A methodological strategy for active learning in multivariate analysis

A methodological strategy for active learning in multivariate analysis

Alba Martinez-Ruiz and Cristian Cardenas-Oviedo

In 2011, the Industrial Engineering program at the Universidad Católica de la Santísima Concepción began implementing a curricular reform based on the CDIO approach. Keeping in mind a future CDIO Certification, we decided to reach level 3 of Standard 8, "teaching and learning based on active experiential learning methods" in several courses. Thus, in a first step toward reaching said level, this paper describes the design and implementation of a methodological strategy for putting active learning activities into practice in the discussion session of the multivariate analysis course of the program.

Learning multivariate methods is a challenging task for undergraduate students. They not only have to know, understand and assimilate those statistical concepts underlying each statistical method, but they also must be able to develop statistical thinking, to interpret statistical results in different application contexts and to acquire the skills to use those statistical software that nowadays are widely used in business and industry. Discussion sessions help students strengthen their multivariate techniques results interpretation skills. To do this, a set of case studies relevant for the industrial engineering program is prepared for each discussion session and distributed to students in advance. Discussion sessions have been traditionally given as recitations but with intensive interaction between lecturer and students.

Literature discusses several methodological strategies that bring active learning into the classroom (Hall et al., 2002, Crawley et al., 2007). From the standpoint of statistics, some authors have developed and implemented active learning techniques in introductory statistical courses. To the best of our knowledge, no experiences have been reported for advanced statistics courses.

We decided to implement active learning through small group work. We redesigned the case studies documents given to the students for the discussion sessions. The new configuration consists in a structured format including: an introduction with instructions and the description of the case study, a data set, and three sections with information. Each section includes tables and figures with the results of the applied statistical methods, the underlying statistical theory, and a set of questions to be answered by students. This new structure also seeks to facilitate understanding the relationships between the mathematical background and results using a real-life and relevant engineering problem. We try to prevent students from mechanizing the interpretation of results. During the discussion sessions, students are organized in groups of three. Each student guides the work of the group in one of the three sections.

To evaluate the impact of this strategy, we designed a survey pre-post intervention as an assessment tool of the active learning method. Preliminary results indicate that 88.64% of the students reported they like to work in small groups with peers to solve the problems posed in the discussion sessions. However, 28.41% reported preferring the traditional lecturing style.

Future work includes the design and implementation of tools for evaluating the "impact on student learning of active learning methods". This seems to us a very challenging task due to the difficulty of separating the effects that multiple factors may have on student performance.

Proceedings of the 10th International CDIO Conference, Barcelona, Spain, June 15-19 2014

Go to top
randomness