Reproducibility simplified for the general audience
I recently gave a talk on reproducibility in science, which you can watch here. While my family and friends enjoyed the presentation, they expressed a desire for a simpler explanation. With this feedback in mind, I’ve attempted to distill the essence of reproducibility into a story that’s easier to grasp.
I’ve always loved cooking, especially how different ingredients mix to create a unique dish. This could be anything from baking a cake to making a traditional South Indian sambar. My grandmother, an amazing cook in our family, once said something interesting about rasam, a popular soup: if you give ten people the same ingredients, they will make ten different types of rasam.
What she meant was simple but important – it’s hard to make the exact same dish every time, and that teaches us something about science too.
Let’s talk about making rasam the same way every time and why it’s important. In cooking, different versions of rasam can be fun. But in science, we want results to be the same if we do the same experiment. This helps us trust the results. Imagine if someone says they made rasam, but it tastes like sambar – that’s a mix-up we want to avoid. If everyone understands the recipe the same way, then they can all make the same rasam. This is similar in science, where we want experiments to give us the same results every time.
“The best way to know if something is true is to check it more than once.”
Reproducibility is a big word that means “can we get the same result (rasam) again?” This is very important in science. If one person’s experiment gives a certain result, other people should be able to do the same experiment and get the same result. This is like ensuring that if you follow the recipe, you should end up with the rasam you wanted to make.
Just like we double-check our answers in an exam, scientists double-check their experiments.
But here’s the issue: not all scientific experiments give the same results when repeated. This is like when someone can’t make rasam the same way twice. This is a problem in science because if results can’t be repeated, it can lead to wasted money and even dangerous mistakes in serious fields like medicine.
I talked about this issue because we need better ways to do science, just like we have clear recipes in cooking. I suggested these three steps we should follow in science, just like in the kitchen:
Clear Counters: Just like you need a clean kitchen to cook well, scientists need to organize their data and materials clearly.
Recipe Book: Writing down every step of an experiment is important. This is like writing down each step in a recipe so someone else can make the same rasam.
Fresh vs. Used Ingredients: In cooking, we keep fresh and used ingredients separate. Scientists should do the same with their data to avoid mix-ups.
I also talked about sharing our recipes, just like sharing our experiments. If someone wants to make rasam, they can follow a recipe. I touched upon a software called R, which will help scientists do reproducible research.
So, understanding and sharing how we do things is key, whether we’re making a rasam or doing a science experiment. This helps everyone know they’re on the right track.
Let’s keep science as straightforward and reliable as a good recipe – that way, everyone can trust the results, just like they trust a good meal.