Organizing Sources for a Literature Review
Organize each literature review source around one traceable record: keep its citation metadata, retrievable file or access path, relevant study characteristics, finding notes with provenance, and consistent codes or tags connected. Use shared fields and a comparison matrix to compare equivalent information, mark missing values explicitly, and move into synthesis only when the source set is retrievable, comparable, and verifiable.

Source organization workflow at a glance
Treat each source as one connected record. Each stage adds a specific value needed for reliable retrieval, comparison, or later verification.
- Identify the sourceStable source identifier + citation metadataOutcome: one distinguishable, retrievable source record.
- Capture relevant evidenceStudy attributes + finding notes + page or location provenanceOutcome: notes can be checked against the source.
- Standardize retrieval labelsConsistent codes, controlled values where useful, and stable concept tagsOutcome: related sources can be grouped without changing their underlying information.
- Align comparable fieldsShared attributes in a comparison matrixOutcome: similarities, differences, and missing information become visible across sources.
- Verify synthesis readinessRecords complete enough for the review + reconciled labels + explicit missingness + source-to-note traceabilityOutcome: the collection can be used for analysis without reconstructing routine context from memory.
The exact fields and level of detail should follow the review question, method, and source types rather than a universal template. Record enough structure for later retrieval and comparison without turning record keeping into synthesis.
Table of Contents
Create a Consistent Source Record and Note-Taking System
Apply the same source record and note-taking system to every included source so its identity, notes, and retrieval path stay connected.
A consistent format reduces lost context and inconsistent capture by keeping information attached to the correct source record.
The exact fields and note length can vary with the review method and source type.
The capture sequence standardizes how each source moves from identification to retrievable notes without extending into coding or synthesis.
Keep each source record tied to one source identifier so citation metadata, study details, findings, and annotations retain clear provenance.
This organization can begin as sources are collected while you search for relevant literature.
Apply the sequence consistently to each included source.
- Establish the source identity. Assign a stable source identifier and keep it associated with the source's citation metadata.
- Create the source record. Record stable fields for the bibliographic details and study details needed by the review, without adding fields that serve no clear purpose.
- Capture notes with provenance. Record factual notes and relevant findings in a consistent format, keeping each annotation connected to the source record and its location or context where available.
- Keep the source retrievable. Save or attach the source file through a consistent retrieval path and label it so the source identifier, record, and file can be matched.
- Repeat and flag gaps. Apply the same capture routine to the next source and mark missing information explicitly rather than inferring or silently filling it.
An incomplete source record should remain visibly incomplete until the missing information can be checked against the source.
Consistent capture keeps source identity, notes, and retrieval information connected while leaving later coding and comparison as separate tasks.
Capture Citation, Study, and Finding Details Consistently
Each source record needs enough citation details, study characteristics, finding information, and provenance to identify the source and compare relevant information later.
Bibliographic metadata establishes source identity, while study-specific fields describe what the source examined and reported.
The exact study fields should follow the review question rather than a fixed universal template.
Separate fields needed for source identity and provenance from study-specific fields that apply only when they matter to the review.
Using the same field labels across source records makes it clear where equivalent information belongs and reduces ambiguity when returning to notes.
| Information class | Fields to capture | Why it matters |
|---|---|---|
| Citation details | Author, year, title, publication details, and identifier where available | Provides bibliographic metadata that identifies the source and supports retrieval. |
| Study characteristics | Research purpose, method, population or sample where relevant, and setting or context where relevant | Describes how and where the research was conducted so relevant study attributes can be compared. |
| Finding details | Main findings relevant to the review question | Records what the source reports without turning the captured finding into an evaluative judgment. |
| Provenance | Page or location reference for important notes | Locates the supporting passage or section so a note can be checked against its source. |
Record a missing applicable value as not reported, and use not applicable when a field does not fit the source or review context; do not infer a value merely to complete the record.
These fields capture descriptive information rather than evidence quality, which belongs with the criteria used to evaluate the sources.
Accurate capture keeps source identity, study context, main findings, and note provenance distinct without turning factual extraction into appraisal.
Code and Tag Sources for Retrieval and Grouping
Codes and tags make source records easier to retrieve and group when each label is stable, interpretable, and meaningful across the relevant sources.
