One aspect of research reproducibility is the ability to recreate research using its description.
Another is in creating research that, other researchers who approach the problem are more likely to come to the same result.
Reproducible research should have a well-defined question and has a research design that actually answers that question.
One of the hardest aspects of doing research, especially for newcomers, is coming up with a good question.
Students naturally have plenty of interests and curiosities, but turning these into a scientific research question is another task entirely.
If we have a well-defined question, it will be easier to create research that answers it, and meaningfully contribute to the field in a way that’s more likely to be reproduced by other research on the same topic.
Consider a student who is interested in whether playing violent video games spurs aggression.
We could hone this down to a bunch of different questions - did violence rise after the release of GTA V? Does forehead temperature rise after a 20-minute violent gaming session? Does it rise more than after a non-violent session?
What if we land on this one: Do people who play shooting games demonstrate more aggression than those who don’t?
We’re likely to get a positive result, if only because of confounders.
Because the result doesn’t actually address the topic of interest, the result is less likely to track other studies in the same area, and may end up entirely nonsensical.
Consider Potential Results
Consider Feasibility
Consider Scale
Consider Design
Keep it Simple!
Honing a good research question is neither automatic nor easy.
In addition to showing students good (and bad!) examples, a checklist approach can help navigate the confusing waters between curiosity and research question.
The result can be a student project that better lends itself to meaningful results and reproducibiltiy.