Measure the exploration factor of research

<p>Assuming that there are <a class="wikilink" href="/problems_with_citation-based_metrics/">problems with citation-based metrics</a> in scientific work, the authors of [<a class="litnote tooltip" href="/literature/@bhattacharya2020">@bhattacharya2020<span class="tooltiptext">Stagnation and Scientific Incentives</span></a>] propose an alternative metric: exploration factor. </p> <p>They argue that it is possible to measure the novelty of research by doing a text-based analysis on the published papers. They can count the number of new terms that appear and on which ideas they build insights. </p> <p>They call this the "edge factor" and could be complimentary to the "impact factor". Perhaps a scientist focuses on novel ideas that are still to immature to generate "impact" but that are part of the <a class="wikilink" href="/scientific_exploration_and_play/">Scientific Exploration and Play</a>. </p> <p>The authors of [<a class="litnote tooltip" href="/literature/@bhattacharya2020">@bhattacharya2020<span class="tooltiptext">Stagnation and Scientific Incentives</span></a>] describe a possible method, also using <a class="wikilink" href="/machine_learning/">machine learning</a>, to identify papers that have a high novelty even though this may not be reflected in the number of citations. </p> <div class="admonition image medium"> <p class="admonition-title">Image</p> <p><img alt="Novelty and impact in pub med publications" src="/images//Screenshot 2021-02-23 at 20.38.06.png" /></p> </div> <p>The idea (shown in the plot above) is that there is a correlation between novelty and number of citations, but by adding another dimension to measuring output, it is possible to justify funding research that not necessarily belong to the Region 1 (i.e. high-impact research.)</p>