A code standardizes a repeatable attribute, while a tag can expose a recurring concept that is useful for grouping related material.
Consistency allows the same labeling scheme to support retrieval without changing the underlying source information.
Controlled study-characteristic codes should use a defined naming convention and, where appropriate, a controlled value set so equivalent attributes are recorded consistently.
Conceptual tags may be more flexible because they label recurring ideas rather than fixed factual attributes, but their wording should still remain interpretable and reusable.
The following criteria keep labels searchable and comparable without creating a code for every minor detail.
- Clear purpose: create a code or tag only when the label supports a defined retrieval or grouping need.
- Stable wording: use the same naming convention for the same attribute or concept across source records.
- Controlled values: define permitted values where a factual study characteristic requires consistent categories, while allowing the level of detail to follow the review question and evidence set.
- Unknown or mixed cases: record an unknown value explicitly and define how a mixed case is represented rather than forcing it into an inaccurate category.
- Reconciled revisions: when a label or controlled value changes, revise the scheme deliberately and reconcile affected existing records so the old and new wording do not fragment retrieval.
A revised label should therefore be applied consistently to earlier records that use the same concept or attribute, rather than changing terminology only for newly added sources.
Unknown and mixed case values should remain distinguishable from confirmed categories, and any revision should retain the intended meaning of the original source information.
This reconciliation keeps retrieval and grouping consistent as the labeling scheme develops.
Apply Consistent Codes to Study Characteristics
Study-characteristic codes should use the same interpretable value system for equivalent factual attributes across sources while preserving legitimate mixed or multiple values.
A controlled code can standardize a study characteristic such as study design, method, population or sample, setting, or publication period without erasing differences that matter to the review.
Consistent values make those attributes usable for retrieval and comparison without treating them as judgments of study quality.
The appropriate controlled values depend on the review question and evidence set rather than one universal taxonomy.
The table shows how factual attributes can retain interpretable values while supporting practical retrieval.
| Study characteristic | Controlled value examples | Retrieval implication |
|---|---|---|
| Study design | Experimental; observational; qualitative; mixed method | Filters sources by the broad design used to investigate the research question. |
| Method | Interview; survey; observation; document analysis; multiple values | Retrieves sources using the same method or an explicitly recorded combination of methods. |
| Population or sample | One defined population; multiple populations; not reported | Groups studies by the population or sample relevant to the review while retaining multi-population cases. |
| Setting | One defined setting; multiple settings; not reported | Supports filtering by study context without assigning an unsupported setting. |
| Publication period | Publication year or a review-defined period | Allows retrieval and comparison by the time boundary chosen for the review. |
When a study uses a mixed method or covers more than one population, preserve an explicit mixed state or multiple values if the coding scheme permits them instead of assigning one inaccurate category.
If a study characteristic is ambiguous, document the ambiguity or the rule used to assign its controlled code rather than guessing.
These codes describe source attributes for retrieval and comparison; they do not rank a study or convert a descriptive characteristic into a quality judgment.
Use Theme and Concept Tags to Connect Related Studies
Theme and concept tags are provisional organizing labels that connect source records around a recurring idea, construct, finding, or debate.
A theme tag or concept tag can group related studies even when their factual study characteristics differ.
These labels expose conceptual relationships across sources without establishing the review's eventual organization.
Name each tag consistently and use synonym normalization when different terms refer to the same concept, so equivalent ideas remain retrievable under a stable label.
Assign multiple tags to a source when it genuinely addresses more than one relevant concept rather than reducing the record to a single interpretation.
For example, the following tags can expose different cross-source relationships without treating the labels as settled conclusions:
- Access barriers: can connect studies discussing obstacles to participation, making a recurring constraint easier to inspect across source records.
- Participant experience: can connect studies reporting relevant experiences or perceptions, allowing related findings to be considered together.
- Implementation debate: can connect studies that examine or contest how an approach is applied, making a shared debate visible across otherwise separate records.
As understanding develops, revision may split an overly broad tag, merge overlapping labels, or clarify a tag name while maintaining consistency across affected records.
Unlike a factual study-characteristic code, an interpretive tag identifies a conceptual connection rather than a descriptive attribute of the study.
Tags can reveal useful groupings, but they do not themselves choose thematic or chronological organization.
