See Manuals methodology

How to formulate a research question?

Formulating a good research question is probably the most difficult part of your research project. Most of the time you start with a topic (energy policy, or human rights) or with a general problem (something is wrong with political participation, this production process, or human resource management). Sometimes the topic or problem you initially started with appears to be different from the one that is interesting. Be clear about the type of research question you want to answer. Several types of questions can be distinguished;

Normative questions are about what is allowed or what is good. These questions should not be confused with conceptual questions or descriptive questions (see below). In most cases normative questions implies philosophical (not empirical) research.

Conceptual questions are about the proper/useful/efficient meaning of words; ‘what is freedom?’, ‘what is equality?’ ‘Which types of markets can be distinguished’. Conceptualization is also a central part of research answering empirical questions, although conceptual questions in empirical research are often discussed and answered without explicitly stating a conceptual question.

Empirical questions are about ‘truth’ and ‘observations’. In science the goal of empirical questions is ‘inference’; generalization. There are several types of empirical questions, ranging from relatively simple to relatively complex.

  • Descriptive questions (what is …). These questions are about describing facts, either at one point in time or over time. Once the concept you want to use is not directly observable (like ‘management style’ ‘political participation’) and/or the units of analysis are sets (like ‘consumers’, ‘firms’ or ‘politicians’) description also involves ‘inference’.
  • Relational questions. These studies involve examining the relationship between different variables. They do not necessarily require that the relationship is causal.
  • Explanatory questions (why is…). These questions are about explaining the causes for something. This requires that the relationship between different variables is studied. However, it is not enough to simply find correlations between variables; answering explanatory questions also requires that the cause precedes the consequence and that there is no third variable responsible for the correlation. In a ‘theory’ causes and consequences are connected by referring to a ‘causal mechanism’.

Applied questions are not aiming at finding or creating generalized knowledge, although they may involve some (mainly descriptive) inference. Applied questions are asked simply because people want to solve a specific social, political or commercial problem. Researchers aiming to answering applied questions, apply existing knowledge to solve a real world problem. Applied research is research using some part of the research communities' knowledge (theories, methods, and techniques) for a specific purpose. Applied research is often opposed to pure research (which is here called ‘empirical research’). Applied research questions can only be answered if empirical research questions have been answered first. We distinguish between different types of applied questions;

  • Predictive questions (what will happen if …). Predictive questions are about things that will happen in the (still unknown) future. Using answers to relational questions (‘correlations’) or (preferably) explanatory questions (‘theories’) it is possible to make predictions that go beyond mythical thinking. If an existing correlation or proposition shows that ‘always when X, Y will occur too’, the observation that X is the case will help you to predict Y.
  • Remedy questions (what is the solution to…). Remedy questions are about finding a solution to a specified problem based on previous research. Generally the solution proposed to such a question will be based around a causal relationship that has been established by existing research; ‘if you use remedy X, under circumstances C, Y will happen’. Your job is to summarize this research and show how it is relevant to the problem at hand. In addition you have to argue that the circumstances (C ) are relevant in your case. It will be necessary to include a much more detailed review of existing research than is done for other types of questions.
  • Design questions (how to…). These questions are about coming up with an effective policy or an effective type of organization with a particular goal in mind. Design questions are similar to remedy questions, although the solution you propose may not necessarily be based on existing literature. Rather, you will be informed by the literature and come up with something new.

Since applied questions differ from empirical questions, students planning to answer a design question are referred to the webpage about applied science.

A special (sub-)category of questions are ‘unanswerable questions’. All types of questions can be ‘unanswerable’ too (given the existing level of (your) knowledge). Although it is difficult to say in general terms which questions are ‘unanswerable’, do not hesitate to admit that they are.


Basic readings

  • Babbie, Earl (2004). The Practice of Social Research (12th edition). Belmont: Wadsworth/Thomson. Chapter 4.
  • De Vaus, David (2001). Research Design in Social Research. London: Sage. Chapter 1.

Additional readings

  • Gerring, John (2001) Social Science Methodology: a critical framework. Cambridge, Cambridge University Press. Part II.