It is widely agreed that science instruction should help students build new knowledge on the foundation of their prior knowledge. Responsive teaching refers to a family of teaching strategies that pursue and build on student ideas. We introduce a particular approach to responsive teaching and examine how it can be supported by the use of computational models. We analyse an 8th grade science teacher’s facilitation of a class discussion near the end of a lesson on sound. We present a moment-by-moment characterisation of her responsive teaching moves, highlighting the ways she used a computational model to help students articulate and examine their thinking. Our findings make empirical contributions to literature concerned with responsive teaching and literature concerned with the role of computational models in constructivist approaches to instruction.

Large Language Models (LLMs) have the potential to fundamentally change the way people engage in computer programming. Agent-based modeling (ABM) has become ubiquitous in natural and social sciences and education, yet no prior studies have explored the potential of LLMs to assist it. We designed NetLogo Chat to support the learning and practice of NetLogo, a programming language for ABM. To understand how users perceive, use, and need LLM-based interfaces, we interviewed 30 participants from global academia, industry, and graduate schools. Experts reported more perceived benefits than novices and were more inclined to adopt LLMs in their workflow. We found significant differences between experts and novices in their perceptions, behaviors, and needs for human-AI collaboration. We surfaced a knowledge gap between experts and novices as a possible reason for the benefit gap. We identified guidance, personalization, and integration as major needs for LLM-based interfaces to support the programming of ABM.

Human-Computer Interaction (HCI) and education researchers have applied Augmented Reality (AR) to support spatial thinking in K-12 education. However, fewer studies support spatial thinking through spatial exploration. Room-scale AR, a recent technology development, brings new opportunities not yet researched. We developed NetLogo AR, an AR authoring toolkit, that allows children to play with, design, and create room-scale AR experiences that combine AR with computational models. To acquire a deeper and more nuanced understanding of children’s interactions with this new technology, we conducted eight-week participatory design sessions with seven children aged 11-13. We analyzed 48 hours of video data, interview transcripts, and design artifacts. Children were enthusiastic and engaged in spatial thinking activities. We affirmed room-scale AR’s role in spatial exploration by comparing it with other supported modalities. Building on existing studies, we propose a new AR design framework around spatial movement and exploration that could help inform design decisions.

Computational modeling plays an important role in scientific research and education. However, many models do not communicate with the real world, which might limit their learning potential. In this proposal, we introduce the design of NetLogo AR, an authoring toolkit that transforms NetLogo agent-based computational models into room-scale AR experiences. We are inspired by studies that integrate real-world environments into computational models, including bifocal modeling and participatory simulation, and those that leverage Augmented Reality (AR) to create authoring tools for children. We describe our design goals that focus on raising the ceiling and lowering the threshold for potential AR designers: researchers, educators, and children. For the IDC conference, we propose an example learning activity to demonstrate the capabilities of NetLogo AR. By incorporating dynamic and unpredictable real-world inputs, NetLogo AR has the potential to inspire future designs of AR experiences and enrich the learning experiences of agent-based models for children.

Agent-based modeling (ABM) has become a major approach to promote computational thinking and complex systems thinking in K-12 education. However, agent-based programming (ABP), the computational foundation of ABM, is less defined and discussed in previous literature. Summarizing previous studies around ABP from computer science and education, we argued for the potential benefits of introducing ABP to youth. Rooted in the interest development theory, we presented the design of a scaffolded agent-based programming space, the Pocketworld Playground (POP), that aims to engage out-of-school online young learners through developing their interest in ABP. The POP was built in Turtle Universe (TU), the mobile incarnation of NetLogo. Using a mixed-methods approach to analyze log data and artifacts created by learners, we found that POP successfully engaged learners with ABP practices; helped develop situational and individual interest; and contributed to TU’s emerging online community. Finally, we discussed design lessons that could benefit other online learning designers.

