Read on Twitter
Universities are not equipping their scientists with the tools or help they need to get science done.
A thread based on a video conf with @sirileknes @Marie_Eikemo & @GuroLoseth
A thread based on a video conf with @sirileknes @Marie_Eikemo & @GuroLoseth
The paperwork involved in a project is huge. Reporting to overseers (uni/funders/govt) requires a ton of work, which eats time from doing the science stuff (more on what that is later) .
The paperwork for projects is hard to out-source, as they are core to the projects and can only really be thoughtfully written by PIs and researchers that grasp the theoretical and practical application of the research being done/suggested.
So much time spent getting all through all the red tape means that when they finally have time and sit down to do the science stuff, the time is limited and their cognitive capacity is spent.
The science stuff are the things that are necessary for career progression and that give feelings of achievement. Like analysing and writing up papers, development bf new skills to apply new and better practices, implement smart workflows that make working easier and reproducible
Not having time for the science stuff, means constantly being behind, and never being able to learn the skills that might have made you more efficient when you did have time to do them. Everybody looses.
Unis get less papers, less grants and less successful researchers. Researcher get less time to do the things that would advance their careers and they end up leaving academia.
So what do we need? Imo, career paths in academia for data scientists and research software engineers.
In terms that are maybe different than industry uses them, but definitely needed.
In terms that are maybe different than industry uses them, but definitely needed.
If you are a researcher and have 3hours per week to actually work with data, spending that time doing data cleaning is a waste. You should be helped by someone who has already been working on system that will make data cleaner for you.
Researchers at uni don't need yet another overseer of beurocratic things that deals with theoretical and overarching systems problems, they need someone to be hands on with them and their data.
A team of data scientists and research software engineers run either institutionally or centrally that can be hired on fixed term basis to groups needing help would alleviate the researchers a lot.
These would then get offices (when we do that stuff again) and sit close to the researchers and help them by developing pipelines and systems for getting their data collection and data in shape already at the onset of a project.
Doing this before data collection etc means the data can be in a tidier format, more securely stored and better logged from the start, saving time at the end when analysis and reporting happens. Cleaning and logging data after completion is a nightmare.
Assuming researchers have the knowledge, time and interest to set up these systems themselves is an institutional oversight. Not everyone is interested in becoming so technically savvy, they just want data to run their analysis on.
A team for hire, trained scientists that don't necessarily want to create their own research groups, with stable permanent positions can be an amazing asset to the academic staff at the university.
Having stable positions would mean these researchers for hire do not need to worry about getting their papers out or getting their grants, but can focus on fascinating the research of other.
This is essentially what @KWalhovd and @LCBC_UiO have done for me. As a staff scientist, most my time is spent creating good data collection routines, cleaning and harmonizing existing data, making systems for long term storing and reproducibility and tracing of the data
Ensuring data quality and making sure the data is usable to the research staff. I get involved in papers and analyses too, time permitting. But making sure everyone else had what they need to do their job, is exciting to me.
Seeing the top notch papers come out of our center, and help disseminate results making R packages and shiny dashboards is really rewarding to me personally.
I am grateful that the center PIs have the foresight to see that the academic staff rigidity is hindering more like me to stay in academia, because there is no room for scientists that are fine with being an 'author in the middle' rather than a first or last.
Having staff like me, and my excellent coworkers doing dev-ops type work for @LCBC_UiO, on permanent positions, means we can focus on fascilitating everyone at the center to Excell, and not focus on our own publication/grant record. This is a benefit to all in and the center.
