It reports its protocol, and it states its organising idea as a typed and testable claim, in Gregor's terms of theory type and against Whetten's two questions. Shepherd and Suddaby supply the tools for building that claim, Boote and Beile the rubric for judging it, and Rogers the case that the effort is worth making.

The sources this unit rests on

Each note states what the cited document did and found. Every DOI was resolved and its record read on 1 November 2026.

• Grant, M. J. and Booth, A. (2009). A typology of reviews: an analysis of 14 review types and associated methodologies. Health Information and Libraries Journal. doi:10.1111/j.1471-1842.2009.00848.x Analysed fourteen review types against SALSA, for Search, AppraisaL, Synthesis and Analysis. It found that few types carry a prescribed and explicit methodology and that many are not mutually exclusive.
• Pare, G., Trudel, M. C., Jaana, M. and Kitsiou, S. (2015). Synthesizing information systems knowledge: a typology of literature reviews. Information and Management. doi:10.1016/j.im.2014.08.008 Built a typology of nine review types on seven core dimensions and applied it to 139 reviews in top information systems journals, where the majority were theoretical, then narrative, and the sample included work labelled hybrid, a category outside the nine types.
• Templier, M. and Pare, G. (2015). A framework for guiding and evaluating literature reviews. Communications of the Association for Information Systems. doi:10.17705/1cais.03706 Distinguish four categories of review and key their guidelines to the type. Their point is that the criteria are not universal: what counts as a good review depends on the category chosen.
• Webster, J. and Watson, R. T. (2002). Analysing the past to prepare for the future: writing a literature review. MIS Quarterly. doi:10.2307/4132319 Argue that an effective review creates a firm foundation for advancing knowledge and facilitates theory development, and place a field's theories on a hierarchy from classification systems, through taxonomies, to conceptual frameworks with propositions.
• Whetten, D. A. (1989). What constitutes a theoretical contribution? The Academy of Management Review. doi:10.2307/258554 Sets out the two questions an editor asks of a conceptual paper, what is new and so what, and states that relationships, not lists, are the domain of theory.
• Sutton, R. I. and Staw, B. M. (1995). What theory is not. Administrative Science Quarterly. doi:10.2307/2393788 Argue that references, data, variables, diagrams and hypotheses are offered in place of theory, and that theory is the explanation of why a relation holds. It is the test a review can be run against.
• Gregor, S. (2006). The nature of theory in information systems. MIS Quarterly. doi:10.2307/25148742 Proposes a taxonomy of five theory types, for analysing, explaining, predicting, explaining and predicting, and design and action, and shows that disagreement about theory often turns on which type the disputants expect.
• Shepherd, D. A. and Suddaby, R. (2017). Theory building: a review and extension. Journal of Management. doi:10.1177/0149206316647102 Review the theory-building literature and organise the tools around the five elements of a good story, conflict, character, setting, sequence, and plot and arc. They add pragmatic empirical theorising, which uses findings to stimulate theory.
• Baumeister, R. F. and Leary, M. R. (1997). Writing narrative literature reviews. Review of General Psychology. doi:10.1037/1089-2680.1.3.311 Argue that narrative reviews can tackle broader questions than empirical reports, theorise after reading without capitalising on chance, and reach any of four conclusions, including that the evidence permits no conclusion.
• Snyder, H. (2019). Literature review as a research methodology: an overview and guidelines. Journal of Business Research. doi:10.1016/j.jbusres.2019.07.039 Distinguishes systematic, semi-systematic and integrative approaches and argues the question and purpose choose among them; an integrative review exists so that a new framework can emerge. She warns that integrative labels are often unearned.
• Wolfswinkel, J. F., Furtmueller, E. and Wilderom, C. P. M. (2013). Using grounded theory as a method for rigorously reviewing literature. European Journal of Information Systems. doi:10.1057/ejis.2011.51 Adapt grounded theory to the review as a five-stage process, from protocol through coding to a conceptual model, and argue it produces theory from a corpus rather than a summary of it.
• Boote, D. N. and Beile, P. (2005). Scholars before researchers: on the centrality of the dissertation literature review in research preparation. Educational Researcher. doi:10.3102/0013189x034006003 Develop a twelve-item rubric in five categories and apply it to twelve dissertations drawn from three colleges. Mean quality was 2.09 on a three-point scale, with individual items down to 1.08 and low average scores on the synthesis criteria.
• Rogers, Y. (2012). HCI theory: classical, modern, and contemporary. Synthesis Lectures on Human-Centred Informatics. doi:10.2200/S00418ED1V01Y201205HCI014 Charts the theoretical development of human-computer interaction in three waves and sets the rhetoric against the reality: what each theory promised, how it was used and which made the most impact.
• Brereton, P., Kitchenham, B. A., Budgen, D., et al. (2007). Lessons from applying the systematic literature review process within the software engineering domain. Journal of Systems and Software. doi:10.1016/j.jss.2006.07.009 Applied the medical systematic review process to software engineering and found it basically appropriate, with protocol validation especially valuable, but the indexing databases inadequate and abstracts too poor to judge relevance from.