Incentives in Science
Perhaps the safest characterization of the discipline of economics is "the study of the ways in which decision-makers respond to incentives".
At the heart of understanding the outcome of the scientific enterprise for society lies the examination of scientists' behavior. Scientists are typically creative and hard-working individuals, who may not be representative of the average person, but they are also strategic human beings who respond to incentives.
In recent years, concerns have been raised that - in several disciplines - published research contains many findings that cannot be reproduced. This indicates that reforms may be needed to improve the standards of research reproducibility. Such reforms have been proposed in several disciplines, most notably the biomedical fields and psychology. But how can we test scientifically the potential effects of such reforms?
Methodological Contributions of Economics
Institutional reforms work by altering incentives. What economics, in particular, can bring to the table is its rigorous methodology in studying incentives and social interactions of agents that act strategically. Very often, these social interactions have unintended consequences at the systemic level. Economic mathematical modelling can help us capture and understand such consequences.
Game theory uses mathematical representations to understand how interactions among different individuals shape economic and social outcomes. For this reason, it is particularly useful for studying the behavior of researchers, often competing with each other for discovering novel effects and for scientific prestige.
Experimental Economics tests theories of social and economic behavior scientifically and it may also be used to examine the effects of different policies. It is therefore particularly useful for serving as a 'wind tunnel' for testing the potential effects of different scientific policies and interventions.
Open Science and Reproducibility
This era of unlimited computational capacity brings opportunities and some challenges for the advancement of scientific knowledge. On the one hand, the ability to conduct calculations at a great scale allows for an unprecedented level of application of scientific principles. At the other hand, it bears the risk of making traditional statistical inference meaningless. The concept of 'scientific discovery' becomes more difficult to interpret in an environment with extremely low cost of statistical experimentation.
Employing meta-analytic techniques to separate wheat from chaff, using Bayesian techniques and redefining statistical significance are prominent ways to address this challenge. However, maybe the most promising avenue ahead is to be transparent ex ante about the main research hypotheses and the statistical methods that are to be used.
Fortunately, technological developments allow for costless sharing of research methods and tools. As this type of discussions mature, tools for sharing material, analysis and coding with the research community will become more popular, and perhaps even part of the technical arsenal to be mastered in the post-graduate curriculum.
Whether these tools and methods will be promptly fostered by the scientific community will depend on the rearrangement of scientists' incentives. Economic theory teaches us that these incentives will have to be structured in such a manner that strategic researchers who wish to advance their careers will choose to use these tools and methods to achieve their objectives. There is great scope for delving deeper into this insight using economic research tools.
Resources:
What, exactly, is Open Science? | The OpenScience Project