Think academic journals look the same ? Well, some do…

We have seen yesterday that finding an optimal strategy to publish is not that simple. And actually, it can be even more difficult in the case the journal rejects the paper (not because it is not correct, but because “it does not fit” with the standards, the quality of the journal, the audience, the editor’s mood, or whatever). The author has basically two choices,

  • forget about the article and move to something else (e.g. start a blog where he/she will be the author and the editor)
  • pretend that the article is worth publishing and then try to find another journal with similar interests


But this last choice is not that easy, since sometimes the author think that this journal was indeed the one that should publish it (e.g. all the articles on the subject have been published in that journal).
So I was wondering if there were clusters of journals, i.e. journals that publish almost the same kind of articles (so that next time one of my paper is rejected by the editor, I just go to for some journal in the same cluster).
So what I did is extremely simple: I looked at articles titles and looked for correlations between words frequency (I could have done that in key words, but I am not a big fan of those key words). I looked at 35 journals (that are somehow related to my areas of interest) and looked at titles of all articles published over the last 20 years. Then I kept the top 1000 of words, and I removed standard short words (“a“, “the“, “is“, etc). Actually, my top words looks like

"models" "model" "data" "estimation" "analysis" "time" 
"processes" "risk" "random" "stochastic" "regression" 
"market" "approach" "optimal" "based" "information" 
"evidence" "linear" "games" "bayesian" "theory" "effects"
"distribution" "multivariate" "tests" "markets" "markov"
"equilibrium" "dynamic" "process" "distributions" 
"application" "stock" "likelihood"

Then, I ran a principal component analysis on my dataset (containing 960 variables – here words – and 35 observations – here journal names).

library("FactoMineR")
res.pca = PCA(MATRICE, scale.unit=TRUE, ncp=5, 
graph=FALSE)
plot.PCA(res.pca, axes=c(1, 2), choix="ind")

The projection of the journals on the first two axis looks like that

Here, we can clearly observe some clusters : on the up-left Journal of Finance and Journal of Banking and Finance (say financial journals) on the top-right Biometrika, Biometrics, Computational Statistics and Data Analysis and Journal of Econometrics (JASA is not far away, i.e. applied statistics journal). And below, on the right, Stochastic Processes and their Applications, Annals of Applied Probability, Journal of Applied Probability, Annals of Probability, Proceedings of AMS and Topology and Applications (ie more theoretical journal).
Note that the projection is rather robust: if I consider my first 200 words, the graph is the same

In order to go further in the interpretation, we can also plot variables, i.e. words from titles,

where we cannot distinguish anything. So if I just look at my top 30, here they are,

On top left we see market(s), risk or information; on top right analysis, effects, models or tests; while below we see Markov or process(es). And we can observe interesting facts: in finance in statistics, we talk about dynamics while in theoretical (mathematical) journal it is about processes.
But the goal was to find cluster, i.e. classes of journals that publish papers with similar titles.

DISTANCE = dist(MATRICE)
cah = hclust(DISTANCE) 
plot(cah)

Here we have

If some classes a rather natural (Journal of Applied Proba. and Advances in Applied Proba.or Economic Theory, Journal of Economic Theory and Journal of Mathematical Economics) some strong correlation are not simple to understand, (e.g. Insurance: Mathematics and Economics and Management Science or Annals of Statistics and the Journal of Multivariate Analysis).
Again, it might be possible to spend hours on the graphs, but if I want – someday – to submit something to one of those journals, I guess I have to stop here, and move to something else…

 


8 thoughts on “Think academic journals look the same ? Well, some do…”

  1. Oui, exemples d’emploi de AFC pour comparer des écrivains ce sont communes chez la littérature de AFC. Je crois qui au-delà que la centrage il y a besoin de employer le distance khi carré aussi. En général, la ACP emploi distance écart type, lorsque AFC khi carré. Je vais lire le doc pour FactoMineR pour avoir a plutôt compréhension de les méthodes de cette paquet.

