Data visualizations are increasingly seen as socially constructed, with several recent studies positing that perceptions and interpretations of visualization artifacts are shaped through complex sets of interactions between members of a community. However, most of these works have focused on audiences and researchers, and little is known about if and how practitioners account for the socially constructed framing of data visualization. In this paper, we study and analyze how visualization practitioners understand the influence of their beliefs, values, and biases in their design processes and the challenges they experience. In 17 semi-structured interviews with designers working with race and gender demographic data, we find that a complex mix of factors interact to inform how practitioners approach their design process, including their personal experiences, values, and their understandings of power, neutrality, and politics. Based on our findings, we suggest a series of implications for research and practice in this space.

In this paper, we seek to understand how grassroots activists, operating within the hegemony of data-centrism, are often disempowered by data even as they appropriate it towards their own ends. We posit that the shift towards data-driven governance and organizing, by elevating a particular epistemology, can pave over other ways of knowing that are central to social movement practices. Building on Muravyov's [102] concept of ''epistemological ambiguity,'' we demonstrate how data-focused activism requires complex navigations between data-based epistemologies and the heterogeneous, experiential, and relational epistemologies that characterize social movements. Through three case studies (two drawn from existing literature and the third being an original analysis), we provide an analytical model of how generative epistemological refusals can support more value-aligned navigations of epistemological ambiguity that resist data-centrism. Finally, we suggest how these findings can inform pedagogy, research, and technology design to support communities navigating datafied political arenas.

AI technologies, despite having well-documented biases and shortcomings, are becoming increasingly pervasive across various aspects of society. AI biases often reflect and interact with broader societal biases, underscoring the need to support children in understanding these biases so that they can identify when they (or others) are being discriminated against by an AI-based system. To explore this learning through a new methodology, we built an interactive system called CLIP4KIDS. We conducted four classroom sessions with 28 fifth graders in the United States and examined our data using qualitative thematic analysis. Students frequently described AI biases in terms of “assumptions” and “stereotypes” and drew connections between historical injustices and present biases in AI models. This work contributes a novel tool for learning about AI biases, an empirical account of children’s experiences, and a theoretical analysis incorporating Vossoughi and Gutiérrez’s framework of critical pedagogy and sociocultural theory.

Designing for values has been a focus of human-computer interaction research, but what happens when value-laden design artifacts are put into practice? Do they exercise their commitment to stated design values? We present four case studies that suggest a gap between the values that technologies set out to support and their performance toward supporting these values in practice. By critically analyzing these case studies, we theorize the phenomenon of value dilution—technical artifacts moving away from values they committed to embody. We hypothesize two significant methodological gaps contributing to value dilution—the static framing of stakeholders and a lack of engagement with politics of values. We argue that addressing value dilution needs to be a long-term and ongoing task in the design and use of technology as values in design are not only embodied, but also they are dynamic, subject to change in how they are enacted.

In this paper, we present a new visual block-based programming system designed for children to process, analyze, and visualize data. We introduce the system and describe how it was used during a series of 7 workshops with 27 children. During the workshops, children played the role of investigators and followed a storyline as part of the system to conduct data analyses to help the story’s protagonist locate a missing family member. We present our findings as a framework of computational data literacy that builds on the dimensions of Computational Thinking proposed by Brennan and Resnick [8], with a focus on aspects that are specific to using programming for data processing, analysis, and visualization. We conclude with a series of recommendations for future designers of systems to support the development of computational data literacy.

Increased technological capacity to collect and use data has created both new possibilities for benefiting individuals and societies, and critical questions of what is acceptable and just [31]. Because early definitions of data literacy have often excluded aspects of power, equity, empowerment, and emancipation, children’s learning experiences have focused more on the potential benefits compared to the critical questions. In this review article, we examine the importance of teaching critical data literacy to children as a key aspect of developing fluency with data. Using constructionist principles [67] as a guiding framework, we synthesize 48 educational research and design approaches that engage youth with data projects. We describe how these projects provide students with information about data’s origins and perspectives, and assist them in identifying, analyzing, and presenting data. Finally, we provide design implications and concrete examples on how constructionist approaches can be utilized for teaching critical data literacy.

