What Is a Literature Review and How to Write One Correctly
What Is a Literature Review
A literature review is the part of your thesis where you organize, compare, and evaluate existing studies directly related to your topic. It answers three questions: what have other researchers studied, which concepts and methods did they use, and what still needs to be clarified in your study.
A literature review is not a list of summaries for individual articles. If every paragraph only says that author A studied X and author B studied Y, without showing similarities, differences, and limitations, the reader still cannot see why your study needs to be done.
In a quantitative thesis, the literature review usually leads to the research model, relationships between variables, and hypotheses. You can think of the reasoning chain as follows:
Existing literature → concepts and relationships → unresolved or missing point → research model → hypotheses → measurement and testing.
The sources you need to read usually include academic articles, methodology books, relevant theses, and the original works behind a model or scale. When you find a source, record the author, year, objective, sample, method, findings, and limitations immediately. This prevents you from reading the same article again when the deadline is close.
If your topic uses latent variables, keep the theoretical review separate from the scale section. The literature review explains why a concept belongs in the model. The scale section answers which items measure that concept. You can also read what is a scale so that you do not mix up these two sections.
Why a Literature Review Matters in Quantitative Research
A good literature review helps you decide on the model before you open SPSS or SmartPLS. If you do not know which theory each variable comes from, you can easily add a variable simply because one paper reported attractive results. That leaves the model without a clear rationale and makes it difficult to explain to the committee why a relationship is included or excluded.
The literature review also helps you identify the unit of analysis, the survey population, and the research context. For example, findings about students' intention to use a banking application do not automatically apply to business customers. You need to read the sample, context, measurement approach, and data-collection conditions before carrying a conclusion into your own study.
When reading sources, pay attention to the following:
- How is the main concept defined, and how is it different from related concepts?
- Which variables are independent variables, dependent variables, mediators, or moderators?
- Which relationships have been tested repeatedly, and which still produce conflicting findings?
- Where did the sample come from, how large was it, and is it suitable for your topic?
- Which questionnaire, scale, and analysis method did the authors use?
- Which limitations could open a direction for a later study?
Mediator and moderator are often mixed up during the literature review. A mediator explains the mechanism through which one variable affects another, while a moderator changes the strength or direction of that relationship. If your model includes either type of variable, read the separate pages on mediator, moderator, and moderating before writing your hypotheses.
The literature review also supports your choice of method. For example, if your objective is to test several relationships between latent variables, explain why you chose SPSS, SmartPLS, or AMOS based on the model structure and analysis objective. Avoid writing that one software package is the “best” choice for every topic.
How Much Literature Review Is Enough
There is no fixed number of pages that applies to every thesis. Quality is judged by relevance, currency, synthesis, and the logical connection to the research model. Prioritize sources that answer your research question directly instead of extending the section with papers that only share similar keywords.
A source is useful when you can answer at least one question: which variable does it define, which scale does it provide, which relationship does it test, what does it find, or which limitation does it identify. The table below is a practical checklist. The thresholds related to sample size and scales apply when you move from reading sources to designing your study.
| Item to check | Reference level or decision | Source |
|---|---|---|
| Observations per item | You can use 5 to 10 observations per item as a reference | (Hair et al., 2010) |
| Regression sample size | You can use n of 50 + 8m as a reference, where m is the number of independent variables | (Tabachnick and Fidell, 2013) |
| Cronbach's Alpha | 0.7 or above is generally accepted; an exploratory scale can be considered from 0.6 | (Nunnally, 1978); (Hair et al., 2010) |
| Corrected Item-Total Correlation | 0.3 or above | (Nunnally and Bernstein, 1994) |
| Factor loading in EFA | 0.5 or above | (Hair et al., 2010) |
| AVE in PLS-SEM | 0.5 or above | (Fornell and Larcker, 1981) |
| HTMT in PLS-SEM | Below 0.85, or below 0.90 for closely related concepts | (Henseler et al., 2015) |
The numbers in the table are not requirements for including a paper in your literature review. You use them when evaluating your own research design and results. When you cite a threshold, give the correct source in your thesis instead of writing only “according to many studies.”
A literature review can be short but convincing when the sources are well chosen and the paragraphs connect. A long review remains weak if it only repeats definitions or cites papers unrelated to the model.
How to Read a Literature Review from Collected Sources
You open your source folder and see dozens of PDF files, but you do not know which ones belong in Chapter 2. Read in three rounds instead of going through every paper from the first page to the last immediately.
The first round is screening. Read the title, abstract, keywords, model, and conclusion. Record the objective, context, sample, and main variables. If a paper only mentions the general topic but is not relevant to the relationships in your model, mark it as background rather than using it as primary evidence.
The second round is question-based reading. For each paper, look for answers in five columns: concept, relationship, sample, method, and finding. Do not copy only the conclusion. A statistically significant result in one context may not be replicated in another because the sample, measurement approach, or market conditions are different.
The third round is comparison across papers. Group papers with the same variables or theory together. Then identify agreement, disagreement, and relationships that have not been tested. This is where the review changes from a “summary of sources” into a research argument.
You can create a literature matrix as follows. This table is illustrative output, not the result of a specific study.
| Author, year | Context and sample | Main variable or relationship | Method | Finding and limitation |
|---|---|---|---|---|
| Nguyen A, 2022 | Application users, n = 280 | Ease of use → intention to use | Linear regression | Positive relationship; sample limited to one city |
| Tran B, 2023 | Students, n = 315 | Trust → intention to use | PLS-SEM | Significant relationship; the role of experience was not examined |
| Le C, 2024 | Retail customers, n = 402 | Service quality → loyalty | PLS-SEM | Results differed by age group; a new context needs to be tested |
The “limitation” column matters more than the “finding” column when you are looking for a gap. A gap should not be written as “no one has studied this topic.” A topic may not have been studied in your chosen context, may not have included a particular population, may not have tested a mediator, or may have produced inconsistent findings. You need to demonstrate this with the sources you have read.
