14 Write the Paper First
(Modified on 2026-08-11)
2,042 words
David Huron
People who are inexperienced with research imagine that research is about poking around through dusty archives, doing fieldwork in a remote culture, running experiments, or analyzing data. The sad truth is that research is mostly about writing. As a researcher, you will need to write research articles, dissertations, reviews, grant applications, and perhaps even books or book chapters. Active researchers spend a lot of time writing grant proposals, writing human subjects (IRB) applications, writing to editors regarding papers, revising manuscripts, responding to reviewers’ comments, communicating with collaborators, e-mailing prospective participants, and many other writing tasks.
Some people don’t like writing. If you don’t enjoy writing, then research may not be a good career choice. Writing can be tiresome. But like practicing an instrument, writing gets much easier with experience. In research, writing is essential.
Over the decades, I have learned one truly important key to doing research well. Most researchers don’t follow this advice, but it is the single most important lesson I’ve learned. It dramatically improves the quality of one’s research, accelerates your productivity, enhances your enthusiasm, and reduces the total amount of work: Write the paper first.
By “write the paper first,” I mean the whole paper—not simply an outline, or a literature review, or the “Methods” section. Before you do anything else, you should write the paper—all the way from title to conclusion. Write the paper as though you have already completed the study. The paper should include figures (that you make up) as well as the statistical analyses (again, entirely made up). Use the capital letter X to mark everything that you’ve made up, including claims about results. Here is a sample:
A study was carried out in which 8X music teachers judged the intonation of 60X recorded amateur performances. The amateur musicians included 38X female and 22X male instrumentalists with an average age of 14.6X years…
Why write first?
There are at least six reasons why you should write the paper first:
1. Doing things right.
In order for a research article to get published, it must be vetted by expert (peer) reviewers in the discipline. As any experienced researcher can tell you, journal reviewers can be quite picky. Reviewers will not approve a submission that does not meet a very high standard of quality. Reviewers read every submission with considerable skepticism. They will question whether your research question is appropriately framed. They will consider whether your literature review is adequate or whether you failed to cite other pertinent prior research. They will question whether your conjecture follows logically from your theory. They will ask whether the conjecture is also consistent with other theories. Most of all, they will critically examine your Method.
- Are the conceptual terms in the hypothesis appropriately operationalized?
- What alternative operationalizations would have been better?
- Should a second or third study have been carried out in which the concepts were operationalized in other (contrasting) ways?
- Was the sample biased in some way?
- Are there better sampling methods that should have been used?
- Should other participants have been recruited?
- Were the instructions to participants carefully phrased?
- Were there appropriate control conditions?
- What variables were left uncontrolled that should have been controlled?
- Should the design have been with-subjects or between-subjects?
- Were the stimuli appropriately generated and presented in random order for each participant?
- Did you use the correct statistical test?
- Are your statistics reported using the proper format?
- Did you correct for multiple tests?
- Have you carefully distinguished a priori from post hoc analyses?
- Were potential demand characteristics identified and efforts made to minimize them?
- Was the conclusion expressed using appropriately circumspect language?
The sad truth is that most submissions to professional journals are rejected because there is some flaw in the design. In many cases, the Reviewers ask the researcher to “Revise and Resubmit.” This can involve re-doing the research with an additional control, or adding another test of the hypothesis. Research articles can sometimes take years before a submission is finally approved by the journal’s reviewers.
Fortunately, if the research is done carefully, reviewers will frequently accept the paper with only minor revisions, or sometimes accept the paper outright without any modifications. Experienced researchers learn that they will save themselves a great deal of frustration if they do the research carefully right from the beginning. If you do the research properly, you will only need to do it once.
Unfortunately, problems can easily pass unrecognized. You’ve probably had the experience of thinking that you have a clear idea of what you are doing, and only later realized that your idea was not nearly as clear as you thought it was. A procedure is not clear until you write it down and think carefully about the logic.
In writing the paper first, you already imagine that you are addressing a skeptical audience of expert reviewers. Think of them as looking over your shoulder while you write. Imagine them asking the skeptical questions: Why are you doing it that way? Have you considered some alternative ways of addressing the same question? There is nothing like writing the Method out completely to make it clear what it is that you are planning to do. Above all, writing is a process of clarification. If you begin the research before you have a clear idea of how you will defend the method to skeptical experts, then you are tempting an ill fate.
2. Channeling your enthusiasm.
Successful researchers are passionate about what they are doing. If you are passionate, then you will find plenty of motivation to do your work. Passion, in turn, is fed by interest. If you are not interested in what you are doing, then the work will become drudgery. It is very common to be passionate about a project when the project first begins. But if the project drags on for a long time, you will inevitably lose some of your original passion. It is important that you don’t squander your passion. In particular, it is important not to leave “hard” tasks to the end of a project. It is disheartening to have to face difficult tasks when your enthusiasm is waning. The hardest thing to do in research is writing-up the final report.
