Opening Remarks
Review Type Fit
“Systematic literature review” is a category containing several different methods, each suited to a different kind of question.
You have a research question and a candidate theory type. The next decision, which kind of review you are about to conduct, is usually made by default: most teams reach for “systematic review” because it is the term they have heard most often, not because they checked whether it is the review type their question actually needs. This unit exists to make that decision deliberately, once, before you spend eleven weeks executing it.
Learning Outcomes
By the end of this unit, you will be able to:
- Select a review type from an eight-family typology based on the knowledge goal your research question actually needs, not on which term sounds most rigorous.
- Align your question type, target theory type, and review type into one consistent decision, and justify the alignment in writing.
- Argue for a literature review as a standalone knowledge contribution, not a preamble to a “real” study.
Review Contributions
Standalone Contribution
A literature review is a knowledge contribution in its own right.
Recall the three routes to contribution from Knowledge, Science and Epistemology: improving explanation, improving evidence, improving method. A review rarely adds new evidence directly. It can improve explanation, by synthesizing what fragmented individual studies only implied separately, and it can improve method, by exposing where a field’s existing measurement or theorizing approaches disagree or fall short.
Rowe (2014) makes the point sharply: a literature review is too often treated as a mandatory, slightly tedious preamble that clears the way for the “real” study. Treated that way, it produces exactly the kind of review nobody wants to read: a list of what has been published rather than an argument about what the field now knows. Your review is the deliverable. It has to earn its place as a contribution the same way a primary study would.
Review Typology
Eight Families, Two Questions
What kind of knowledge goal does the review pursue, and how comprehensively does it sample the literature to pursue it?
Paré et al. (2015) organize the landscape of literature reviews into a typology built around the goal a review pursues, not the topic it covers. Broadly:
- Narrative review: provides a broad, qualitative overview of a topic, often to orient readers new to a field. Comprehensiveness and a documented search protocol are usually secondary to readable synthesis.
- Descriptive (or mapping) review: assesses the extent, range, and nature of research activity in a domain, often quantitatively, how many studies, using which methods, across which years, without necessarily synthesizing their findings into a single argument.
- Scoping review: maps the key concepts and evidence in a domain, often as groundwork for a later, more targeted review, useful when a field is too young or too heterogeneous for narrower methods to work yet.
- Critical review: evaluates the quality of existing research, aiming to expose conceptual or methodological weaknesses in a body of work, not just summarize it.
- Realist review: seeks to explain, for complex, context-dependent interventions, what works, for whom, and under what circumstances, rather than asking a single yes/no effectiveness question.
- Qualitative systematic review: applies systematic, documented search and appraisal methods, the hallmark of “systematic” reviews, to synthesize qualitative or conceptual findings, not to pool numerical effect sizes.
- Meta-analysis: statistically pools quantitative results across comparable studies to estimate an aggregate effect. Requires a body of comparable, quantitative primary studies to work at all.
- Umbrella review: synthesizes existing reviews rather than primary studies, useful when a field already has multiple reviews that themselves need reconciling.
Core Message
Review type follows from the knowledge goal the review is meant to serve.
The single most common mistake at this stage is picking “systematic review” because it sounds rigorous, then discovering three weeks into search and screening that the actual goal, understanding a young, fast-moving, conceptually unsettled phenomenon, was better served by a scoping review or a critical review. Ask first: what does a good answer to my research question actually look like? An aggregate effect size? A map of what has been studied and what hasn’t? An account of what works under which conditions? Only then pick the review type that produces that shape of answer.
IS Review Genre
Webster & Watson (2002) set the standard for what a rigorous literature review looks like in information systems specifically.
Two of their contributions matter directly for your decision today, and you will meet both again in later units: organizing a review’s presentation around concepts, not the sequence of studies that produced them (covered fully in SLR IV: Analysis and Synthesis), and using backward and forward search as complementary strategies for building a search’s coverage (covered fully in SLR II: Search Strategy and Protocol). Both are genre conventions specific to how IS, as a field, expects a review to be conducted and reported, adopting them is part of writing for your target audience, not an arbitrary add-on.
