This is the second part of 7 part series discussing ideas presented in “How Learning Works: 7 Research-Based Principles for Smart Teaching”.

In the second chapter of “How Learning Works” authors show that not only what students learn but how they organise their knowledge has profound implications on their learning. Knowledge can be organised in ways that either do or do not facilitate learning, performance, and retention.

Knowledge organisations of experts and novices are very different. The former organise knowledge in a complex network that connects the facts, concepts, procedures within the domain. The latter have not yet developed such connected or meaningful organisations. They tend to build sparse, superficial knowledge structures. The teacher should not assume that students will organise new knowledge in a way that she does. She should provide students with a structure for organising new information. This structure should not only support the creation of a high number of quality connections but also fit the task the new organisations need to support.

What are the actions a teacher can take to improve the way students organise their knowledge?

Understand how you organise your knowledge. This can be done by creating a Concept map of the topic you want to teach. I find this exercise very insightful. Often it shows me the gaps in my own knowledge or makes me realise that connections I considered obvious are not so obvious after all and much harder to explain. This will push me to look for better more clear connections. I also try to think of the connections with concepts outside of the topic. Comparison of maps around different concepts may also show some repeating patterns which can be very useful when teaching.

Introduce knowledge organisations that support the way they will be used. If your goal is to introduce students to the data analysis, do not only show the functions and operations that are used in the process. Explain all the steps like data import, cleaning, reshaping, visualisation and group operations accordingly. Point out why the order of steps is important. Show how reshaping the dataset into the dataframe will facilitate visualisation or what happens if one visualises the dataset with errors or missing values. This will make much easier for students to retrieve and apply new skills when faced with the whole data analysis task.

Novices tend to organise their knowledge in silos, mirroring the way it was presented to them. Using the same example as above, if you teach those different steps of data analysis as separate lessons without pointing out how they are interconnected you cannot expect students to make the connections themselves. They may learn all the necessary steps of data cleaning or all the different ways to visualise them but may have problems performing the whole task. You can emphasise those connections even more by presenting an outline of the course and outlines of the lessons, pointing out how they fit together. In Software Carpentry workshops each lesson has a list of objectives. It is easy to skip those, but your teaching will have a much higher impact if you take the time at the beginning of the lesson and go through them. Even though I know that making connections between different lessons is important, tend to forget about it while teaching. I do make connections between concepts explicit, for example between head in shell and head() function in R or filter and select with dplyr and SQL but I rarely take time at the beginning of the workshop to explain how all the lessons fit together, draw a bigger picture. I think it is very important thing to do.

Presenting with contrasts and boundary cases will allow for the formation of more sophisticated knowledge organisations. For example, when teaching lists and generators in Python, showing both the differences and common features will help students understand deeper how they work (especially generators, which I think are harder to grasp).

Authors suggest asking students to draw concept maps at the beginning and in an ongoing manner to monitor how their knowledge organisations change. I think this not only helps teachers to see how students are improving their understanding of the concepts but also gives students time to process and organise all new information and transform it into knowledge. As I mentioned in the previous post it’s common to use concept maps at the beginning of the DC spreadsheet lesson. Unfortunately, there is not much time to do it more frequently during the workshop.

Similarly to concept maps, sorting different problems, concepts or situations into categories will help students establishing more robust knowledge organisations and unveil cases when students focus on superficial similarities instead of deep features.


How students organise their knowledge has a great impact on how efficiently they can use it. Teachers cannot assume students will figure out efficient knowledge organisations and should help students with this task.

There are few techniques or approaches mentioned in this chapter, that I rarely use. I believe I can improve my teaching by taking time in the beginning of the workshop to outline how all the lessons fit together, what is the bigger picture. I also want to try out introducing more use of concept sorting.

I’m curious what are your thoughts on knowledge organisations.