Oct 19 2011 Published by under Uncategorized

So today I've been making some revisions on a paper we have been struggling to publish for over a year. Granted it's been a low priority publication and the student who did the work is long-gone, but still, its not too bad. What struck me about the last round of reviews is that it was sent to ONE reviewer, who proceeded to write a five page critique with more than eighty, specific comments. Typically papers go out to two or three reviewers, each with a handful of major comments and a few minor ones. So I didn't know what to do with eighty comments. Mind you this was also not a particularly high-caliber journal, to put it generously. The comments ranged from the petty:  "I can't find reference X, could you cite something newer?", or "I don't like your axis labels in figure 2B"; to the somewhat useful: "statistical test X would be more appropriate that statistical test Y"; to ones which to me bring up a grey area in terms of what to do with them, and that if I were to follow them I think the paper would be worse off. These usually have to do with writing style.

Most scientific papers follow the same basic organization: summary, introduction, methods, results and discussion. But how one treats these sections varies a lot from person to person. In my view, the most straightforward approach is also the most dogmatic and makes for some pretty dry, snooze-inducing reading. In this approach, after your introduction, the results are basically a pure description of the experimental findings, with no room for interpretation or background, these are saved for the discussion and introduction. The results become a list of findings to be discussed later in the paper. In  a second approach, which is the one I take, I split the results into sections describing the different experiments. Then for each subsection I first describe the rationale for the experiment, with a citation or two if needed and recapitulate briefly the method to be used. This may sometimes spell out expected outcomes as in "if such a hypothesis is true, we would expect that this and that would be observed when we manipulate these other things." Then I go through the actual findings. Finally, at the end of each subsection I summarize the interpretation of the data as in "the results from this experiment suggest that blah, blah, blah." This helps the reader understand not only why you did the experiment you just did, but also how you interpret it, and facilitates the transition to the next logical experiment. In this way you hold the reader's hand through the results and build your conclusion as you go along, adding helpful bits of interpretation. Then in the discussion, you summarize your findings and their interpretation and go on to discuss the larger context of your study. I find that this type of results section makes for much easier reading and better flow.

Now back to the reviewer with the eight-hundred comments. In many of the comments he or she kept saying "this line belongs in the discussion" or "move this to the methods". In other words "please make this paper so dry and boring so that my ego can be stroked and my terse writing style will prevail". They must have made about fifteen or so of these types of comment. And here is the crux of the dilemma, is it worth bowing to a reviewer's demand which you know will make your paper worse? Are stylistic comments valid criticisms? As for me, we're not sending the paper back to that crazy journal, so I'm just going to ignore the comments, its not even worth trying to address them all. How about you, reader, what do you do in this situation?

11 responses so far

  • NatC says:

    um...I'm really interested to hear what people do.
    we have gone the route of answering everything, often by saying "since reviewer thought X was unclear/poorly written, we've rewritten that section so that it now says..." (not always doing exactly what they say to do, but trying to address whatever underlying concern we think they are having). Occasionally explaining why we disagree.
    But it makes the responses longer than the paper, and drives us to drink.
    I *really* want to hear what other people do!

  • Zen Faulkes says:

    I think stylistic comments are valid criticisms, but they're criticisms that are best taken with a grain of salt. When I make those kinds of criticisms, I usually phrase it to indicate that it’s a suggestion.

    My attitude when I get such comments is similar. Sometimes, there are useful things that can be gained from those comments. But I try not to be fanatical about it.

  • DrugMonkey says:

    1) hahaaha, oh that is a classic reviewer!

    2) Results are properly a vehicle for results and stats...keep it to the Discussion and Intro loquacious Joe.

  • namnezia says:

    NatC: Even though we don't always make the changes to the paper the reviewer wants, we always address them in the rebuttal letter, which makes for a pretty long, tedious document. And you have to fill it with gratuitous praise for the reviewer's intellect and such.

    DM: I don't mean a LOT of discussion in the results, just a sentence usually suffices. In fact some journals allow you to combine the results and discussion sections, and then ask for a short conclusions section.

  • anon says:

    If many of the comments are along the same line (writing style, minor changes), I think one response to all of them will do. Example: "we tried to present the results as suggested by the reviewer, but found that it worked better the way it is.."
    Something like that.

  • WeiterGen says:

    I'd contact the editor, telling her that you would address these specific relevant points of the referee and ignore all others. If not ok for the journal, then do not resubmit the manuscript. If it was a high impact journal I would probably obey, though.
    Writing, just as reviewing a manuscript is challenging.

  • becca says:

    Check with the editor. If the reviewer is both providing good scientific critiques (even if not exclusively so) and they happen to be right about what the overwhelming convention of the field and journal's stylistic conventions are, it may be worth going with.
    I view this issue as very akin to the first person active vs. more passive writing convention. What is, intrinsically, the 'best' writing style goes so strongly against the grain of the field that it impairs understanding. If you interpret in the results, you are more likely to get dinged for overinterpreting, because of the strong expectation that results are descriptions. Also, your discussion sounds repetitive.

  • WhizBANG says:

    First, respond to the useful critiques before submitting to another journal, since crazy reviewer may review for the next one as well. If s/he does and you haven't changed a damn thing, the outcome isn't pretty.

    Things that are solely style of the writer should not enter into the critique, as long as everything is in the correct section (results in results, interpretations in the discussion) and the information they wish to convey comes across clearly. If an author wants to do the whole paper in passive voice it doesn't matter. It may be boring and have more words than it needs, but that is not a criteria for acceptance at this point in time.

  • physioprof says:

    If it's a shitteasse journal, then just tell the editor you are not responding to anything other than substantive scientific concerns, and if that is not acceptable, then you'll take the fucken thing elsewhere.

Leave a Reply