Publication doesn't need to run serially!
We can work on several projects in parallel!
When we work in parallel, we can always make progress on at least one project, even if other projects are stuck, stalled, or failing.
My personal workflow is parallel and distributed (delegated).
Personally, I tend to have a workflow with 8+ projects at the same time, each in different phases of development.
- two in outline/design phase
- two in development/programming/ethics/logistics phase
- two in data-collection (delegated to RAs)
- two in data-analysis
- two in writing/editing/submitting/revising
With this many projects, there will always be something to work on.
I couldn't allocate more time to these projects than I already allocate.
For example, if I design a project and pass it off to my supervisor for review, I am waiting for a reply. Rather than waiting, working on nothing, I can turn to a different project where I can make progress.
Likewise, I cannot make data collection go faster than it will go.
I delegate my project to my research assistants and they collect data: students sign up and it takes as long as it takes. In the meantime, I work on other projects. If I didn't work on a different stage in a different project, I would be waiting, which would waste my time.
Likewise-likewise, all the way through the process.
If I submit a paper, the next step is to wait to hear back from reviewers, which can take months. In the meantime, I start writing the next paper. Based on what? Based on the queue: the prior study's data was collected and now it is ready for analysis and writing. Additionally, I want to make sure my RAs are always collecting data so I'm always making sure to program new experiments. To do that, I always need to be designing new experiments. What do I design? That is informed by the results that I just wrote up in my most recent paper! I also sprinkle in random novelty projects.
Personally, I like having multiple projects at each stage so I can work on whichever stage I feel like. For example, if I'm writing a paper on Tuesday, but then I come to a block on it, I can put it aside on Wednesday and work on the other paper I'm writing. Or I can pick up data-analysis for a day instead. I can go with my whims and it feels more like play than "work".
This workflow is in addition to the uncounted number of potential studies in "brainstorm" phase and a number of other projects where I've collected data but I have side-lined them for various reasons (or just lost interest).
And these are only counting my first-author projects.
If we add in collaborations...
What about quality?
There is no reason to think that a person who produces high-quality papers must necessarily produce fewer papers.
The task of publishing papers does not get linearly more difficult.
If anything, it gets easier to publish the more you publish!
- My writing skills are in constant use so these skills are honed and maintained.
- I notice patterns so I can navigate journal submissions efficiently.
- I think about experimental design in terms of publishing: I am designing to publish a coherent study.
- I read reviewer comments so often that I imagine potential reviewer comments while writing and designing.
- I design experiments that will be interesting no matter what. I often aim to get multiple papers from each project thanks to efficient experimental designs (not "salami publishing").
Writing for publication is a skill.
If you do it often, you will sharpen your skills.
If you do it rarely, you will get rusty.
Index
Return to Graduate Psychology
Jump to Reviewing papers