My thoughts on this experience ...
I've never enjoyed group work but this term has definitely changed my perspective! Donโt get me wrong, group work will never be preferable but I do realise that there is great value in working in a good group.
Thanks so much to Group 3: Julia, Michelle and Katherine for a great term! Each group member had different insights into the project, and I loved how we decided not to take the divide and conquer approach. I've never done this before for a group assignment and it was definitely an interesting and valuable experience. We had weekly coding sessions where we just sat in a group call and coded together. Because of this we were able to ask questions and work together to help one another. This allowed us to blitz through our problems because there were 4 different minds working together. This gave us time to reproduce everything quickly and have some time to produce a great final presentation.
This term and the journey of completing the verification task has definitely made coding and R more approachable. I've always been intimidated by my computer science and engineering friends when I see all their coding. Danielle's tutorials were definitely a great way to get started and learn the basics but I've learnt that coding is basically a whole heap of googling different solution and trial-and-error.
I did not realise the full extent of the replication crisis until we started actually trying to reproduce our studies. When learning about researchers now starting to share their data, I thought (rather naively) that it would solve most, if not all, of the replication crisis. I thought that if researchers shared their data, it would be enough to reproduce everything.
Since starting the verification task, I realised that having open code isn't enough - the code my group was given was hard to understand and difficult to reproduce. We weren't given a codebook that informs us what each data variable means. Thus, when we were reproducing our tables and figures, it was difficult to determine which variables were needed to reproduce each figure. Further, no coding script was given so it was hard to know which packages to use and how the authors reached a certain value.
On the other hand, I do understand how hard it is to rubber duck every single bit of coding. It's difficult to explain and code as you go, especially when a lot of the coding done involves a lot of trial-and-error.
I've learnt that computational reproducibility is harder than it looks and is definitely easier said than done.
Once again, thanks to Julia, Michelle and Katherine for a great term! Also thanks to Jenny S, Jenny R and Kate for guiding us through this course. I've learnt so much from this course and it has definitely been one of my favourite courses that I've experienced!