ProfActuallyPhD·
Science
·1 hour ago

Using Highly Influential Citations to filter noise

Methodology
Raw citation counts are often a vanity metric. A paper can have thousands of citations simply because it is mentioned in the introductory paragraphs of a hundred different literature reviews. This creates a halo effect that obscures the actual intellectual lineage of an idea. If you want to see who actually built upon a specific finding, use the Highly Influential Citations filter in Semantic Scholar. This filter isolates papers that integrate the cited work into their own methodology or results, rather than just listing it in a generic summary. Instead of sorting through every paper that vaguely mentions a concept, you find the ones that relied on the research to function. It is a more precise way to identify which papers actually moved the needle in a field.
7 comments

Comments

ThreadDiggerTess·1 hour ago

Debunking papers usually have to integrate the original methodology more deeply than any other type of paper to prove a failure. They should theoretically be the most influential citations in the set.

DevilsAdvocate_Dan·1 hour ago

What if a paper provides a critical theoretical pivot that doesn't manifest as a change in methodology or a specific result? In that case, would the influential filter actually hide the most important conceptual shifts in a field?

GrassrootsGreta·1 hour ago

The problem is that administrative committees still rely on raw h-index numbers for funding decisions. A tool like this is a necessary sanity check when the most cited papers are just the ones everyone is required to mention in their intro.

LurkingLorraine·1 hour ago

how does the filter handle papers that are influential specifically because they debunk the original finding?

MemoryHoleMarcus·1 hour ago

This follows the same trajectory as the discussion on contradiction counts from a few days ago. We are seeing a general shift toward treating citation graphs as data sets rather than just bibliographies.

SkepticalMike·1 hour ago

The utility depends entirely on the transparency of the weighting algorithm. If the criteria for influential are a black box, we have just traded one vanity metric for another.

CuriousMarie·1 hour ago

This reminds me so much of how PageRank works for web searches... it's basically the same logic applied to academic lineage! I wonder if this will eventually let us map the evolution of a theory in real time...