In general, aim to do the hardest tasks when your enthusiasm is greatest. In empirical research projects, there are two points in the project where the researcher is most excited. The first point is when you first dream-up your project. Take advantage of this burst of excitement: sit down and begin writing the paper as though the research is already finished. In particular, aim to write-out (and defend in prose) the complete Method section. In addition, you must write the Results section. This may sound impossible and even unethical, but recall that the Results section often consists principally of your data analysis. It is important to have a good plan for how you will analyze the data you collect.
The second point of high enthusiasm in research occurs when you have finally got your data. As you carry out your data analysis, return to the paper and revise all of the passages marked “X”.
- Instead of 38 female instrumentalists, you had 29: change 38 to 29.
- Instead of 60 recorded performances, you used 50: change 60 to 50.
- This second burst of enthusiasm will carry you through the final revision of your paper, which will then quickly find its way to submission for publication.
3. Increasing the likelihood of publishing negative results.
Research sometimes fails. You may have a nice theory, a good clear method, collect very clean data, do careful statistics, and discover that you have negative results: the collected data are not consistent with your hypothesis. You stuck your neck out, and it got chopped off. Negative results are often disappointing, but they are important nonetheless.
Negative results are not just disappointing for the researcher: they are also often disappointing for the journal reviewers. If reviewers (like you) favor your hypothesis, they will be unhappy that the results turned out to be negative. As a result, they may become hyper-critical of your method. They may seize on one or another methodological detail and reject your article because it could have been done more carefully. Negative results should be published more frequently. If you have written the paper before doing the research, then you will have a stronger method, and reviewers will be more likely to accept your paper for publication, even though they are disappointed by the results. Remember, it’s not research if you don’t invite failure. And failures are important ways in which we allow the world to tell us that our ideas are problematic.
4. Clarifying the distinction between a priori and post hoc tests.
Recall that the rhetoric of science is the rhetoric of prophecy. That is, we present a hypothesis in the form of a prediction, and then collect data to test this prediction. Of course, when you already know what the data look like, it’s easy to formulate a hypothesis that appears to “predict” what the data will look like (hindsight is not 20/20). When researchers don’t write the paper first, they may inadvertently come to believe that a post hoc hypothesis was an a priori hypothesis. Writing the paper first makes it clear what you were predicting before the data were collected. In short, writing the paper first helps keep researchers honest.
5. Anticipating the statistical analysis.
Many people collect data and then have trouble figuring out what sort of statistical analysis should be done. At this point, they commonly seek out a statistical consultant who might try to help them resolve the problem. It is possible to collect data that simply cannot be analysed using common statistical procedures. Statistical consultants may wrestle for months, trying to help resolve a difficult situation. Statistical consultants are delighted when researchers approach them first and ask for advice. A statistician can help you determine the best way to collect your data so that the statistical analysis will be straightforward. Do not collect data without a clear idea of how you plan to analyse it. Writing the paper first will help you see the big picture and will ensure that you collect data that can be analysed.
Ronald Fisher, one of the greatest statisticians in history, described the situation as follows:
To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of.
6. Avoiding repeated IRB applications.
In carrying out research with human participants, it will be necessary to get Institutional Review Board (IRB) approval. The application process is commonly fairly tedious, and it can take several weeks or months to get approval. You will want to start the IRB application process as soon as possible in order to allow you to begin collecting your data.
In the IRB application, you will need to get approval for each aspect of your interaction with human participants, including recruitment, advertising, instructions to participants, data collection procedure, debriefing, storage of data, and so on. If you make any change to your procedure (for example, re-wording the instructions to participants), you will need to go back to the IRB in order for them to approve the modification. This is tedious and time-consuming. You don’t want to submit your IRB application until you are fully convinced that you have a fully refined method. By writing the paper first, you will have a detailed procedure which you have argued carefully: the written-out procedure can simply be imported to the IRB application, and you can feel confident that you won’t have to return with a later request to modify your application.
For most of us, we start many more things in life than we complete. (It is typically easier to start things than to complete things.) One of the best reasons for writing the paper first is that pushes most of the work to the beginning. Consequently, it becomes easier to finish projects than to start them. This dramatically improves the quality of one’s research, reduces the number of abandoned projects, accelerates your productivity, enhances your enthusiasm, and reduces the total amount of work. As I noted at the beginning, write the paper first is the single most important lesson I’ve learned as a researcher.
Mantras
There is no simpler way to improve the quality and efficiency of research than writing the paper first—i.e., thinking through and planning everything in advance. That doesn’t mean it’s easy to make yourself do it! It’s not research until you write it down.