RDM introduction#
Objectives📍
What Research Data Management (RDM) means and why it is important
The FAIR principles
The Research Life Cycle
RDM & Open Science
Overview of RDM components
Introduction of our example project
Summary#
Fig. 1 Research Data Management: making it possible to retreive data from last year. The Turing Way project illustration by Scriberia. Reused under a CC-BY 4.0 license. Original version on Zenodo. http://doi.org/10.5281/zenodo.3695300.#
Research Data Management (RDM) entails practices of storing, annotating, and re-using research data. We are all doing it every day: we all store, sort, and annotate our data in our own way, namely in the way we’ve always been doing it and how we feel the most comfortable with. This might be sufficient, you think - at least that’s how it feels in the moment. However, leaving RDM to be an individual issue results in several restrictions for the individual as well as for science per se. When getting back to your data after some time [1] and try to find, read, and understand it, you might realize that your way of storing, sorting, and annotating your data isn’t so intuitive after all. Do you find everything you need in a short time? Do you understand the different folders and files instantly? Does your code still run without errors? Do you know which was the latest version of your project? Maybe…or maybe not? So, if you yourself, the master of your data, are unable to confidently reuse your data…what about others?
The field of RDM provides us with recommendations, schemes, and standards which eventually will save all of us a lot of time and memory while also fostering collaboration, preventing data loss, and improving the overall quality of science.
The Research Data Lifecycle#
RDM techniques are aiming to support our research work in all stages of the so-called Research Data Lifecycle:
Fig. 2 Research Data Lifecycle. The Turing Way project illustration by Scriberia. Reused under a CC-BY 4.0 license. Original version on Zenodo. http://doi.org/10.5281/zenodo.3695300.#
Depending on your career stage, you might not have been confronted with every single stage of this cycle yet. However, each step of the research life cycle needs to be considered for RDM. Developing a research idea is exciting and fun but your idea will only come true when you care just as much about the other stages in the research data lifecycle. Here’s what you will encounter during the different stages:
design research
plan data management
plan consent for sharing
plan data collecting, processing, protocols, and templates
explore existing data sources
collect data
capture data with metadata
acquire existing third party data
enter, digitize, transcribe, and translate data
check, validate, clean, anonymize
derive data
describe and document data
manage and store data
analyse and interpret data
produce research outputs
cite data sources
establish copyright
create user documentation
create discovery metadata
select appropriate access to data
publish/ share data
promote data
migrate data to best format/media
store and back up data
create preservation documentation
preserve and curate data
conduct secondary analysis
undertake follow-up research
conduct research reviews
scrutinize findings
use data for teaching and learning
The FAIR principles and practices#
Behind every RDM technique stands the FAIR principles. FAIR means that research and data should be
Findable: Annotating your research and data with proper Metadata goes a long way. This enables to search for specific keywords and find the respective research items.
Accessible: Data users must know how to access the data. This starts with simple metadata about hard- and software requirements and involves information about potential authentication and authorisation procedures.
Interoperable: Reuse of data and collaboration can only work if the data is provided in a way that it can be easily integrated with other data or easily operated with other applications.
Reusable: Again, sufficient metadata and requirements for using your data in different settings have to be given to make your data reusable.
Here’s what you should know about FAIR:
By making your data FAIR you expand the life-cycle of your data and contribute to sustainability in the digital world, which brings benefits for your own research as well as the whole science community.
FAIR does not equal open. Following the FAIR principles does not necessarily mean that you have to share your data with everyone. Your data can meet the FAIR principles but still be shared under certain restrictions.
The intention of the FAIR principles is not to have everybody apply all the principles to all their data perfectly right away. Implementing just a few of the recommended steps helps you and the science community to get one step further toward open science. The important thing is simply to start somewhere to get used to practicing the FAIR principles on your data regularly. Step by step.
The FAIR principles are guidelines, not a standard.
RDM & Open Science#
Fig. 3 Private Data. The Turing Way project illustration by Scriberia. Reused under a CC-BY 4.0 license. Original version on Zenodo. http://doi.org/10.5281/zenodo.3695300.#
Personally, I think that there are two misunderstandings when people think of Open Science.
First, just putting everything you have in your project folder on some repository and sharing it with the world does not make your research “open”. If your data isn’t FAIR, people won’t be able to derive any more information from it than you’ve given in a conventional paper. People might not be able to open your files, don’t know what your data is about without sufficient metadata, cannot run your code due to proprietary software you used etc. etc.
Second, making your research open does not mean that you have to share everything you have with everyone in the world. Sometimes this is simply not possible, for example when it comes to personal data. This is a good thing, people deserve their right to having their private data protected. What you can still do then though, is to share comprehensive metadata to let people know what your data is about and grant them authorized access (i.e., Data User Agreements). If you then additionally share your code, other researchers can still reuse your resources for reproducibility and/or replicability purposes.
Thus, RDM and Open Science go hand in hand: without proper RDM your research won’t be open.
Applying RDM standards to your work and contributing to Open Science benefits you by increasing your data quality, accountability, reputation in the science community, time efficacy, visibility & transparency and many more:
Fig. 4 Reasons for sharing your work. The Turing Way project illustration by Scriberia. Reused under a CC-BY 4.0 license. Original version on Zenodo. http://doi.org/10.5281/zenodo.3695300.#
RDM components#
As described above, RDM aims to support you in all stages of the research data lifecycle. However, there are some components of RDM which are especially important and which we will get to know over the course of this school:
Planning & Design: Making a Research Data Management Plan (rdmp) and pre-registering your study
Data collection: Organizing your data and annotating it with metadata
Analysing data: using open source software and version control as well as ensuring reproducibility of your research
Publishing & sharing data: Popular repositories, data privacy, copyright, and licenses
Our Example Project for the Week#
The Choice Reaction Time Task (Deary & Liewald, 2011) [2]#
In a choice reaction time task, participants must make a decision related to the stimulus (such as what shape it is) and respond accordingly. E.g., participants must click left if the stimulus is blue and right if the stimulus is red.
During our Psychopy Module we will develop the code for such a choice reaction time task (ChoiceRTT). It won’t be exactly the same as in the example by Pavlovia (we’ll play around with different kinds of stimuli) but the principle stays the same. Hence, it doesn’t hurt to make yourself familiar with the basics of the experiment as you will need it for upcoming tasks during the school:
Task
Read the introduction of the paper by Deary & Liewald (2011) to gain new or refresh knowledge about choice reaction times.
Task
Go to the Pavlovia demo-experiments website and do the ChoiceRTT task.
Task
Go to the Pavlovia GitLab and take a peak at what input and output data for the experiment exists.