At school, if you studied science, you probably came across the concept of a research hypothesis. This is a proposed explanation for a phenomenon, developed so that you can investigate it through experiments or research. It is likely that in doing experiments in science lessons, you started with a hypothesis, and then investigated it.
This approach is strongly encouraged in scientific research. It is also encouraged in academic social science research. However, in more informal settings, the approach is less rigid—and we think perhaps may be the worse for that.
In market research in particular, we often see clients launching themselves into survey or questionnaire design without a clear hypothesis to investigate. This means they often struggle to analyse their findings. They have lots of interesting data at the end, but no real answers to their questions.
Research design and philosophy
Let’s start by saying that it is perfectly legitimate to start research without a hypothesis—if that is appropriate.
Sometimes, researchers start by collecting data, and looking for patterns in those data. From this, they develop theories and hypotheses. This approach is called inductive research. It starts from the specific (the data) and works its way up to the general (the theory). It is ideal when you have no idea what patterns you might see in your data, or what you might be looking for: in other words, when the purpose of your research is to develop a theory. It is often used to answer questions that start ‘how…?’ or ‘why…?’, such as ‘how can we support young homeless people?’.
Deductive research, by contrast, starts with a general theory, and gathers specific data to test whether that theory seems to be true. It is therefore most often used when you have an idea about what is going on, and want to test it. Deductive research studies include exploring whether increased sentences act as a deterrent in criminal justice.
An inductive approach is therefore perfectly legitimate in research. Why, then, are we arguing that it should be avoided in market research?
Interesting data—or important data?
The answer is that in practice, life is seldom that simple. What often happens in research is that you test a theory (deductive approach) and find that your data do not quite fit. You therefore have to look at your data again and develop a new theory that fits those data (inductive)—and then go back and test it again (deductive).
It is this re-testing that is key to understanding why you need to develop strong hypotheses in market research. In academia, every outcome is intrinsically interesting because it adds to the broad state of knowledge about the topic. You can publish an article proposing a new theory, and suggesting that you will test it in future, and you can consider that the end of your research project.
In business, however, you only get value from your research once you know that your theory is accurate. You need to be able to apply it in practice, and that means you have to test it. You may need to start with a very broad-brush approach to data collection and hypothesis development—but you must end up with a specific test of your theory before you can apply it more generally.
Developing a strong research hypothesis in market research has several advantages. First, it avoids the problem of reaching the end of your research only to find that you have learned nothing very new, or that you have no actionable data. It enables you to focus on the issues that are most important to you, and therefore get the most out of your investment in research.
Second, it makes it much easier to develop questionnaires or surveys. You can focus on the questions that are really necessary to test your hypothesis. Most people have fairly short attention spans, so you want to keep market research surveys short if possible. It also means that you have an easy framework for evaluating your research: against the hypotheses.
Developing a strong hypothesis
It can be challenging to develop a strong hypothesis. However, it is well worth the investment of time, because it forces you to focus on what you really want to get out of the research, or the end product. Remember that you do not have to be sure that your hypothesis is correct. It only needs to be testable, and to focus on what matters.