Can You Write a Paper in a Day?

In the spirit of a hackathon, we, a team of researchers, launched an ambitious project to write a complete academic paper in a day. While facing challenges such as time constraints as well as difficulties in collaborative simultaneous writing, by the end of the day, we kind of succeeded in drafting a paper. Our experience highlights that a project, such as this one, might be feasible, although there is definitely room for improvement in future collaborative writing efforts.

Rather than expecting a fully-fledged paper at the end, the emphasis should be on the hackathon as an experiment fostering a group experience. It is first and foremost a challenge that we pose to ourselves individually and as a group to find new collaboration avenues and against-the-clock writing strategies.

Sometimes you have to try to find out, and so we did, to figure out whether you can write a full paper in a single day. Inspired by hackathons, we gathered seven researchers who were brave enough to give it a try – five second-year PhD students, one postdoc, and one project leader.

The plan was simple: One day, a blank page, and that it would be a quantitative paper using the Migration-Mobility Survey. We chose the Migration-Mobility Survey because it contains many questions related to migration, so we expected to find common ground. The idea was a day would not be feasible for a qualitative study. Beyond that, there was no plan.

Process

Armed with coffee and snacks, we started at 9:10 a.m. with a blank page. First, we needed a research question, for which we split the team into pairs, and after some 20 minutes, we discussed the different ideas in the plenum. As we were all aware of the ticking clock, we converged on a topic. Unsurprisingly, there were many ideas, but beyond the constraints – “that’s not feasible with the data at hand” and “that’s already been done” – we wanted to, despite our time restrictions, work on something with social relevance.

Subsequently, we continued to work individually or in pairs as needed, with regular discussions in the plenum, typically every half an hour to an hour. In the plenum, we made the necessary decisions and kept each other informed about our respective progress and challenges, giving considerable autonomy to each participant. At the end of each plenum, we decided when to convene again. The process was organic, and we managed without clear leadership, although the project leader moderated the discussion, a bit like chairing a panel where all presenters kept to their time and comments were short and to the point.

Working Against the Clock

How did we fare? Fundamentally, the format seemed to work, and we succeeded in the sense that by 18:30 we had a draft paper with all the sections somewhat completed. Some of us were surprised at how well we managed to sift through relevant literature and build a coherent theory based on existing work on the topic; others felt like we were taking shortcuts. Even though we had a seasoned quantitative team, the available time to understand a new data set and write the appropriate code was very short. Therefore, two participants were working on the quantitative analysis, which was necessary to progress quickly enough.

What challenges did we come across? One was availability during the day, which we simply ignored at the start. Specifically, one of the participants had to supervise an exam and joined us later, one had an emergency appointment with the dentist and disappeared for a while, and another had to leave earlier. There was no time to delve into these points. Energy was no major issue: the buzz of jointly writing against the clock did wonders. Awareness of the limited time available also helped us to stay focused and avoid digressive discussions that we social scientists may be prone to.

Challenge of Sharing the Working Draft

Another challenge we did not anticipate at all was the experience of writing in the same document at the same time. While discussing questions of terminology as we faced them, and developing a policy of editing, rather than commenting, was definitely useful, the concurrent writing (we used Google Docs) meant that sometimes the text evolved so quickly that it kept moving around the screen. This might be something we would get used to, but several of us found it distracting. The solution was to work on a separate document for specific chunks, but this required coordination to ensure nobody else was working on the same paragraphs.

Getting it Wrong

There were many aspects where we failed. First up, the time pressure led us to hasty interpretations of results and serious misinterpretations that we had to fix after the hackathon. This may be a bit like those late-night essays that seem to make sense at the time, but not entirely so when we get them back graded. Based on theoretical considerations, we decided to construct scales combining different variables. This implied recoding many variables – that took a lot of time – and we had to fight temporary frustration when one of the scales failed empirically. In the interest of time, we decided against pre-registering the analysis (writing the pre-registration is also writing), but in hindsight, a pre-registration would have prevented some of the hasty (mis-)interpretations subsequently.

Fixing the theory part and bringing it back in line with the properly interpreted results was also a major effort that we had to put in the days after the hackathon. Those involved in the writing of the theoretical framework had to consult additional literature and rewrite the literature review.

Getting it Right

Writing a “paper” in a day is feasible, in the sense that all the sections were written in coherent English. We are sharing our experience to encourage others to try this format: for the experience of collaborative writing, not as support for over-productivism in academia that none of us condones. What we did not figure out is how to handle editing the draft paper in the context of a hackathon. Carefully editing a paper takes time, and it also requires a bit of distance from the text. Would an editing-athlon, or several days, be the solution? It is something to try out, especially as it may keep up the energy post-hackathon when otherwise most are also engaged in other activities. Perhaps the solution is splitting the tasks into several hackathons: finding the topic, doing the literature and theory part, the writing, and the editing. Racing against another team might be fun, but racing against the clock was motivating enough in our case.