I finished preparing a math paper, and am now trying to decide where to submit it.

I read How do you judge the quality of a journal?, but it seems that the answers there mostly deal with how to judge if a journal is good at all.

The journals I am considering (see list below) are all very good. But how can I know how do very good journals rank?

I know my paper is not at the level of Annals, so I considered the following journals: Selecta mathematica, Duke, Compositio Mathematica, American journal of mathematics, Advances in mathematics, International mathematical research notices, and Crelle journal.

How can an author compare the quality very good journals such as these? Is there any reasonable way to use public data in order to sort them from best to worse?

Note: this question is looking for an answer about a general method, not necessarily a particular ranking of those particular journals.

  • 1
    Hi, and welcome to Academia.SE! I'm going to edit your question slightly to emphasize the general "how to compare good journals" over the "these specific journals", to try to ensure that it doesn't get closed as being too specific or a "shopping question".
    – jakebeal
    Jun 27, 2016 at 12:58
  • 2
    I am asking how to compare very good journals. Of course, this will help me to decide where to submit.
    – user57267
    Jun 27, 2016 at 13:11
  • 4
    You may want to add Math Annalen?
    – T K
    Jun 27, 2016 at 13:12
  • 2
    Out of curiosity, do none of these options stand out as being the most cited in your paper? Presumably if a journal excels at subject X to the point where most of your research into the subject comes from there, you would want to publish in X, at least to first order.
    – user4512
    Jun 27, 2016 at 19:06
  • 10
    Acta is generally thought of as the very top tier (with JAMS, Annals, Inventiones, though everyone disagrees on the order). So its somewhat better than Duke and clearly better than the others mentioned in this post. It has somewhat of a lean towards analysis. Jun 28, 2016 at 10:37

7 Answers 7


Is there any reasonable way to use public data in order to sort them from best to worse?

Not really. To start with, there isn't a remotely well-defined notion of ranking. For example, it depends on the subfield (some journals attract better papers in certain areas than others), it can vary over time, and it depends on the specific goals you have in mind. More importantly, people just don't always agree: I know some people who would argue that one of the journals you listed is clearly a better choice than the others, but they wouldn't all choose the same one.

The explicit rankings I've seen do a poor job of accounting for these factors. At best they are mediocre (perhaps valuable as a first cut for distinguishing between journals at very different ranks, but useless for fine distinctions). At worst they are positively misleading.

Fortunately, choosing between well-known journals like these is not so difficult. You should look them over and try to gauge which ones look to you like they are publishing the most important papers in your area, while also asking around to get other people's opinions. (Ask mentors of yours, ask your friends or collaborators, ask colleagues at tea.) If there's a clear consensus among the people you ask and it's consistent with your own impressions, then you're done. If there's a clear consensus that disagrees with your impressions, then you need to think and talk a little more. If there's no clear consensus among several options, then it doesn't really matter which of them you choose. After all, journal prestige matters only to the extent it is perceived to matter by the community; if the community can't make up its mind, then there is no right choice.

Carrying this out takes a little time, but it's worth it since you learn more about how things are perceived in your research area.

  • Could you explain the sentence "and it depends on the specific goals you have in mind"?
    – user57267
    Jun 27, 2016 at 14:18
  • 6
    One case is that journal A is older and has greater name recognition than the new journal B, but those who know about B consider it superior. In five years B will probably be recognized as better to have on your CV, but it might attract less attention than A would if you apply for jobs this year. Another is that some journals publish papers of a broader range of quality than others (even with the same average quality). Different people in different career circumstances may have different concerns about the extent to which a journal sets a high lower bound for the importance of their papers. Jun 27, 2016 at 15:10
  • 1
    I guess for instance taking journal A = Journal of Algebra, and journal B = Algebra & Number theory might be a current concrete example.
    – the L
    Jun 27, 2016 at 16:31
  • 2
    @the L: I would say you have a concrete example from about five years ago, but it makes the point nevertheless. Also I think your choice illustrates what Anonymous M was saying about quality: the best papers from A seem roughly comparable to the best papers from B (i.e., certifiably strong, important works), but A publishes a wider range of papers than B. Jun 27, 2016 at 19:08
  • 1
    I am glad to here that @PeteL.Clark (since I hava a paper in ANT). I wasn't sure to what extent people are aware of its quality.
    – the L
    Jun 27, 2016 at 19:12

I agree with everything in Anonymous Mathematician's answer, but are some additional thoughts.

