Artificial Intelligence without BIAS (NoBIAS)
Abstract
Very recently it was proposed that we report a self-citation index s rather than rely on curated scorekeeping. Here, a scholar has an index s if he or she has published s articles each of which has received at least s self-citations.
Defining s similar to h makes it easy for people to understand, and importantly, pairing h and s indices, unlike excluding, appreciates the worth of self-cites, while at the same time it deters excessive tendencies. Reporting, unlike excluding, will allow us to see clearly how the different academic disciplines are resorting to self-citing, thereby making excessive behavior more identifiable, explainable, and accountable. Transparency in this context is not punitive but revelatory. Showing an s-index should push academic institutions and selection committees to think more carefully and less rigidly about the h-index.
A necessary step towards using the metric is validation. Here we propose to use the s-index to measure how self-citation patterns vary according to different fields, academic ages, countries, and institutions.
In this project we want to apply the self-citation index to a large fraction of Web of Science data in the KB infrastructure. The project proposes the following work packages:
- Identification of self-citations in WoS data
- Implementation of a self-citation index in the KB infrastructure
- Systematic analysis of self-citations of different disciplines and journals within WoS (WoS subject area)
- Systematic analysis of self-citations of a random sample of researchers in different career stages
Runtime
2020-01-01 – 2023-12-31Funding
Horizon 2020