    REPONSE: donc on considère deux variables: le journal (avec plusieurs modalités qui sont les différents journaux) et le titre (avec plusieurs modalités qui sont les mots). On peut alors effectivement utiliser une distance du chi deux pour mesurer la proximité entre modalités. Le tableau de contingence est alors ma matrice… Je tenterais de regarder lundi (je crois que la base est resté sur mon disque, au bureau…). Merci pour l’idée ! Mais en fait, je me demande si ce n’est pas ce qui était fait dans les graphiques au dessus… car les dessins ressemblent à des graphs d’AFC…

  2. J’ai trouvé la référence que je crois explique la raison parce que l’emploi de ACP c’est ne pas a bonne idée:
    http://www.utdallas.edu/~herve/abdi… regardez la section 4.1 où l’auteur explique ça. Pour moi il semble le même type de problème que vous êtes a tenter analyser.

    REPONSE: le papier est intéressant et amusant. J’aime bien l’idée d’utiliser la ponctuation pour comparer des écrivains…  et comme le notent les auteurs, “rather than the absolute number, the “style” of the authors, is, in fact, expressed by the relative frequencies of their use of the punctuation marks“. Mais ça ne dit pas que l’ACP est une mauvaise idée, juste que l’ACP non centrée réduite est une ‘bad idea‘. Ce avec quoi je suis d’accord. C’est pour ça que j’ai centré et réduit, précisément pour éviter l’effet volume. Donc je ne vois toujours pas pourquoi ça serait une mauvaise idée….

  3. Thanks to @coulmont (and his blog, here), I discovered an extremely interesting tool in R to illustrate connections on a graph (so far, I have to admit that I do not clearly understand what is plotted). Here is the graph based on the first 50 most popular words in titles,

    library(sna)
    gplot(MATRICE[,1:50],gmode="twomode",
    displaylabels=TRUE,edge.lwd=0,edge.col="light blue")

    and here on the top 100,

    (I can mention a nice document online here about the syntax of that function – but not the detail of what is inside). On the graph above, we can clearly see some regions, with statistical journals on the upper right, theoretical economic journals in the lower left, mathematical journals in the lower right, and applied economic journals in the upper left.

    If we look now at words, in the corner of statistical journals (opposite to theoretical economics one) we see “bayesian“, “estimating” or “testing“, “nonparametric” and “semiparametric” techniques; if you mention a “theorem” in the title, then you publish in mathematical journals as well as the names of some weird “space“… etc.

    One can also observe that “methods”, “empirical” and “modeling” are between finance/risk journals and statistical ones; and that “asymptotic” “properties” mean that you go from statistical to mathematical journals…

    So obviously, one can get easily interpretation with those graphs. If anyone knows what is behind, the blog is yours ! And thanks @coulmont for the function !

  4. Juste un détail, mais je suis perplexe …
    La façon dont vous décrivez la construction de la matrice de données MATRICE, où des cas sont les revues et les variables sont les mots, m’a conduit à comprendre les entrées dans le tableau sont les fréquences, pas des variables continues.
    Ainsi, l’analyse appropriée aurait été une AFC, et non le PCA.
    Ai-je raté quelque chose?

    RÉPONSE: c’est un comptage de mots, donc en un sens c’est continu (ou au moins quantitatif)… mes variables sont les nombres d’apparition de mots, et les observations les journaux. D’où mon idée de faire une ACP. Avec une AFC, si ma mémoire est bonne, on regarde si pour deux variables, des modalités sont répondues ensemble par un même individu… Ici j’ai du mal à voir ce que seraient les variables, les modalités, et les individus… Pour être plus précis, la base que j’ai construite ressemble à celle là (que je retravaille ensuite pour enlever les mots courants)

