Writing for impact: how to craft papers that will be cited

For the past few years I’ve co-taught a professional development course for doctoral students on completing a thesis, getting a job, and publishing.  The course draws liberally on a book I co-wrote with the late Duncan Fuller entitled, The Academics’ Guide to Publishing.  One thing we did not really cover in the book was how to write and place pieces that have impact, rather providing more general advice about getting through the peer review process.

The general careers advice mantra of academia is now ‘publish or perish’.  Often what is published and its utility and value can be somewhat overlooked — if a piece got published it is assumed it must have some inherent value.  And yet a common observation is that most journal articles seem to be barely read, let alone cited.

Both authors and editors want to publish material that is both read and cited, so what is required to produce work that editors are delighted to accept and readers find so useful that they want to cite in their own work?

A taxonomy of publishing impact

The way I try and explain impact to early career scholars is through a discussion of writing and publishing a paper on airport security (see Figure 1).  Written pieces of work, I argue, generally fall into one of four categories, with the impact of the piece rising as one traverses from Level 1 to Level 4.

Figure 1: Levels of research impact

Figure 1: Levels of research impact

Level 1: the piece is basically empiricist in nature and makes little use of theory.  For example, I could write an article that provides a very detailed description of security in an airport and how it works in practice.  This might be interesting, but would add little to established knowledge about how airport security works or how to make sense of it.  Generally, such papers appear in trade magazines or national level journals and are rarely cited.

Level 2: the paper uses established theory to make sense of a phenomena. For example, I could use Foucault’s theories of disciplining, surveillance and biopolitics to explain how airport security works to create docile bodies that passively submit to enhanced screening measures.  Here, I am applying a theoretical frame that might provide a fresh perspective on a phenomena if it has not been previously applied.  I am not, however, providing new theoretical or methodological tools but drawing on established ones.  As a consequence, the piece has limited utility, essentially constrained to those interested in airport security, and might be accepted in a low-ranking international journal.

Level 3: the paper extends/reworks established theory to make sense of phenomena.  For example, I might argue that since the 1970s when Foucault was formulating his ideas there has been a radical shift in the technologies of surveillance from disciplining systems to capture systems that actively reshape behaviour.  As such, Foucault’s ideas of governance need to be reworked or extended to adequately account for new algorithmic forms of regulating passengers and workers.  My article could provide such a reworking, building on Foucault’s initial ideas to provide new theoretical tools that others can apply to their own case material.  Such a piece will get accepted into high-ranking international journals due to its wider utility.

Level 4: uses the study of a phenomena to rethink a meta-concept or proposes a radically reworked or new theory.  Here, the focus of attention shifts from how best to make sense of airport security to the meta-concept of governance, using the empirical case material to argue that it is not simply enough to extend Foucault’s thinking, rather a new way of thinking is required to adequately conceptualize how governance is being operationalised.  Such new thinking tends to be well cited because it can generally be applied to making sense of lots of phenomena, such as the governance of schools, hospitals, workplaces, etc.  Of course, Foucault consistently operated at this level, which is why he is so often reworked at Levels 2 and 3, and is one of the most impactful academics of his generation (cited nearly 42,000 time in 2013 alone).  Writing a Level 4 piece requires a huge amount of skill, knowledge and insight, which is why so few academics work and publish at this level.  Such pieces will be accepted into the very top ranked journals.

One way to think about this taxonomy is this: generally, those people who are the biggest names in their discipline, or across disciplines, have a solid body of published Level 3 and Level 4 material — this is why they are so well known; they produce material and ideas that have high transfer utility.  Those people who well known within a sub-discipline generally have a body of Level 2 and Level 3 material.  Those who are barely known outside of their national context generally have Level 1/2 profiles (and also have relatively small bodies of published work).

In my opinion, the majority of papers being published in international journals are Level 2/borderline 3 with some minor extension/reworking that has limited utility beyond making sense of a specific phenomena, or Level 3/borderline 2 with narrow, timid or uninspiring extension/reworking that constrains the paper’s broader appeal. Strong, bold Level 3 papers that have wider utility beyond the paper’s focus are less common, and Level 4 papers that really push the boundaries of thought and praxis are relatively rare.  The majority of articles in national level journals tend to be Level 2; and the majority of book chapters in edited collections are Level 1 or 2.  It is not uncommon, in my experience, for authors to think the paper that they have written is category above its real level (which is why they are often so disappointed with editor and referee reviews).

