Category: Research Methods

You would like to apply the case study research methodology to you next academic paper or thesis?

Then you should stop everything else right away, because in this article you will get a super fast and effective beginner’s tutorial on how to conduct a case study.

In only 6 easy-to-follow steps you will learn the basics of case study research and how to apply them.

What is the case study research methodology?

Case study research is often used in social sciences. It investigates a current phenomenon that can be observed in our world (as opposed to, for example, historical events or natural laws).

This phenomenon is always anchored in a specific context, which must be taken into account throughout the entire case study. A possible context can be an organization, a country, or even a single person.

The case needs to provide the context in which the phenomenon under investigation can be observed.

In a case study, the researcher has no influence on the events (as opposed to, for example, an experiment in a laboratory). Rather, data is collected in the field or from third parties about that case to analyze the phenomenon and to arrive at theoretical and/or practical conclusions.

Both qualitative approaches (such as interviews and grounded theory) and quantitative methods (such as surveys and statistical tests) can be used in data collection and analysis. The special feature is that the research focuses only on one specific “case”.

Comparative or multiple case studies are special forms of case study research that relate and contrast several cases to each other.

A case study always answers a specific research question, which is best started with “how” or in rare cases also with “why”.

Who should use a case study research methodology?

As mentioned earlier, a case study is a methodology that is popular in the social sciences. Those include economics, psychology, political science, and so on.

Natural sciences and humanities do not fall under this category. However, as interdisciplinary research and teaching are almost everywhere nowadays, it is not impossible that case studies can also be used here. Case studies are therefore a quite common and widely used research methodology.

Critics of this method claim that case studies are too “soft”. This means that they have little explanatory power due to their descriptive research design. This can be countered by collecting unique data and analyzing it empirically, resulting in a “harder” case study.

Moreover, critics would claim that a lack of generalizability is a limitation of case studies. This is true, but only for statistical generalization. Other forms of generalization are possible with case studies but are often not considered.

Differences exist here both between different disciplines and cultural backgrounds. For example, case studies in European management literature must be quite “hard”. In the United States, on the other hand, case studies are often written quite “soft” and rely on the storytelling and interpretative abilities of the authors.

Especially for a dissertation, a case study is a great option. Depending on your data collection possibilities and methodological training, a dissertation can move freely on the spectrum from “soft” to “hard”.

Conducting a Case Study in 6 Steps

Now that you have all the background information, let’s move on to the 6 steps you can follow to write a case study.

I mainly rely on the work of Robert K. Yin and the 2014 version of his book “Case Study Research: Design and Methods”.

You can find the book linked below the video and in any well-organized university library.

Whenever you want to use the case study research methodology in academia, you should refer to at least one source in your methods chapter that has established generally accepted rules for the process. In Yin’s book, you will also find an overview of the most important sources for each research discipline.

Planning your case study #1

First and foremost, you need to decide that you want to conduct a case study. But that’s not enough. You should carefully consider why a case study is preferable to other methods.

• Why is a literature review, a survey, or an experiment unsuitable?
• What are the advantages of a case study in your situation?
• Is a case study even possible with your resources?

You should have an answer to these questions and discuss them with your supervisor. Planning also includes formulating a research question.

To conduct a case study you need a relevant research question. Start with the question word “how” and proceed slowly. There are two possibilities:

• Case-specific research question (e.g., “How does Volkswagen respond to hate speech on Twitter in the wake of the Dieselgate scandal?”)
• Generic research question (e.g., “How do large companies respond to hate speech on social media?”)

Both approaches are possible and have their advantages and disadvantages. The research question should always be discussed with your supervisor.

Setting up the research design #2

Now it’s time to set up your research design. The crucial questions here are:

Which method(s) can be used to answer my research question?

And:

What data do I need for that?

In my example, I could proceed as follows: I construct my case study backwards. I could answer my research question by identifying various strategies in the Twitter replies of the VW Group.

I could do this by collecting a dataset of relevant tweets (e.g., using the hashtag “Dieselgate”) and applying qualitative content analysis.

Can I answer the question differently?

Sure. In theory, I could also interview VW employees and have them answer the question.

Which approach you choose also depends on the possibilities you have to obtain data.