There has been a growing interest in the use of computer-based models of scientific phenomena as part of classroom curricula, especially models that learners create for themselves. However, while studies show that constructing computational models of phenomena can serve as a powerful foundation for learning science, this approach has struggled to gain widespread adoption in classrooms because it not only requires teachers to learn sophisticated technological tools (such as computer programming), but it also requires precious instructional time to introduce these tools to students. Moreover, many core scientific topics such as the kinetic molecular theory, natural selection, and electricity are difficult to model even with novice-friendly environments. To address these limitations, we present a novel design approach called phenomenological programming that builds on students' intuitive understanding of real-world objects, patterns, and events to support the construction of agent-based computational models. We present preliminary case studies and discuss their implications for STEM content learning and the learnability and expressive power of phenomenological programming.

While the Next Generation Science Standards set an expectation for developing computer science and computational thinking (CT) practices in the context of science subjects, it is an open question as to how to create curriculum and assessments that develop and measure these practices. In this poster, we show one possible solution to this problem: to introduce students to computer science through infusing computational thinking practices ("CT-ifying") science classrooms. To address this gap, our group has worked to explicitly characterize core CT-STEM practices as specific learning objectives and we use these to guide our development of science curriculum and assessments. However, having these learning objectives in mind is not enough to actually create activities that engage students in CT practices. We have developed along with science teachers, a strategy of examining a teacher's existing curricula and identifying potential activities and concepts to "CT-ify", rather than creating entirely new curricula from scratch by using the concept of scale as an "attack vector'' to design science units that integrate computational thinking practices into traditional science curricula. We demonstrate how we conceptualize four different versions of scale in science, 1. Time, 2. Size, 3. Number, and 4. Repeatability. We also present examples of these concepts in traditional high school science curricula that hundreds of students in a large urban US school district have used.

National and state educational initiatives are increasingly prioritizing computer science and computational thinking as valued sets of skills and practices. However, despite ongoing efforts to broaden participation in computing, the field faces increasing underrepresentation of women and other racial and ethnic groups. In this paper, we argue that physical computing can provide rich, varied entryways into computing practices, and that kits that support social or collaborative use may be more compelling for underrepresented groups. We present the design of a scaled-up physical computing kit, called StegaCircuits, which we created to foster more social, exploratory introductions to computing. We present preliminary data collected during user tests at a Maker Faire that suggest StegaCircuits supports social introductions to physical computing through multi-user interaction. Finally, we highlight promising facets of the design that may positively impact cognitive and emotional responses to computing fields.

For many years, beginning computer science students have been asked the seemingly simple question: what makes one algorithm better than another? Usually, this concept of "best-ness" is introduced with the many algorithms that are used to sort data. Thus, students are overwhelmed by the intricacies of sorting while simultaneously trying to understand one of the greatest questions in computer science. In this work we take inspiration from embodied cognition and describe a new, constructionist curriculum for approaching algorithmic efficiency. In order to help students construct understandings of both sorting algorithms and algorithmic efficiency, we present a curriculum that focuses around participatory simulations (PartSims) where students "play as" datapoints and experience sorting algorithms from an entirely different perspective. Instead of first experiencing the usual top-down approach to sorting where datapoints are manipulated by an "observer," each student follows a simple set of rules which, collectively, end in an emergent sorting algorithm. In this way, instead of talking in the abstract about the different pros and cons of sorting algorithms, students have different physical experiences across the different algorithms. With this embodied, grounded experience, students then construct for themselves metrics for comparing the different algorithms by evaluating the differences in their physical experiences. This two-week curriculum is targeted towards beginning computer science students (high-school or collegiate level) and utilizes the Parallax hackable Electronic Conference Badge and NetLogo to track the students experience in the PartSims throughout the curriculum.

The last ten years have seen a proliferation of introductory programming environments designed for learners across the K-12 spectrum. These environments include visual block-based tools, text-based languages designed for novices, and, increasingly, hybrid environments that blend features of block-based and text-based programming. This paper presents results from a quasi-experimental study investigating the affordances of a hybrid block/text programming environment relative to comparable block-based and textual versions in an introductory high school computer science class. The analysis reveals the hybrid environment demonstrates characteristics of both ancestors while outperforming the block-based and text-based versions in certain dimensions. This paper contributes to our understanding of the design of introductory programming environments and the design challenge of creating and evaluating novel representations for learning.