Does this basic taxonomy of impact work in practice? 

I’ve not done a detailed empirical study, but can draw on two sources of observations.  First, my experience as an editor two international journals (Progress in Human Geography, Dialogues in Human Geography), and for ten years an editor of another (Social and Cultural Geography), and viewing download rates and citational analysis for papers published in those journals.  It is clear from such data that the relationship between level and citation generally holds — those papers that push boundaries and provide new thinking tend to be better cited.  There are, of course, some exceptions and there are no doubt some Level 4 papers that are quite lowly cited for various reasons (e.g,, their arguments are ahead of their time), but generally the cream rises.  Most academics intuitively know this, which is why the most consistent response of referees and editors to submitted papers is to provide feedback that might help shift Level 2/borderline Level 3 papers (which are the most common kind of submission) up to solid Level 3 papers – pieces that provide new ways of thinking and doing and provide fresh perspectives and insights.

Second, by examining my own body of published work.  Figure 2 displays the citation rates of all of my published material (books, papers, book chapters) divided into the four levels.  There are some temporal effects (such as more recently published work not having had time to be cited) and some outliers (in this case, a textbook and a coffee table book) but the relationship is quite clear, especially when just articles are examined (Figure 3) — the rate of citation increases across levels.  (I’ve been fairly brutally honest in categorising my pieces and what’s striking to me personally is proportionally how few Level 3 and 4 pieces I’ve published, which is something for me to work on).

citations by level

So what does this all mean? 

Basically, if you want your work to have impact you should try to write articles that meet Level 3 and 4 criteria — that is, produce novel material that provides new ideas, tools, methods that others can apply in their own studies.  Creating such pieces is not easy or straightforward and demands a lot of reading, reflection and thinking, which is why it can be so difficult to get papers accepted into the top journals and the citation distribution curve is so heavily skewed, with a relatively small number of pieces having nearly all the citations (Figure 4 shows the skewing for my papers; my top cited piece has the same number of cites as the 119 pieces with the least number).

Figure 4: Skewed distribution of citations

Figure 4: Skewed distribution of citations

In my experience, papers with zero citations are nearly all Level 1 and 2 pieces.  That’s not the only kind of papers you should be striving to publish if you want some impact from your work.

Rob Kitchin

8 thoughts on “Writing for impact: how to craft papers that will be cited

  1. Chris Hables Gray

    this is a very strong article that all grad students (perhaps even professors) should read. But it does strike me as somewhat naive about how knowledge is accepted. It could use a little Thomas Kuhn to remind people that often great new ideas are rejected because they don’t fit the dominant paradigm. I don’t accept that “generally the cream rises.”I agree with Kuhn (and many others) that often this is not until the old paradigm dies. And in the real world of academia, ideas are often rejected because they don’t follow the genre rules of the disciplines being addressed. Or they are written by outsiders, people of color, women, and those from less elite institutions. And it is a little weird to think that every problem needs new theory. Yes, when the old theories don’t work make new theories, but maybe we can understand the mania for airport security without inventing any new theories of governance, for example. — Chris Hables Gray

  2. Pingback: Writing for impact: how to craft papers that will be cited | Trurl and Klapaucius

  3. Rob Kitchin Post author

    Chris, where we differ is that I don’t believe the paradigm theory works fully practice, especially outside of the sciences. In many disciplines, notably in the social sciences and humanities, there is little evidence of paradigms presently operating, with a diverse set of philosophical approaches employed, and a plurality of minor theories and epistemological/methodological approaches. Further, I think that paradigmatic accounts of scientific progress produce overly sanitised and linear stories of how disciplines evolve, smoothing over the messy, contested and multiple ways in which academic knowledge production unfolds in reality. That’s not to say that different academics or groups do not try to police theoretical developments, but that it in a globalised and plural scientific landscape with dozens of disciplinary journals, competing interests and social media it is almost impossible to do that. In such a landscape if an idea is well conceived, well argued, well written and compelling it will get published and over time engaged with. The days of narrow disciplines, with strong paradigmatic gatekeepers is over in my view.

