In June 2017 I was invited to give a workshop on academic publishing for the IMISCOE PhD network, together with Sawitri Saharso, editor-in-chief of Comparative Migration Studies. We had invited questions in advance and received many more than we had time to discuss during the workshop. In this blog post, I address some of those questions.
Q: Are there standard elements one should always include in an article, such as introduction, theory, method, literature review, data analysis, and conclusion?
There is a standard structure, sometimes referred to as IMRAD, that is common in some journals and disciplines. It is an acronym for ‘Introduction; Methods; Results and Discussion’. But most social-science journals will be open to a variety of structures. In fact, IMRAD is clearly modelled on quantitative or experimental studies where you can present ‘results’. Personally, I prefer a structure that reflects the contents of the specific article. In other words, ‘Introduction’ and ‘Conclusion’ might be the only generic headings. Depending on the article, a review of existing research could be part of the introduction, or it could be a separate section. In that case, the heading should perhaps not be ‘literature review’ but rather a relevant thematic description. With regards to methods, it also depends on how much needs to be said. Sometimes I have had a section called ‘Context and methods’ that gives a brief description of the empirical context and also accounts for the data collection process and analysis. In other cases, what needs to be said about methods can fit towards the end of the introduction. With quantitative papers, however, proper descriptions of the data, variables and analytical choices usually require a separate section.
Here are two examples of outlines from my own articles. The first one is based on ethnographic fieldwork in Cape Verde and the Netherlands, but focused on conceptual issues and drawing upon a lot of secondary data from elsewhere. There is no ‘Introduction’ but a section on ‘Asymmetries of long-distance closeness’ that addresses the article’s core theme. Within this section, I account for the empirical context of my own data and the conceptual distinction between migrants and non-migrants. The next three main sections represent the three core elements in the theoretical framework that I develop.
The human dynamics of migrant transnationalism
Published in Ethnic and Racial Studies
- Asymmetries of long-distance closeness
- Dutch–Cape Verdean transnationalism
- Migrants and non-migrants
- Transnational moralities
- Repaying the gift of communality
- Moral economies of ingratitude
- The pro-migrant turn in discourse
- Practicing transnationalism in a moral context
- Information and imagination in transnational relations
- Migrants’ information about life in the place of origin
- Non-migrants’ information about life abroad
- Information and imagination
- Transnational resource inequalities
- Material resource inequalities
- Mobility resources and their yields
The other example is from an article that is entirely based on quantitative analysis of survey data. It has a more traditional structure with more generic headings. The section ‘variables and measures’ is the only one that is divided into sub-sections. Perhaps because I am a mixed-methods academic, my quantitative articles tend to have thorough discussions of concepts, variables, and their relation to questionnaire formulations. Quantitative specialists would perhaps cover these issues more quickly and instead have discussions on mathematical modelling strategies.
Capacity and desire to remit: comparing local and transnational influences
Published in Journal of Ethnic and Migration Studies
- Theoretical framework
- Context and data
- Variables and measures
- Economic integration
- Socio-cultural integration
- Ties to the country of origin
- Background variables
- Results and discussion
In order to develop a good structure for any type of article, it helps a lot to use built-in heading styles. If you write your text in Microsoft Word and apply Heading 1 and Heading 2 to your headings, you can show the Navigation Pane in order to view an outline of the article, just like the examples above. This also makes it easy to jump between sections by clicking the headings.
As a reviewer, I sometimes receive manuscripts where the headings have disparate manual formatting that makes it hard to figure out what the main sections and sub-sections of the article are. This could easily result in errors if the article is published, or it might be a symptom of the author not having thought clearly about the structure.
Many publishers now also use articles’ structure to help readers navigate the text online. For instance, Oxford University Press displays the main headings of an article in the right-hand margin, just like the navigation pane in Microsoft Word. This makes the structure even more important for the reader experience.
Q: It seems a bit discriminating towards non-English speakers that they have to pay for proofreading. Is it really necessary, or can a paper in slightly broken English also be published?