The number of students taking high school computer science classes is growing. Increasingly, these students are learning with graphical, block-based programming environments either in place of or prior to traditional text-based programming languages. Despite their growing use in formal settings, relatively little empirical work has been done to understand the impacts of using block-based programming environments in high school classrooms. In this article, we present the results of a 5-week, quasi-experimental study comparing isomorphic block-based and text-based programming environments in an introductory high school programming class. The findings from this study show students in both conditions improved their scores between pre- and postassessments; however, students in the blocks condition showed greater learning gains and a higher level of interest in future computing courses. Students in the text condition viewed their programming experience as more similar to what professional programmers do and as more effective at improving their programming ability. No difference was found between students in the two conditions with respect to confidence or enjoyment. The implications of these findings with respect to pedagogy and design are discussed, along with directions for future work.

This paper presents Frog Pond, an interactive code-first learning environment about biological evolution. We deployed Frog Pond as part of a six-day curricular unit on natural selection implemented in six 7th grade science classes. Here we describe a case study of two students, Charlie and Aaron who participated in the unit. Comparing pre- and post- interviews in which they were asked to design a program for a hypothetical simulation of evolution, we found that both students shifted from an event-based programming approach to a rule-based approach. Both students also drew upon their experience with Frog Pond to explain an evolutionary phenomenon. However, the level of sophistication of the two students' explanations varied along with the aspects of Frog Pond they drew upon. These findings have implications for design improvement to better support students' understanding of evolution.

We present an interface for programming relationships between two or more NetLogo [18] models running concurrently. The interface is designed specifically to help high school aged novices explore and define computational relationships between agent-based models, and to investigate how prompting learners to reason about the relationships between complex systems may change how they reason about the systems individually.

Blocks-based programming tools are becoming increasingly common in high-school introductory computer science classes. Such contexts are quite different than the younger audience and informal settings where these tools are more often used. This paper reports findings from a study looking at how high school students view blocks-based programming tools, what they identify as contributing to the perceived ease-of-use of such tools, and what they see as the most salient differences between blocks-based and text-based programming. Students report that numerous factors contribute to making blocks-based programming easy, including the natural language description of blocks, the drag-and-drop composition interaction, and the ease of browsing the language. Students also identify drawbacks to blocks-based programming compared to the conventional text-based approach, including a perceived lack of authenticity and being less powerful. These findings, along with the identified differences between blocks-based and text-based programming, contribute to our understanding of the suitability of using such tools in formal high school settings and can be used to inform the design of new, and revision of existing, introductory programming tools.

Blocks-based programming environments are becoming increasingly common in introductory programming courses, but to date, little comparative work has been done to understand if and how this approach affects students' emerging understanding of fundamental programming concepts. In an effort to understand how tools like Scratch and Blockly differ from more conventional text-based introductory programming languages with respect to conceptual understanding, we developed a set of "commutative" assessments. Each multiple-choice question on the assessment includes a short program that can be displayed in either a blocks- based or text-based form. The set of potential answers for each question includes the correct answer along with choices informed by prior research on novice programming misconceptions. In this paper we introduce the Commutative Assessment, discuss the theoretical and practical motivations for the assessment, and present findings from a study that used the assessment. The study had 90 high school students take the assessment at three points over the course of the first ten weeks of an introduction to programming course, alternating the modality (blocks vs. text) for each question over the course of the three administrations of the assessment. Our analysis reveals differences on performance between blocks-based and text-based questions as well as differences in the frequency of misconceptions based on the modality. Future work, potential implications, and limitations of these findings are also discussed.