Preparing for data collection #3

Now it’s time to prepare. Just follow these three steps:

Create a literature review

Before you do any research, you have to read. Conduct a thorough literature review that reflects the current state of research.

(And if you are a bit more advanced:)

Is there a theory that can explain your case study?

In this case, establish a theoretical background. You do this by focusing on a theory that helps you understand the phenomenon under investigation. You then discuss your results in relation to this theory.

Identify data sources

Where do I get my data from?

Which interview partners do you need, which social media platforms, which company data? Which archive reports?

Contact the right people

Now all you have to do is get access to the data. Write to interview partners, call archive owners, and so on. Create a table with all your data sources for better overview and keep a diary of your progress.

• Whom have you already contacted?
• Who responded positively?
• Were there any rejections?

This way you can meet your desired timeline and optimize your project management.

Data collection #4

The fourth step according to Yin is actual data collection. Again, this can look completely different depending on your research design.

If you conduct interviews, I just apply the method as you would normally do. Remember that a case study is a methodology and not a method.

This means that you are flexible in the methods you choose.

At this point, the literature review should be completed and already written up.

Data analysis #5

The most work when conducting a case study awaits you in the analysis. In “hard” case studies, of course, a little more than in “soft” ones. When analyzing your data, follow a few guiding questions.

• How can your data be described?
• Do the data have special characteristics?
• What patterns can be identified here?

Collect your results digitally and make enough backups. Nothing is more annoying than losing days of work. Use software wherever possible, because you are not the first person to conduct such an analysis. Smart software solutions make pretty much every research method easier.

Interpreting the results #6

Finally, you filter the important results from the unimportant ones and present them “from general to specific” in the results section of your manuscript.

These 3 elements are essential for an outstanding case study:

• Figures (e.g., flowcharts, bar charts, or pie charts)
• Tables (e.g., with absolute or relative values of your analysis; results of statistical calculations such as frequencies or correlations)
• Explanatory text between the visual elements that shows the reader which of the results are particularly noteworthy

In another chapter, you discuss the results in relation to:

• Your specific case
• General conclusions or implications (to theory)

Note that the results of a case study are not generalizable in a statistical sense. However, other generalizations are possible if your reader is willing to make some judgement calls.

For examples, this means that you should not draw conclusions about all other car manufacturers from VW. However, you can advise the reader to transfer the findings onto a another case if they are willing to accept that this case is similar enough to VW.

Moreover, case studies are great if researchers want to develop new theory. This is why case study research methodology is often combined with techniques from grounded theory.

Have you ever wondered what the distinction between dependent and independent variables is?

Then you’ve stumbled upon digital gold.

In this article, I will explain to you shortly but precisely what the difference between dependent and independent variables is and what function they play in your quantitative research design.

If you’re still interested after that, I will go a bit more in-depth and explain why this designation of variables in the context of survey studies and other methods is often not correct and how to correctly describe them.

Why do you need dependent and independent variables in a quantitative research design?

In a quantitative research design, your goal is to test a theoretical relationship. One of the building blocks of theory constructs that consist of variables.

In order to test a hypothesis in a quantitative research design, you must first determine the variables of that hypothesis and ensure that you can measure them.

As the name suggests, variables can change. They can experience various forms of change, for example, changing human behavior such as the tendency to choose more organic fruit at the supermarket.

Similarly, a variable can vary by location, such as in counties with the highest subsidies on organic fruit. Moreover, a variable can change over time, such as a fruit vendor’s profit per quarter.

Independent variables are the variables that are manipulated or changed in order to observe the effect on the dependent variable. In our example, the independent variable would be the type of fruit (organic vs. non-organic) and the dependent variable would be the number of fruits sold.

Variables in hypotheses

A hypothesis typically includes two variables and their relationship to each other. It’s about how one variable affects the other, i.e. the hypothesis expresses a relationship between cause and effect.

H: Eating a banana immediately after exercise increases muscle regeneration.

In this hypothesis, eating a banana is the cause. This is the independent variable.

Increased muscle regeneration is the expected effect. This is the dependent variable.

Independent Variables

Ok, and why is the first variable now independent?

That’s because this variable can be varied arbitrarily. The variable could also be “drinking a protein shake”. Or: eating two bananas.

In that sense, this variable does not depend on other variables – hence independent variable.

