Thursday, December 13, 2012

Acceptance Rates and Impact Factors

For those of you who do publish at computer science conferences*:

1) On your CV/Biosketch/Website/etc. list-of-publications, do you include conference acceptance rates?

2) If you do, what  is the threshold for you to mention it? (e.g., 50%, 30%, 10% ?)

For those of you who don't publish at computer science conferences:

0) Hey, why aren't you publishing at CS conferences? We're cool people, and we start counting at zero.  

1) On your CV/Biosketch/Website/etc. list-of-publications, do you include journal impact factors?

2) If you do, what is the threshold for you to mention it? (e.g., do you list IFs for startup journals)?

And for anyone willing to share their field / subfield, I'd be interested to hear that as well. I'm planning to assemble this information into a longer post on the topic in a few weeks.

* This includes anything which has archival proceedings, like some workshops, symposia, summer/winter/fall/spring schools, etc. 


  1. 0. Because I'm not a computer scientist (Totally down with starting at zero, though!)
    1. So far, no.
    2. N/A

    Since my answers were pretty useless, I'll offer a tidbit (tidbyte?) that might be more helpful: If you want to know the IF for a journal that isn't covered in the official rankings, you may be able to find it here:

  2. Before getting tenure I listed acceptance rates to help show out-of-area people that conferences can be selective venues. But I always thought this was a little silly and crass and have since removed them.

    Another thing that helps out-of-area people is to give them a word count for your papers. Some people in my department think a conference paper is 4-6 pages whereas a 14-page SOSP or ASPLOS paper might contain more content than many people's journal papers.

  3. I just went through the tenure process in databases. I was told it was best to have acceptance rates for conferences to save the reviewers the time of looking it up and (perhaps more importantly) for the people outside my area to see how tough they were. After looking at some other CVs in the area, I listed anything less than 50%.

    I didn't do impact factors since I failed to find any meaningful criteria. I looked up one thing and it had some journal I'd never heard of as higher than a really good one.

  4. 1) Not in my human-readable CV. But the bureaucracy-readable "bio-data form" my university uses for promotion cases requires them.

    2) 0% — If they want 'em, they get 'em all.

    Also for journal papers:

    1) No. Not even my university bureaucrats want impact factors.

    (I'm a theoretical computer scientist.)

  5. I am not in CS, but I have seen plenty of academic CS CVs. They all include conference acceptance rates.

    In my field, for tenure dossier, I was supposed to list the most important journals in the field and their impact factors in a table. For startup journals, just say IF N/A. I don't list this stuff on CV or the website.

  6. 1. Yes
    2. 40%

    (computer architecture)

  7. I only included conference acceptance rates on the CV that I submitted with my tenure dossier, for the reasons anon@9:37 mentions above. (And I only included the acceptance rates that I could calculate or find.) All of them were below 40%, but I guess I would have included acceptance rates above this too if I'd had that data. I'm at a SLAC so the level of competitiveness of the conference is not a huge factor; as long as it's reasonably competitive and peer-reviewed it's fine. (i.e. publishing in conferences with 10% acceptance rate is not required for tenure and promotion). My subfield is networks.

  8. I always thought putting acceptance rates on a CV was tacky. Then I started sitting in on tenure meetings and seeing how much the acceptance rate matters---it's a pretty good rough guide when you have to do out-of-subfield evaluations. And since I found that acceptance rates below 20% were considered very impressive, and most conferences in my field have acceptance rates in the 15-20% range, I've started putting them on any CV meant for internal review. I still don't put them on my for-the-public CV though.

    I don't have a threshold. I report them for full conference papers, no matter how high they are. I don't bother with poster papers and workshop papers.

  9. I used to be in computer engineering (which, like CS, was all about the archival conferences). Now I'm in bioinformatics, which is all about journal publication (the major conferences make sure that accepted papers are co-published in a journal special issue).

    Since the problem is discipline-wide, we did not put acceptance rates in the CV, but the department letter always explained the archival conference culture.

  10. In my subfield of CS there is rather low correlation between acceptance rates and quality, so I simply use a A, B, C ranking for each conference (we all know which handful of conferences are A in most subfields). If the bureaucrats want to confirm the numbers, presumably external reviewers can be asked to comment on the ratings used.

  11. Thanks all for your comments so far, this is interesting.

    Anon @8:23am -- How do you treat conferences where you have a paper accepted as a "full" paper (i.e., 8+ pages double column), but you happen to get assigned to present it as a poster instead of as an oral presentation? It's still archival, and ACM DL/IEEExplore doesn't note any distinction with it, the only difference is how it was "delivered" at the conference itself.

    Or, alternatively, how do you treat workshops that accept full papers, are competitive, and publish their proceedings along with their affiliated main conference, and there's no distinction in IEEExplore/DL?

    Anon @4:03pm, I like the A/B/C distinction a lot. I think that would help solve much of the aforementioned ambiguity. Particularly how in some subfields there are workshops that are top notch, but because they are called workshops people outside the subfield may think they're "just workshops".

  12. Instead of "poster papers" I should probably have said "short papers". In my subfield there's usually no distinction between "poster" and "short", though it does occasionally happen that a full paper is presented as a poster. In that case I list it as a full paper and give the acceptance rate.

    I've never had the other scenario come up. I'm not sure what I'd do in that case.. maybe start giving acceptance rates for all workshops that I have the data for? In my field workshops tend to be uncompetitive; acceptance rates are so high that organizers usually don't even tell us what they are.

  13. It is important to distinguish the "implied audience" of a CV. Sometimes (eg tenure case documentation) a CV is for people out-of-field, and then any verifiable information that clarifies the relative standing is useful. In Australia, in 2010 as part of a research evaluation process, the government produced a huge list, ranking all journals and conferences (in all fields) into categories (journals were A+, A, B or C; conferences were A, B or C). These rankings were abandoned in the evaluation done in 2012, but CVs for outsiders still usually mention the rank of each publication outlet.

  14. 1) no, but I started doing it for the internal reports for performance management (for the same arguments as mentioned in the previous comments)
    2) no cut-off point; if I know it, I include it.

    either way, ACM and Springer's LNCS/LNAI/etc have a cap on acceptance rate (around >30-33%), so having a paper in one of those proceedings already communicates a certain selectiveness (unlike the IEEE ones, which depends on the conference).

    my field is a subfield in AI

  15. I'm in theoretical CS and don't list any impact factors or acceptance rates. Moreover, I have never seen a CV which does (but I am not senior enough to have to evaluate people that are not in TCS).

  16. I'm not in computer science and no-one in my field puts IFs on their CV. I do list my total citations and h-index though.

  17. I'd really appreciate if you share some links to other resources that have data about this subject of course in case you happen to know any.