Background and context. Concept inventories (CIs) are a widely used tool in STEM education that can help instructors identify specific misconceptions students hold about key concepts. Over the past several years, much research has been published contributing to CIs in computer science education.

Through a mixed-method analysis of data from Scratch, we examine how novices learn to program with simple data structures by using community-produced learning resources. First, we present a qualitative study that describes how community-produced learning resources create archetypes that shape exploration and may disadvantage some with less common interests. In a second quantitative study, we find broad support for this dynamic in several hypothesis tests. Our findings identify a social feedback loop that we argue could limit sources of inspiration, pose barriers to broadening participation, and confine learners’ understanding of general concepts. We conclude by suggesting several approaches that may mitigate these dynamics.

Why do some peer production projects do a better job at engaging potential contributors than others? We address this question by comparing three Indian language Wikipedias, namely, Malayalam, Marathi, and Kannada. We found that although the three projects share goals, technological infrastructure, and a similar set of challenges, Malayalam Wikipedia’s community engages language speakers in contributing at a much higher rate than the others. Drawing from a grounded theory analysis of interviews with 18 community participants from the three projects, we found that experience with participatory governance and free/open-source software in the Malayalam community supported high engagement of contributors. Counterintuitively, we found that financial resources intended to increase participation in the Marathi and Kannada communities hindered the growth of these communities. Our findings underscore the importance of social and cultural context in the trajectories of peer production communities.

Wikipedia edit-a-thon events provide a targeted approach toward incorporating new knowledge into the online encyclopedia while also offering pathways toward new editor participation. Through the analysis of interviews with 13 edit-a-thon facilitators, however, we find motivations for running edit-a-thons extend far beyond adding content and editors. In this paper, we uncover how a range of personal and institutional values inspire these event facilitators toward fulfilling broader goals including fostering information literacy and establishing community relationships outside of Wikipedia. Along with reporting motivations, values, and goals, we also describe strategies facilitators adopt in their practice. Next, we discuss challenges faced by facilitators as they organize edit-a-thons. We situate our findings within two complementary theoretical lenses-information ecologies and public pedagogy to guide our interpretation. Finally, we suggest new ways in which edit-a-thons, as well as similar peer production events and communities, can be understood, studied, and evaluated.

A core aim for designing constructionist learning systems and toolkits is enabling "wide walls"-a metaphor used to describe supporting a diverse range of creative outcomes. Ensuring that a broad design space is afforded to learners by a toolkit is a common approach to achieving wide walls. We use econometric methods to provide an empirical test of the wide walls theory through a natural experiment in the Scratch online community. We estimate the causal effect of a policy change that gave a large number of Scratch users access to a more powerful version of Scratch data structures, effectively widening the walls for learners. We show that access to and use of these more powerful new data structures caused learners to use data structures more frequently. Our findings provide support for the theory that wide walls can increase engagement and learning.

Although informal online learning communities are made possible by users' decisions to share their creations, participation by females and other marginalized groups remains stubbornly low in technical communities. Using descriptive statistics and a unique dataset of shared and unshared projects from over 1.1 million users of Scratch-a collaborative programming community for young people-we show that while girls share less initially, this trend flips among experienced users. Using Bayesian regression analyses, we show that this relationship can largely be attributed to differences in the way boys and girls participate. We also find that while prior positive feedback is correlated with increased sharing among inexperienced users, this effect also reverses with experience or with the addition of controls. Our findings provide a description of the dynamics behind online learners' decisions to share, open new research questions, and point to several lessons for system designers.

In this paper, we present Scratch Community Blocks, a new system that enables children to programmatically access, analyze, and visualize data about their participation in Scratch, an online community for learning computer programming. At its core, our approach involves a shift in who analyzes data: from adult data scientists to young learners themselves. We first introduce the goals and design of the system and then demonstrate it by describing example projects that illustrate its functionality. Next, we show through a series of case studies how the system engages children in not only representing data and answering questions with data but also in self-reflection about their own learning and participation.

