Computing integrated activities scored for programming concepts

Educators across disciplines are implementing lessons and activities that integrate computing concepts into their curriculum to broaden participation in computing. Out of myriad important introductory computing skills, it is unknown which—and to what extent—these concepts are included in these integrated experiences, especially when compared to concepts commonly taught in introductory computer science courses. Thus, it is unclear how integrated computing activities serve the goal of broadening participation in computing. To address this deficit, we compiled a database of 81 integrated computing activities, constructed a framework of fundamental programming concepts, and scored each activity in the database for the presence of each concept. The dataset also includes different activity features, including discipline, programming language, student age, and duration of activity. 

Methods
Selection Criteria: Features and Limitations
Non-CS Disciplinary Learning Objectives
The first selection criterion for activities to include in the analysis was the inclusion of learning objectives in a discipline other than computing. No restrictions were placed on which other disciplines qualified, and we found activities from language arts, math, science, art, music, foreign language, history, social studies, and even spatial skill development for young children. One indirect benefit of requiring non-computing disciplinary learning objectives was that many included activities have substantive lesson plans. These lesson plans make the activities more accessible to teachers by including TPACK-related information, such as disciplinary learning objectives for the activity. As a result, the authors recognize the limitations of requiring non-computing learning objectives but also that it provides a level of authenticity and accessibility for the included activities.

One of the major sources of computing integration activities affected by this requirement was the ScratchEd website. Scratch is a popular language for computing integration activities, aided by an extensive repository of student- and teacher-created projects that users are encouraged to remix into their own projects. The thousands of programs in this repository are of widely varying complexity and quality, and most of them are listed with a topic but without explicit learning objectives. To draw from this wealth of activities without comprehensively including projects, we identified lists of vetted computing integrated activities using Scratch to include in the analysis. These lists were "Integrated Scratch Programming in the Curriculum," "Scratch Projects Across the Curriculum," "From Music to Math: Scratch Across Every Subject," and "Scratch Cross-Curricular Integration Guide." Similarly, resources related to the Snap! language had plentiful examples of projects across disciplines with limited explicit non-CS disciplinary learning objectives.

Block-Based Programming Languages
Because computing integration activities are becoming popular, an initial search revealed too many activities to score in one analysis. To narrow the scope of the analysis, the next selection criterion was that the activity had to use a block-based programming language. This criterion has benefits and limitations. One of the main benefits for the goal of the current analysis was that block-based activities include a range of concepts, regardless of their syntactic or semantic difficulty (Grover & Basu, 2017; Papadakis et al., 2014). This benefit means that concepts that best serve the activity can be included for learners with little to no programming experience (Weintrop & Wilensky, 2018). The associated limitation, however, was that concepts are also restricted by the blocks that are built into the language. Most popular languages use a low-floor, high-ceiling design that includes blocks for all concepts that would be taught in an introductory programming course, though (Grover, 2021; Weintrop & Wilensky, 2015). Another limitation was that prominent, text-based integration activities, such as Bootstrap’s curricula in Algebra and Physics, are excluded.

This selection criterion also notably excluded commonly used science simulation platforms, like NetLogo and PhET. These platforms include a large range of simulations for scientific phenomena and other models beyond science. While the simulations allow users to easily access the source code, the primary interface does not include the program used to create the simulation. In addition, the source code, except for some adapted NetLogo simulations, is text-based. Though the programs are heavily commented to make them understandable, they do not meet the inclusion criteria for the current dataset. More programming-centric and block-based options for scientific simulations, like StarLogo Nova, were included.

Access
Accessibility of the activities was the final criterion for inclusion. Following the accessibility criteria used by Lin and Weintrop (2021), we included activities only if they could be found online, were free of cost, did not require a physical device like robotics toolkits, and were updated recently enough that they ran on current versions of languages and operating systems. The requirement to be found online is not expected to substantially narrow the analysis because Lin and Weintrop found that 90% of block-based programming languages ran in a web browser. Exclusion for use of physical devices is a corollary to the requirement to be free of cost. We felt that these criteria would result in a dataset that had the broadest and most equitable applications because many public schools in low-income areas in the US cannot afford physical computing or robotics kits.

Search Criteria
Users need to recognize that the current dataset was based on a review of computing integration activities but not a systematic review. Unlike systematic literature reviews of scholarly work on a given keyword or topic area, there were no databases of indexed computing integration activities that span our inclusion criteria. Some repositories for certain languages exist, such as ScratchEd’s repository of Scratch projects and the Exploring Computational Thinking repository of Pencil Code and Python activities. However, computing integration activities are not published through a central organization, so they can be difficult to find.

In lieu of a systematic review, we attempted to build a database that represented activities from a variety of disciplines, student ages, designers, and languages. To create this database, we included any activities that we were already aware of, such as Action Fractions, links from lists of computing integration activities, such as "Scratch Projects Across the Curriculum," links from CSforAll’s curriculum directory, and a general Google search for "‘integrated computing’ activities" and "‘computational thinking’ + programming" or "‘computational thinking’ + coding." We examined the first 100 returns for these searches. However, many of the activities found through Google search were excluded based on our criteria, primarily for not including non-CS learning objectives.

We included activities as whole units, whether they were single-class lessons or extended curricular units that included multiple lessons, like Coding as Another Language. Treating individual lessons from curricular units as individual activities would have created an over-representation of extended units (e.g., 72 lessons for the Kindergarten, 1st, and 2nd grade curricular units from Coding as Another Language instead of 3 activities). Our database included 81 activities from the following sources:

• CANON Lab

• Code.org’s CS Connections

• Code.org’s Hour of Code

• Coding as Another Language curriculum

• CS+ units from University of California San Diego

• CSforALL’s Curriculum Repository (including 144 curricular units at the time of searching)

• CT4Edu

• Everyday Computing

• Exploring Computational Thinking

• Google search

• Google’s CS First

• Integrated computing activities from Georgia State University

• Project GUTS

• ScratchEd

• The Tech Interactive

• TVO Learn

• UCL Scratch Maths

We analyzed the distribution of these activities’ characteristics based on primary discipline, student age, programming language, and minimum time to complete. Based on discipline, we recognized that we had only two from history or social studies and searched for additional activities. While we found many projects on ScratchEd’s website, they did not meet the selection criteria. Required courses, including Language Arts, Math, and Science had a sufficient number of activities, matching their representation in the school day. We also had a wide range of activities based on student age and minimum time to complete, so we did not search for any additional activities based on these characteristics.

To explore the representation in our database based on programming languages, we used the categories identified by Lin and Weintrop (2021) to ensure coverage of different types of block-based languages. The database has activities from Pencil Code (i.e., block-based implementation of a text-based language), Scratch (i.e., multimedia focused on animations and storytelling), AppLab (i.e., mobile app development), StarLogo Nova (i.e., simulations), and ScratchJr (i.e., pre-reading language). We decided against requiring languages from Lin and Weintrop’s other categories for data science, physical computing, and task-specific languages because they did not match our inclusion criteria. We explored other common languages to include, like Alice, Snap!, and App Inventor, but we did not find activities that matched our criteria.