First, there is no object well-ordering in terms of quality, in the sense that everyone will agree upon it. This is a necessary consequence of there not being a well-defined measure of quality for individual papers, variation among referees and editors, and that the value of a paper often can't be properly assessed until years after publication.

In math, unlike some other fields, there are typically a lot of journals that would be a good fit for your article, and in most situations it shouldn't make too much difference picking one over the other. I personally don't have strict mental ranking of journals in my field in mind when I submit, but some vague notion of tiers and it's more like, okay these are all in the same tier and I'll just pick one of them. (I also usually have no sense of how good my papers are when I submit them.)

I would never think, oh, we should hire this person instead of that one because one has a paper in American Journal and the other has a paper in Advances. (Note: I would not put all the journals you listed in the same tier.) I just think, oh this person has 1-2 papers in great journals and 2-3 papers in good journals, and the rec letters are stellar.

There are some situations where rankings of journals are directly used for evaluating people (either impact factor, or Australia's letter grades, though these do not line up with most people's notions of quality), though these are relatively rare in the US, and I would only worry about this if your colleagues tell you too.

So, I would talk to some colleagues in your field, to see if they have further input. A couple of other things you should consider are what kinds of paper the journals in your field have been publishing recently (look through some recent issues as well as at the editorial board---I often choose a journal just based on an editor--an editor who can better appreciate your work can both better choose an appropriate referee and is more likely to push for your paper to be accepted), and the general operation of the journal (e.g., some are much faster than others---see the AMS data).

  • I can confirm that this answer also holds in many other fields (e.g. life sciences).
    – Bitwise
    Jun 27, 2016 at 18:14
  • 2
    To be honest, considering the situation of Adv. Math., having a paper in AJM is more convincing.
    – user92646
    May 15, 2018 at 23:35

One important factor which the other answers haven't mentioned is the editorial board. Given two roughly interchangeable journals, I'd expect a better experience at the one with an editor whose expertise is closer to the paper. You're more likely to get an unfair referee report if the editor isn't close enough to the subject matter to pick appropriate referees.

  • Yes editorial board is important
    – learner
    Aug 8, 2023 at 6:08

There are many different journal rankings or numerical scores that you can compare. Out of the ones I've seen, the one that corresponds closest to my own subjective impression of journal strength is definitely the Mathematical Citation Quotient (MCQ). The MCQ is listed on the MathSciNet page about each journal. I think that's a good starting point.

That said, I'll agree with what Anonymous Mathematician said: most people don't pay much attention to various journal rankings. Instead you continuously update your impressions of different journals based on discussions with colleagues (at tea, during meetings, while ranking job applicants' CVs, etc) but also based on where you see papers get published: if you see a paper appear in a journal you would've thought too strong then you'll think the authors got lucky, but if it happens several times with the same journal you'll downgrade your opinion of the journal.

PS - about the journals you asked about - my ranking would be Duke > Crelle, Compositio, Advances, Amer J Math > Selecta > IMRN. The difference between the strongest and weakest of those you listed is significant - I'd put Duke at 6th or 7th best out of all math journals, and IMRN outside the top 20.

  • 6
    I am very curious if others here agree about your assessment of Amer J Math, as I heard that it is regarded as a very good journal.
    – the L
    Jun 28, 2016 at 8:34
  • 2
    @theL I was also surprised about the ranking of AJM. I'd consider AJM at about the same level as Advances and Compositio. OTOH, I would not have cared much about Selecta, which appears to be quite a bit better than I would have though. I assume this may be some subfield bias.
    – quid
    Jun 28, 2016 at 18:40
  • I think I agree now and edited the answer. Apr 19, 2017 at 9:40
  • 3
    Personally, I think Advances is clearly much worse than it used to be. Publishing 500 papers a year is making it from a top journal to a 2 class journal.
    – user92646
    Apr 27, 2018 at 5:30

I would not bother too much about this question. Both as an author and as a member of a hiring committee I just look at four categories: The four top journals, the may be 10-15 very good journals, the many journals that are well known, and the many journals which are obscure. The only discussions which might actually happen in a hiring committee is whether a certain journal is obscure or not. Here opinions are especially divided if a journal is only known within a certain community, but does not say so in its name.

