<!-- canonical: https://catalini.com/research/negative-citations/ -->
<!-- source: catalini.com · author: Christian Catalini · license: all rights reserved -->

---
title: "The Incidence and Role of Negative Citations in Science"
authors: ["Christian Catalini", "Nicola Lacetera", "Alexander Oettl"]
venue: "Proceedings of the National Academy of Sciences"
year: 2015
area: science
featured: false
links:
  pdf: "https://www.pnas.org/content/112/45/13823"
preview: "/images/papers/negative-citations.png"
abstract: "The first large-scale evidence on negative citations in science. They are common enough to matter, tend to come from scholars who are intellectually close but geographically distant, and target visible, influential work rather than weak papers — suggesting they are part of how science polices quality at its frontier."
---

Whereas colocation - whether temporary or not - is disproportionately responsible for idea recombinations between scientists that belong to the same communities or institutions, diffusion through publication should theoretically be less constrained by pre-existing relationships. Nevertheless, since journal rankings and citation counts are extensively used to allocate attention, diffusion through peer-review is still subject to substantial frictions.

"negative citations may indeed play aspecial role in science"

In the paper we test how effective citations are at carrying information by focusing on citations that limit or challenge past findings. After identifying such "negative citations" using natural language processing techniques, we explore whether papers that are negatively cited actually incur a citation penalty or not. Once we accurately match them with control papers of similar quality, we do not observe a decay in their citation rates, possibly because most negative citations go unnoticed, take a long time to diffuse, or are part of the natural evolution of science. In the paper, we discuss how online repositories such as Google Scholar or PubMed could use our approach to facilitate the discovery of related but conflicting information, accelerating the process through which scientific knowledge is improved upon over time.

[Download the paper](https://www.pnas.org/content/112/45/13823.full.pdf)
