Systematic review psychology dissertation research uses transparent, reproducible methods to find, appraise, and synthesise evidence for a focused psychological question. It is not a longer version of a conventional literature review. A systematic review follows a prespecified protocol, documents every search and selection decision, evaluates study limitations, and connects the synthesis directly to the research question.
This guide explains how to plan a feasible review, choose the right review type, write eligibility criteria, search psychological databases, screen records, extract data, assess risk of bias, synthesise findings, and report the process clearly. It also shows where a meta-analysis may be appropriate and where a careful narrative or qualitative synthesis is the stronger choice.
What makes a Systematic Review Psychology Dissertation different?
A traditional dissertation literature review usually supports a broader empirical project. It organises theory and prior findings, develops a research gap, and justifies a new study. A systematic review is itself the research method. Its dataset consists of eligible studies, and its conclusions depend on how those studies were identified, selected, appraised, and combined.
That distinction changes the standard of documentation. Readers should be able to understand what was searched, when it was searched, which records were excluded, what information was extracted, and how the final synthesis was produced. Transparency matters because small decisions can alter the evidence base.
The existing guide to writing a psychology literature review explains thematic organisation and gap development. Use the present guide when the dissertation question will be answered through a formal evidence-synthesis design rather than new participant data.
Check whether the project is feasible
A systematic approach does not automatically make a question suitable for a dissertation. A question can be too broad to search, too narrow to yield enough evidence, or so heavily reviewed that another synthesis would add little. Begin with a scoping search before finalising the title.
Search at least one major psychology database and a multidisciplinary source using a few central concepts. Look for recent systematic reviews, common outcome measures, likely study designs, terminology differences, and the approximate volume of records. This preliminary work is not the final search. Its purpose is to test feasibility and refine the question.
Ask five feasibility questions
- Does the question address a meaningful uncertainty in psychological evidence?
- Are enough potentially eligible studies available for a defensible synthesis?
- Can the databases, full texts, software, and methodological support be accessed?
- Can screening, appraisal, extraction, and writing be completed within the deadline?
- Will the review offer a clear contribution rather than repeat a current review?
Suppose you are interested in digital interventions for social anxiety among university students. A review of every digital mental-health intervention, outcome, population, and study design would be unmanageable. A question focused on app-based cognitive behavioural interventions, validated social-anxiety outcomes, and post-secondary learners may be more feasible, provided the scoping search finds a sufficient evidence base.
Table of Contents
Choose the right review design
“Systematic review” describes a family of methods, not one fixed template. The review type should follow the objective and available evidence. A statistical meta-analysis is a synthesis technique, not a synonym for systematic review.
The table introduces common options. Confirm the choice with your supervisor and applicable institutional guidance before locking the protocol.
| Review type | Best suited to | Typical output | Main caution |
|---|---|---|---|
| Effectiveness review | Whether an intervention changes an outcome | Narrative synthesis or meta-analysis | Design and outcome comparability matter |
| Association review | Relationships between psychological variables | Structured synthesis or pooled associations | Confounding and measurement vary |
| Qualitative evidence synthesis | Experiences, meanings, or processes | Interpretive themes or concepts | Methodological coherence is essential |
| Scoping review | Mapping concepts, evidence types, or gaps | Evidence map and descriptive account | It usually does not estimate effectiveness |
| Measurement review | Properties of psychological instruments | Evidence on reliability, validity, and use | Appraisal must fit measurement studies |
The JBI Manual for Evidence Synthesis provides methods for different evidence-synthesis questions, while the Cochrane Handbook gives detailed guidance for intervention reviews. These are methodological resources, not checklists to copy without adapting them to a psychology dissertation.
Turn the topic into an answerable review question
A review question must define what evidence could answer it. Frameworks help make each component explicit. PICO covers population, intervention, comparator, and outcome. PECO replaces intervention with exposure. SPIDER can support qualitative or mixed evidence by considering sample, phenomenon of interest, design, evaluation, and research type.
Choose the framework that represents the question accurately. Do not force an experiential question into an intervention template simply because PICO is familiar.
