Scientific research is often about finding the answers to pressing questions. Equally so, it is often about testing ideas with no specific question in mind. Through these two different approaches, science can cover the theoretical and practical ground it needs to produce solutions to society’s problems.
The distinction between exploratory and confirmatory research goes back to John Tukey’s 1980 argument that science needs both types of research approach, and it is now central to the case for preregistration.
But what is the difference between exploratory and confirmatory research, and how do the two work together in the research cycle? And what goes wrong when the line blurs?
What is exploratory research?
Exploratory research is all about, well, exploration, as the name says. This may involve testing new ideas, patterns, or hypotheses in the planning and data collection phases of research.
The purpose of exploratory research is to investigate a scientific problem that is less understood or covered in the available literature. This may be because the subject is an emerging focus or has been previously underacknowledged. Ultimately, exploratory research looks to build a foundation from which further analysis can investigate such subjects.
Typical methods of exploratory research include qualitative and interpretative/observational work.
What is confirmatory research?
In contrast, confirmatory research is about testing a specific hypothesis in advance, with a fixed study design and analysis plan.
Confirmatory research is typically conducted when there’s research around a subject, and for this reason it usually follows exploratory research.
Importantly, confirmatory research has stricter “boundaries” than exploratory research. Whilst the latter is more open to experimentation, confirmatory research has a stricter responsibility to science: to validate the legitimacy of a study’s specific research question and structure. Therefore, parameters outlined in advance must be followed throughout the course of analysis.
The key differences
Here are the key differences between exploratory and confirmatory research.
| Study element | Exploratory | Confirmatory |
| Aim | Exploring and testing ideas, patterns, or hypotheses | Testing a pre-conceived hypothesis with a fixed study design and plan
|
| Starting point | Emerging or understudied problem or subject | A priori hypothesis, a study/group of studies, or a dataset |
| Flexibility | Open, unstructured | Structured, pre-defined |
| Type of conclusions | Preliminary or inconclusive | Definitive |
| How results should be reported
|
Qualitative and interpretive analysis | Quantitative and statistical analysis |
How the two work together in the research cycle
Exploratory and confirmatory research are not oppositional methods of analysis. In fact, they are complementary parts of the research cycle.
Exploration generates new hypotheses; in turn, these new ideas or questions can be tested by confirmatory analysis using pre-defined methodological structures and the new data gathered through these types of study.
And, in a cyclical way, by answering questions to pressing scientific issues, confirmatory analysis then frees up more time for researchers to explore new ideas, helping the research cycle sustain itself.
What goes wrong when the line is blurred
Although exploratory and confirmatory research are distinct methods of analysis, there are instances in scientific research conduct where the two become problematically blurred. This is seen in the phenomenon known as HARKing.
HARKing, or “Hypothesizing After the Results are Known”, is the practice of presenting a post hoc hypothesis as if it was an a priori hypothesis. Researchers may do this to make the data and results of their study conform to the presented hypothesis (and therefore appear as statistically significant). Such practice can be symptomatic of the wider issue of publication bias in scientific publishing, or the tendency of journals and publishers to prioritise and publish studies demonstrating positive over negative or inconclusive outcomes.
But HARKing means that exploratory and confirmatory research become mixed and confused. Testing different hypotheses belongs to exploratory research; gathering data on one pre-defined hypothesis belongs to confirmatory research. But changing your hypothesis after gathering your results blurs the lines between the two types of research and runs the risk of reporting false positives. And, ultimately, this distorts the scientific literature.
How to keep them separate in practice
Preregistration is the main way to keep your exploratory and confirmatory methods separate. This is the formal process of specifying your research plan in explicit detail in advance of performing the study. This can be seen as “locking in” your hypothesis.
Another way to keep them separate is to post your research as a preprint first. On Preprints.org, researchers can share exploratory findings as preprints, clearly describe their exploratory nature, so others can test them. Furthermore, the Preprints.org Hypothesis article type gives authors a dedicated format for proposing new ideas for others to investigate.
When a confirmatory follow-up is complete, versioning and linked preprints create a transparent trail from initial exploration to confirmation. Importantly, Preprints.org welcomes the results whether or not the hypothesis is supported.