As organizations increase their capacity to gather more data, the differentiator will be having people who can ask the right questions to make the most of such data. For this reason, fostering critical thinking in students is essential in teaching statistics.
In the era of big data, automation and the fourth industrial revolution, how can we adapt our teaching so that students will be able to address future challenges successfully? The capabilities of computers and their algorithms are undoubtedly becoming better and more useful to human beings. In fact, there is a widespread fear that machines could replace us in the jobs we currently hold (World Economic Forum, 2018). However, according to experts, the tasks performed by computers are oriented towards very specific activities, in such a way that the intelligence needed to face unknown and unexpected situations in which it is necessary to evaluate what is happening, why it is happening and what the consequences are of taking certain measures in this regard, is a task that is still pending (Smith, 2018).
In this regard, I believe that fostering skills in students, such as critical thinking, is essential in teaching statistics, because, as organizations increase their capacity to gather more data, the differentiator will be having people who can ask the right questions to make the most of such data. Although current discussions on the concept of "data" focus on the technology and artificial intelligence facet, it is the human side that will continue to be the biggest differentiator for teams and organizations, according to Chamorro-Premuzic (2018).
“The use and application of Data Science could be understood as an art, since handling so much information requires intuition, qualitative vision, understanding and interpretation of the information available on a real problem.”
In statistics, Data Science is an established term to refer to the scientific discipline that involves mathematics, computer science, operations research, applied sciences and statistics (Weihs and Ickstadt, 2018). In general, large volumes of data, known as "Big Data", are handled to extract knowledge derived from the processing, analysis and interpretation of the same.
Data Science also uses programming languages and artificial intelligence to analyze multivariate databases that contain data in image, text, sound and measurement formats, among others. In the business world, where information is constantly generated, data modeling through Data Science allows directors to make timely decisions, monitor the quality of products and services continuously, and forecast sales.
Some authors such as Boire (2018), consider that the use and application of Data Science could be understood as an art, since handling so much information requires intuition, qualitative vision, understanding and interpretation of the information available on a real problem that requires the creative integration of advanced knowledge (statistical and computational) to enable its modeling.
Other researchers comment that the skills that stand out most in this field of action are the formulation of productive questions, thinking computationally and analytically, visualizing and summarizing data, and communication and narration or effective argumentation (Dichev and Dicheva 2017). The latter, narration or effective argumentation is one of the skills least often inculcated in students.
“Narration or effective argumentation is one of the skills least often inculcated in students. Is one of the skills that graduates in any profession should have, because they need to communicate clearly, giving reasons that justify what they are expressing verbally or in writing.”
Effective argumentation is one of the skills that graduates in any profession should have, because they need to communicate clearly, giving reasons that justify what they are expressing verbally or in writing. Pasquier, Rahwan, Dignum and Sonenberg (2007) maintain that both argumentation and justification and even explanation can be viewed as processes by which a person shows someone else why their viewpoint is coherent.
In Data Science teaching, Parker (2018) maintains that effective argumentation is one of the least developed skills in students and that it must be acquired. Thinking of developing this argumentation skill, apart from statistical and computational skills, Parker (2018) planned a course in which her students had to formulate several research projects on real problems, in which they were asked to indicate the problem, collect information, analyze it and discover the relevant findings of the analysis in the context of the problem. In these projects, students needed to use technological tools and document the findings of the statistical analysis in written reports and present them orally.
Parker suggests that students can develop not only the skills specific to Data Science, but also other skills, such as effective argumentation, teamwork, etc. As a result, a recommendation for teachers of any discipline is that they should focus not only on developing the specific skills of that disciplinary field, but also on others that contribute to the comprehensive development of future professionals.
In turn, in statistics education, the need to foster inferential reasoning (e.g. Bakker & Derry, 2011; Makar & Rubin, 2014) in students has been highlighted. This type of reasoning is required in quantitative research to generalize toward a population when the data from a representative sample are available.
Therefore, this reasoning should be fomented in students from all specializations. In itself, inferential reasoning is defined by Inzunza (2013) as the process of going beyond sample data to draw conclusions from a universe that has not been fully explored, and, consequently, the conclusions reached are inaccurate.
Education in statistics researchers have exported several strategies to help their students develop inferential reasoning, such as informal inference, considered as inference carried out without using formal statistical inference methods (e.g. hypothesis testing and confidence intervals), as well as random sampling simulation. For example, in Ben-Zvi’s proposal (2018), his students had to experience three research cycles during an introductory course: the first using exploratory data analysis (graphical analysis), the second with informal inference and the third generating inference using simulation.