This poster presents an environment and set of pedagogical strategies designed to explore how best to use blocks-based programming tools to prepare learners for future, text-based programming languages. Starting with the snap! programming environment, we added the capability to view the JavaScript equivalent of any blocks-based script authored inside the environment. Additionally, when students define behaviors for new blocks, they do so in JavaScript. This makes it possible to compose blocks-based scripts alongside text-based JavaScript programs and have the two run side-by-side. This environment was used during the first 5-weeks of an introductory programming class at the high school level as part of a quasi-experimental study investigating the relationship between programming modality and emerging student understanding. Teachers of the course used the blocks/text hybrid features in various ways to support learners developing an understanding of programming concepts and laying the foundation for future text-based instruction. These strategies included having students compose programs with graphical blocks then view the equivalent JavaScript, prompting class discussion on similarities and differences between the two modalities; having students write pseudocode for their blocks-based programs before comparing the pseudocode to the JavaScript; and finally, having students implement their algorithm directly in JavaScript, using blocks as a resource to reference proper syntax.

Understanding processes of evolution and natural selection is both important and challenging for learners. We describe a "codefirst" learning environment called Frog Pond designed to introduce natural selection to elementary and middle school aged learners. Learners use NetTango, a blocksbased programming interface to NetLogo, to control frogs inhabiting a lily pond. Simple programs result in changes to the frog population over successive generations. Our approach foregrounds computational thinking as a bridge to understanding evolution as an emergent phenomenon.

In this paper we present a novel approach to the design of game-based learning environments in which the content to be taught is embodied by the opponents the learners compete against as they play. By providing the player with the resources to make sense of the concepts exemplified by their opponents, as well as the tools needed to incorporate the concepts into their own gameplay strategy, players are challenged to learn from their opponents in order to advance in the game. This paper introduces RoboBuilder, a blocks-based, program-to-play game that uses this design strategy to introduce programming novices to core computer science concepts. Along with more fully developing this design principle, we provide evidence from a preliminary study conducted with RoboBuilder of players learning from their opponents to create winning strategies that use the concepts designed into the opponents they are facing.

RoboBuilder is a blocks-based, program-to-play game designed to introduce students to core aspects of computational thinking in a fun and engaging environment. The game employs a constructionist design to challenge players to invent and implement strategies to control an on-screen robot using a specially designed visual programming language. During the game, players' robots compete against a series of progressively more challenging opponents in one-on-one battles. Through playing the game, players construct working programs, providing learners with the experience of reifying their own ideas using a computational medium, a practice central to our notion of computational thinking. This poster presents the design rationale for RoboBuilder and discusses key aspects of the game that contribute to giving learners a positive, hands-on introduction to core computational thinking skills including computationally expressing ideas, algorithmic thinking, and debugging.

In this paper we present a case study of children's collaborative behavior around a multi-touch tabletop interface. The study includes data from four sessions with four children over a period of three weeks. The children in our study exhibited a diverse set of collaborative behaviors including territorial control of screen real estate, conflict over interface elements, and turn taking behavior, all of which seemed related to specific aspects of the interface design. Most notably, we observed conflict relating to a graphical toolbar that the children could drag around the screen. After observing this conflict, we redesigned the interface so that children were forced to use a tangible object (a wooden block) to make the toolbar appear on the screen. This tangible object seemed to help the children resolve their conflict and to promote spontaneous turn taking behavior. This paper is an effort to understand why the graphical toolbar alone seemed to spur conflict and why the introduction of a tangible object seemed to help children resolve the conflict on their own

One method of viral marketing involves seeding certain consumers within a population to encourage faster adoption of the product throughout the entire population. However, determining how many and which consumers within a particular social network should be seeded to maximize adoption is challenging. We define a strategy space for consumer seeding by weighting a combination of network characteristics such as average path length, clustering coefficient, and degree. We measure strategy effectiveness by simulating adoption on a Bass-like agent-based model, with five different social network structures: four classic theoretical models (random, lattice, small-world, and preferential attachment) and one empirical (extracted from Twitter friendship data). To discover good seeding strategies, we have developed a new tool, called BehaviorSearch, which uses genetic algorithms to search through the parameter-space of agent-based models. This evolutionary search also provides insight into the interaction between strategies and network structure. Our results show that one simple strategy (ranking by node degree) is near-optimal for the four theoretical networks, but that a more nuanced strategy performs significantly better on the empirical Twitter-based network. We also find a correlation between the optimal seeding budget for a network, and the inequality of the degree distribution.

