This week was focused on getting introduced to the team and the expectations of the project. I was also given information regarding the process of building an autograder. After I got settled in with the online research course, I moved towards understanding the autograders that were previously built. I read through the documentation that was provided by my mentor and learned the testing framework, pytest, that we would be working with. Soon, I set up my Coursera environment as I would coding my first autograder here with my two teammates. This autograder was focused on using user input with the random library to produce an output string. The first step was to create sample student solutions where some were correct, while most were incorrect. Throughout the week, I created test cases for each student solution to evaluate how the autograder would perform. While, I was working on this autograder, I was reading through previous papers on autograders and open-ended assignments to get an idea of what has attempted. Also, the online research course introduced me to how to properly read a research paper and how to work well in a team.

I have read two papers this week:

• “Attitudes Surrounding an Imperfect AI Autograder” by Silas Hsu, Tiffany Wenting Li, Zhilin Zhang, Max Fowler, Craig Zilles, Karrie Karahalios

This paper focused on analyzing how students percieved incomplete autograders in their classes. They used surveys and interviews to gather their perspectives and folk theories. It was noticed that students viewed the autograders negatively and assumed that that the autograder was trying to just match their code to a solution. This mostly stemmed from their frustration and confusing on how the autograder actually worked. Therefore, the researchers recommended that instructors inform students how the autograder works so students can better cater their code to these autograders.

• “CS for All: Catering to Diversity of Master’s Students through Assignment Choices” by Sohail Alhazmi, Margaret Hamilton, Charles Thevathayan

This paper focused on helping students better learn new CS concepts in an introductory coding course. It was noticed in a class that the assignments did not align with student interests and as a result, many students felt bored and did not learn the concepts. Therefore, the researchers decided to design 5 assignments for students to choose from, each with varying difficultly and topic. Some of these assignments were group based, while some were individual based. It was revealed that giving students a chance to choose their assignment, improved their enjoyment and they understood the concepts better. So, the researchers recommended that implementing open-ended assignments that students can choose from will benefit their learning.