Introduction: This dataset compiled from a Portuguese higher education institution uses predictive analytics to assess student performance. A dataset (acquired from several disjoint databases) related to students enrolled in different undergraduate degrees that speculated from journalism, education, social service, technologies, nursing, design, management and agronomy. Our dataset consisted of 36 variables that contributed to predicting students academic success into categories predicting dropout, enrollment, and graduation rates. This was designed to help educators recognize patterns within student performance and is used to forecast student academic success. Among the predictive variables are demographic factors such as nationality, gender, and family income, as well as parent qualifications. We specifically looked at the fathers' qualifications and this raised our interest of how do uncontrolled and inherited variables such as parent qualifications unfairly impact a students predicted success? This brings us to our question of how do structural variables like parent qualifications influence a students academic success? This, we believe issues reinforcement of social inequality to the opportunity of academic success in contributing to a lot of the students academic standing and opportunity.
Anatomy of Formats: In using this method we focus on not only the data, but how the structure of the dataset determines which aspects of student identity and experience are seen and which are left out. This directed our attention to focusing on the variables seen such as parental qualification, as well as, the factors absent such as motivation, emotional well-being, and social support. These factors being more difficult to measure and quantify, but crucial to academic success.
Reading Datasets: Using connotative reading which goes beyond surface level definition and constantly shifts definitions for variable impacts, we referred to the term "academic success" and how the meaning has evolved. Traditionally, academic success refers to a students participation within educational courses. In contemporary context, this meaning has expanded. Success in education increasingly depending on financial resources and requiring more support for enrolling, maintaining, and achieving educational success.
Our Work: We conducted research by graphing our datasets to fully realize what were more the important variables, as well as, any factors in absence. In our research we noticed some missing factors, we thought about what other factors could potentially be absent that could affect overall outcome of each dataset. We thought about what exactly the "parental qualifications" meant for the graphs, a factor that was a bit too vague, how qualified as a parent someone would need to be in order to affect a students academic outcome; this question intrigued us and we decided to push further, trying to look deeper.
Findings: Financially supportive fathers with heavy educational qualifications lead to a greater chance of success in students with more room for academic investment and influencing academic achievements. Using Koopman's method, we examined the fathers qualification construct situated at the meso level between the individual student and the broader institutional or systematic context. Revealing the students academic outcome is strongly influence by factors beyond their control. With students chance of academic success based on some uncontrolled factors such as family qualifications, this puts the student at risk for success without regarding their individual performance or academic potential. This then raised some more questions:
What is the employment rate of students who dropped out?
Is academic success in this dataset a measure of institutional achievement?
Is "academic success" strictly graduating? (rather than dropping out)
These misunderstandings in the equation makes it more difficult to base any claims on these concrete variables. As far as we know, according to the data, the individuals academic success is quantified exclusively by graduation, ignoring an individuals GPA (implying that an all D's passing student who graduates is considered "better" than an all A's dropout student). Through applying Koopman's method we can identify and interoperate the missing layers and particularly the macro of this analogy, but in our case the institutions definition of success, or any economic systems influencing student financial burdens. These variables are missing details of the dataset and become even more visible when absent as they play a crucial role in shaping a students educational experience and how we understand and assess academic success.
Through these many variables, this dataset attempts to predict student dropout and success rates. Utilizing their known background information from parental occupation and their qualifications, we look at how it affects the students opportunity of academic failure or success. We have looked through 370 variables that contribute to the students academic outcome. These variables were set to give a good prediction of students at an early stage of academic risk and aims to reduce academic failure or risk.