Why are we still manually conducting literature reviews for academic papers?
The missed potential of scientific research
In a recent interview during a 24-hour challenge of the Joint Chiefs of Global Tax Enforcement (J5) the FIOD provided insight into which software system they use to find meaningfull connections within data (Wouters, 2018). This automated version of “connecting the dots” has proven itself in finding relevant information. For example, it was successfully used on the Panama papers. The concept may be familiar to you, as it has been used in many tv shows to solve crime or a mystery by following the ‘red thread’. Research is in that sense very similar to investigative work, because both trying to find proof or connections can be a daring undertaking without guaranteed results.
The human element
So, while investigating for your own thesis, how can you be sure that a primal source you find is actually a primal source?
For “there are an enormous number of spillovers with no citations”
Jaffe, Trajtenberg & Henderson. (1993, p. 584).
The assumption is that, while manually performing the task, you will not always be able to know. This is mainly caused by missing links. The traditional way therefore has some limitations, including the fact that for the most part only contributing papers predominantly get published. But what is the risk of never knowing which research initiatives did not make it?
Imagine a neural network showcasing all research initiatives related to several key-words based on associations between them (co-occurrence grouping) or based on predictions of connections between key-words (link prediction). If you would only include positive contributions, how efficient would it be to try and find new/undiscovered topics? It is thought provoking isn’t it? Because it means that the general database is being undermined by not cataloging papers that did not achieve significant results. Time constrained, as most individuals are, it is knowing which subjects not to pick that might just be as valuable as the other way around.
From an artificial intelligence point of view, recommending interesting subjects would be advantageous. Utilizing existing technology to connect the red thread automatically would evolve the way research is performed. The amount of information that could be linked is enormous, think about drafts, papers, books and databases. To question whether research is performed efficiently is an essential part of the improvement of our understanding. During this day and age, we are able to step beyond the isolated bubble of historic mediums and contribute to research fields in our own way.
We cannot continue to store relevant research information separated from the rest of the world. The goal should be to remove these barriers and use individual best practices. Eventually we would form an overarching group that thinks ahead in ways that we cannot achieve independently. Current projects on providing open access, with no strings attached, include:
- CORE: allows it’s users to search more than 135 million open access articles.
- ScienceOpen: offers open access to more than 28 million articles in all areas of science.
- Directory of Open Access Journals: offers two million articles from 9,519 journals.
To pick a subject
Unfortunately, a system as described above has not yet been created for research topics. But you can take comfort in the fact that the overall ‘investigation’ is still performed as a team, where you and your colleagues equally contribute to the advancements of this world. What the future holds? I can only dare say that connectivity will improve our way of preforming research.
Jaffe, A. B., Trajtenberg, M., & Henderson, R. (1993). Geographic localization of knowledge spillovers as evidenced by patent citations. the Quarterly journal of Economics, 108(3), 584.
Wouters, N. (2018, November 08). Witwassers in het vizier na nachtelijk overleg opsporingsdiensten. Retrieved from nos.nl: https://nos.nl/artikel/2258414-witwassers-in-het-vizier-na-nachtelijk-overleg-opsporingsdiensten.html
Scribendi. (2018). 101 Free Online Journal and Research Databases for Academics. Retrieved from www.scribendi.com: https://www.scribendi.com/advice/free_online_journal_and_research_databases.en.html