Selection Mistakes
The eight families in Paré et al. (2015) are easy to pick for the wrong reason.
- Defaulting to a qualitative systematic review or meta-analysis because the terms sound most rigorous, before checking whether enough comparable studies actually exist.
- Running a meta-analysis on a corpus that turns out too heterogeneous in measures and designs to pool into one effect.
- Treating scoping and narrative reviews as a lesser, less careful option, when a young or fast-moving field is exactly the case they are built for.
- Picking a review type before the research question is settled, then reshaping the question to fit the method already chosen.
Each mistake here lets a review type get picked by convention or reputation before the knowledge goal has actually been checked against it. A corpus with too few comparable quantitative studies cannot support a meta-analysis, regardless of how rigorous the label sounds. A young, conceptually unsettled field is usually served better by a scoping or narrative review than by a systematic review promising a comprehensiveness the field cannot yet deliver. Catching these mistakes before search begins is cheaper than discovering them after three weeks of screening.
Alignment
Three Decisions, One Matrix
Research question type × target theory type × review type. They have to line up.
You have already made two of the three decisions this alignment exercise checks:
- Question type, from Good Research Questions: a “what/who/where” question, or a “how/why” question?
- Theory type, from What is Theory: which of Gregor’s five types does your question aim at?
The third decision is today’s: which review type actually produces an answer of the shape your question and theory type demand? A “what/who/where” question aiming at an analysis-type theory is usually well served by a scoping or descriptive review: you are mapping a landscape. A “how/why” question aiming at explanation-and-prediction usually needs a qualitative systematic review or, if enough comparable quantitative studies exist, a meta-analysis: you need to synthesize an explanation, which goes beyond mapping what exists. A question aiming at design-and-action needs a review that surfaces design principles across studies, which rarely fits a meta-analysis at all.
Question Framing
A Vague Question
“How does AI affect the workplace” is not yet a question a review can search on.
A sentence can sound like a research question and still be too fuzzy to search on: which AI systems, affecting whom, compared to what, in which setting? PICOC decomposition pulls those parts apart before search begins, so the search string, the inclusion criteria, and the eventual write-up all answer the same specific question rather than three slightly different ones.
PICOC
PICOC breaks a fuzzy question into five parts a search can act on.
PICOC, Population, Intervention, Comparison, Outcome, Context, originated in evidence-based medicine as a way to structure a clinical question before searching the literature for an answer. Petticrew & Roberts (2006) adapt the same decomposition for social-science and management reviews, and IS reviews use it the same way, to sharpen a review’s guiding question before search begins.
Five Elements
- Population: the unit or phenomenon the review is about, employees, development teams, organizations, individual users.
- Intervention: the technology, practice, or condition whose effect or role the review examines.
- Comparison: the baseline or alternative condition the intervention is read against; not every review needs one.
- Outcome: the effect, behavior, or state the review is trying to explain, predict, or map.
- Context: the setting that bounds the review, an industry, a national context, a time period, an organizational size class.
Not every review fills in all five elements. A scoping review, for instance, may carry no comparison at all, because it maps a landscape rather than testing an effect. The decomposition earns its value by forcing the question into specific parts, whether or not every slot ends up filled.
Worked Example: Algorithmic Management
Population
Gig-economy platform workers.
Intervention
Algorithmic task assignment and performance monitoring.
Comparison
Human-supervised task assignment.
Outcome
Worker autonomy and job satisfaction.
Context
Ride-hailing and delivery platforms, since 2020.
The fuzzy version, “how does algorithmic management affect workers,” could support almost any search string. The PICOC version tells a search which terms to combine, and tells a reader, before they open the review, what kind of answer to expect.
Homework
Write your review type decision.
Argue your review type, the knowledge goal it is meant to serve, and how it aligns with your question type and target theory type.