Systematic Literature Review. Exclusion and Inclusion Criteria (S4.1)

Systematic literature reviews rise or fall on study selection: once the search is done, researchers must screen thousands of hits down to a defensible set of studies using clear inclusion and exclusion criteria. Because it’s impossible to read hundreds or thousands of papers in full, the process starts with title and abstract screening to remove clearly irrelevant records, then moves to full-text review to confirm whether each potentially eligible study truly meets the predefined criteria. A key point is that exclusion decisions should be based on the study’s characteristics—such as whether it contains the required type of evidence—rather than on convenience or preference.

Documentation is treated as non-negotiable. Every inclusion or exclusion decision needs to be recorded, often in a spreadsheet or table, and reported transparently using a PRISMA-style flow diagram (Preferred Reporting Items for Systematic Reviews and Meta-Analysis). The goal is to make the selection trail auditable: how many records were found, how many were screened, how many were excluded at each stage, and how many ultimately entered the analysis. Screening ideally involves two independent reviewers to reduce bias and improve reliability; when that isn’t feasible, the work can be done by one person but later checked by independent reviewers or experts. Disagreements are resolved through discussion, typically by involving a third reviewer.

The criteria themselves must be tightly aligned with the review’s research objectives. Inclusion criteria define what qualifies a study—examples given include focusing on a specific population, using a particular methodology (such as survey data collected via questionnaires), restricting the time window (e.g., 2001–2025), limiting language (e.g., English), and specifying publication type (e.g., journal articles). Exclusion criteria define what disqualifies studies, such as excluding non-peer-reviewed work (conference papers, viewpoints, editorials), studies lacking primary data, studies relying only on secondary data, or studies with an irrelevant geographic focus. The transcript also highlights a conceptual alignment test: if the review targets outcomes of servant leadership, studies that only examine antecedents (without servant leadership outcomes) should be excluded.

Before full screening, criteria can be pilot-tested on a small subset (around 50 studies) to check clarity and practicality, then refined if the rules prove confusing or too restrictive. All criteria should be justified and explicitly documented.

After selecting studies, the next step is assessing methodological quality and risk of bias. Standard tools are recommended based on study design—such as CASP (Critical Appraisal Skills Programme), tools for systematic reviews like MSAR, and Joanna Briggs Institute instruments. Risk of bias is evaluated across domains including selection bias, performance bias, detection bias, and reporting bias. Some tools produce ratings (high/medium/low), and studies with high risk of bias may be excluded from synthesis or included with caution, with the concern clearly flagged in the analysis.

Concrete examples illustrate how published reviews report screening outcomes. One PRISMA-based ESG review reports a final set of 85 included articles and notes that reference lists were checked for additional eligible studies. Another example describes database searches (including ProQuest, Springer, SAGE, and others), the number of non-duplicate citations screened, exclusions at title/keyword and abstract stages, and the final count included after abstract review. A further example shows that some reviews describe inclusion/exclusion criteria without a diagram, though a flowchart is preferred for reader clarity on what was removed and why.