Building reproducibility in science with LEGO

If you want your research to be reproducible, you need to document your methods carefully. In practice, however, this is not always easy. During a LEGO workshop, TU/e students experience how difficult it can be to replicate someone else's work when important information is missing.

LEGO was probably the favorite toy of many architecture students. Today, they get to put their building skills to the test in Vertigo during a workshop organized by the Research Data Management team (RDM).

This team within LIS supports researchers in carefully managing, documenting, and archiving research data. The workshop gives aspiring researchers a chance to experience how important it is to properly record data and methods. After all, this is the only way someone else can reproduce your research.

Groups

During the workshop, several groups are tasked with building a LEGO car. As they work, they must write down the steps they take. They then leave these instructions for the next group, which has to recreate the car.

Two of the three groups receive tools to help them document their steps. One group gets an overview of the available LEGO bricks, while the other receives both this overview and a template for writing step-by-step instructions.

The third group has to manage without these tools.

Without tools

One of the participants is PhD candidate Krista van Zandwijk. She does not receive any tools. All she has in front of her on the table is a bag of LEGO bricks and a general instruction sheet.

After giving it some thought, she decides to write down the instructions “like a kind of recipe.” She creates a catalog of all the LEGO bricks used, draws the car from different angles, and puts together a step-by-step plan for assembling it.

Writing all of this down takes a lot of time, and without the tools, it is more difficult to leave structured instructions for the next group.

Planning

When researchers document their data, methods, and the choices they make, others can reproduce their work. This allows them to see how the results were obtained and check whether the reported effects actually exist.

It helps to make a plan in advance for how you will collect and document this data.

Not as straightforward as it seems

After half an hour, the groups switch places. Using the information left behind, they eagerly set about rebuilding the cars.

It soon leads to confused looks and questions: should this red brick go at the front or the back? Does the color of the brick matter? Which exact brick is being referred to? The participants begin to realize how much knowledge they had taken for granted and therefore had not written down.

Van Zandwijk manages to recreate the car built by the other group, which did have tools, without making any mistakes. However, the group working from her instructions is unable to do the same. This demonstrates why good, structured documentation is essential for making research reproducible.

This article was translated using AI-assisted tools and reviewed by an editor.

Share this article