Research: A Mixed-Methods Approach to Characterizing Engineering Students' Computational Habits of Mind

Purdue University, West Lafayette IN

Investigators

Abstract

Finding solutions to complex problems in our society requires engineering graduates who not only possess the technical knowledge and skill set of a discipline, but also a professional mindset. "Habits of mind" are modes of thinking required to become effective problem solvers capable of transferring such skills to new contexts. Such skills form an essential part of engineer's professional mindset. This project studies undergraduate engineering students' habits of mind related to computing. The novelty of the proposed approach is that the data collection and analysis methods integrate machine learning with methods traditionally used in education to analyze data. The guiding education research question of the project is: What computational habits of mind do engineering students exhibit at different points of their academic development? The goal of the effort is to provide a better understanding of how these patterns relate to academic achievement, which in turn informs learning pathways toward computational proficiency in engineering. The project follows an Explanatory Sequential Mixed Method Design to study engineering students' computational habits of mind by integrating machine learning in a mixed method research. High-dimensional quantitative data is collected for each learner and machine learning is used to find patterns corresponding to groups of similar learners. By dividing the learners into similarity groups, these patterns provide a higher detail of understanding than global analysis methods. The reasons behind the patterns will be identified by collecting and analyzing qualitative data. Building on the machine learning results, an automatic subject selection strategy is used to reduce the number of learners considered to a manageable size for the qualitative part of the research. The goal of this learner reduction strategy is to obtain a small but statistically meaningful sampling for each group of subjects in order to focus resources for the most costly and time consuming part of the research. By doing so, mixed method research is shown to be scalable to large and complex datasets. The rationale for this project is that its successful completion: a) addresses challenges in cultivating a culture of lifelong learning among professional and future engineers; b) provides an understanding of how patterns of habits of mind relate to academic achievement, which in turn informs learning pathways toward computational proficiency in engineering; and (c) demonstrates how mixed method research can be scaled up to study very large and complex multidimensional datasets through the use of machine learning. This effort supports the need for the foundation in engineering students' education to incorporate necessary skills such as creativity, ingenuity, professionalism, persistence, and willingness to take calculated risks. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.