Introduction

The COVID-19 pandemic has thrust online learning environments into the global spotlight. The sudden shift to emergency remote teaching \citep{Hodgesdifferenceemergencyremote} has forced educators and students to adjust their educational routines using digital platforms and even pedagogical approaches with which they were previously unfamiliar \citep{Quintana_2020}. Schools and universities have confronted this new reality, while lifelong learners have sought new online options for advancing their education.
One of the outcomes of these events has been an unprecedented rise in enrollments in massive open online courses (MOOCs), leading to a series of publications on the subject \citep[see][]{lohr2020}. In this paper, we explore the impacts of the pandemic on one MOOC: Problem Solving Using Computational Thinking. We give special attention to the educational benefits of a pedagogical approach that is rarely used in courses of this scale: project-based learning (PBL). We identify particular benefits of this approach that current events have highlighted, and we offer implications for teaching computational thinking in online settings.

Course Description

Audience and Scope

The Computational Thinking MOOC was developed by an interdisciplinary team from a large public university in the Midwestern United States. The course was designed for a target audience of pre-college learners and early college learners who intend to pursue STEM careers and would thus need to develop fluency in the computational tools used in STEM. It aims to equip students with the modes of thinking needed to set up problems and potential solutions as a foundation for being able to eventually use computational tools and programming to address those problems.
Given this perspective, the course is based on one particular definition of computational thinking (CT). CT has been described and discussed in some form for decades \citep[see][]{Papert_1996}, but it has received increased attention in recent years, with an especially sharp rise in relevant scholarship since 2015 \citep{Hsu_2018}. The definition of CT adopted in this course draws heavily on the work of \citet{Wing_2006}, who presents CT as the practice of conceptualizing problems, complementing and combining mathematical and engineering thinking. \citet{Wing_2006} argues that CT should not become a synonym for computer programming and notes that there are a range of skills necessary for CT, including the ability to define problems, reformulate seemingly intractable problems into solvable ones, use abstraction and decomposition when approaching a complex task, and use massive amounts of data and computation for problem solving.
While the Computational Thinking MOOC similarly emphasizes the problem-solving aspect and many of the skills that Wing highlights, it does so without requiring learners to make use of actual data or computer programming. In the course introduction, CT is defined in the following approachable way:
“Before you can think about programming a computer, you need to work out exactly what it is you want to tell the computer to do. Thinking through problems this way is Computational Thinking. Computational Thinking allows us to take complex problems, understand what the problem is, and develop solutions. We can present these solutions in a way that both computers and people can understand.”
The goals of the MOOC, therefore, focus on helping learners to specifically define and decompose problems through abstraction while teaching them to use insights from similar problems in other domains to guide potential solutions.

Project-Based Pedagogy

This MOOC was designed around a “project-based learning” (PBL) approach, which integrates instructional activities within projects motivated by students’ own interests and contexts \citep{Krajcik_2005}. PBL is popular in K-12 classrooms, but there are questions of its utility in other settings, especially in MOOCs, which typically adopt a didactic, lecture-based format with instructional videos, quizzes, discussions, and graded assignments. However, some early work has found positive attitudes among learners in a project-based MOOC and has emphasized the importance of learner autonomy in these contexts \citep{Barak_2017}
The Computational Thinking MOOC centers much of its pedagogy around a final project in which learners identify a problem to solve computationally, and then use the knowledge and techniques they learned throughout the course to iteratively develop an algorithmic approach towards a solution. They are asked to submit both a graphic organizer that displays the multiple iterations of their work and a diagrammed algorithm of their final solution for peer evaluation. This approach affords learners a level of flexibility rarely found at the scale of MOOCs since they are able to select the topic that will define their final projects.

Using Case Studies to Shape Student Work

Using case studies is a way to ground CT in real-life scenarios. This provides a concrete, actionable foundation for CT that can more effectively lead to learning, retention, and application \citep{Weintrop_2015}. For this purpose, the Computational Thinking course revolves around a series of three case studies.
Most pertinent here is the case on epidemiology, which was incidentally developed before the events surrounding the global pandemic. This case presents a large, complex problem: how do we prepare for the seasonal flu and make sure we are ready for the next pandemic? The expert who presents this case—an associate professor of epidemiology—breaks this down into a more specific sub-problem and then walks learners through an algorithm that centers around four categories of people: vaccinated, susceptible, infected, and recovered. Finally, a series of computational modeling tools specific to epidemiology are offered to learners interested in diving more deeply into this problem.

Enrollment and Engagement Trends

As a reflection of the recent spike in attention that online education has garnered, the Computational Thinking MOOC experienced a sharp increase in enrollment around the time that many countries began to experience the full force of the COVID-19 pandemic. We found that most new learners who reported their educational attainment and employment status hold a bachelor’s degree and are unemployed (see Figure \ref{544146}). As expected, this suggests that the rise in overall enrollments is likely due to a combination of more people working from home, universities shifting to remote teaching, and rising unemployment rates.