When synthesizing, write by issue or variable group, not in the order in which you downloaded the papers. For example, the first paragraph can explain studies of service quality, the next can compare the effect of service quality on satisfaction, and the final paragraph can explain why your study selects a specific context.
What to Do When Your Literature Review Is Not Good Enough
If your literature review is currently a chain of summaries, work through the following sequence.
Step 1: Reconfirm the research question
Write the research question as one sentence with a clear variable and population. “Factors affecting purchase intention” is too broad. “Factors affecting online cosmetics purchase intention among consumers in Ho Chi Minh City” helps you screen sources and identify which papers actually fit.
Step 2: Group sources by concept and relationship
Create groups for independent variables, dependent variables, mediators, moderators, and context. Each group needs a definition, scale, empirical findings, and limitations. If a paper belongs in more than one group, record its contribution to each group instead of copying the same paragraph.
Step 3: Recheck the methods section of the main sources
Check the sample, data-collection approach, number of items, and analysis method. If you adopt a scale, you need to know which population used it and how it was adapted. Do not take a questionnaire from another paper and claim that it fits your context without checking the language and content.
Step 4: Write a synthesis paragraph after each group
After two or three sources, write a comparison paragraph. It should answer whether the findings are consistent, where the differences may come from, and what your study adopts or tests again. If you cannot identify a gap, you have not read enough of the limitations and contexts of the studies.
Step 5: Connect the gap to the model
The gap must lead to a specific choice: add a variable, change the context, test a population group, examine a mediator, or compare findings. Then write hypotheses grounded in theory and earlier evidence. Do not add a hypothesis simply to give the model more arrows.
The committee often asks, “Where is your research gap?” because many papers use the phrase without evidence. Point out which source did what, which source did not do something, and which part your study addresses. This is easier to defend than claiming that your topic is completely new.
How a Literature Review Differs from Theory and Research Background
A literature review and a theoretical framework are related, but they do different jobs. The theoretical framework presents the concepts, principles, or models that underpin the study. The literature review focuses on studies that applied those concepts, the results they obtained, and what still needs to be tested.
Research background is often used to refer to studies conducted domestically and internationally. This division works when your university requires separate sections. Whatever label you use, the content still needs analysis and comparison rather than two lists of sources.
You should also distinguish a literature review from the reference list. The reference list records the sources cited. The literature review is the argument in the main text, where each source helps explain a concept, support a relationship, or clarify a gap.
If your study uses TAM, TPB, UTAUT, or SERVQUAL, present the origin of the model first, then show how previous studies adapted it in different contexts. Naming a model without explaining its connection to your study variables leaves the theoretical framework disconnected.
Common Literature Review Mistakes
The first mistake is using an irrelevant source simply because it contains the same keyword. A paper about purchase intention may study a completely different product, population, and distribution channel from your topic. Check the relationships between variables and the context, not only the title.
The second mistake is citing a secondary source as if you had read the original. If a paper repeats a definition from another work, find the original source when possible. If you cannot access the original, write cautiously and do not assign a broader conclusion than the source you actually read supports.
The third mistake is mixing findings from the literature review with findings from your own data. Chapter 2 presents evidence from previous sources. Cronbach's Alpha, EFA, regression, or path coefficient results from your sample belong in the results section after you run the analysis.
The fourth mistake is deleting a source because its relationship was not statistically significant. Conflicting results help you explain why the relationship needs to be tested again. Keeping only studies that support your hypotheses makes the review unbalanced.
The fifth mistake is using AI or a search tool to generate citations without opening and checking the original paper. An incorrect author name, publication year, or finding can undermine the entire argument. Save the PDF, DOI, or publication details before adding a source to your thesis.
Frequently asked questions
What is a literature review in a thesis?
It is the section that synthesizes and evaluates studies related to your topic, including concepts, models, scales, findings, and limitations. Its purpose is to support your model, hypotheses, and choice of method.
How many sources are enough for a literature review?
There is no fixed number that applies to every topic. Prioritize sources that are directly relevant and report the sample, method, and findings, then check your department's or supervisor's requirements. A long list without comparison is weaker than a smaller group of sources that you analyze clearly.
What should you focus on when reading research papers?
Record the concept definitions, variables and relationships, context, sample, scales, analysis method, findings, and limitations. Read the methods and limitations sections carefully instead of relying only on the abstract and conclusion.
How should you write a research gap?
Use evidence from the sources you read to identify what has not been studied, which findings are inconsistent, or which context has not been tested. Then connect the gap to a specific choice in your model and hypotheses.
Can you use another person's thesis as a literature review source?
You can use it to find terminology, models, scales, and original sources, but prioritize original articles and appropriate academic sources. Read the sources cited in that thesis instead of copying its argument or reference list.
Should a literature review be organized by author?
You can start by taking notes by author, but the final draft should be organized by concept, variable group, or research relationship. Each group should end with a comparison and a connection to your study.
How to write this in your thesis
Open your Word file and literature matrix now. Recheck each source using five columns, concept, relationship, sample, method, and limitation, then write one synthesis paragraph that leads to your model. If you need support running the analysis after this step, see M4 data analysis in DoThesis.