Remixing as a path to Computational Thinking

Abstract

Section titled “Abstract”
  • remix as a tool for learning
  • tested via multimedia Scratch projects
  • exposure to wider range of computational thinking concepts increases likelihood of using these concepts
  • shows that young people learn through remix

Introduction

Section titled “Introduction”
  • online community for Scratch projects that allows unrestricted reuse and modification
  • licensing can be explained in a child-friendly way
  • learning by downloading, studying and modifying others’ programs
  • ‘appropriation or remixing’
  • described as an important literacy or skill for young people by Bruckman [7], Jenkins [20], Ito [18], Lessig [25], Manovich [28], and others.
  • a strong current in constructionist approaches to learning [34]
  • put computational thinking in a social context
  • this paper is an empirical test of these theories

Background

Section titled “Background”

  • ‘reworking and combination of existing creative artifacts’

  • a culturally-significant shift in creative collaboration brought on by the Internet

  • remixing “a built-in feature of the digital networked media universe” (Manovich)

  • “example of the power and potential of a new mode of social production ” (Benkler)

  • normative position that participation in cultural production is prima facie socially beneficial (Lessig)

  • a low-cost and accessible form of cultural and social participation

  • path toward high quality information goods through the mass aggregation of many small contributions

  • similar, in both process and effect, to Wikipedia or free/open source software

  • remixing leads to learning as a form of legitimate peripheral participation (Jenkins)

  • detractors say remixes are largely derivative, uninteresting, and of poor quality (is this a problem if learning takes place? and ors this detract from the cases that are more interesting or higher quality? And would this be the case with more cultural normalisation of remix?)

  • other studies on remixing look at what is remixed, by who, on authorship and credit, on remix quality and on copyright law

  • Constructionist theories hold that we learn best by designing and constructing “public entities,” and that this learning is even more meaningful and effective in a social context

  • research on learning in Scratch has focused on social interactions and relationships within the community as well as on appropriation

  • Wing defines computational thinking as “the thought processes involved in formulating problems and their solutions so that the solutions are represented in a form that can be effectively carried out by an information-processing agent”

  • building blocks like sequences, loops, conditionals as well as higher methodology like debugging

  • previous quantitative studies of learning in informal environments by looking at user “trajectories” (Yang)

  • learning as growth in the ‘cumulative repertoire of programming tokens’ like vocabularies

  • measuring breadth of types of concepts applied

  • using new concepts for the first time in new projects if they’ve remixed projects with them

  • Scratch online community stores extensive meta-data including whether projects are created through remixing

  • algorithmic analyses of each project

  • measuring multiple methods of parallelism, event handling etc

  • using a continuous time-survival approach using the Cox proportional hazards model

  • A limitation of our hazard models is that a Scratch user may copy a stack of blocks that includes a block we are looking for without understanding the underlying concept behind the block. In robustness checks, we attempted to address this by using an alternative threshold for demonstration of a CT concept in each of our hazard models that ignores any demonstration of a CT concept if it occurs in a script that is identical to one used in a previous remix. […] In all six cases, the results are still positive, statistically significant, and only moderately reduce the parameter estimates.

  • also controlled for initial block repertoire

  • the design of Scratch may make the CT concept of parallelism emerge unintentionally - scratch projects often have multiple sprites - discarded parallelism between sprites

  • questions of generalizability are both important and difficult to answer

  • remixing may be more effective at promoting engagement with some concepts, like loops, than others, like operators and data