What happens to computer science research after it is published? Tracking CS research lines

Abstract

Are computer science papers extended after they are published? We have surveyed 200 computer science publications, 100 journal articles, and 100 conference papers, using self‐citations to identify potential and actual continuations. We are interested in determining the proportion of papers that do indeed continue, how and when the continuation takes place, and whether any distinctions are found between the journal and conference populations. Despite the implicit assumption of a research line behind each paper, manifest in the ubiquitous “future research” notes that close many of them, we find that more than 70% of the papers are never continued.

Author‐based analysis of conference versus journal publication in computer science

“…In other words, the collaboration in conference articles does not necessarily lead to that in journal publication, or vice versa. Especially if we assume that coauthors from a conference article would work together for developing their article into a journal submission, the mean proportion of overlapping coauthors for conferences (that is, 39.81%) indicate that many conference articles may not be extended into journal articles, possibly supporting the findings of previous studies (Bar‐Ilan, ; Vardi, ; Wainer & Valle, ).…”

“…If we assume that a conference article is turned into a journal article with exact or similar title words (for instance, Bar-Ilan, 2010), this observation implies that many CS conference articles may not be extended into journal articles (or vice versa). Previous studies found that the ratios of conference articles that transit into journals have been 25~33% in computer science (Bar-Ilan, 2010; Montesi & Owen, 2008;Wainer & Valle, 2013), which is comparable to the mean ratio of overlapping title words for conferences (26.21%) in this study. A note is that these preceding studies used articles as a unit of analysis, not individual authors.…”

“…In a study of sampled CS scholars, around 25~33% of CS‐related conference articles were found to lead to journal publications (Bar‐Ilan, ), supporting the similar findings from an interview with 22 editors of 13 CS journals and 122 authors (Montesi & Owen, ). Recently, a study surveying 200 articles reported that about 26% of conference articles were extended or republished in journals (Wainer & Valle, ). A similar conference‐to‐journal transition ratio (30%) was reported for research publication in the field of computer vision (Eckmann, Rocha, & Wainer, ).…”

The first impression of conference papers: Does it matter in predicting future citations?

“…According to Wainer, Przibisczki de Oliveira, and Anido (2011), among all articles published by the ACM in 2006, 39% of the references were conference papers and 41% of the references cited more than 10 times were conference papers. Several studies examined differences between conference papers and journal articles: the differences in citations (Freyne et al, 2010;Vrettas & Sanderson, 2015), the patterns of reproducing conference papers for journal publications (Wainer & Valle, 2013), and the patterns of citing references (Wainer et al, 2011). Most studies confirmed the value of conference papers in comparison with journal articles.…”

A scientometric analysis of 15 years of CHINZ conferences

“…The citation and self-citation rates show CHINZ as roughly comparable to the much larger APCHI conference. Wainer and Valle [19] report approximately half of the computer science papers sampled from the ACM DL from 2003 were uncited in the ACM DL: their conference subset was 44% uncited. The 48% uncited rate for a small regional conference such as CHINZ (Table 3) seems roughly consistent with their results: particularly with the limited audience for CHINZ 2001-3.…”