Exploring Practices to Support Commenting and Debugging in Early-Years Tangible Programming

• In Event: Power in the Plurality of Pedagogies and Partnerships in "Computer Science for All"

Abstract

OBJECTIVE
One important programming skill and greatest struggle beginning programmers have is debugging their code (e.g., Perkins & Martin, 1986; Wing, 2006). Commenting on (explaining) code can help students interpret code; however, beginning programmers often fail to properly comment on their code (Mason, 2003) or may introduce comments that do not match the code, leading to additional bugs (Tan, Yuan, Krishna, & Zhou, 2007). Previous studies focus on the strategies and comments that debuggers successfully use (McCauley et al., 2008), but little research pinpoints the strategies of unsuccessful debuggers. Therefore, understanding beginning debuggers can help educators design activities to promote debugging practices. In this project, we explored whether free-play (play) versus goal-directed play (goal) helps students, females in particular, to more effectively comment on and debug their code after engaging in a tangible programming game (OSMO’s Coding Awbie).

THEORETICAL FRAMEWORK
Ginsburg (2006) asserted that “play can produce learning” (p. 145). For example, young children playing with tangible coding blocks improved their sequencing ability (Sullivan, Bers, & Mihm, 2017). When students identify bugs in their programs, they begin to modify their conceptions of programming (e.g., Vosniadou, Vamvakoussi, & Skopeliti, 2008), and explaining bugs could help students develop larger categories of bugs (Griffin, 2016). Therefore, students’ commenting and debugging practices should reflect their new conceptions of programming.

METHODS AND DATA
Overall, we recruited 40 males and 29 females (32 first- and 37 third graders). Within each grade, we randomly assigned pairs to a play-first or goal-first group. We included all-female, all-male, and mixed gender pairs. After the midtest, students switched groups. See Table 4 for details.

We compared students’ performance gains on the commenting and debugging items across tests (see Figure 1). We coded their debugging responses to determine their strategies and practices.

RESULTS
Overall, females had greater gains when they played, and females who started out playing for the first three sessions made the greatest progress in terms of the commenting and debugging items (see Table 5).

From pretest to posttest, 24% more females in the play-first group correctly debugged the programs without having to rewrite the program, compared to 6% in the female goal-first group; 10% in the male, play-first group; and 19% in the male, goal-first group (see Table 6).

Based on our debugging codes, students had difficulty with the debugging tasks because they did not know the meaning of coding pieces (e.g., thought loop meant going around in a circle and jump 1 meant move one space instead of jumping over a space) or due to counting errors. Students sometimes double-counted a square or continued counting when having the rabbit change directions.

SIGNIFICANCE
The females in the goal-first group may have experienced stereotype threat when forced to confront how well they knew the code they were using during the sessions; their performance on debugging decreased or was low compared to when they freely played. Even at an early age, we need to make an effort to diffuse stereotype threat and give students time to play.

Authors

• Laura Bofferding, Purdue University

• Sezai Kocabas, Purdue University

• Mahtob Aqazade, Rice University

• Lizhen Chen, Ohio University

• Ana Maria Haiduc, Purdue University West Lafayette