Build a Comparison Matrix From Your Source Records
A comparison matrix aligns the same relevant fields across source records so similarities, differences, and missing information can be inspected across studies.
It is an organizing and comparison aid rather than a substitute for analytical reading.
Each entry should remain traceable to the source record from which it was taken.
Choose only comparable fields that serve the review question, such as method, population or context, core finding, theme or concept, and a documented qualification.
Keep each cell concise enough to scan while preserving enough detail to avoid misleading one-word summaries.
The following matrix uses hypothetical entries only to illustrate how selected fields can be aligned; the actual columns and values should come from the review's source records.
| Source | Method | Population or context | Core finding | Theme or concept | Qualification |
|---|---|---|---|---|---|
| Illustrative Study A | Interview | Community participants | Reported barriers affected participation | Access barriers | Context-specific finding |
| Illustrative Study B | Survey | University setting | Participants reported differing levels of access | Access barriers | Population differs from Study A |
| Illustrative Study C | Not reported | Regional program | Implementation varied across locations | Implementation | Method missing from the source record |
The comparison matrix can expose a shared theme or concept, a difference in method or population or context, or missing information that needs checking.
Those visible relationships can guide later analysis, but the matrix does not perform synthesis or establish the final argument.
Preserve traceability by keeping each matrix entry linked to the corresponding source record and by leaving unsupported information explicitly missing rather than inferring it.
Use Shared Fields to Keep Studies Comparable
A shared field should be used across studies only when it enables a meaningful comparison tied to the review question.
The criterion is whether the field distinguishes relevant similarities or differences across the included studies, not whether the information can simply be recorded.
Essential fields support the current comparison directly, while optional fields apply only when the study type or review question makes them informative.
- Research aim: include it when differences in what studies investigate affect interpretation; it enables comparison of the purposes addressed by related studies.
- Method: include it when research approaches are relevant to the review question; it enables comparison of how evidence or observations were produced.
- Population or sample: include it when participant characteristics or sampled groups affect the comparison; it enables distinctions between findings drawn from different populations or samples.
- Setting: include it when context may distinguish otherwise related studies; it enables comparison across locations, institutions, environments, or other relevant settings.
- Concept: include a measured or discussed concept when it recurs meaningfully across sources; it enables comparison of how related studies address the same or closely aligned construct.
- Main finding: include it when the reported result bears directly on the review question; it enables comparison of agreement, difference, or qualification across source findings.
Optional fields, including relevant limitations, should remain optional when they are not consistently meaningful across the included study types.
If a source does not state information required for a shared field, record it as not reported; if the field does not apply to that source, record it as not applicable rather than inventing a standardized value.
For example, a raw note such as “participants were first-year university students” can become a standardized population or sample value such as “first-year university students” when that distinction serves the comparison.
Shared fields should therefore prioritize comparability over completeness for its own sake.
Distinguish an Evidence Table From a Synthesis Matrix
Institutions may use the terms evidence table and synthesis matrix differently, so purpose and structure matter more than terminology alone.
An evidence table is primarily source-centred, recording extraction and summary information for individual studies, while a synthesis matrix is primarily relation-centred, organizing sources across a shared theme or question, concept, or comparable dimension.
The same criteria can therefore be used to compare what each format primarily organizes and how it represents a cross-source relationship.
| Comparison criterion | Evidence table | Synthesis matrix |
|---|---|---|
| Primary purpose | Records extracted study details and concise source summaries. | Aligns information so relationships across sources can be inspected. |
| Unit of organization | Primarily source-centred, with each study treated as the main record. | Primarily relation-centred around a theme or question, concept, or shared comparison dimension. |
| Typical content | Study characteristics, methods, findings, and documented qualifications. | Comparable evidence from several sources arranged against shared concepts, questions, or dimensions. |
| Cross-source relationship | Relationships may be visible when rows are compared, but they are not usually the primary organizing logic. | Cross-source relationships are made explicit by aligning studies within the same comparative frame. |
The two formats may overlap, and a hybrid format can contain both source-centred extraction fields and relation-centred comparison fields.
For example, the same study may appear as one factual row in an evidence table and as one comparative entry under a shared concept in a synthesis matrix.
Neither format automatically performs critical synthesis; each organizes evidence in a way that can support later analysis.