This is actually quite a big question that touches upon many aspects of writing and publishing. Moreover, it doesn’t just concern non-native writers of English.
Most journal editors and publishers would want the articles to be free of language errors. However, the publisher generally won’t have a budget for correcting errors in a poorly written manuscript. So, as an author, it is your responsibility to, at least, ensure that there are as few grammatical and spelling errors as possible. The language functions in Word actually go a long way towards ensuring that.
The bigger issue is the quality of the writing. A manuscript can be free of errors, but still written in a way that makes it difficult to understand, poor at conveying the substance, and tedious to read. Such weaknesses can get in the way of having the manuscript accepted, also if the writer is a native speaker. A good language editor can help improve the quality of the writing, beyond correcting errors. This is the difference between language editing and proofreading.
Beyond correcting errors, it is somewhat artificial to separate the quality of writing from the quality of the analysis. As Eric Hayot has expressed it, ‘you cannot know what your ideas are, mean, or do until you set them down in sentences.’ In other words, academics need to craft good sentences.
My colleagues and I always budget for language editing in our projects, regardless of who the authors are. But the end result will obviously be best if the manuscript contains good sentences before reaching the language editor. Personally, I work to continuously improve my writing. There are three ways of doing it: (1) Reading books (and blogs) about writing; (2) reading high-quality non-fiction, such as The New Yorker, in order to absorb and learn, and (3) thinking specifically about the quality of language as I write and edit my own manuscripts. Two blogs worth reading, by the way, are and Patter and Explorations of Style.
Q: When can I use ‘I’ in the text of a journal article?
My impression is that few reviewers or editors today would have any principled objections to seeing ‘I’ used in a manuscript. If you describe research that you have carried out, for instance, it sounds a lot better to write ‘I conducted semi-structures interviews’ than ‘Semi-structures interviews were conducted’. Sometimes it also works well to refer to ‘I’ as the author, for instance writing ‘I will return to this point in the conclusion’.
Less inexperienced academics tend to often use formulations like ‘my results show that…’. I can’t say it’s wrong, but, in my mind, it easily undermines the authority of the author. It can make the whole exercise sound a bit like a student assignment. The same is true for the over-use of ‘I’ in sentences that describe straight-forward tasks where it makes little difference who carried them out. It can easily give the impression of saying ‘look what I did!’ instead of concentrating on what’s useful for the reader.
For instance, if I write about the analysis of survey data, I probably wouldn’t write ‘I converted the Likert scale for subjective-wellbeing into a dummy variable’. First, this is a technical operation where my contribution as a person is quite meaningless (unlike in conducting interviews, for instance). Second, I find that with quantitative analysis, the important point is to properly explain the variables that are used in the analysis, not to account for the work that was carried out. So, I might instead have written something like ‘Subjective-well-being is measured as a dummy variable, converted from the Likert scale in the questionnaire’ and then explained how the specific answers were grouped.
Q: How can I make sure that my work is read by fellow academics? How should I (not) promote my articles?
First, you can do a lot at the writing stage. In particular, crafting good titles, abstracts and opening paragraphs can make a huge difference. I quickly scan hundreds of article titles every year, for instance when I receive content alert e-mails from journals, and a good title can make me stop and click, even if it is not about something that I particularly need for my research at the moment. Then, the abstract decides whether I download the full text or not. You might also look into how titles and abstracts can help potential readers find your work if they do relevant searches.
Second, you can promote your article once it is published. You can, for instance, e-mail it to people you’ve cited, and say that you have appreciated their work. I also sometimes receive e-mails from authors who think I might be interested, regardless of whether they’ve used my work or not. Both junior and senior migration scholars do this.
Publishers now have guides for authors on how they can promote their work. For instance, see what Taylor & Francis and Elsevier recommend. These include guidelines on how to share manuscripts that don’t have Open Access licenses. You should not post the final PDF from the journal on your web site or upload it to networks such as Academia.edu or ResearchGate. (You can still register your article on those sites, though, and link to the publisher).
You can also promote an article through social media. If you use Twitter, the impact will be greater if you don’t just tweet the title, but formulate something attention-grabbing that others might retweet. Especially if you don’t have many followers of your own, you depend on others retweeting.
