Guest Editorial

All research, whether qualitative or quantitative, basic or applied, discipline-specific or interdisciplinary, is based on theory (Flynn, Sakakibara, Schroeder, Bates, & Flynn, 1990). Theory helps us to understand why things happen when much of our data tells us only what is happening. Whether the theory develops as a logical extension of related concepts (deductively) or in response to the observation of an unexplained phenomenon (inductively), the rigor that supports theory building affords a level of interpretation and generalizability that can have a snowball effect on later progress in research and implementation.

Building theory does come with substantial costs, but they are worthwhile. First, to build upon theory, researchers must be familiar with the latest literature on related topics. Processing and incorporating this information serve to improve our work by incorporating multiple viewpoints and factors into a nuanced understanding of phenomena and the variables that affect them. Second, theory building requires isolating or simultaneously examining multiple facets of key variables. Everything else must remain constant or be constantly mutable to understand the effect that variables have. This rigorous approach means effecting only incremental changes to learning environments over time. Slow progress can be difficult to accept from a practical standpoint in which we are preoccupied with helping the students we know personally as quickly as possible. However, building theory allows us to understand how to better help our future students and many others, accelerating our progress toward computing literacy for all.

In this special issue, we aim to reduce the costs for researchers who are studying how novices learn programming. We have selected papers that provide comprehensive literature reviews across several areas of interest, including K-12, higher education, sociocultural and cognitive factors. The literature reviews inform rigorous empirical studies and systematic literature reviews that build theory about novice programmers. By developing sound foundations, the theory-focused papers in this issue provide a strong basis for future work in both theory and design.

In the first article, Concepts before Coding: Non-Programming Interactives to Advance Learning of Introductory Programming Concepts in Middle School, Grover, Jackiw and Lundh explore methods of teaching programming concepts without requiring programming. They identified four concepts that middle school students tend to struggle with and developed non-programming activities, both digital and unplugged, to help students learn the concepts before applying them in Scratch. The concepts are variables, expressions, loops and abstraction (VELA). The authors used mixed methods to examine the effect of six VELA activities on students in three middle school classes. They found that the VELA activities consistently improved learning for students, regardless of grade, gender or prior academic preparation. They argue that non-programming-based approaches to learning programming can be appropriate, and preferable, for programming novices.

In the second article, Teaching Computer Programming with PRIMM: A Sociocultural Perspective, Sentance, Waite and Kallia applied Vygotsky’s sociocultural theory to programming education in secondary education. This paper contributes uniquely to theory-based conversations, which tend to disproportionately focus on cognitive factors. They discuss the development and testing of their PRIMM (Predict, Run, Investigate, Modify, Make) approach to teaching programming. PRIMM was implemented in 13 schools in the UK for 8–12 weeks, and the authors used a mixed-methods approach to evaluate the effect on students and teachers. Overall, students who learned with the PRIMM approach performed better than students in the control group. Teachers also reported several benefits of using PRIMM, especially that it helped them to differentiate instruction for students with various levels of skill.

In the third article, Block-based versus Text-based Programming Environments on Novice Student Learning Outcomes: A Meta-Analysis Study, Xu, Ritzhaupt, Tian and Umapathy take a systematic and numeric approach to the comparison of using block-based and text-based programming languages for novice programming. They examined 52 effect sizes comparing block- to text-based languages for both cognitive and affective student learning outcomes. The authors found a small effect size for cognitive outcomes and a trivial effect size for affective outcomes. They argue that more systematic evaluations should continue as the research continues to compare block- to text-based programming languages and that more research is needed on hybrid environments.

In the fourth article, A Theory of Instruction for Introductory Programming Skills, Xie, Loksa, Nelson, Davidson, Dong, Kwik, Tan, Hwa, Li and Ko propose and test a theory of teaching introductory programming skills that identifies and sequences fundamental skills that are often not explicitly taught to novices: tracing code, writing syntax, comprehending templates and writing code with templates. The authors explored the effects of explicitly teaching students these skills in a mixed-method study and found that they improved exercise completion rates, improved understanding and decreased errors compared to traditional instruction that implicitly teaches the skills. They conclude that explicitly teaching these skills and structuring instruction to support the development of these skills reduces cognitive demand and improves learning.

In the fifth article, CS1: How Will They Do? How Can We Help? A Decade of Research and Practice, Quille and Bergin describe the iterative development and evaluation of a model for predicting student success in CS1, called PreSS (Predicting Student Success). This work is 13 years in the making, including rounds of model development and longitudinal, multi-institutional validation. The product of these years of refinement is a web-based tool, PreSS#, that can predict student success with 71% accuracy in the first weeks of CS1. The goal of the model is to detect students who are at-risk of dropping out or failing CS1 early enough to intervene.

In the final article, A Systematic Literature Review of Student Engagement in Software Visualization: A Theoretical Perspective, Al-Sakkaf, Omar and Ahmad contribute to the debate about the efficacy of software visualization. They consider the role of student engagement as a moderating variable between software visualization and academic achievement. To do this, the authors considered 58 papers about software visualization and analyzed the 18 that included a theoretical framework. From their review, they make recommendations for the design of software visualization tools and future research.

As guest editors for this special issue, we are very pleased with the submissions we received, and this final set selected for publication. Each paper contributes a unique position in the discourse on the theory of the novice programmer. We feel this collection of manuscripts furthers our knowledge and advances the discipline, and we hope you agree.