Use Reference Management to Keep Sources Traceable
Reference management supports traceability by keeping citation metadata, the source file, a stable identifier, and working notes retrievable as connected parts of the same source record.
Its purpose here is functional rather than a choice between software brands.
Maintaining those connections reduces orphaned files and uncertain citations, but reference-management software does not replace source evaluation or synthesis.
A repeatable workflow should establish the citation record first, verify metadata accuracy before relying on imported details, and then maintain the connections needed for later retrieval.
The source file should remain attached to or reliably locatable from the record, while an identifier or collection provides a stable organizational reference.
An annotation or note should retain enough connection to its source to preserve traceability, and duplicate record control should prevent parallel records from fragmenting the same source information.
- Create or import the citation record. Enter or import the available citation metadata so the source has a defined record.
- Verify the metadata. Check imported author, title, publication, date, identifier, and other relevant details against the source, and reconcile a duplicate record before relying on the record.
- Connect the source file. Attach the full-text file where the tool supports file attachment, or maintain a reliable location that can be traced from the citation record.
- Organize the record consistently. Use a stable identifier, collection, or equivalent organizational convention so the source remains distinguishable and retrievable within the source set.
- Maintain retrievable notes. Keep each relevant annotation or note connected to its source and preserve it in a form that can be retrieved or included in an export where the chosen tool and export format support that function.
Portability, file handling, and export options vary by tool and configuration, so corrected metadata and source-to-note connections should not depend on an assumed feature.
For example, a source remains traceable when its verified citation record identifies the corresponding PDF and a working note can be followed back to that record, regardless of the particular reference manager used.
The practical requirement is consistent retrievability across the reference-management workflow.
Keep Citation Records, Files, and Notes Connected
A source is traceable when its citation record, source file, and derived notes can be reconnected and checked without relying on memory.
A stable identifier and retrievable file path or file attachment preserve the connection between the record and the source, while note provenance connects extracted information back to its origin.
The verification points below test whether retrieval and provenance remain intact from the citation record through the source file to the notes.
- Stable identifier: Confirm that the citation record retains an identifier that can still distinguish the source after routine organizational changes.
- Retrievable source file: Confirm that the source file can still be located through its file attachment, naming convention, or documented storage path without assuming one storage model.
- Note provenance: Confirm that each relevant note identifies the source record from which the information was derived, so a claim can be checked against its source.
- Page or location reference: Where the source format permits a stable location reference, record the relevant page reference or equivalent location so the supporting passage can be verified.
- Metadata correction: After a metadata correction, confirm that the updated citation record, stable identifier, source file, and associated notes remain synchronized and retrievable.
A renamed file, metadata correction, or detached note should trigger a traceability check rather than an assumption that the existing connections remain intact.
For example, if a PDF is renamed, verify that the citation record still locates the correct source and that its notes and page references still identify material from that source.
After a change, reconcile the affected connections so the writer can retrieve the source, verify the note, and reconstruct its provenance.
Check the Organization Before Moving Into Synthesis
The source set has synthesis readiness when its records are sufficiently complete for the review, labels are reconciled, comparison fields are usable, missingness is explicit, and traceability supports verification.
Organization readiness means the source set can be retrieved, compared, and checked without reconstructing missing context from memory.
This is a collection-level readiness check, not an evaluation of the final synthesis.
The system-level check should verify whether the organized evidence base can support later comparison without reopening routine organization work.
Each readiness condition should expose either a usable state or a specific issue that remains unresolved.
- Complete records: Verify that source records meet the minimum completeness requirements defined for the review; otherwise, later comparison may rely on incomplete context.
- Reconciled labels and codes: Verify that equivalent labels and codes have been reconciled across the source set; otherwise, related evidence may be separated by inconsistent terminology.
- Usable matrix fields: Verify that matrix fields contain comparable information where applicable; otherwise, cross-source comparison may be incomplete or misleading.
- Explicit missingness: Verify that unavailable or inapplicable information is visibly identified rather than silently left ambiguous; otherwise, a missing value may be mistaken for an overlooked entry.
- Source-to-note traceability: Verify that relevant notes can be traced back to their source records for verification; otherwise, the origin of a claim may be difficult to reconstruct.