Facebook is also relevant. I think there’s been a change over the past few years in that more academics use Facebook for work as well as for personal things. I often see people sharing their new articles. I have academic Facebook friends that I haven’t yet met in person, and I have many that I only see at conferences. If you don’t want to overwhelm your family and other friends with article promotion and other academic stuff, you can use friend lists to not share everything with everybody.
Q: How can I be a balanced reviewer when reviewing someone else’s work?
I would say mainly through practice. Say yes if you are invited to review, and appreciate that you are learning as well as contributing. Some journals will provide guidelines after you accept to review a paper, but many will not. You can have a look at the general advice for reviewers at publishers’ web sites, for instance at Sage or Taylor and Francis. You can also find a big collection of online resources on the Peer Review Week resources page.
Some manuscripts clearly fall short of being publishable, be it because of the quality of data, the connections to existing literature, the clarity of the writing, or something else. But if a journal accepts only 20 percent of the submissions, the cut-off point might be a matter of the difference between very good and truly excellent articles. This difference often comes down to taste and perspective—in other words, it is subjective. Editors therefore prefer to have three reviewers and to decide on the basis of all three assessments. Knowing this takes some pressure off you as an individual reviewer. In most cases, you will receive a copy of the editor’s e-mail to the author, including all the anonymized reviews. Reading the others is also a good way of learning.
As a relatively inexperienced reviewer, you should be prepared to spend some time on writing a conscientious review and providing constructive comments. Since you are also an author, it should be easy to imagine what kind of feedback is helpful and what is not. As long as you provide your assessment of the article’s strengths, weaknesses, and potential for improvement, it will be valuable for the editor in making a decision.
As an experienced reviewer, I also make an effort to be constructive and respectful. But I feel that my main responsibility is towards the journal and the editor rather than towards the author. If I think that an article should be rejected, I will explain properly what I think the shortcomings are, but I will not spend a lot of time providing detailed feedback on how to improve the text. It’s a better use of my time, I think, to say yes to invitations to review as often as I can.
Q: How do journal editors, reviewers and employers distinguish between theoretical, empirical or review papers?
This differs between disciplines and employers, of course, but my main point would be that all articles are valuable. In migration studies, journals generally don’t publish ‘review papers’ without any original analytical contribution. Since theoretical papers also need to build upon the work of others, the distinction between ‘review papers’ and ‘theoretical papers’ is not always clear. But they would differ from empirical papers in saying something more general and not being based on primary data. If you manage to get such a paper accepted in a good journal, that is excellent! With respect to employers, the typical danger is that your expertise seems too narrow and specialized. One way of adding breadth is to have one or more articles that address broad themes, perhaps in a theoretical way.
Q: How much should I polish a paper before submitting to a journal? Is it OK to use the review process to get feedback on a mediocre manuscript even if a rejection is the most likely outcome?
I would say that it’s not OK. From a pragmatic perspective, a mediocre paper is likely to get a desk rejection before being sent to review, so it might not help you either. From an ethical perspective, it’s not OK to abuse reviewers’ time. The academic community struggles with a shortage of people who are willing to spend time on reviewing. If reviewer capacity is a scarce resource, it should be devoted to reviewing the papers of authors who have already done their best.
Sometimes, as a reviewer, I read papers that need a type of input that a reviewer can’t give. For instance, some manuscripts appear to be a graduate student’s first submission to a journal and it is clear that they haven’t quite managed the journal article genre yet. Ideally, their supervisor should have provided guidance. If the supervisor fails, the author might need to seek help elsewhere, or work independently with book such as Writing for Peer Reviewed Journals. As a reviewer, I can try to define the problem, but I can’t hand-hold the author through fixing it.
Q: If you do a PhD by publication, is it best to aim for lower-ranked journals in order to get articles published in time for submitting the dissertation?
