Product Information:-

  • Journals
  • Books
  • Case Studies
  • Regional information

How to... undertake case study research

Options:     Print Version - How to...  undertake case study research, part 3 Print view

Data collection

If case study research is never a soft option, the collection and analysis of data can be particularly challenging.

The data collection process demands that the researcher be actively involved, asking the right questions which link to those that are central to the study, in a manner which does not alienate the subject. He or she needs to be a good listener, paying attention not only to what is said, but also to what is not said explicitly, perhaps indicated by mood or body language.

He or she needs to pay attention to multiple sources of evidence, and be able to handle complexity, and the possibility that new information may lead in a new direction, while at the same time not losing sight of the original research questions.

Preparation for data collection

The amount of preparation required for this stage will depend on the complexity of the research, and on your experience of this type of data collection. A large project with multiple cases, or a complex single case with much data collection, for example, may require a team of research assistants.

There are five aspects to preparation:

1. Ethical guidelines and protection of human subjects

Case studies involve research about humans and their actions, and you need to take special care to protect your subjects from any harm resulting from your research.

You need to gain their informed consent, ensuring that they understand the purpose of your study and are not deceived; will come to no harm as a result of your investigations; and that their privacy and confidentiality is protected. Especial care needs to be taken with vulnerable groups, such as children.

Your institution may well have a review board or committee with guidelines you need to adhere to, and whose approval you need to seek.

2. Training for case study research

If you are not part of a team, you may find it beneficial to attend a course on research methods for case studies, perhaps held by your institution or by an independent body.

You can also talk to experienced case study researchers you know in your department or through your network, and look at relevant scholarly articles. In some cases, the online versions of these articles may contain supplementary documents on which the research was based, such as the protocol.

3. Development of the case study protocol

A protocol is a blueprint for a research instrument: for example, a survey questionnaire, or a set of questions for an interview.

However, a case study protocol is more than an instrument: it is a combination of project outline and management document, research instrument, table as shell for data collection, and guidelines for the report.

Rahim and Baksh (2003, p. 32) define the case study protocol as follows:

"A case study protocol is a record (normally a document) that contains the methods, procedures and general rules that will be followed in using instruments of data collection. It is used to improve the reliability of case study results."

In summary, the protocol should cover the following:

  1. A project overview, including background, main issues to be explored, and a statement which can be presented to those from whom one seeks access, or help with the project.
  2. Field procedures, in other words, guidelines as to how the case study will be conducted. These should cover how the subjects will be accessed (including ethics), the schedule, together with such practical issues as how to handle document collection, etc.
  3. Case study questions: these are the original research questions which should guide the project, and should not be confused with the interview schedule. They should however act as a guide for any questions the researcher poses to human subjects, or which are used in the perusal of documents.
  4. Guidelines for writing the report, which may include a tentative outline.

4. Final selection of case(s)

You should have decided on the type of case, and type of design, and how your case(s) relate to the population studied, etc., at an earlier stage.

The final screening of cases involves selecting the case(s) which best fit your design, where you can obtain the best data, and where you can obtain easy access to documents, people and other information.

5. Pilot case study

If you conduct a pilot case study, this will help you refine your plans for collecting data, for example the type of questions you might want to ask your interview subjects.

The main types of data

Case study researchers are often advised to include more than one source of evidence, in order to facilitate triangulation and increase the richness and multifacetedness of their study. Choosing more than one method also has the advantage that one method's weakness can be balanced out by another's strengths.

The most usual sources of information are:

Documentation and archival records

The most important use of documentation lies in:

  • Providing background to the case.
  • Corroborating, or contradicting, evidence from interviews or other sources.
  • Providing inferential information, for example, about networks, which can be deduced from distribution lists etc. (Yin, 2003, location no. 2168).
  • Helping make sure that people's names are spelt correctly.

See "How to... use secondary data and archive material", especially "Using archival data" for more details


Because of the human element of case studies, interviews are one of the most important methods of case study research, and are almost always an element of the research design. See "How to... conduct interviews".

Direct and participant observation

Direct observation occurs when the researcher observes but does not participate. This way of collecting data is very powerful, because the researcher is unobtrusive, and can therefore freely observe behaviour which is not "edited", as it might be in a laboratory setting, or when the interviewer's questions frame the response (Woodside, 2010, p. 406).

Participant observation involves actually being part of the culture. A notable use has been the study of different cultural or social groups, or urban neighbourhoods (see "How to... use ethnographic methods and participant observation").

Physical artefacts

Physical evidence from objects, including technological devices, tools, instruments, works of art, videos etc.

Visual data collection

Because visual communication precedes verbal, visual data collection methods are a powerful way of helping individuals retrieve unconscious thoughts.

Woodside (2010) provides detailed accounts of several other methods: for example storytelling, visual narrative art, conversational analysis, and forced metaphor elicitation technique. What these methods have in common is the fact that they are designed to probe below the surface of what is being said, in other words to look at unconscious processes.

The three main principles of data collection

The following principles will help increase construct validity and reliability.

Principle 1. Use multiple sources of data

There are numerous ways that data can be collected. While many cases use just one source (usually the interview), it is considered good practice to obtain data by several methods.

This helps with triangulation, where different lines of enquiry converge, with the findings of one set of data corroborating another. As Rowley puts it (2002):

"Triangulation uses evidence from different sources to corroborate the same fact or finding."

Furthermore, different data sources can support one another with complementary strengths: for example, document analysis is good for establishing facts, whereas interviews enable the researcher to probe.

Woodside (2010, p. 107) maintains that a multiple methods approach is essential in order to elicit both conscious and unconscious thinking processes. Methods can include interviewing, and observation, with the researcher concurrently recording his or her thoughts.

This methodological richness is one of the great strengths of case studies as a method.

Principle 2. Create a case study database

According to Yin (2009, location no. 2495) the data from the case study is often merged with the report, whereas it should be stored in such a way that it can easily be accessed and viewed by another researcher. This will increase transparency and hence the reliability of the research.

The database may include the following:

  • Notes, from interviews, observation, document analysis, etc.
  • Documents collected during the case study, which should be accompanied by an annotated bibliography.
  • Tabular materials, including survey and quantitative data.
  • Any narratives composed by the researcher during investigations.

Principle 3. Maintain a chain of evidence

The case study database should contain plenty of evidence that is then referred to in the report. It is also important to have clarity in terms of the trail of evidence that leads from the research questions to the conclusions, and back again.

Opoku and Williams (2011, pp. 256-257) commented on how they maintained a chain of evidence:

"Using multiple researchers, we continually communicated our methodological decisions to respondents and recorded the data mechanically by using a tape recorder. This helped to enhance the reliability of this study. We have also tried to improve the reliability of this study by documenting all the necessary steps which were followed in order to conduct it ... The interviews were transcribed, analysed, and validated with the interviewees. Provisional research conclusions were subsequently sent back to the respondents via e-mail and the resulting feedback discussion was annotated."

Surviving complexity

It will be clear from the above that data collection for case studies is highly complicated. Each type of data collection requires its own skill set which the researcher will have to master.

Another problem is that there is no end point for data collection, unlike, for example, the survey, where the parameters of the investigation are fixed by the number of responses (assuming the researcher considers the response rate to be satisfactory).

With case study evidence coming from so many sources, there is no such fixed cut-off point, so it can be difficult to know when to stop.