We propose a new class of crossover operators for genetic algorithms (CrossNet) which use a network-based (or graph-based) chromosomal representation. We designed CrossNet with the intent of providing a framework for creating crossover operators that take advantage of domain-specific knowledge for solving problems. Specifically, GA users supply a network which defines the epistatic relationships between genes in the genotype. CrossNet-based crossover uses this information with the goal of improving linkage. We performed two experiments that compared CrossNet-based crossover with one-point and uniform crossover. The first experiment involved the density classification problem for cellular automata (CA), and the second experiment involved fitting two randomly generated hyperplane-defined functions (hdf's). Both of these exploratory experiments support the hypothesis that CrossNet-based crossover can be useful, although performance improvements were modest. We discuss the results and remain hopeful about the successful application of CrossNet to other domains. We conjecture that future work with the CrossNet framework will provide a useful new perspective for investigating linkage and chromosomal representations.

Van Valen first discovered the "Red Queen Effect" (RQE), where two species can dramatically co-evolve their phenotypes over time but never gain a relative advantage [2]. In the ideal version of the RQE, regardless of the actual values that the species evolve to obtain, they have not moved in relation to each other. Though previous models of the RQE exist, we developed an agent-based model (ABM) which has a base ontology more similar to real world coevolutionary systems than equation-based models (EBMs). For instance, this model contains spatial information and an individuallevel reproduction mechanism. Yet this model recreates traditional EBM results. For instance Dieckmann et al show that there are three possible outcomes of competitive coevolution: predator dominance, prey dominance and evolutionary cycling (RQE) [1]. By reconceptualizing this EBM using an ABM, we make it easier for students and researchers to understand, manipulate, and modify this model [4]. The model is written in the NetLogo agent-based modeling environment [3]. The model is initialized with 150 predator and 150 prey agents. Predator agents have a resistance level r and prey agents have a poison p. The agents are initially randomly distributed on a toroidal real-valued 35 by 35 grid. The initial resistance and poison values for the predators and prey are drawn from normal distributions with means µr and µp and a standard deviation of 1. During a model timestep, each agent moves one unit at a random heading. If at the end of its move a predator is within 1 unit of a prey, then it will challenge the prey. The predator will compare its resistance value to the prey's poison value and which ever agent has the larger value will win the challenge and the other agent will be killed. At the end of an agent's turn if the total number of agents is less than the maximum carrying capacity the agent will reproduce with a 50% probability. The new agent's initial poison / resistance will be drawn from a normal distribution with the parent's poison / resistance as the mean value and a standard deviation of 1.

A considerable body of research has demonstrated the benefits of modeling a variety of scientific phenomena via the multi-agent approach and its vast potential in education. However, most multi-agent simulations are purely virtual, with no connection to the physical world. To address this issue, we designed a platform to enable students to seamlessly link computer models and sensors, in real time. The platform is designed for learners to validate, refine, and debug their computer models using real-world data, in a Constructionist fashion. Also, the platform broadens the types of possible scientific explorations and makes available a low-cost version of apparatus only available in advanced research laboratories. We will focus on technical and educational aspects of this development, describing preliminary pilot studies in which we analyze projects built by students.

Multi-agent modeling has been successfully used in several scientific fields, oftentimes transforming scientists' practice and mindsets. Educational researchers have also realized the potential of this modeling approach for learning. Studies have suggested that students are able to understand concepts above their expected grade level after interacting with curricula developed using multi-agent simulation. However, most multi-agent models are exclusively 'on-screen', without connection to the physical world. Real-time model validation and real-world sensing are very challenging to accomplish with extant modeling platforms. As an attempt to address this issue, we designed a technological platform to enable students to seamlessly connect multi-agent models and electronic sensors, in real time. The platform is designed for learners to validate, refine, and debug their computer models using real-world data. This paper focuses on the technical and pedagogical aspects of this project, describing pilot studies which suggest a real-to-virtual reciprocity that catalyzes further inquiry toward deeper understanding