Example question map
Consider: “What is the association between sleep regularity and depressive symptoms among university students?” The population is university students, the exposure is sleep regularity, and the outcome is depressive symptoms measured with a defined instrument or acceptable family of measures. Eligible designs might include longitudinal and cross-sectional observational studies, but the protocol should state whether they will be synthesised together or separately.
If your question is still vague, use the guide to psychology dissertation research questions to check scope, constructs, ethics, and answerability.
Write a protocol before the final search
The protocol is the review’s decision record. It prevents the inclusion criteria, outcomes, or analyses from drifting in response to convenient findings. At minimum, state the rationale, objectives, eligibility criteria, sources, draft search strategy, screening process, extraction fields, appraisal method, synthesis plan, and procedure for recording amendments.
The PRISMA-P guidance supports transparent reporting of systematic review protocols. Some projects may also be eligible for registration in PROSPERO. Registration requirements and scope can change, so check current eligibility rather than promising that every psychology dissertation can be registered.
Record amendments honestly
A protocol does not forbid sensible changes. It makes them visible. If a pilot search reveals that an outcome term is too restrictive, document what changed, when it changed, why it changed, and which stages had already begun. Avoid presenting a post hoc decision as if it had been planned from the start.
Define inclusion and exclusion criteria
Eligibility criteria translate the question into repeatable selection rules. Define eligible populations, constructs, interventions or exposures, comparators, outcomes, study designs, settings, dates, languages, publication types, and evidence formats. Each restriction needs a methodological or feasibility reason.
A common mistake is to use vague rules such as “relevant psychology studies” or “high-quality papers.” Relevance and quality must be operationalised. Quality is usually assessed after inclusion because excluding a study on an undefined impression can introduce selection bias.
| Criterion | Clear specification | Weak specification |
|---|---|---|
| Population | Enrolled post-secondary students aged 18 or older | Young people |
| Construct | Sleep regularity measured across at least three days | Sleep problems |
| Outcome | Depressive symptoms assessed with a named validated scale | Mental health |
| Design | Cross-sectional or longitudinal observational studies | Empirical research |
| Publication | Peer-reviewed articles and eligible dissertations | Reliable sources |
Language and date limits can be necessary, but they may exclude relevant evidence. State the practical reason and discuss the resulting limitation. Do not imply that English-language publication is a marker of study quality.
Build a reproducible search strategy
A strong search balances sensitivity and precision. Sensitivity reduces the chance of missing eligible studies; precision limits irrelevant results. Construct the search from concept blocks rather than writing the entire research question as one sentence.
Combine controlled vocabulary and keywords
Database subject headings capture indexed concepts, while title-and-abstract keywords capture newer terms, spelling variants, and language used by authors. For sleep regularity, keywords might include regularity, variability, consistency, social jetlag, and sleep timing. For depressive symptoms, include the construct and relevant indexed terms without inserting the names of every possible scale unless justified.
Use OR within a concept block and AND between concept blocks. Apply proximity operators, truncation, and field codes only after checking the syntax of each database. A strategy copied from PsycINFO into PubMed or Scopus may fail silently because the platforms use different commands.
Search more than one appropriate source
PsycINFO is central for many psychology questions, but it should rarely stand alone. Depending on the topic, suitable sources may include MEDLINE or PubMed, Scopus, Web of Science, CINAHL, ERIC, SPORTDiscus, or specialised registers. Search choices should follow the discipline, population, and evidence type.
Record the database and platform, full search string, filters, date searched, coverage period, and number of results. Save or export the strategies. Search reference lists and citations where appropriate, and explain how non-journal evidence was handled.
Manage records and remove duplicates
Export complete citation information into reference-management or review software. Preserve a raw export before deduplication. Automated matching can miss duplicates when titles, author names, or conference records differ, so combine software rules with a documented manual check.
Assign every record a stable identifier. Keep decision fields separate from bibliographic fields, back up the library, and protect any unpublished material shared by authors. A systematic review may not collect participant data directly, yet it still requires responsible data management.
Screen studies consistently
Screening usually occurs in two stages: title and abstract, then full text. Pilot the criteria on a sample before formal screening. The pilot reveals ambiguous terms, hidden assumptions, and criteria that two readers may interpret differently.
