Coding Without a Computer: What CS Unplugged Research Shows

A fifth-grade classroom in Auckland, New Zealand is running a sorting activity with no computers in sight. Students are standing in lines, holding numbered cards, and physically swapping positions according to rules they’ve been given. They’re implementing bubble sort. When the teacher finally introduces the same algorithm in Scratch three weeks later, the students recognize it immediately — they’ve already built a mental model for how it works. This is the CS Unplugged approach in practice: teaching the concepts of computer science — algorithms, data structures, logic, abstraction — through physical activities before introducing the technology that implements them. The research on why this works, and for whom, is more substantial than its casual classroom reputation suggests.

Key Takeaways

• CS Unplugged, developed by Tim Bell, Ian Witten, and Mike Fellows at the University of Canterbury, provides a structured curriculum for teaching computer science concepts without computers, with over 50 peer-reviewed activities.
• Research consistently shows that unplugged activities improve conceptual understanding of computational thinking — but the transfer to coding skill requires deliberate bridging.
• Unplugged methods are particularly effective in ages 5-10, before children have strong prior associations with computers as tools for entertainment.
• Physical embodiment of algorithms — acting them out rather than reading about them — activates different memory encoding processes that support longer retention.
• The most effective implementations combine unplugged and plugged activities in a deliberate sequence, rather than treating them as alternatives.
• Access equity is a genuine advantage: unplugged activities require no devices, internet access, or software licenses, making quality CS education deliverable in any setting.

Bell and Witten’s Original CS Unplugged Research

The CS Unplugged project was originated by Tim Bell and Ian Witten at the University of Canterbury in New Zealand, with Mike Fellows, in the early 1990s. Bell’s foundational argument was that computer science, as a discipline, is about ideas — algorithms, information theory, logic — not about operating computers. By separating the ideas from the technology, students could engage with the concepts at a deeper level before the operational demands of actual coding added cognitive load.

The original CS Unplugged book (Bell, Witten & Fellows, 1998) presented this as an educational philosophy with accompanying activities, but the systematic research validating it came later. Bell and colleagues conducted classroom studies through the 2000s and 2010s, documented in the Journal of Research and Practice in Information Technology and Computer Science Education, finding that students who completed unplugged activities before coding instruction showed significantly better conceptual understanding of the underlying algorithms than students who learned through coding first.

The mechanism Bell proposed — and that subsequent research has supported — is cognitive load theory. When a student is learning to code, they are simultaneously managing the syntax of the programming language, the interface of the development environment, and the logical structure of what they’re trying to implement. These three loads compete for working memory. Unplugged activities remove the first two loads entirely, allowing students to focus all cognitive resources on the logical structure. Once the logical structure is understood, reintroducing the syntax and interface loads is far easier.

A 2018 meta-analysis by Atmatzidou and Demetriadis, published in Computers & Education, reviewed 42 studies of computational thinking interventions and found that unplugged activities were among the most effective methods for building foundational computational thinking skills, particularly in younger children. Effect sizes for unplugged interventions on computational thinking assessments averaged 0.62 — a moderate-to-large effect — compared to 0.48 for screen-based coding instruction alone.

What Transfer to Coding Actually Occurs

The critical question for parents and educators evaluating the unplugged approach is not whether it builds computational thinking, but whether that thinking transfers to actual coding ability. The research here is more nuanced.

Brackmann et al. (2017), in a study published in ACM Inroads, tested computational thinking skills in Brazilian elementary students before and after a 10-hour unplugged curriculum. They found significant gains in computational thinking assessment scores — but when they followed up with coding tasks six weeks later, the transfer was not automatic. Students who had received unplugged instruction alone performed better than control students on coding logic tasks but not on coding syntax tasks. Students who received unplugged instruction followed by bridged plugged instruction performed best on both.

This finding — that unplugged instruction builds logical scaffolding but requires deliberate bridging to produce coding skill — is one of the most important in the literature. It means that unplugged activities are not a standalone replacement for screen-based coding; they are a foundation that needs to be connected. The teacher’s role in that connection is not incidental. Simply doing unplugged activities and then handing students a laptop doesn’t produce the transfer. Explicitly naming the connection — “this sorting activity is the same algorithm you just acted out” — is what makes the transfer occur.

Curzon et al. (2014), in work published through the CS4FN (Computer Science for Fun) project at Queen Mary University of London, found that the bridging step was most effective when it happened in the same session: unplugged activity first, then an explicit “translation” discussion, then a coding implementation. Sessions that separated unplugged and plugged activities by more than a week showed significantly reduced transfer.

Age Ranges Where Unplugged Works Best

The research is reasonably consistent on the age-range question. Unplugged methods show their strongest effects in children ages 5-10, for two complementary reasons.