Dependent Variables

The second variable, which represents the effect, is called dependent because the value of this variable depends on the cause.

In reality, however, independent and dependent variables are often not as clear-cut as they may seem. In many real-world situations, multiple variables can be both independent and dependent at the same time, depending on the specific research question and the level of analysis.

For example, in a study looking at the relationship between income and education, income could be considered the independent variable at the individual level, but when looking at the relationship at the societal level, education could be considered the independent variable.

Additionally, it’s important to note that the cause-and-effect relationship between independent and dependent variables can be difficult to establish as it may be influenced by other factors. That is why we need experiments.

Dependent and independent variables in experimental designs

These terms originated in the context of scientific experiments. To test the example hypothesis, you could set up an experimental design that examines a sample of athletes. Under supervision, each participant receives a banana – this is how the independent variable is measured.

Then they can let off steam during the workout and afterwards their muscle regeneration, the dependent variable, is measured.

In this experiment, the independent variable can now be varied, this is also referred to as “manipulation”.

3 example experiments

Example #1

For an experiment, the temperature inside a car is changed. People sitting in the car indicate how they feel at each temperature. Temperature is the independent variable. The dependent variable is the reported well-being of the occupants.

Example #2

You want to investigate how smartphone usage affects heart rate. The independent variable is smartphone usage and the dependent variable is heart rate.

Example #3

You want to find out how time spent working from home affects the work performance of your employees. In this example, the independent variable is time spent working from home and the dependent variable is work performance.

Dependent and independent variables in cross-sectional studies

In an experiment, data is collected at different times. This allows the study director to manipulate the independent variable.

In studies that only collect data from different individuals at a single point in time, this is not the case. This is also referred to as cross-sectional studies. An example of this is an online survey.

Here, variables cannot be manipulated and thus no causal relationships can be tested. The terminology of independent and dependent variables would therefore be incorrect. Nevertheless, everyone knows what is meant when you talk about it, but if you want to be completely correct, you can use

Predictor variable or prognostic variable instead of independent variable and

Response variable instead of dependent variable in speaking or writing.

After all, predictions about variables are also made in cross-sectional studies – only causality is not assumed.

If you’re interested, the difference between causality and correlation has been discussed in another video that you can find linked on the top right.

For experiments, this designation also works. You could therefore theoretically always use the designation predictor variable and response variable – independent and dependent only in the context of experiments (Field, 2015).

Measuring

Of course, the method is of the utmost importance here. For a quantitative study design, as already mentioned, experiments or standardized surveys such as online surveys can be used. But also collecting sensor data or other measurements or collecting documents, texts or social media data can be the basis for a quantitative research design.

Each method now produces data that has different levels of measurement or scale levels. These scale levels, if you will, decide the quality of your variables and what statistical operations are available to you to test your hypothesis.

Reliability and Validity of measuring variables

It is important to note that in order to make accurate inferences and conclusions, variables must be measured in a reliable and valid manner. Reliability refers to the consistency of measurement, while validity refers to the accuracy of measurement. For example, if you are measuring the number of organic fruits sold in a supermarket, it is important that the counting method is consistent and accurate.

If your variables are metric scaled, i.e. they consist of numerical values (e.g. number of bananas eaten), then the relationship (depending on the data, causal or correlational) between the dependent and independent variable can be calculated using a regression analysis.

Variables for regression analysis

Exactly how to do this is a topic for another video – but at least you now have a small glimpse of what you can do after determining and measuring your variables.

I hope this was helpful and if you want to delve further into this topic, I recommend the textbooks by Andy Field.

Are you trying to wrap your head around empirical research methods? You keep reading about quantitative and qualitative research methods but what is the difference?

Don’t panic, you will get an easy answer to this question over the next few minutes.

In this video you will learn about the 5 crucial differences between quantitative and qualitative research methods. So not only will you be able to distinguish these two approaches in your sleep, but hopefully you’ll get some valuable ideas for choosing the method for your next scientific work.

What is empirical research?

Before we dive into the differences between qualitative and quantitative research methods, we need to clarify what they have in common. They are both part of what is called empirical research. This means, in a very simplified way, the systematic collection of data to gain knowledge.

In a figurative sense, the “experiences”, i.e. the data, are used to derive new knowledge. This always involves some sort of observations in the real world or with real people. For example, social sciences but also the natural sciences are dominated by empirical research.