Best practice often uses two independent reviewers, but dissertation resources may make full duplicate screening difficult. Do not pretend that single-reviewer screening is equivalent. Agree on a feasible safeguard with your supervisor, such as independent screening of a substantial sample, duplicate review of uncertain records, or verification of exclusions. Report exactly what was done.
Use one defensible exclusion reason
At full-text stage, record the primary reason each report was excluded. Use a hierarchy so similar cases receive consistent decisions. “Not relevant” is too vague. Better reasons include ineligible population, wrong study design, outcome not reported, or duplicate report of the same study.
The PRISMA 2020 statement includes reporting items and flow-diagram templates that help readers trace records from identification to inclusion. PRISMA guides reporting; it does not replace a method for conducting the review.
Extract data with a piloted form
Extraction should capture everything needed to describe studies, judge limitations, and complete the planned synthesis. Pilot the form on several different studies before extracting the full set. A form that works for a simple experiment may fail for longitudinal, multilevel, or qualitative research.
| Domain | Example fields | Why it matters |
|---|---|---|
| Study identity | Authors, year, country, reports linked to one study | Prevents double counting |
| Design and sample | Design, setting, sample size, recruitment, attrition | Defines applicability and bias |
| Measures | Construct, instrument, timing, scoring | Tests outcome comparability |
| Results | Effect estimate, uncertainty, themes, quotations | Supplies synthesis evidence |
| Analysis | Model, covariates, qualitative approach | Supports interpretation |
| Notes | Missing data, author contact, protocol deviations | Preserves decision context |
Distinguish reports from studies. One longitudinal project may produce several papers from overlapping participants. Treating each paper as an independent sample can exaggerate the amount of evidence.
Assess risk of bias, not reputation
Risk-of-bias assessment asks whether design and conduct could systematically distort a study’s findings. It is not a score for how prestigious the journal appears, and it is not the same as reporting quality. Choose a tool that matches the included designs and the causal or descriptive claim being assessed.
Apply the tool at the appropriate outcome level, justify any adaptations, and explain how judgements will influence the synthesis. Avoid converting complex domains into a single percentage without a defensible rationale. A study can be strong in one domain and weak in another.
For psychology examples, relevant concerns may include selective outcome reporting, unvalidated measures, high attrition, inadequate control of confounding, lack of blinding where feasible, interviewer influence, or weak coherence between qualitative data and interpretations.
Plan the synthesis before seeing the pattern
The synthesis method should match the question, study designs, measures, and degree of heterogeneity. Start by grouping studies into meaningful comparison sets. Describe samples, constructs, interventions, outcome timing, and risk-of-bias patterns before attempting a pooled estimate.
When meta-analysis may be justified
Meta-analysis can estimate an average effect and its uncertainty when studies address sufficiently comparable questions. It requires compatible effect measures, independent data or correctly handled dependence, and a model that reflects expected variation. Statistical software cannot solve conceptual incompatibility.
Plan how to handle multiple outcomes, repeated time points, several intervention groups, cluster designs, adjusted and unadjusted estimates, and missing statistics. Explore heterogeneity using prespecified, evidence-based analyses rather than many underpowered subgroup tests.
When narrative synthesis is stronger
A structured narrative synthesis is appropriate when measures, designs, populations, or reported statistics cannot be combined responsibly. “Narrative” does not mean listing one study after another. Organise evidence by question-relevant groups, compare direction and magnitude, examine methodological differences, and explain how study limitations affect confidence.
Qualitative evidence synthesis
Qualitative synthesis may aggregate findings or develop a more interpretive account, depending on the selected methodology. Preserve context, examine researcher interpretations, and maintain a clear audit trail from primary findings to review themes. Do not count how many studies mentioned a theme as if frequency alone established importance.
The site’s psychology dissertation data analysis guide gives broader advice on quantitative, qualitative, and mixed-methods analysis. In a review, every analysis must additionally account for study-level dependence and variation across evidence sources.
Interpret findings without overstating certainty
A systematic process can still produce uncertain evidence. Interpretation should consider risk of bias, inconsistency, imprecision, indirectness, measurement quality, publication processes, and the limits of the search. Avoid turning “absence of convincing evidence” into “evidence of no effect.”
