10

I'm working on a problem A, and recently have two small results:

  • The first result is using technique X for A1, which is a particular instance of A. A1 is a long-term problem, but nobody has used X for it before, and I achieve slightly better performance with it. In other words, it is only an incremental result.
  • The second one is solving A2 using technique Y. A2 is totally new problem that nobody has touched before. However, the method is extremely expensive, and only works on toy programs.

Since both two results have clear weaknesses, and we don't know how to improve them, my advisor is suggesting me to put the two results in one paper to submit to a major conference, since each of them is not enough contribution.

It is only a suggestion and I can decide whether to do it, there is no problem with the advisor. I'm only confused if it is a good idea to do so. My concerns are:

  • A1 and A2 are just loosely related. It is difficult to tell a story that connects these two problems.
  • The first method is only applicable for A1, the second method is only applicable for A2, while they are different instances of the problem A. Putting them together poses another weakness: the usability of each method.

The other option is of course: submitting them separately to lesser conferences. We only want to do this if the first option is not promising.

Some context: my advisor is a leading expert on X and Y, but knows very little about A (but wants to explore). On the other hand, I did my PhD on A, and only have some minimal background on X and Y. I'm a new postdoc, hired to work on A using the tools for X and Y that my advisor had developed. Therefore, none of us are 100% sure about what to do in this situation.

  • 4
  • 1
    Could you say what field you're working in? In my field (mathematics), this strategy would be a poor one: if you have two results that individually could not be published in a certain venue, then unless they fit together synergistically to make a stronger result there is no way that a paper which contains the two results together could be published in that venue. Also I'm confused as to why two small results would be suitable for publication in a major conference. Finally, you have a PhD and an advisor? I'm a bit confused about that. Is your advisor a coauthor? – Pete L. Clark Oct 1 '15 at 7:51
  • Let me follow up by saying that there is no inherent problem in publishing a paper containing loosely related results: that happens commonly enough; one just does not (in my experience, in my field) gain any advantage from doing so. In my experience, papers are evaluated primarily by the strength of the best result in the paper and secondarily by how much space they spend establishing that result. – Pete L. Clark Oct 1 '15 at 8:03
  • So if you have two minor results of about the same strength and length, you might as well try to publish them together: it's less fuss for you. The other alternative is to tuck a minor result into a paper which has other, better results. As long as this doesn't take up an inordinate amount of space, referees are unlikely to object to the inclusion of the result, even if they don't care about it at all. This is a good way to make sure that all of one's work gets published; certainly it's not as good if you want separate credit. But if it's truly minor, how much credit do you want? – Pete L. Clark Oct 1 '15 at 8:05
  • 1
    @PeteL.Clark I'm in Software Engineering, which is easier and less rigorous than maths. I'm a new postdoc, and my advisor is my boss. Maybe I used incorrect words due to my poor English. – qsp Oct 1 '15 at 12:04
15

Just reading and trying to understand your question made my head spin, most likely due to use of generalization (X, Y, A1, A2), but not only that. Therefore, IMHO, the safest approach to solving this dilemma is to separate results into two papers. The rationale for this approach is the following.

  • Each paper would have a clear and, thus, easily understandable scope, goals, methodology, results and their interpretation. That would significantly improve readability and overall quality of each paper and chances to form a positive impression.

  • Since the problems A1 and A2 are "loosely related", shoving them into a single research study simply doesn't make sense from an architectural / logistical / layout perspective. It would be very difficult to present two sub-studies in a synergistic manner ("tell a story", as you put it).

  • Based on the above-mentioned points, I don't see how combining the studies would make the resulting "study" of higher quality, thus, warranting a publication in a more major conference. Quite the opposite most likely would be true (due to coherence and, thus, readability issues).

  • Your advisor's recommendation in regard to improving your work is quite strange, to say the least. It is based on the assumption that combining two studies of different scope/topic with weaknesses would make the resulting "study" less weak. If anything, it would make it even weaker, since weakness is one of multitude of characteristics, exhibiting additive properties.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.