In contrast, a mathematician or philosopher, who reaches insights only by thinking and logical reasoning, is not considered empirical. That does not mean one is better or more scientific than the other. They are just different ways to gain scientific knowledge.

In philosophy of science, the question about how researchers can gain knowledge in the first place is subject to epistemology. Look up this term if you want to dive deeper into this.

The two paradigms (quantitative and qualitative)

The distinction between quantitative and qualitative research is the result of more than one century of debates between researchers about what science is and how research should be conducted.

In a nutshell, two positions emerged based on different philosophical assumptions about how the world is made up (in philosophical terms, this is called ontology). The more sciency scientists were convinced that there is an objective reality that is independent from us humans (in philosophical terms, this world view is called realism). This reality can be measured by using numbers and statistics (in philosophical terms, this world view is called positivism). This was the starting point for quantitative research.

As this thinking was dominant in the academic landscape of the early 20th century, emerging disciplines adopted the same principles. For example, psychologists studying the inner workings of the human mind applied the same natural science-based approach.

Later, this natural science approach to social science led to the emergence of other positions that were not at all happy with this type of thinking.

“Wait a minute,” they said, “don’t you think social phenomena are constructed differently by different groups and individuals?”

They had a point.

A stream of research emerged that was not interested in statistically computing the personality traits of teenagers or how the management of an IBM computer worked (the relevant philosophical terms here are constructivism and interpretivism).

One milestone in advancing the qualitative paradigm was Glaser and Strauss’ (1967) Grounded Theory Methodology. It reversed the deductive logic of testing theory to an inductive logic of generating theory. Until this day, the Grounded Theory Methodology is an important part of qualitative research.

The 5 Differences between quantitative and qualitative research

Now, let’s look at 5 different characteristics of both approaches.

#1 Object of study

Quantitative research investigates large samples and assumes more of an outside view

Qualitative research investigates small samples and assumes more of an inside view

#2 Types of data

Quantitative research relies on numerical data (“hard” and replicable)

Quantity as a characteristic means measuring the amount of a certain unit. Quantitative data is therefore available in large quantities in the best case. But not only the quantity is decisive, but also the quantifiability. This means that these data are unambiguous.

If a test person answers question 1 in a questionnaire with the answer option 3, then this data is unambiguous.

A measurement method of a sensor or the number of comments on a Twitter message is also unambiguous. In the language of statistics, different types of data can be categorized as nominal, ordinal, and metric.

And this brings us to the great advantage of quantitative data: You can apply statistical techniques to analyze it.

Qualitative research relies on linguistic data (“soft” and unique)

Qualitative data are always context-dependent, i.e. the basic conditions of your data collection environment must be taken into account during the investigation. This is particularly important in case studies.

The great advantage of qualitative data is that they can be rich. As a researcher you really get underneath the surface level and can attempt to explain “why” things are the way they are.

#3 Analytical Logic

Quantitative research typically measures things and follows a deductive logic

Qualitative research typically interprets things and follows an inductive logic

Deductive logic means using a general (mental) model to form a specific conclusion. (For example, using a theory to derive hypotheses)

Inductive logic means starting from a very specific example or case and forming a general conclusion. (For example, using interview data to generate a theory)

#4 Methods (Examples)

Quantitative research uses methods such as surveys, experiments, or simulations

Qualitative research uses methods such as interviews or observations

#5 Theoretical Contributions

Quantitative research typically contributes to theory by testing propositions or hypotheses

Qualitative research typically contributes to theory by introducing new concepts, models, or mechanisms

How to choose between quantitative and qualitative methods

If you are faced with the challenge of conducting empirical work, then choosing the right research design is one of the most important decisions of all. It all comes down to your research objectives, i.e. answering your research question(s).

Which research method can best achieve this goal?

This question is the basis of your consideration. I would like to make it clear that scientific work does not have to be empirical. In many disciplines this would not make sense at all.

Moreover, most scientific disciplines welcome both quantitative and qualitative research. In fact, many researchers see great value in combining quantitative and qualitative research methods. This is called “Mixed Methods”.

The next step you can take is grabbing Creswell’s (2014) classic textbook and dig deeper into what you just learned.