This is a tricky question! I think that a PhD by publication is generally a very good model, but I am sceptical of letting publication of the papers determine the evaluation. In my opinion, the ‘PhD by publication’ or ‘article dissertation’ model should mainly be about the format of the text, i.e. that it is written in the form of self-contained articles, perhaps with diverse audiences in mind. The editorial process is also a very valuable learning experience, but it would be best to keep it separate from the evaluation of the dissertation and not focus on whether or not the papers have been accepted.
Many universities will accept dissertations that consist of articles with different publication status. As a PhD candidate, it’s good to be able to submit the dissertation when the content is of PhD quality without having to wait months (or years) for reviewers and editors to do their part. If the article ends up being published much later, but with additional improvements and in a highly-ranked journal, that’s great. Whether this strategy is feasible for you depends on the rules and expectations in your own university.
Q: How can I tailor my work to different audiences? How much of previously used material should be re-utilized in a different way for a paper for a different audience?
Tailoring your articles to different audiences mainly has to do with how they are framed and which type of journals they are submitted to. By framing, I mean the essence of what the article is ‘about’. For instance, many PhD projects in migration studies address a particular theme within a particular empirical context – say, entrepreneurship among Bangladeshis in Italy. Rather than writing three or four articles about entrepreneurship among Bangladeshis in Italy, it could be good to foreground empirical, thematic or theoretical elements in different ways, and submit them to journals that correspond to each type of framing.
For instance, one article could focus on how the specific economic and political context in Italy shapes immigrant entrepreneurship, and be submitted to Modern Italy. Another could address implications for entrepreneurship theory and be submitted to Journal of Entrepreneurship. Of course, each such specialized framing might require engaging with a new literature, so you will have to judge how far you want to go, and in how many different directions.
With the PhD candidates that I have supervised, it has often been the case that all their articles could potentially be submitted to interdisciplinary migration studies journals such as JEMS. Then it has made sense to see which articles might be well-suited for journals with other audiences, just in order to diversify. We have had these discussions on the basis of article plans that have evolved in the course of analysis and writing.
If the framing is sufficiently different, it is not a problem to re-use the same data for many articles. It is often a bigger problem that the data is not fully utilized, i.e. that more articles could easily be written without problematic overlaps. Having said that, it is not OK to squeeze as many articles as possible out of a single dataset in order to boost your publication record. Each article must make a substantial, original contribution.
The Management and Organization Review has published an article on this, discussing the questions that arise for authors who write several articles from the same (quantitative) dataset. A key piece of advice from that article is always err on the side of transparency. If you have already published or submitted other articles based on the same data, you should consider citing them. For instance, you could write ‘The gender dimensions of these processes have been analysed elsewhere (Author 2016)’ and just list ‘Author 2016’ (with no further details) under A in the list of references. What you should avoid, is seeming deceptive by hoping that reviewers or editors won’t discover that you have published or submitted something very similar elsewhere. If the article draws upon a working paper or a previously published version in a different language, you should ask the editor beforehand if the journal will still consider your article.
Q:When is it useful to publish something as a working paper?
Working papers present a dilemma. On the one hand, working papers are a great way on engaging with an academic audience and getting your results disseminated quickly. They can be widely read and cited, especially if they are in a well-established series. For instance, the International Migration Institute at Oxford has had a very successful working paper series that has strengthened the institute’s position.
On the other hand, many journals will not accept your article if it has already been published as a working paper. In some disciplines, notably economics, journals will happily accept manuscripts that have been posted online as working papers. This is not the norm in other social sciences, but some scholars argue for a change in that direction.
While journals insist of original, unpublished manuscripts, it can solve the problem if your working paper is substantially different from the article. You might, for instance, publish a long working paper with the intention of creating a shorter version with a different title that can be submitted to a journal. You must make it transparent, though, for instance by including a note explaining that a longer, previous version has been published as a working paper.
In the past, a big argument in favour of working papers was the speed of publication. It was common for accepted journal articles to wait in line for several years before getting a slot in an issue of the journal and becoming available to readers. That’s changed because articles are now usually published online soon after being accepted. It could still take a long time before they are allocated to an issue and officially published, but readers can access and cite them in the meantime.