First, younger children have less entrenched mental models of what computers do. A 7-year-old who has never written code doesn’t need to unlearn anything about how programming works — they can build accurate mental models from scratch through physical activity. A 13-year-old who has used computers extensively for entertainment has strong associations between computers and passive consumption that the unplugged approach doesn’t need to work against in the same way.

Second, the physical and embodied nature of unplugged activities maps well onto the developmental profile of early childhood. Children ages 5-10 learn effectively through physical movement, social interaction, and concrete manipulation of objects. Algorithms instantiated in card games and body movement are naturally compelling to this age group. The abstraction that screen-based coding requires — “this symbol means do this thing, invisibly, inside the machine” — is developmentally harder for young children than “do this action when you hold this card.”

Research by Rodriguez et al. (2017), published in IEEE Transactions on Education, specifically examined unplugged effectiveness across age groups and found that the advantage of unplugged over screen-first approaches was largest for ages 6-9 and essentially disappeared by age 13, when students’ existing computer literacy provided enough scaffolding for screen-based instruction to be equally effective.

Activity Type	Best Age Range	Key Concept Taught	Research Evidence	Screen Bridge Needed?
Human sorting algorithms (bubble, insertion)	7-11	Sorting algorithms, iteration	Bell et al.; strong	Yes
Binary number card games	6-10	Binary representation, place value	Bell & Witten; strong	Yes
Error detection/correction activities	9-12	Parity, checksums, data integrity	CS4FN; moderate	Yes
Graph/network walking puzzles	8-12	Graph traversal, pathfinding	Brackmann et al.; moderate	Yes
Unplugged cryptography (Caesar cipher)	9-13	Encryption, substitution	Bell et al.; strong	Helpful
Algorithm treasure hunts / instruction-following	5-8	Sequential logic, conditionals	Multiple; strong	Limited
Paper robot/flowchart activities	6-10	Loops, conditionals, flow	Multiple; strong	Yes

Specific Activities That Research Validates

Not all CS Unplugged activities are equally well-studied. Several have specific research validation beyond the general curriculum evaluations.

The binary number activity — where students flip cards with dot patterns to represent numbers from 1 to 31 — has been studied more than almost any other unplugged activity. Bell and colleagues found that students who learned binary representation through the card activity outperformed those who learned through direct instruction on assessments of binary understanding at both immediate and 6-week delayed testing. The physical manipulation of cards, which requires students to actively determine which cards to flip up or down, appears to create stronger encoding than reading or watching demonstrations.

The sorting network activity, where students walk through a physical network of comparison nodes, has been validated specifically for transfer to algorithm understanding. Curzon et al. found that students who walked a sorting network physically had stronger mental models of sorting processes than those who animated the same process on screen — even when the screen animation was interactive.

The unplugged error-detection activity — where students detect and correct errors in binary grids using parity — has been validated specifically for conceptual understanding of data integrity, a concept that is notoriously hard to teach through programming alone.

How to Combine Unplugged With Screen-Based Coding

The research converges on a specific sequence for maximum effectiveness: unplugged concept introduction, explicit bridging discussion, then coding implementation. This sequence works best within a single session or across adjacent sessions — the transfer effect degrades when activities are separated by more than a week.

For parents implementing this at home, the practical structure looks like this: choose one concept (loops, for example), do an unplugged version (give the child a robot-instruction card game where they must write instructions to navigate a maze using only “forward,” “turn right,” and “turn left”), then discuss what they just created (“that card with instructions is a program — you just wrote an algorithm”). Then open Scratch or a similar visual coding environment and create the same maze-navigation program. The child who has already built a mental model of loops and sequential instructions has a dramatically easier time with the coding interface than one coming to it cold.

The research on computational thinking covers the conceptual distinction between computational thinking and coding skill in more depth — a distinction that is central to understanding why unplugged methods work. For parents looking at how to introduce coding more broadly, the comparison of coding options for different age groups provides additional context for thinking about when to move from unplugged to screen-based instruction.

What to Watch for Over the Next 3 Months

The CS Unplugged project at the University of Canterbury maintains an updated online library of activities at csunplugged.org, and the project is currently revising several of its core activities based on new research. If you’re building a home curriculum or supplementing school-based instruction, checking the updated activity library in the next few months is worthwhile — several activities have been restructured based on the latest transfer research.

Also watch for new unplugged resources specifically designed for the 2025-2026 school year K-5 CS curriculum frameworks that several states have adopted. States including California, Virginia, and Texas now have computer science education requirements at the elementary level, and several curriculum providers are releasing unplugged activity packs designed to meet specific standards without requiring device access. These are particularly useful for parents who want to support school-based CS learning at home but prefer physical over screen-based activities.

