Designing for deeper learning in a blended computer science course for middle school students – Grover, Pea & Cooper (2015)
My research skills clearly peaked at the end of our article review period. Of all the papers I read over the past 5 weeks, this one had the most solidly designed study. And (excitingly!), it’s directly applicable to my teaching content. It was also 40 pages long and went very in depth into the details of their experiment, so I’ll do my best to not get lost in the weeds.
Researchers developed a 7-week curriculum for middle school students entitled “Foundations for Advancing Computational Thinking” (FACT), whose goal was to “prepare and motivate middle school learners for future engagement with algorithmic problem solving” (pg.199). Sounds boring, but this is actually very important in building capacity for future computer science work in secondary school and beyond. Algorithmic problem solving (specifically “serial execution, looping constructs, and conditional logic” (pg.201)) is transferrable between programming languages and is foundational in the development of computational thinking. The other goals of the study were to change the students’ perception of CS, and to encourage “deeper learning” (pg. 201).
A quick note on this study’s definition of “deeper learning”: this concept is concerned not just with content but also a student’s ability to problem solve, collaborate, communicate, and engage in self-directed learning (pg.204). Deeper learning extends beyond the cognitive domain and works to include important skills from the intrapersonal and interpersonal domains. Researchers choose a “deeper learning” framework because of its focus on the transferability of skills as students learn in one setting and are able to apply it in another.
Transferability of skills was actually built into the assessments used to collect data for the study. During the 7-week course, students learned the basics of algorithmic problem solving using the very kid-friendly Scratch platform. Scratch uses block-based coding that allows students to focus on the problem and not stuck looking for syntax errors (*Disclaimer: I’ve had really good luck using Scratch in my own classroom). Usually Scratch is used for game creation, but for this course it was used as a space to test algorithms with a variety of learning goals. At the end of the course students were then given the “preparation for future learning (PFL)” assessment in which students had to apply their computational thinking knowledge developed using block-based code to text-based code, specifically Pascal and a “Java-like” language (pg.201).
The FACT course was piloted in two iterations at the same middle school. The first iteration was a more traditional face-to-face course that used online tools, while the second iteration was delivered entirely online through the OpenEdX MOOC platform. Researchers used the feedback from the first iteration to significantly inform the design of the second iteration. Findings were collected through pre & post assessments, PFL, final projects, and interviews.
They did not run a control group (one not exposed to FACT), so the findings for this study can really only be compared between the two iterations or discussed as a whole. Overall, they found that the results from the students participating in the MOOC iteration had similar-to-better understandings of algorithmic structures. Both groups of students also demonstrated their knowledge more effectively in the final project and interview than they did in the post assessment. The separate “PLF” test left the researchers feeling “cautiously optimistic” although they felt that the test itself was too hard (pg.222). Students were able to transfer some of their skills to text-based problems, but struggled with loops and variables, which also showed on their post assessments. The open-ended questions on the post assessment also revealed that students gained a better understanding of the breadth of topics in computer science and its opportunities for problem solving and creativity.
At the time of publishing, this study was one of the first to have developed an online introduction to CS course that provided empirically positive results in the learning gains of middle school students (pg.224). We all anecdotally support middle school students building up their computational thinking, but it’s important to have the data. At this age students are going through some serious cognitive development and it’s critical to slip in some analytical reasoning to support their future STEM studies. Let’s get more pre-teens practicing their algorithmic problem solving skills!
Grover, S., Pea, R., & Cooper, S. (2015) Designing for deeper learning in a blended computer science course for middle school students, Computer Science Education, 25(2), 199-237, DOI: 10.1080/08993408.2015.1033142