Technology
Learning Outcomes
By the conclusion of the master’s degree program, students should be able to:
- demonstrate effective use of current technologies
- explain the impact and potential of current and emerging information technologies for management and delivery of services
- organize and provide digital resources and services
- evaluate and select appropriate technology for specific information services or applications
- analyze issues, concepts, and challenges related to the impact of technology on information ethics and policy
Description
I demonstrated my proficiency in technology through a project I completed for ISCI 709, Fundamentals of Data and Digital Communication. In Part I of the assignment, I described the contents of three datasets. Two datasets had primarily numerical data while the third contained lexical data.
In the second part of the assignment, I demonstrated my knowledge and facility with Python to analyze one of the datasets. The dataset I chose to analyze is called U.S. Education Datasets: Unification Project. This dataset was compiled using data from the U.S. Census Bureau, the National Center for Education Statistics (NCES), and the National Assessment of Educational Progress (NAEP), and provides an overview of financial and student achievement information in the U.S. from 1986 to 2019. To better manage the dataset, I omitted several columns with some information on expenditures and revenues. I analyzed the dataset using basic statistical functions. The preliminary results indicated a weak positive association between education expenditures and NAEP test scores, suggesting that beyond a certain point, additional spending has little impact on student achievement. State comparisons revealed that South Carolina was below average in both academic achievement and education expenditures compared to national averages. Despite spending a higher percentage of its revenue on education, the per capita spending in South Carolina was still below the national average.
Analysis
After searching for datasets to use for this project, I picked two datasets with lots of numbers that I knew I’d be able to quantitatively analyze, and one dataset of words, the Shakespeare dataset. I was originally going to pick the Shakespeare dataset for Part II of the project because I find text analysis fascinating, and as an academic librarian, I plan to support students and researchers in the humanities, who often need to analyze textual information. However, the requirements for the project fit more neatly with a dataset of numbers, so I decided to analyze the U.S. Education Datasets: Unification Project dataset for Part II. I chose this dataset because I am interested in education policy and assessing student performance. My goal was to explore the relationship between educational expenditures and student outcomes.
Some information was missing from the dataset, which slightly skewed my results. I wish I had better handled the missing data by imputing data via a method such as replacing the missing data with averages. I chose to simplify the dataset by focusing on total expenditures and revenues, omitting columns with subtotals. I wish I had been able to plot student achievement alongside the more granular financial data, but that was beyond the scope of this project.
I absolutely did not expect the finding that, after a certain point, higher spending does not necessarily translate to better outcomes. That really went against my preconceived ideas, providing a valuable example of why making decisions based on actual data is so important. However, my analysis was preliminary and I wasn’t able to determine any trends in achievement correlated with spending over a certain amount. I was also really surprised that South Carolina spends a higher percentage of its revenue on education compared to the national average. That finding was especially disappointing considering South Carolina still lags behind in per capita student spending and academic achievement.
As for technical insights, I significantly improved my understanding of Python while doing this assignment. I had some familiarity with Python before taking this course, but my knowledge of Python was reinforced and extended in ISCI 709. I had a great deal of difficulty using Python for data preprocessing; while I do still want to improve my Python skills, I have found that, in practice, any tool that will efficiently and effectively do the job is the right tool. If I had to redo this project, I would use Excel for data preprocessing as I am more familiar with Excel.
Reflection
The person who compiled the U.S. Education Datasets: Unification Project dataset also created another dataset with more information, including demographic information, that I chose not to use. Reflecting on the project, I wished that I had chosen to analyze the larger dataset. I was initially daunted by the amount of data and thought it would be unmanageable, but after becoming more familiar and comfortable with the data, I regretted my decision. The larger dataset could have provided insights into the relationship between race, gender, education spending, and test scores, information that is critical to achieving diversity, equity, and inclusion goals.
After completing the project, in addition to improving my Python skills, I found that I understood more about the complexities of educational data analysis, the importance of thorough data cleaning, and the nuanced relationship between educational spending and student achievement. Encountering issues with missing data and Python syntax errors highlighted the importance of meticulous data handling and the necessity of solidifying my coding skills. In future projects, I will approach data preprocessing with more robust strategies as well.
My experiences in ISCI 709 impacted future projects by causing me to adopt a more critical approach to data interpretation and spurring me to explore alternative hypotheses when faced with data that defies my preconceptions. Working on this project also helped to highlight the importance of effective communication of data insights. Creating visual representations, such as scatterplots, helped illustrate complex relationships within the data. I have found great value in using visual aids to enhance data narratives, ensuring that insights are accessible to diverse audiences.
Since working on this project, I have completed an internship during which I analyzed data generated by a university-wide survey at the University of Michigan. Even though we used different tools than the software I used in ISCI 709, the foundational data analysis skills I learned in class were invaluable in providing context that helped me succeed in a highly technical internship.