You have everything in place: manuscript, figures, scripts or analysis files, and data. You are done. Well, there are still steps to possibly level up your research game! The codeword is open science, a set of practices to make research more transparent and accessible.
A good place to start is making your data open. While this practice is a requirement at the Western Norway University of Applied Sciences (HVL), it is important to consider that it’s not just an administrative hurdle to overcome. Instead, open data has two main advantages for yourself and others.
- It makes your research more transparent and accessible.
- It gives you a certain level of security over data loss and forgotten details.
Advantage 1: accessibility & transparency
Accessibility is not only helpful when aiming to publish in many journals which now require open data (as long as data are not sensitive). Additionally, it will give others the chance to reproduce and replicate analyses or to form a better understanding of the study for judgments about the study quality. Your data might also be directly used in further studies, for example in meta-analyses, which are used to analyze the evidence from a selection of studies.
Furthermore, prior to publication, a properly prepared dataset can be given to an independent team to check whether your analyses hold when being executed by others. This can be done between co-workers or teams from completely different universities. The checks can also be testing the understandability of your data in the context of the manuscript.
Thinking fully within the framework of wanting to advance science, open data seems to be a no-brainer: it speeds up follow-up research by skipping waiting times to request access to the data and might circumvent paywall problems (data being hidden behind a payment-based subscription). This entails the opportunity for others to check the robustness of your findings and give your work more attention within the field by publishing on it.
Advantage 2: protection from data loss
Usually, researchers juggle several projects at the same time. Over time it is easy to lose the overview of all the projects and details within them. So, imagine you get a great idea about building upon a study you have already conducted. When you revisit your data you notice that you cannot remember all the variable acronyms you chose for the spreadsheet and for some reason there seem to be some columns missing.
Storing your data online at www.dataverse.no or other repositories can protect you from such problems. Data deposition requires a minimum effort of uploading the data which you already use, preferably in two pieces: the raw data, and the processed data. In addition, a descriptor file containing variable names and how the data were processed (e.g., how exclusions were made) can be a lifesaver for your future self. If it is not a lifesaver, then you will at least have everything tidily organized in one place, and exchanging your computer for a newer model will not be a headache anymore.
It is never too late to start a new habit. In the case of open data, minimal efforts of a few minutes per data set can save hours to days of work and prevent data loss. And it gets just better: at your local HVL library or dataverse.no, there will be people helping you with all steps and even curating your dataset to make maximize processability. How to now resist engaging in the open data reform?
PhD student at HVL