So if you are aiming for the second tier, you should look at the article and compare it to articles in those journals. Apart from editors and cited/influential articles which are already mentioned, you should also check whether your article fits into the usual length range of the journal.


Don’t forget to look into the journal backlogs. Getting accepted at a top journsl is not going to be as useful to your career if the acceptance happens only after you do your next job hunt, submit your next grant proposal, or go up for tenure. The AMS published them:



I have a bone to pick with some of the posters here.

First, let me make clear that I agree with certain general premises. You should definitely take into account whether a particular journal is geared towards certain fields. As an example, if you have a truly outstanding paper in algebra, you would do well not to consider submitting to Acta Mathematica. Said journal, while top 5 overall, is geared towards analysis rather than any other field. And you should look for trends, and look at compositions of editorial boards.

Next, let me address a minor point: it is a little silly, in my opinion, that Annals is considered the best of them all, when it is abundantly clear (to me at least) that Publications Mathematiques de l' IHES is really the best. I think that Annals is simply more geared towards papers which finish-off problems, and is thus more fanciful than other top 5 journals.

Now, as to my main grievance: I find it annoying that people think that quality cannot be measured. Yes, in art for example, it is virtually impossible to measure the quality of a piece by any scientific method. Ranking the quality of journals, however, is a very different game. While we may disagree how to measure quality in principal, I put forth the proposition that the only reliable such measure is the article influence score. This thing takes into account not only number of citations, but also where these citations come from. If you look at impact factors only, you might get an approximately accurate picture of the scene, but you are liable to make mistakes in your judgement, if it is based solely on that. An example here is Aequationes Mathematicae. This journal is average at best, yet it has an impact factor of 1.0, which places it at number 58 out of the 312 math journals in the Thompson Reuters database. Its article influence score, however, is 0.465 which puts it at place number 201.

Alright, so here's a list of the top 20 journals based on their article influence score:

enter image description here

P.S. IMRN is nowhere near the top: it is currently sitting comfortably at place 39.

  • 3
    I find this answer way over-simplified. For example, how is it clear to you that the journal ranked 1 in this list is really the best? If you look at the data for 2014 it was ranked 5 with a huge gap up to number 1, so either it has improved immensely in a very short time (I assume these are the newest numbers even though you don't give a source for them), or the measure is very volatile and thus not very useful in the long run. Feb 7, 2017 at 10:36
  • 1
    I only meant that it is currently the best, though I should have been more precise about this; agreed. My point with Annals is that it is usually not the best, year-by-year, but people tend to think that it is. It usually ranks number 3, with most first places between Acta Mathematica and Journal of the American Mathematical Society. The data I have provided is from Thompson Reuters, 2015 Journal Citation Reports. Why do you think my answer is over-simplified though? Do you disagree with the premise that quality can be measured?
    – the_fox
    Feb 7, 2017 at 19:29
  • 4
    I disagree with the premise that quality can be measured as a single number, especially within the full field of mathematics rather than a subfield. Feb 7, 2017 at 19:31
  • 1
    Without wanting to get into a philosophical dispute, I venture that you have an issue with the distinction between 'truth' and 'approximate truth'. I take it as granted that virtually all human knowledge, what we consider and trust as true, is only approximate. So I consider the premise that journal quality admits quantification as a statement that admits a very good approximation. If you want to argue that it's impossible to distinguish between the top 5 journals, for instance, ok, I'll go with it. But I think that "accuracy" (however you interpret it) of the article influence score is good.
    – the_fox
    Feb 7, 2017 at 19:41
  • 6
    I had not actually downvoted until now. But your comments here make it clear that the answer really was as oversimplified as it felt, and that you would rather delve into weird philosophical asides (with a light condescending tone) than provide any actual arguments. Feb 13, 2017 at 9:46

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .