Global Academic Publishers Logo
GAP Insights · Publishing Guides · 2026

How to Write a Case Study in Research: A Practical Guide

By Muaz Dave · 7 November 2026, 9:54 AM PKT

How to Write a Case Study in Research: A Practical Guide

Introduction

In March 2024, Dr. Ayesha Khan, a medical officer at a 40-bed district hospital in Jhang, noticed something odd: her ward had added a second triage nurse in January, and waiting time had fallen from 95 to 40 minutes — but only on Mondays and Thursdays. On other days, patients still waited over an hour. That is the question this guide answers.

A case study is a close investigation of a single case — a person, a group, an organisation, a clinic, or an event — studied in its real-life setting, using several sources of evidence. It is not an anecdote: done well, it follows a clear design, collects evidence systematically, and reports findings honestly, including the limits of what one case can prove. The sections below work through Ayesha's ward and show each decision you will face, from choosing the case to writing the paper.

One Clinic, One Question: A Worked Example

Example — the ward that improved only twice a week. Ayesha's ward sees about 120 outpatients a day. In January 2024, the hospital added a second triage nurse, Sana, to the morning shift. By March, the register showed waiting time down from 95 to 40 minutes on Mondays and Thursdays, but stuck at 80 to 90 minutes on other days. Her question became: "How did the second triage nurse change patient flow in this ward, and why did the change appear only on some days?"

Notice what the question does: it names one case, asks how and why, and points at the evidence needed — the ward register, the duty roster, interviews with the two nurses, and observation of the morning queue. Ayesha did not study all six wards, and she did not try to prove that adding a nurse always cuts waiting time. She studied one ward deeply enough to explain a puzzle.

What Counts as a Case Study — and What Does Not

A case study is a research design, not a format. It has four marks: a defined case with clear boundaries, a how or why question, more than one source of evidence, and a systematic analysis linking evidence to conclusions. If any is missing, the work is something else — useful, perhaps, but not a case study.

Three confusions are worth naming. A case report describes one patient's condition and treatment; it rarely states research questions, so it is not a case study. A single story told from memory has no design and no evidence trail. And a small survey of 20 people studies a sample, not a case in depth. Check your plan against the four marks above before you write a word.

Pick the Design Before You Pick the Case

There are three basic designs. A single-case study looks deeply at one case — choose it when the case is unusual, critical, or revealing, like the first clinic to cut waiting times. A single case gives detail a larger study would miss, but it cannot show whether the same pattern appears elsewhere.

A multiple-case study looks at two or more cases chosen to compare or confirm each other. Cases may be similar, so a repeated finding strengthens the result, or different, so the contrast explains why outcomes varied. Ayesha could have added a second ward that also gained a triage nurse but showed no improvement; the contrast would have made her explanation far stronger. Multiple cases take more time and more data, and reviewers usually find them more convincing. A third option, the embedded design, studies one case with several units inside it — one hospital with three wards, for example.

Writing the Research Question: How and Why

Case study questions ask how or why. "How did the ward reduce waiting times?" and "Why did the new referral system fail in this hospital?" are good questions because they ask for an explanation of a process. Questions that ask how many or how often — "How many patients waited over an hour?" — are better answered by a survey, so keep your questions matched to the design.

Write one main question and two or three sub-questions that break it into parts. For Ayesha's ward: main question, "How did adding a second triage nurse change patient flow between January and March 2024?"; sub-questions, "Which steps in the morning routine changed, and on which days?", "What did the nurses and the duty doctor do differently on Mondays and Thursdays?", and "What kept the change from appearing on other days?" Each sub-question decides what data to collect — the second one needs interviews, not only register numbers. If a question cannot be answered with evidence you can actually gather, rewrite it.

Setting Boundaries So the Case Does Not Swallow the Study

Decide the boundaries before you collect data: the time period, the place, the people, and the events that belong inside the case — and write those limits down. Ayesha's boundaries were her hospital's outpatient ward, the morning shift from 8 a.m. to 1 p.m., the period from 1 January to 31 March 2024, and the staff on duty. The evening shift and the inpatient wards were outside the case.

A case without clear boundaries grows endlessly — new documents, new interviews — until the study can never be finished. State the boundaries in the methods section, and name the cases you did not study and why: a reader who knows what you left out can judge what your findings are worth.

Three Sources, One Finding: Building Your Evidence Base

Use more than one kind of evidence. Documents show what was decided and when; interviews show how people understood those decisions; observation shows what actually happened — often different from both. Each source has its own bias, so use several and compare them. A case study built on one interview is an opinion piece, not research.

Example — when the sources disagreed. In Ayesha's ward, the register said the second triage nurse started on 2 January; Sana said she started on 9 January, after a week's leave. The duty roster agreed with Sana — and also showed that on Mondays and Thursdays the senior doctor, Dr. Bilal, ran the morning clinic, while a junior doctor covered the other days. The disagreement was itself the finding: the improvement tracked Dr. Bilal's days, not the new nurse's start date. Across all twelve weeks, waiting time was under 45 minutes on 22 of 24 Mondays and Thursdays, and under 60 on only 3 of 36 other mornings. One source alone would have given her the wrong story.

Keep a case study database: all evidence in one organised place, with a note on where each item came from and when it was collected, kept separate from your interpretation — so every claim in the findings can be traced back to a document, an interview, or an observation.