As contemporary youth learn, play, and socialize online, their activities are often being recorded and analyzed. What should young people know about these data collection and analysis efforts? Although critiques of these new forms of data collection and analysis have grown increasingly loud, the voices of users, and particularly youth, have largely been absent. This paper explores the critical perspectives of youth who are programming with public data about their own learning and social interaction in the Scratch online community. Using a bottom-up approach based on ethnographic observation of discussions among these young users, we identify a series of themes in how these youth critique, question, and debate the implications of data analytics. We connect these themes-framed in terms of critical data literacies-to expert critiques and discuss the implications of these findings for education and design.

Education research suggests that learning in one's local language can have a positive impact on learning outcomes. We offer a quantitative test of the association between local language use and the rate at which youth learn to program. Using longitudinal data drawn from five countries and over 15,000 users of Scratch, a large informal learning community, we find that novice users who code with their programming language keywords and environment localized into their home countries' primary language demonstrate new programming concepts at a faster rate than users from the same countries whose interface is in English. We conclude with a discussion of the implications of our findings for designers of online learning systems.

Quantitative studies of learning using block-based programming languages in informal environments have relied on identifying the presence or absence of individual visual blocks in learners' projects. Many important programming concepts (e.g., initializing a variable) involve the combination of several blocks. In this poster, we present a technique that uses a statistical method from epidemiology called "survival analysis" to model the rate at which programmers begin to use new code patterns. By analyzing data drawn from the trajectories of over 90,000 users from the Scratch online community, we demonstrate the potential of our approach. In particular, we model when users are at higher and lower levels of "risk" of demonstrating two particular code patterns -- variable initialization and counting collisions. We show that learning of these patterns is associated with behaviors like viewing the source code of other projects, remixing, and commenting. We explain how our method can be extended to help understand predictors of skill acquisition in informal environments more generally and how it can inform the design of more effective learning support structures.

This paper contributes to a growing body of work that attempts to measure informal learning online by revisiting two of the most surprising findings from a 2012 study on skill progression in Scratch by Scaffidi and Chambers: users tend to share decreasingly code-heavy projects over time; and users' projects trend toward using a less diverse range of code concepts. We revisit Scaffidi and Chambers's work in three ways: with a replication of their study using the full population of projects from which they sampled, a simulation study that replicates both their analytic and sampling methodology, and an alternative analysis that addresses several important threats. Our results suggest that the population estimates are opposite in sign to those presented in the original work.

Theorists and advocates of “remixing” have suggested that appropriation can act as a pathway for learning. We test this theory quantitatively using data from more than 2.4 million multimedia programming projects shared by more than 1 million users in the Scratch online community. First, we show that users who remix more often have larger repertoires of programming commands even after controlling for the numbers of projects and amount of code shared. Second, we show that exposure to computational thinking concepts through remixing is associated with increased likelihood of using those concepts. Our results support theories that young people learn through remixing, and have important implications for designers of social computing systems.

We present the Scratch extension system, a toolkit that enables anyone to extend the vocabulary of the visual Scratch programming language through custom programming blocks written in JavaScript. The extension system is designed to (i) enable innovating on the Scratch programming language itself, in addition to innovating with it through projects, and (ii) enable the creation of new interest-driven pathways into Scratch programming. In this paper, we describe some of the prior work done in this space, our design and implementation, open questions and challenges, and some preliminary outcomes.

Being able to store and access data online enables a wide range of creative possibilities, starting from surveys to collaborative art, from global high-score-lists for games to real-time chat-rooms. While end-user tools in these categories are increasingly becoming available to children, what is still missing is the opportunity for children to program and create such systems. Causes behind this lack of opportunity include, among other things, high barriers to entry due to complex client-server technologies, as well as hard to understand topics such as access control, etc. This paper presents Cloud data-structures -- a feature in the online visual language Scratch 2.0 that enables children to programmatically store and retrieve data online. While standard data-structures are stored in memory, for Cloud variants, all operations (and data) on the data-structure are additionally sent to remote servers over the Internet. This has two consequences for a given Scratch 2.0 project: (1) Cloud data-structures are persistent across multiple execution instances, and (2) they are shared between simultaneous instances. This paper describes the motivations behind, and the design of Cloud data-structures, along with case studies describing projects created by children with this system, with a focus on the learning outcomes.