"Toucan Has Failed Me": Peer-to-Peer Playful Discourse During Debugging

• In Event: Orchestrating Collaborative Play for Diverse Participation Across Contexts and Disciplines

Abstract

Purpose and Framework
Students often construct causal explanations of failure (e.g., that quiz was too hard). External and stable attributions (e.g., lack of effort) have been tied to low motivation (Graham, 2020). However, in natural conversation, students propose chains of internal and external attributions (Authors, 2017), and in play-based spaces, motivation can derive both from players blaming external avatars (Chase et al., 2009) and from game characters derisively blaming internal, “stable” player traits (Juul, 2009). We explore here how playful discourse during the programming of virtual characters shapes causes of failure and impacts debugging.

Methods and Data Sources
This research takes place in free-of-charge, two-week summer coding workshops (M-F 9am - 4pm) for 5th - 10th grade students (n=120) who attended title 1 schools or demonstrated financial need. Classes were led by undergraduate computer science majors. We focus on PixelBots, in which students wrote JavaScript to control the movement and painting actions of an animal avatar around a two-dimensional grid. Using laptop screen recordings and GoPro recordings, we marked, transcribed, and interpreted moments of debugging using conventions of interaction analysis (e.g., Goodwin, 2013). We revisited a case study of one 7th grade female student over five sessions (6 hours and 40 minutes of coding) and a subset of playful failure explanations from 5 other case studies (3 girls, 2 boys).

Results
Students explained programming failures by characterizing their own actions (“I’m doing something so dumb”) alongside multiple external spaces: the computer, the PixelBot (“Toucan has failed me”), the computer (“The computer’s glitching”), and instructors. Many of these attributions took place in the context of playful dialogue. For example, two seventh grade girls in conversation across two coding sessions spoke in a funny voice (“Oh my gah, what’s wrong?”), read an error message with a robotic voice (“It’s been declared!”), talked sarcastically about incorrect syntax (“Linew? Good name”), sang pop songs about failure (“When you try your best but you don’t succeed”), asked iPhone’s Siri for incongruously unspecified help (“Siri, how do you code?”), talked about line limit task constraints as judgmental (“Who cares? Like it’s judging us.”), pretended to destroy the PixelBot (“Bop it. Pull it. Kill it.”), and trivialized errors (“Silly me,” “Whoopsie daisy”). This playful talk occurred throughout our five case studies and even became a week-long classroom meme with the “Curse of Robert,” a notion that an instructor’s presence around code caused bugs.

Significance
On the basis of these observations, we argue that playfulness is an overlooked yet critical feature of some peer-to-peer discourse around failure. Our interaction analyses show that these playful discourse moves often draw in a peer who provides a debugging resource, create social connections between students seated at the same table, create an epistemic climate that takes the edge off of wrong code (see Chase et al., 2009 and Ryoo, 2013), motivate the application of productive debugging practices, and normalize that learning to communicate with computers is challenging.

Author

• David DeLiema, University of Minnesota - Twin Cities