Analysing the Evidence: Pattern Matching in Practice

Organise the evidence around your research questions, not around the order in which you collected it. The most practical technique is pattern matching: write down the pattern you expected to see before you look at the data, then check whether the evidence matches it. Ayesha expected waiting time to fall every morning after Sana joined. The register showed a different pattern — improvement only on Dr. Bilal's days — so she built a second explanation: on those days, Dr. Bilal reviewed the triage queue himself at 8:30 a.m.; on other days, the queue waited for the junior doctor's 10 a.m. round. Each part was tested against the roster, the register, and two short interviews.

For interview material, code the data by theme and record how the themes were developed, so the reader can see the analysis was systematic. For numbers, report them plainly: "waiting time fell from 95 to 40 minutes on Mondays and Thursdays" is a finding; "the new nurse caused the fall" is an interpretation that needs the triangulation behind it. Make your reasoning visible.

Writing the Case Description So a Stranger Can Follow It

The case description is the heart of the paper: a clear, ordered account, not a pile of notes. Use a short timeline to show the sequence of events, and quote documents or interviews briefly to let the evidence speak. Ayesha's description opened with the ward in one paragraph — 40 beds, 120 outpatients a day — then walked through the twelve weeks in order, with the register figures at each step. Choose detail that matters to the question; leave out the rest.

Be precise about names, dates, and numbers, and anonymise people and places where ethics require it — saying so in the methods section if you change a detail. A description that is vague about who did what and when is hard to trust, so check every factual claim against your case study database before you submit.

Discussion: What One Case Can and Cannot Claim

Connect the findings back to the research question and to earlier work on the same topic. Say what the case adds, what it confirms, and where it differs from what others found. Ayesha's discussion did not claim that adding a triage nurse cuts waiting time everywhere. It claimed something smaller: in this ward, the improvement depended on a senior doctor clearing the triage queue early. That is a claim one case can support.

Frame claims carefully — what the case suggests and illustrates, not universal rules. The honest limit of a case study is also its strength: one case studied deeply gives rich, specific findings a large study cannot match. Use that depth to explain the mechanisms and the context, and end by saying what a larger study should test next.

Ethics: Consent When Everyone Knows Who You Mean

Get ethics approval from your institution's committee before collecting data, and get informed consent from every person you interview or observe. Explain how their information will be used, who will see it, and how their identity will be protected. In a small ward, anonymity is fragile — a colleague may still recognise the clinic from the bed count and the dates — so discuss this risk with participants before they agree, and let them read the parts of the draft that describe them.

If a participant withdraws consent, remove their material and record what was removed and why. Ayesha's paper used pseudonyms for both nurses and the two doctors, noted the change in the methods, and gave the draft to Sana and Dr. Bilal to check. That member checking caught one wrong date and added a detail she had missed about the 8:30 a.m. queue review.

Five Mistakes Reviewers Spot in the First Two Pages

1. A case chosen for convenience. "We studied our own ward because it was nearby" tells the reviewer the case cannot answer the question — give a reason tied to the question. 2. One source of evidence. A paper built on three interviews and nothing else is an opinion piece; show documents, numbers, and observation too. 3. No boundaries. If the reader cannot tell where the case ends, state the time, place, and people up front.

4. Description with no analysis. A timeline of events is not a finding — say what the pattern means and show the evidence. 5. Overclaiming. "This proves triage nurses cut waiting times" from one ward will be rejected in a sentence. State the limits plainly; a modest claim backed by a visible evidence trail beats a grand claim backed by none.

Case Study Designs Compared

DesignBest forMain data sourcesMain risk
Single-case (holistic)A rare or revealing caseInterviews, records, observationFindings may not transfer
Single-case (embedded)One case with several unitsDocuments, interviews, measuresUnits treated as separate cases
Multiple-caseComparing or confirmingSame protocol across casesTakes more time and data
ComparativeExplaining different outcomesMatched cases, same questionsCases may not be comparable

Further Reading and Useful Resources

These authoritative external resources explain the standards and tools mentioned in this guide:

Frequently Asked Questions

What is a case study in research?
A case study is an in-depth investigation of one case — a person, group, organisation, or event — studied in its real setting using several sources of evidence. It answers how and why questions, and it reports what happened, why it happened, and what it means.

What is the difference between a case study and a case report?
A case report, common in medicine, describes one patient's condition and treatment. A case study is a research design: it states research questions, uses several sources of evidence, analyses them systematically, and discusses what the case adds to knowledge — and what its limits are.

How many cases do I need?
One case is enough when it is unusual, critical, or revealing. Two or more cases are better when you want to compare outcomes or confirm a finding across settings. Choose the number that answers your question, not the largest number you can manage.

Can a case study use quantitative data?
Yes. A case study can include numbers, such as counts, rates, or measurements, alongside interviews and documents. The design is defined by studying a case in depth, not by the type of data, so mixed evidence is welcome — and numbers often make the case description clearer.

How do I choose a case?
Choose a case because it can answer your research question — because it is unusual, typical, or critical for a theory. Explain your choice in the paper, set clear boundaries for the case, and say which cases you did not study and why.

What makes a case study trustworthy?
Several sources of evidence, a clear chain from evidence to conclusions, and honest reporting of limits. Triangulation, member checking, and a case study database all help readers trust that the findings come from the evidence, not from the author's opinion.

Submit your manuscript  ← Back to GAP Insights

More Resources for Researchers

Research Resources Directory

Useful publishers, databases, funders, tools and institutions for researchers:

Publishers

Databases and Indexes

Preprint Servers and Repositories

Ethics and Reporting Standards

Funders

Research Tools

Health and Veterinary Organisations

Veterinary Schools

Leading Universities

Major Journals