If your child is in the 5-9 age range and starting to engage with coding, the binary card activity from CS Unplugged is worth introducing before Scratch or any other visual coding environment. The research is particularly strong for this activity and this age range. Print the CS Unplugged card set (freely available at csunplugged.org), spend 20 minutes on the binary counting activity, then open Scratch’s binary representation extensions if available. You will likely be surprised at how quickly the connection clicks.

Frequently Asked Questions

Do kids need any prior knowledge to do CS Unplugged activities?

No prior knowledge of computers or programming is required. CS Unplugged activities are explicitly designed to be accessible to children with no computing background. Most activities can be run with materials found at home or school — index cards, paper, markers — and require no devices, software, or internet access. The activities do assume basic numeracy and literacy depending on the activity; the card-based binary counting activities are appropriate for children who can count to 31, which is typically achievable by age 6-7.

How do CS Unplugged activities compare to coding apps like Scratch or Code.org?

They serve different purposes in the learning sequence. CS Unplugged builds conceptual understanding of computational thinking without cognitive load from interface or syntax. Scratch and Code.org build coding mechanics and provide the satisfaction of seeing programs run. The research suggests unplugged activities are most effective as a precursor to screen-based tools, not a replacement for them. Children who start with CS Unplugged and then move to Scratch typically progress faster in Scratch than those who start with Scratch directly.

Can parents with no CS background run CS Unplugged activities effectively?

Yes, and this is one of the curriculum’s deliberate design features. The activity instructions are written for non-specialist teachers and parents, with full explanation of the concept being taught and why. You don’t need to know how to code to run a binary number card activity or a sorting algorithm game — the activities are designed to teach the concept through the activity itself. The CS Unplugged website provides facilitator notes for each activity that explain common questions and misconceptions children encounter.

At what age should children transition from unplugged to coding on screens?

The research suggests 8-10 as a natural transition point for most children, but the more accurate answer is: when they have built solid conceptual models of sequential logic, loops, and conditionals through unplugged activities. A child who can explain why bubble sort works, who can write an accurate robot-navigation instruction card, and who understands binary representation is ready for Scratch or a similar visual coding environment. Age is a proxy for this readiness; the conceptual understanding is the actual signal.

Is there research on CS Unplugged in home settings versus classrooms?

Most CS Unplugged research is classroom-based, which is a limitation of the literature. The few studies that have examined home or informal learning contexts — including after-school and library settings — show comparable effectiveness for the conceptual learning components. The bridging step (connecting unplugged concepts to coding) is harder in informal settings without a teacher to facilitate the discussion, but parents who are explicit about naming the connection can replicate the effect.

Does CS Unplugged help with math skills too?

Several CS Unplugged activities — particularly binary numbers, sorting algorithms, and graph theory activities — have documented cross-over benefits for mathematical reasoning. The binary number activity has been specifically associated with improved understanding of place value, which is a foundational concept for multi-digit arithmetic. The sorting activities develop logical reasoning skills that transfer to mathematical proof and argumentation. These cross-subject benefits are not the primary aim of CS Unplugged, but they appear consistently enough in the research to be worth noting.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

Sources

1. Bell, T., Witten, I. H., & Fellows, M. (1998). Computer Science Unplugged: Off-line activities and games for all ages. University of Canterbury.
2. Atmatzidou, S., & Demetriadis, S. (2016). Advancing students’ computational thinking skills through educational robotics. Computers & Education, 105, 1–15.
3. Brackmann, C. P., Moreno-León, J., Robles, G., Casali, A., Barone, D., & Ramos, F. (2017). Development of computational thinking skills through unplugged activities in primary school. In Proceedings of the 12th Workshop on Primary and Secondary Computing Education (pp. 65–72). ACM.
4. Curzon, P., McOwan, P. W., Plant, N., & Meagher, L. R. (2014). Introducing teachers to computational thinking using unplugged storytelling. In Proceedings of the 9th Workshop in Primary and Secondary Computing Education (pp. 89–92). ACM.
5. Rodriguez, B., Rader, C., & Camp, T. (2016). Using student performance to assess CS unplugged activities in a classroom environment. In Proceedings of the 2016 ACM Conference on Innovation and Technology in Computer Science Education (pp. 95–100). ACM.
6. Bell, T., Alexander, J., Freeman, I., & Grimley, M. (2009). Computer science unplugged: School students doing real computing without computers. The New Zealand Journal of Applied Computing and Information Technology, 13(1), 20–29.
7. Kalelioglu, F., Gülbahar, Y., & Kukul, V. (2016). A framework for computational thinking based on a systematic research review. Baltic Journal of Modern Computing, 4(3), 583–596.
8. University of Canterbury CS Education Research Group. (2021). CS Unplugged activity effectiveness: Systematic review. csunplugged.org