A fundamental component of scholarly research is reading, and citing, other people’s work. One effective way to get more context and a deeper understanding of a subject is to look at the citations to, and from, an article.

Science is one long trail of citations. Apart from the very first academic article, published in January 1665, every article has had earlier papers to draw on, to agree with, or to refute. And even the first article must have been based on other people’s thinking. Science works like this:

Smith says A causes B, then Brown disagrees and says, C causes B. Then, a third researcher, Jones agrees with Brown, and we reach a new consensus; or Jones comes up with an entirely new theory. In this way, a trail of related ideas is created. Scientific knowledge is not just about documenting and sharing observations made in the lab; it evolves when researchers read the claims of others, examine their arguments, and develop those theories or produce new ones.

Backward and forward citations

For any article, there are two types of citation. In almost every paper, the author will cite some sources and after its publication, other researchers may come along, like the article, and cite it. These are forward citations – references later in time, from other articles. While backward citations are references to articles published before this article and cited by it.

Both types of citation are useful. Backward citations are valuable because they can help you assess the quality of an academic article by looking at which other articles it cites. Which papers did the authors draw some of their assertions from? Which papers did they disagree with? A reasonable number of citations suggests that the author of the paper has done their homework: they have checked to see what else been written on the subject. After all, if someone has already made the point you are trying to make, there’s no need to write a new article.

We can also assess an article by seeing how often another researcher has read it. For a printed book or article, there is no way of knowing, of course, who has read it. But one great advantage of digital publication of scholarly literature is that it’s easier to measure how often an article has been both read as well as cited, and therefore how influential it is.

A quick way of assessing the quality and veracity of an article is to consider both how many articles it cites and how many article cite it. You would naturally be suspicious about an article that cites no sources, and, roughly speaking, the more articles that have cited this paper, the more interesting or significant it might be. But a simple count of citations doesn’t give you the whole story. Authors might cite an earlier article because they think it’s misleading. And their work might be cited because others are calling its findings into question.

It’s much more helpful to a researcher to be able to understand what each citation is saying. A quick glance at citations suggests that they fall into a small number of categories:

• Building on previous research
• Discussing prior work
• Confirming earlier findings
• Differing from previous work

But classifying each one could get very time-consuming, very quickly. So, is there an easier way?

Citations in context

Here is where Scholarcy’s Comparative analysis feature can help. For every article, Scholarcy not only links straight to the cited source, it also sets that cited source in context, showing how the author might agree with or offer a counterpoint to earlier studies in the field. Here’s an example from an article on post-operative pain management:

For forward citations, there are several tools that can help. If you want to see who has cited an article, PubMed lists all citations of any article:

Similarly, Google Scholar can show forward citations – simply click on the “cited by” icon to see all the articles:

Putting research into context

In this way, a citation trail can be built that shows where any article sits in the wider body of research. This trail may be very simple, with just one paper leading to another, but more typically, science proceeds in a more collaborative way. Several papers indicate a trend in thinking, which slowly becomes the orthodoxy. For example, the association between smoking and lung cancer, one of the most famous discoveries of the 20^th century, was not the result of a single academic article. There were several research strands in parallel, but a milestone article from 1954 by E C Hammond and D Horn reported on a study of almost 200,000 males to determine the link between smoking and lung cancer. This famous article was cited no fewer than 464 times. This is exceptional; a study of articles in Web of Science showed the average number of citations per article was just over 10.

Some articles never get cited at all; an article in Nature (2017) found that around 4% of life science articles published in 2006 had never been cited. The good news is that the proportion of articles published each year that are never cited is getting lower.

By exploring the trail of citations, we can get a rapid overview of where research is going and decide for ourselves which articles are the really innovative ones. And with the help of Scholarcy, evaluating the nature of those citations becomes more manageable.