- Unresolved items: Verify that pending issues are clearly flagged with the action still required; otherwise, an unresolved item may be treated as settled when the source set moves forward.
Correctable issues such as inconsistent labels, misplaced entries, or unflagged missingness can be reconciled within the organized source set when the necessary information is already available.
If required information is absent or a note cannot be verified from its current record, return to the original source rather than infer the missing context.
Recheck the affected readiness condition after the correction so the unresolved status is either retained or cleared.
Once these organization checks pass, the source set is sufficiently retrievable, comparable, and verifiable to close the organization task.
The next task is to synthesize the evidence.
Confirm Every Included Source Has a Complete, Traceable Record
An included source has a sufficiently complete record when it contains the minimum information needed to identify, retrieve, contextualize, and trace that source for the current review.
Completeness depends on the review's actual comparison needs rather than a universal field count.
A complete record should preserve enough citation metadata, source location, study context, and provenance for later verification without requiring every possible field.
- Citation metadata: Confirm that the record contains the identifying citation details required to distinguish the source; missing identifiers reduce traceability and can make retrieval uncertain.
- File or access location: Confirm that the source file or a reliable access location can be retrieved from the record; an unlocatable source prevents later verification.
- Essential study attributes: Confirm that the study attributes needed for the review's comparison are recorded where applicable; missing relevant attributes reduce comparability across sources.
- Finding notes: Confirm that important finding notes retain provenance to the source, including a page or location reference where needed; unlocated notes weaken verification of claims.
- Codes or tags: Confirm that applicable codes or tags are recorded consistently when the review uses them; if they do not apply to the source, mark that state explicitly rather than forcing a value.
- Matrix entry: Confirm that the source has a matrix entry where it participates in the review's comparison; if the source is outside a particular matrix field, record the relevant missingness state rather than inventing equivalence.
A source record is complete for the review when its required information and any unresolved states have an explicit completeness status.
Correct Duplicate Records, Inconsistent Labels, and Unlinked Notes
Duplicate records, inconsistent labels, and unlinked notes are different organization errors even when they disrupt the same workflow.
A visible symptom may result from more than one likely cause, so diagnose the record-keeping failure before applying a corrective action.
The table below separates the symptom, likely cause, check, and narrowest correction needed to restore traceability.
| Symptom | Likely cause | Check | Corrective action |
|---|---|---|---|
| Duplicate record or duplicate source file | The same source may have been imported, entered, or stored more than once under different metadata or identifiers. | Compare citation metadata, identifiers, attached files, annotations, and notes before deciding that the entries are duplicates. | Merge or remove only after confirming equivalence, preserving the most complete metadata, annotations, and source links. |
| Inconsistent label or conflicting code | Synonyms, renamed codes, or unreconciled terminology may have created separate labels for the same concept. | Compare label definitions and the records using each term to confirm whether the labels represent the same concept or genuinely different meanings. | Reconcile equivalent labels under a consistent code or retain separate labels when the distinction is meaningful, then update affected records. |
| Unlinked note or orphaned note | The note may have been copied, moved, or created without preserving its source identifier or provenance. | Check the note text, page or location reference, surrounding annotations, and candidate source records to verify its origin. | Relink the note only when its source can be verified; otherwise, flag it for verification rather than assigning an uncertain source. |
| Broken file reference | The source file may have been renamed, moved, detached, or referenced through an outdated path. | Check the citation record, stable identifier, attachment, file name, and documented storage location to locate the correct source. | Restore the valid file reference or attachment and verify that the citation record and related notes still point to the same source. |
| Mismatched matrix entry | The matrix entry may refer to the wrong source record, an outdated identifier, or unreconciled source information. | Trace the matrix entry back to its source identifier and compare it with the current record, notes, and relevant field values. | Correct the entry to match the verified source record and recheck the affected comparison fields for consistency. |
Apply the narrowest corrective action that resolves the verified cause without discarding valid annotations, metadata, or provenance.
Before merging, deleting, relinking, or rewriting an entry, confirm the relationship among the affected records so the repair does not create a second error.
After the correction, verify that the source can be retrieved, its labels are consistent where intended, its notes retain provenance, and any matrix entry points to the correct record.
The repair is complete only when traceability works again across the affected part of the organization system.