How to use artificial intelligence strategies to ease student insecurities

People are beginning to feel uneasy about whether college students are learning and developing well, but there are few clear numbers to identify "important differences". Numeric situations usually refer to situations where there is a "right" or "wrong" answer (similar to the on/off button on a laptop). If the data are not interpreted correctly, some conventional statistical procedures may support the idea that a "difference" may be found in an analysis comparing the scores of two groups of students (e.g., yes/no p ≤ .05?). However, no single finding can be convincing because student learning and development is a complex process that goes far beyond the sophistication of numerical analysis.

The purpose of data analysis is to identify patterns and anomalies in student learning and development. Student learning and development is a gradual process that requires the comprehensive consideration of multiple factors. As a result, universities and higher education institutions are adopting artificial intelligence and “simulation” strategies to analyze data to gain a more comprehensive perspective. These simulation tools can create virtually unlimited options, between nothing and everything in between, to help institutions better understand student learning and development.

Even considering whether different student subgroups have more similar rather than different scores is a simulated situation because we realize there is no one right answer that applies to all students on campus. In order to explain why students' learning and development are so complex, we need to expand our perspective and understand the influence of all relevant factors, including but not limited to aspects of students' background, culture, education and family life.

Therefore, we need a deeper understanding of students' learning and development processes, rather than just relying on the results of conventional statistical procedures. By employing artificial intelligence and simulation tools to analyze data, we can gain a more complete and comprehensive view of student learning and development.

Developmental science, including developmental psychology, cognitive science and neuroscience, does not only focus on the "age and stage" development of children, but focuses more on exploring the "trajectory" of students. Changes in these trajectories are determined by many factors, not just those predicted by immutable demographic characteristics and past academic performance. Development trajectory is a student's life path influenced by the past, present and future, which determines the student's future development direction. Therefore, understanding changes and factors in student trajectories is critical to developing individualized education and development plans.

We examined fifteen longitudinal datasets that combined different computer information systems and performance-based assessments to collect data on student learning and development. These datasets date from 2007 and each longitudinal dataset contains more than 1.9 million individual data points. By using machine learning techniques and AI cognitive analytics, we built predictive models to identify patterns and anomalies in the data collected about student success in these longitudinal cohort studies. We also used SPSS statistical software for linear and binary logistic regression analyses, and AMOS for structural equation modeling. By using different analytical methods, we confirmed the findings and arrived at the same findings, increasing confidence in the findings.

In our research, we found that changes in student trajectories can be seen as a conscious deviation, and students can make changes in their expected life paths through self-adjustment. For example, a student may be placed on a trajectory leading to college success but decide to reorient themselves toward a different trajectory that leads to dropout. Our research also shows that changes in students' trajectories are determined by multiple factors, such as students' personality, family environment, education level, psychological state, etc. Therefore, developing personalized education and development plans needs to comprehensively consider these factors to help students find the trajectory that best suits them and realize their greatest potential.

Similar results were obtained using techniques such as machine learning, AI cognitive analysis, and traditional statistics. In a 2017 paper, "Using Support Vector Machines to Predict Student Graduation Outcomes," it was introduced how to apply machine learning technology to predict student graduation. The paper leveraged more than 100 features to build a predictive model, including a set of factors to measure student learning and development. The research results confirm the conclusion of AI cognitive analysis: students' admission background does not determine their future, but their learning and development experiences after admission are more important in predicting academic achievement and graduation. Applying AI strategies provides the most useful information. Student trajectories are complex, but AI can handle this complexity.

Collecting data on student learning and development helps advance “simulation” thinking because developmental science considers all of a student’s experiences on campus and over time within the same framework. AI strategies are useful when analyzing all the “fragmented” data on student learning and development.

The above is the detailed content of How to use artificial intelligence strategies to ease student insecurities. For more information, please follow other related articles on the PHP Chinese website!