    > B=read.table("http://freakonometrics.blog.free.fr/public/data/BASETITRE-DECOMPTE.txt") > head(B) of the  and in.   a for. with  on  to models model an data estimation
    1 1212 629 490 461 285 720 307 163 130 206 127 42 131 184 2 351 228 230 212 82 115 121 51 47 1 15 23 1 0 3 388 244 197 179 97 244 158 55 51 132 72 21 35 107 4 1350 707 439 264 242 456 184 299 99 10 10 36 1 1 5 700 332 425 380 167 585 292 74 188 207 117 33 192 129 6 792 675 1038 504 161 150 89 125 169 18 45 34 30 4 analysis time from processes risk random stochastic regression s market
    1 72 37 46 78 5 108 33 165 55 0 2 4 6 6 1 3 7 13 0 12 17 3 7 85 6 55 7 13 25 61 8 0 4 4 16 21 3 0 18 13 0 101 0 5 109 73 56 26 17 28 12 115 25 2 6 22 19 185 2 48 8 0 5 77 89 using approach optimal by based information evidence linear games bayesian
    1 23 36 9 49 72 19 0 135 1 0 2 4 12 6 11 6 38 5 8 305 2 3 18 10 3 14 28 7 0 31 0 0 4 7 14 3 44 2 1 0 22 1 0 5 66 51 10 40 52 12 4 69 0 22 6 16 27 15 19 15 27 49 2 19 2 under two theory effects distribution multivariate tests markets markov
    1 92 35 23 13 113 233 61 0 0 2 12 25 9 2 0 0 1 7 0 3 26 7 30 13 25 13 62 0 4 4 14 29 16 0 9 0 1 0 4 5 25 26 8 48 18 45 34 0 2 6 25 11 45 56 12 0 0 59 0 equilibrium non dynamic process distributions application stock likelihood
    1 0 33 3 28 143 23 0 72 2 38
    27 9 5 1 1 0 0 3 0 2 14 14 13 12 1 27 4 2 40 0 7 4 5 0 2 5 0 0 17 22 22 38 1 39 6 19 0 15 5 2 10 20 0

    … etc ….

      absolute validation spatially sediment reply regions quantiles predicting
    1 1 7 0 0 0 6 8 1 2 0 0 0 0 2 0 0 0 3 5 1 0 0 0 1 3 0 4 2 0 0 0 0 1 0 0 5 2 9 5 0 0 5 1 0 6 0 1 1 0 1 0 0 0 levels hyperbolic ergodic JOURNAL
    1 1 1 2 Journal of Multivariate Analysis
    2 0 0 0 International Journal of Game Theory
    3 1 0 1 Econometric Theory
    4 1 22 3 Proceedings of the American Mathematical Society
    5 6 1 0 Journal of the American Statistical Association
    6 3 1 0 American Economic Review

    Mais encore une fois, je connais peu l’analyse des données donc toutes les idées sont les bienvenues !

  5. Nice topic ! Is it possible to have some details about how you retrieved the titles from these journals ?

    RESPONSE: sure, with Scopus (you can also program a robot to look at titles in other websites, but it is more complicated, and you can easily get into trouble, but I’ll write more about that soon), then I use something like that

    > title=LISTE[i]
    > title
    [1] "Limiting dependence structures for tail events with applications to credit derivatives"
    > S=strsplit(tolower(title)," ")[[1]]
    > S
    [1] "limiting" "dependence" "structures" "for" "tail"
    [6] "events" "with" "applications" "to" "credit"
    [11] "derivatives"
    > rev(sort(table(S)))
    S
    with to tail structures limiting for
    1 1 1 1 1 1
    events derivatives dependence credit applications
    1 1 1 1 1
  6. I am not that surprised by the association between Annals of Statistics and JMA. Both journals publish papers in [hardcore] mathematical statistics, the former being more prestigious.

  7. Fun!! I love looking at journals and who publishes where. I am assuming you scraped the titles off of the web pages somehow? I wonder how different it would be if you used article keywords.

    RESPONSE: the only problem is that it is a bit more complex to get (since the number of key words is not the same). So it is a technical problem…

  8. Merci pour ce billet très brillant comme d’habitude! j’aimerais savoir si c’est aussi le cas des journaux d’économie agricole et d’environnement (AJAE, ERAE, JEEM, JAE, EE)… j’ai eu un papier refusé récemment et j’aimerais savoir où l’envoyer…

    RÉPONSE: bon, si je commence comme ca, je n’ai pas fini…. mais ca peut être marrant, effectivement…. par contre demain matin j’ai cours, il faudra attendre un peu (l’accès aux données se fait via les abonnements de l’université). Sinon merci pour les compliments (que je ne mérite pas franchement…. faire 6 lignes de codes c’est assez facile)

Leave a Reply

Your email address will not be published. Required fields are marked *