Did you know what “exploratory research” is? Did you know that it is one of the most useful tool to be still more efficient in science? Did you know that the kind of tools it requires are absolutely different from those used in confirmatory research?
All that is exactly what I’d like to talk about today.
Two complementary approaches
First things first. This is my attempt to summarize both concepts:
Exploratory research is the stage of the research process that aims at connecting ideas as to unveil the “why”s of potential cause/effect relationships. This occurs when researchers get started at understanding what they are actually “observing” when in the process of building cause/effect models.
Confirmatory research (a.k.a. hypothesis testing) is where researchers have a pretty good idea of what's going on. That is, researcher has a theory (or several theories), and the objective is to find out if the theory is supported by the facts.
How they complement each other?
The essence of all this is that exploratory and confirmatory research are two complementary components of the same goal: to discover relevant findings in the most efficient, reliable, replicable, applicable manner.
They are complementary because researching always starts with a complex unknown issue. An issue that must be observed even before finding which hypotheses must be further confirmed.
As J.W. Tukey states in his renowned work, we need both exploratory and confirmatory approaches because ...
“Finding the question is often more important than finding the answer”.
Another key question now is whether exploratory analysis task should be (or not) worried or not about false alarms. In that sense, let me show you what political scientist and statistician Andrew Gelman of Columbia University in New York and author of one of the most renowned blogs on the matter says in this excellent article recently published in Nature magazine:
“… In this approach, exploratory and confirmatory analyses are approached differently and clearly labelled. … Researchers would first do two small exploratory studies and gather potentially interesting findings without worrying too much about false alarms. Then, on the basis of these results, the authors would decide exactly how they planned to confirm the findings.”
The exploratory analysis task should thus provide potential relationships and novel relevant questions that feed the classical confirmatory process focused on minimizing type II error, that is, failing to assert what is present, a miss.
Want to read more...?
Fruitful Collaborations: real success cases at applying automated EDA.
How To Design an Efficient EDA Study : the 3 key ingredients of the EDA recipe.
5 kinds of exploratory questions you may be asking yourself : from our experience ...
Data Exploration in Biomedical Research : promising approaches to the use of EDA published in papers.
Advanced Characterization of Experimental Groups using Automated Exploratory Data Analysis : how to identify significant differences between the study groups.
Preparing Your Data For AutoDiscovery: Table-Like Structures : some funny superfast tricks to prepare your data just by using Microsoft Excel.