CS + X

It is hard to believe this is the last week with the students and the last week of the Upward Bound program! The goals for this week were to have students complete the silly sentences linguistic project and to create and finish a presentation for their celebratory, end-of-program banquet.

To start, I walked the students though a linguistics and computer science “Daily Dojo”. To recap, the “Daily Dojos” are brief Google Colab activities that I created to help students keep their Python skills fresh and to explore fun and relevant applications of programming and linguistics. Each Daily Dojo has a brief introduction of the topics as well as three “menu” choices for students titled “Illini”, “Orange”, and “Blue”. Students are meant to read through each activity and select the one that matches their skill level and interest; this is a form of differentiated learning where the three levels of difficulty are meant to present the information in ways to meet students where they are and then build upon those skills. This is one of my favorite tools to use because it is appropriately challenging for everyone.

The students had the most fun with the Word Cloud generator:

from wordcloud import WordCloud
import matplotlib.pyplot as plt

def create_word_cloud(text):
    # Create a WordCloud object
    wordcloud = WordCloud(width=800, height=400, background_color='white').generate(text)

    # Display the generated word cloud
    plt.figure(figsize=(10, 5))
    plt.imshow(wordcloud, interpolation='bilinear')
    plt.axis('off')
    plt.show()

This program counts word frequency in the input text and then creates a multi-colored word cloud of the most used words. I chose this package because it eliminates particles and determiners in favor of the verbs, adjectives, nouns, and more “interesting words”. The students enjoyed entering song lyrics and poetry and visualizing their content.

Tuesday’s goal was to complete the project, so after the dojo activity we did brief review of the English grammar concepts to make sure that the students retained the concepts from last week. Once we had working examples of determiner, proper nouns, common nouns, and verbs, we started independent work time. Sophia and I took a facilitator role of making sure students were on task and answering any questions that came up. The most frequent questions we had were about debugging the program as students were still learning how to program and would forget quotation marks on either ends of their strings.

On Thursday, we finished up the project and divided the presentation work among the students. They were nervous about presenting, so we made sure to help them create the best presentation they could as well as practice. There are a lot of people that make Upward Bound possible, so there were many names for the students to remember and pronounce correctly. They were able to add slides about themselves and snippets of the code they wrote for the project, and I hope they had a successful presentation!

Study Skills and Belonging

The team and I were able to meet this week to discuss the results of the linear regression of predicting final grades based on resources utilized. After training, the model was able to predict 21% of the variance, which means there is still a lot of unexplained phenomena and randomness that effected the results. The coefficients of each of the resources used was interesting because the highest coefficients were for the old examlets and tutorial problems.

I found this especially compelling because nearly all of the students in the recorded stressed the importance of completing practice examlets in preparation for the tests. Students mentioned that working through these old examlets was crucial to their understanding of the examlet layout and format as well as testing their application of concepts. Two students even mentioned that in preparation for an examlet they will complete every single old examlet available, which can be upwards of eight! The tutorial problems as a high coefficient was also interesting because several of the students in the interviews mentioned that the tutorial problems are often more challenging than what is asked on the homework and even on the examlets. Other students mentioned that because they were so challenging, they do not utilize them in their studies. A conclusion to be drawn is if students are able to correctly finish the tutorial problems, they are demonstrating confident understanding in the course material. So it appears that if students are using practice examlets and tutorial problems, it is more likely that they will finish the course with a higher grade.

Conversely, the lowest coefficients were external resources (such as YouTube and chatGPT) and posted solutions; so students who reported using these resources are more likely to have lower grades at the end of the course. I feel that this is critical information because students who use outside resources are their primary form of learning and examlet preparation are not engaging with the course material as much as a high grade earning student. They may find information that is incorrect on these external sites or create associations that are not correct within the criteria of the course. The posted solutions are a resource that is the answer to the tutorial problems mentioned above. If students are spending a lot of time looking at the posted solutions, they are not necessarily trying the problem themselves, which we are seeing is imperative to their understanding.

General Reflections

As Upward Bound draws to a close, I am considering making adjustments to the curriculum for next year’s cohort. I think the project is most effective whene it is worked on in small bursts over the period of several days. The students in general benefit from new activities after about twenty minutes to keep their attention focused and refreshed. It is difficult in six sessions to cover the skills needed for introductory Python programming, introductory concepts to linguistics, and to produce an artifact of their learning. The Daily Dojos were a refreshing activity and also covered a wide variety of topics, which I feel is more appropriate for such a short program. I’ve also learned a lot about what it means to be a computer science teacher and how it is similar and different than my time as a music teahcer, and the importance of careful and considerate curriculum scope and sequencing in planning. Building successes with students is important for them to feel confident moving forward into harder material.

I am excited by the results of the linear regression in the CS173 analysis. While I know that the model does not explain a lot of the variance in the data, it is still exciting to see that there are trends emerging from the data. The results of the high coefficients for old examlets and tutorial problems and the low coefficients for external sources and posted solutions feel promising as we look to provide advice for future CS173 cohorts. Students have different ways of acquiring information and unique ways of learning by finding different combinations of materials. Perhaps the folks who are utilizing external resources prefer to listen to someone explain the concepts and we could recommend that they use the lecture videos and attend office hours instead. I believe it is important to consider the human experiences behind the data points, so as not to minimize students’ study behaviors, but instead offer alternatives.