Friday, February 11, 2011

Impact factor: Conferences vs. Journals

This week, my boss (Hermann Hellwagner) entered my office with the following paper:

Jilin Chen and Joseph A. Konstan. 2009. Conference paper selectivity and impact. Commun. ACM 53, 6 (June 2009), 79-83. DOI=10.1145/1743546.1743569

Reading the abstract ...
Studying the metadata of the ACM Digital Library (, we found that papers in low-acceptance-rate conferences have higher impact than those in high-acceptance-rate conferences within ACM, where impact is measured by the number of citations received. We also found that highly selective conferences—those that accept 30% or less of submissions—are cited at a rate comparable to or greater than ACM Transactions and journals.
... I got confirmation that submitting papers to high-quality conferences is still a good thing and - next to that - I was wondering whether someone has done something similar for the IEEE Computer Society (doing a quick/simple search in CSDL, I couldn't find it).

Many (young) researchers are confronted with the requirement (coming from their institution, Ph.D. supervisor, etc.) submitting papers to journals rather than conferences. This is often justified with the higher impact factor of journals (compared to conferences). Another reason - I think - is that submitting papers to journals is much cheaper than conferences due to travel expenses. However, I believe attending a high-quality conference and presenting a paper there, discussing with other researchers on-site/face-to-face and socializing with them (having dinner etc) is much more worth, specifically when someone enters a new domain (btw. successful journal publications will then follow automatically, no worries).

While I certainly agree with the conclusions from the article mentioned above, I think one important aspect is missing. If
papers in highly selective conferences—acceptance rates of 30% or less—should continue to be treated as first-class research contributions with impact comparable to, or better than, journal papers
then what's the reason of having journal publications (or journals at all ...)? Both conference proceedings and journals are accessible through digital libraries offering search facilities better than the office desk with a huge stack of paper. Thus, having it printed somewhere can no longer be the reason. Ah yes, there was this other issue with the travel expenses which is acknowledged but probably it's worth thinking about new ways of remote participation (e.g., I'm attending ACM MMSys'11 end of February which enables that) or other means of interaction, e.g., through social networks. The latter is something that is currently being explored within the IEEE Computer Society, specifically within Computing Now but also the newly established Special Technical Communities (STCs) as mentioned by Sorel Reisman in his IEEE Computer president's message from January 2011.
By the end of 2011, we'll have launched pilot Special Technical Communities in social networking, cloud computing, gaming, education, software engineering, and green computing.
If you're interested in joining the first one - social networking - please contact me via
Post a Comment