Literature review examples and format: what a finished review looks like

What are the main ways to structure a literature review?

There are four, and the headings show which one a review uses.

Go thematic unless you have a strong reason not to. A committee that reads “Objective 2: to examine the effect of X on Y” in your synopsis expects a section on X and Y in the review.

Mixing is normal. What doesn’t work is arranging by author (“2.3 Studies by Indian authors”, “2.4 Studies by foreign authors”), which I still see every year and which almost always produces summary. If you have a separate framework chapter, keep the review’s theory section short.

What does a literature review outline look like for a PhD thesis?

Here is an outline for a management thesis on remote work and employee well-being in Indian IT firms. The topic is made up; the shape is typical of what goes to a DC or RAC meeting.

1. 2.1 Introduction. Scope, databases and years searched, and the order of sections. About a page.
2. 2.2 Remote work: definitions and development. Short and roughly chronological, including how the pandemic changed the research.
3. 2.3 Theoretical perspectives. Job demands–resources theory and one competing view, with the studies that apply them to remote work.
4. 2.4 Remote work and job autonomy. First theme, tied to Objective 1.
5. 2.5 Remote work, work–family conflict and well-being. Second theme, tied to Objective 2.
6. 2.6 The role of organisational support. Third theme, the proposed moderator.
7. 2.7 Studies in the Indian context. Indian articles and Shodhganga theses, and how their samples differ.
8. 2.8 Summary and research gap. What the evidence shows, where it is thin, and the gap your objectives address.

Sections 2.4 to 2.6 map onto the objectives one to one, so an examiner can check coverage in a minute. Give Indian studies their own section only if your argument depends on how they differ. Our PhD thesis structure guide shows where this chapter sits; in a synopsis the same outline shrinks to a few pages (see the research proposal guide).

Literature review examples: weak summary vs good synthesis in three fields

These passages are illustrative. The citations are placeholders and the findings are invented to show the form, so don’t quote them as evidence.

Example 1: management (remote work and well-being)

Weak version:

Author A (Year) studied 200 IT employees in Pune and found that remote work increased job satisfaction. Author B (Year) conducted interviews with managers in the UK and found that remote workers felt isolated. Author C (Year) used a survey in Bengaluru and found that working from home increased work–family conflict for women employees.

Stronger version:

Evidence on remote work and well-being points two ways, depending on who is studied. Surveys of employees in large IT firms report higher job satisfaction among remote workers (Author A, Year; Author D, Year), whereas interview studies report isolation and blurred boundaries (Author B, Year). The difference may partly reflect method, since surveys capture average attitudes while interviews draw out strain that averages hide. Gender adds a further complication: the one Indian study to separate men and women found work–family conflict rising only for women (Author C, Year), a split the satisfaction surveys could not detect because they did not report results by gender.

What changed: the paragraph opens with a claim, groups studies by finding, offers a reason for the disagreement, and ends on a weakness that already points at a gap.

Example 2: engineering (fault detection in solar PV plants)

Weak version:

In [1], the authors proposed a threshold method for fault detection. In [2], a random forest model was used and gave high accuracy. In [3], drone thermal imaging with a CNN was used to find hot spots.

Stronger version:

PV fault detection has moved from fixed thresholds towards learned models, but the evidence for learned models is weaker than their reported accuracy suggests. Threshold rules are simple to deploy and detect open-circuit faults reliably, yet they confuse partial shading with faults [1]. Machine-learning classifiers report much higher accuracy [2], but that study trained and tested on simulated I–V curves, so its performance under dust, soiling and temperature swings in the field is unknown. Image-based methods [3] have been validated in operating plants, although they rely on drone surveys that small rooftop installations are unlikely to fund.

The stronger version compares every method on one question (does it work in the field?), and that comparison is what makes it a review.

Example 3: health (self-care in type 2 diabetes)

Weak version:

Diabetes is a major health problem in India. Many studies have been done on self-care. Author E (Year) found that education improves self-care. Author F (Year) also found that education improves self-care. Author G (Year) found that family support is important.

Stronger version:

Structured education improves self-care in type 2 diabetes in most trials (Author E, Year; Author F, Year; Author H, Year), but all three recruited from urban hospital clinics and delivered education through doctors or diabetes educators, which limits what they tell us about rural patients. The fewer, mostly cross-sectional studies of rural patients suggest that family support predicts self-care more strongly than knowledge does (Author G, Year). Whether community health workers can deliver education that also draws in the family has not, to our knowledge, been tested.

The weak version opens with a statement readers already accept and repeats a finding. The stronger one treats agreement as a starting point and asks who the agreeing studies left out. Our guide to writing a research gap statement shows how that final sentence becomes a formal gap.

How do you open and close a literature review?

The opening states the scope and the route; it doesn’t argue anything yet. For the remote-work thesis:

This chapter reviews research on remote work and employee well-being published between 2010 and 2025, mainly on knowledge workers. Sources were found through Scopus, Web of Science and Shodhganga (search terms in Appendix A). Section 2.2 traces how remote work has been defined; Section 2.3 sets out the theoretical perspectives used in this study; Sections 2.4 to 2.6 review evidence on autonomy, work–family conflict and organisational support; Section 2.7 considers Indian studies; and Section 2.8 identifies the gap this study addresses.

Many drafts open instead with a paragraph on the topic’s national importance. Cut it; that belongs in Chapter 1.

The closing section pulls the themes together and states what remains unknown:

Taken together, the studies reviewed show that remote work is associated with greater autonomy, but its effect on well-being depends on work–family conflict, which falls unevenly on women. Organisational support appears to soften that conflict in Western samples, but no study has tested this in Indian IT firms, where household structures and working hours differ. This study therefore examines whether organisational support moderates the relationship between remote work and well-being among Indian IT employees, and whether that effect differs by gender.

The last sentence reads almost like an objective, and it should. If your supervisor can predict your objectives before turning the page, the chapter has done its job.

What is a literature survey, and how is its format different?

“Literature survey” is what many Indian engineering and computer science departments call the literature review, in theses and M.Tech project reports alike. It is the same chapter. The differences are in form: citations are usually IEEE style (numbers in square brackets, in order of first citation; see our IEEE formatting guide), sections are more often organised by method, and many surveys add a summary table of prior work.

The table doesn’t replace the prose. The commonest problem in engineering surveys is a chapter that is almost all table, with a line under each row saying “[2] used random forest and got good results.” Examiners want the paragraph in Example 2, with the table as its summary.

How do you format a literature review in APA style?

Your university’s thesis format overrides APA on margins, fonts, spacing and numbering. Within those limits, an APA 7 literature review should show these points:

• Chapter title as an APA Level 1 heading (centred, bold), each theme as Level 2 (flush left, bold), sub-themes as Level 3 (flush left, bold italic). Most theses never need Levels 4 and 5.
• Several sources supporting one point go in a single parenthesis, in alphabetical order and separated by semicolons: (Author A, Year; Author B, Year; Author C, Year).
• Past research is reported in the past tense or present perfect (“researchers found” or “researchers have found”), used consistently.
• Double spacing, 1-inch margins and a 0.5-inch first-line indent, unless your university’s thesis format says otherwise.
• A reference list titled “References”, entries in alphabetical order with a hanging indent, and every in-text citation matched by an entry.

The grouped-citation rule and the tense advice come from APA Style’s own pages on citing multiple works and its student paper checklist.

Our APA 7 formatting guide has the detail; for Harvard, Chicago or Vancouver, the citation styles comparison shows what changes.