    I agree that the situation is not linear and straightforward – that I am presenting some magic rule that works in all cases. What gets picked up and cited are subject sometimes to luck and to fad and fashions, and citation can also be self-reinforcing (as an author gets cited it pushes people to their other work), and sometimes very worthy and useful pieces get overlooked. Nor am I denying that exclusionary pressures are at work – the most obvious being the dominance of the English language (which I’ve written about if you’re interested — Kitchin, R. (2003) Disrupting and destabilising Anglo-American and English-language hegemony in Geography. Documents d’Anàlisi Geogràfica 42: 17-36. Reprinted in Social and Cultural Geography 6(1): 1-16). However, I do think the general rule I am setting out here works – if you write Level 1 or Level 2 pieces, do not expect to be cited very much. That’s not to say that these pieces are not useful, just that they will not have very much wider impact.

    Personally, I think it is self-evident that we continue to need new ideas, theory, ways of making sense of the world as we patently have not created a body of knowledge/theory that adequately explains how the world works. If we did all academics could give up on that front and swap to policy and application. Also, in my view, there is certainly a lot of useful theoretical work to be done in understanding airport security and how it fits into a wider shift in the nature of governance in Western democracies post 9/11.


  4. Stacy Konkiel

    While I understand that researchers need citations to get ahead, I found the underlying assumptions of this article to be a bit cynical. Should we craft our papers for citations? Or should we craft them to add knowledge to and advance our field of study?

    1. Rob Kitchin Post author

      Stacy, there is no intentional cynical undertones to this piece. A very common complaint from early career researchers is how difficult they find it to get their papers accepted into international journals. And for established academics it is that their work is little read or referenced. I am trying to explain what editors are looking for and why some papers struggle to get published and why many papers are rarely read and cited by the author’s peers. This advice is given because most academics hope that their work will have some kind of impact and make a difference to the thinking and practice of others. Whilst the advice might be read as being cynically careerist, what I’m trying to do is help colleagues realise that hope (which little doubt has the supplemental value of aiding one’s career). If you are adding new knowledge and seeking to advance your field you are on the right path. The kind of addition though is important. Those papers that are most read and referenced are those that do not simply recycle an idea or apply it to a new empirical case or nudge it along, but ones that provide fresh insight. That is the general take home I am trying to get across.

  5. Gary Holden


    Only had time to skim your piece (late for meetings). You say:
    “I’ve not done a detailed empirical study, but can draw on two sources of observations.”

    We’ve done a study related to your essay. I can send a copy to you/anyone who is interested and doesn’t have access (just email me).

    Holden, G., Rosenberg, G., Barker, K., & Onghena, P. (2006). An assessment of the predictive validity of impact factor scores: Implications for academic employment decisions in social work. Research on Social Work Practice, 16, 613-624.

    Objective: Bibliometrics is a method of examining scholarly communications. Concerns regarding the use of bibliometrics in general, and the impact factor score (IFS) in particular, have been discussed across disciplines including social work. Although there are frequent mentions in the literature of the IFS as an indicator of the impact or quality of scholars’ work, little empirical work has been published regarding the validity of such use. Method: A proportionate, stratified, random sample, of n = 323 articles was selected from 17 Web of Science listed social work journals published during the 1992 to 1994 period. Results: The relationship between journals’ IFSs and the actual impact of articles published in those journals (predictive validity) was r = .41 (short term) and r = .42 (long term). Conclusion:
    The practice of using the IFS as a proxy indicator of article impact merits significant concern as well as further empirical investigation.

    This study was done to draw attention to the inappropriate practice of judging the impact of your tenure candidates work, on the basis of the IFS of the journals in which they they had published.

    Regards, gary

  6. Jane Gray

    While I agree we should all be striving to move towards your Level 4 (though with the caveat that social science needs contributions at all four levels), I’m a bit dubious that this translates to increased citation rates in a straightforward fashion. I don’t think the relationships in your Figures 2 and 3 (which look to me surprisingly weak) would stand up very well to a multivariate analysis across the population of scholars.

    There’s actually rather a lot of evidence that citation rates vary according to things like the tiles you choose for your articles, what you put in your abstract, whether or not you are in a discipline where self-citation is acceptable and normal, and whether or not you publish as a member of a team:

    If you want to get a lot of citations, it would appear that your best bet is to write a review of the literature:

Leave a Reply