For example, five small cross-sectional studies may show similar associations between sleep irregularity and depressive symptoms. That pattern may justify further longitudinal research, but it cannot establish that irregular sleep causes depression. The conclusion should match the underlying designs.
Write the dissertation as a research report
A review-based dissertation still needs a coherent introduction, method, results, discussion, and conclusion. The chapter labels may vary, but the logic should remain visible.
- Introduction: define the problem, existing knowledge, uncertainty, and review objective.
- Method: report the protocol, eligibility, sources, searches, selection, extraction, appraisal, and synthesis.
- Results: show the flow of studies, characteristics, bias judgements, and synthesis findings.
- Discussion: answer the question, compare prior evidence, explain limitations, and give proportionate implications.
- Conclusion: state what the evidence supports, what remains uncertain, and what should happen next.
Use the APA Journal Article Reporting Standards where they fit the included evidence and synthesis type. Also check the psychology dissertation structure guide to keep chapters aligned.
Common systematic review mistakes
Calling one database search systematic
A repeatable search of one source may still miss much of the evidence. Justify database coverage and supplementary searching from the topic, not convenience alone.
Changing criteria after reading results
Unreported changes allow preferred findings to shape inclusion. Use a protocol, date amendments, and separate planned from exploratory analyses.
Using PRISMA as the method
PRISMA improves reporting completeness. The actual review methods must come from a suitable methodological framework and be justified for the question.
Equating quality with a numerical score
Summed scores can conceal which bias matters for a particular result. Present domain-level judgements and use them in the interpretation.
Pooling studies because software permits it
A forest plot can be technically produced from conceptually incompatible studies. Decide comparability before calculation, and explain exclusions from each synthesis.
A practical dissertation workflow
- Define the uncertainty and conduct a limited scoping search.
- Choose the review type and question framework.
- Write and approve a protocol.
- Pilot eligibility criteria and the search strategy.
- Run all final searches and preserve the exact outputs.
- Deduplicate records and pilot screening.
- Complete title, abstract, and full-text screening.
- Pilot extraction and risk-of-bias assessment.
- Describe studies before conducting the synthesis.
- Interpret results in proportion to the evidence.
- Complete the reporting checklist and decision audit.
- Proofread the saved dissertation against the protocol.
Build time for inaccessible full texts, author queries, duplicate reports, extraction disagreements, and search updates. Screening hundreds of records usually takes longer than expected, especially when criteria are still changing.
Frequently asked questions
Does a systematic review psychology dissertation need a meta-analysis?
No. Meta-analysis is appropriate only when the studies and effect estimates can be combined meaningfully. A transparent narrative or qualitative synthesis may answer the question more accurately.
How many studies should a systematic review include?
There is no universal minimum. The number depends on the question, evidence base, and synthesis method. Feasibility should be tested before the protocol is finalised, while conclusions should reflect the amount and quality of included evidence.
Can one student screen every record?
Resource constraints sometimes make single-reviewer work unavoidable. Discuss safeguards with your supervisor, pilot the rules, independently check a defined sample or uncertain cases, and report the limitation honestly.
Is ethics approval required?
Reviews of public research reports may not require the same participant ethics process as primary studies. However, institutional rules differ, and unpublished data or author contact can introduce additional responsibilities. Obtain a formal determination from the appropriate institutional route.
Can dissertations and unpublished studies be included?
They can be eligible when the protocol justifies their inclusion and applies consistent appraisal. Including non-journal evidence may reduce publication bias, but availability, duplication, and reporting completeness require careful handling.
When should the search be updated?
Update the search when a substantial interval has passed between the original search and submission, or when required by the chosen guidance. Report both dates and merge new records through the same documented process.
Conclusion
A systematic review psychology dissertation succeeds when its question, protocol, search, selection, appraisal, synthesis, and conclusions form one transparent chain. The goal is not to collect the largest number of papers or produce a complex statistical model. It is to answer a focused question with methods that another researcher can understand, scrutinise, and repeat.
Start with feasibility, record decisions before seeing the results, and choose the synthesis that the evidence can support. If you need help refining the question, checking the protocol, or improving the clarity of your review chapter, seek ethical guidance that preserves your authorship, institutional rules, and responsibility for every research decision.