The similarity between Parker’s and Ben-Zvi’s proposals lies in the fact that in both of them students need to acquire course-specific skills, as well as other skills. To develop them, both Parker and Ben-Zvi exposed their students to activities or projects in which they have to put the skills into practice during the course in the same way as they would have done when learning an art (e.g. how to play a musical instrument, paint, sculpt). It must be noted that this article includes Parker’s and Ben-Zvi’s proposals solely to exemplify how these researchers aimed to develop certain skills and reasonings, and to give a brief explanation of the strategies they used. However, there are other investigations and proposals for teaching statistics in introductory curses, in which diverse tools and statistical software have been used.
As a teacher of the course Probability and Statistics, I have participated in the design of a course that sought to introduce inferential ideas, preliminary ideas on the formal concepts of statistical inference (e.g. p-value, confidence intervals and statistical significance, hypothesis testing) at the beginning of the course and then develop them during throughout the same (Ruiz, Albert, Tobías, Villarreal, 2014).
I then studied inferential reasoning through argumentation in written reports and oral presentations on comparing inferential situations in real contexts (time between heartbeats in patients with heart disease who use a pacemaker vs. those who do not use one, PM10 pollution in spring-summer vs. fall-winter) and in which the data that needed to be analyzed were provided. They have been generated in teams, using the statistical software Minitab. In an ongoing research project, we are interested in analyzing how students validate and support their statistical inference, observing whether they manage to integrate the contextual reasons and statistical reasons, as Bakker, Kent, Derry & Noss (2008) suggest. To analyze the solidity and validity of inferential reasoning, a conceptual framework was proposed that integrates Toulmin’s argumentation model and the essential components in inferential reasoning (see Gómez-Blancarte and Tobías-Lara, 2018).
I would like to invite my colleagues to read up on what experts in the disciplinary area we teach are doing. To achieve this, I suggest creating educator collaboration networks to share our ideas and experiences.
About the author
María Guadalupe Tobías Lara (firstname.lastname@example.org) conducts research in educational statistics and teaches the course Probability and Statistics at Tecnológico de Monterrey.
Albert, J. A., Ruiz, B., Tobías, M.G., Villarreal, O. (2014). Proyecto: Transformando la Educación Estadística desde un Enfoque Inferencial. Fondo NOVUS. Instituto Tecnológico y de Estudios Superiores de Monterrey. 2014-2015 Call
Bakker, A., Kent, P., Derry, J., Noss, R., and Hoyles, C. (2008). Statistical inference at work: Statistical process control as an example. Statistics Education Research Journal, 7(2), 130-145. [Online: http://iase-web.org/documents/SERJ/SERJ7(2)_Bakker.pdf]
Bakker, A., & Derry, J. (2011). Lessons from inferentialism for statistics education, Mathematical Thinking and Learning, 13(1-2), 5-26. doi: 10.1080/10986065.2011.538293
Ben-Zvi, D. (2018). Three paradigms to develop students’ statistical reasoning. In M. A. Sorto, A. White, & L. Guyot (Eds.), Looking back, looking forward. Proceedings of the Tenth International Conference on Teaching Statistics (ICOTS-10), Kyoto, Japan. Voorburg, The Netherlands: International Statistical Institute.
Boire, R. (2018). “The art of Data Science”, Predictive Analytics Times. [https://www.predictiveanalyticsworld.com/patimes/the-art-of-data-science/9475/]
Chamorro-Premuzic, T. (2018). 3 Ways to Build a Data-Driven Team. Retrieved from https://hbr.org/2018/10/3-ways-to-build-a-data-driven-team
Dichev, C., & Dicheva, D. (2017). Towards Data Science Literacy. Procedia Computer
Science 108, 2151-2160.
Gómez-Blancarte, A., and Tobías-Lara, M. G. (2018). Using the Toulmin model of argumentation to validate students’ inferential reasoning. In M. A. Sorto, A. White, & L. Guyot (Eds.), Looking back, looking forward. Proceedings of the Tenth International Conference on Teaching Statistics (ICOTS-10), Kyoto, Japan. Voorburg, The Netherlands: International Statistical Institute. 2018).
Inzunza, S. (2013). Un acercamiento informal a la inferencia estadística mediante un ambiente computacional con estudiantes de bachillerato. E-journal AMIUTEM, 1(1), 60-75.
Makar, K. & Rubin, A. (2014). Informal statistical inference revisited. In K. Makar, B. de Sousa, & R. Gould (Eds.), Proceedings of the 9th International Conference on Teaching Statistics. Flagstaff, Arizona, USA: International Statistical Institute.
Parker, H. (2018). Cultivating creativity in data work. Keynotes in In M. A. Sorto, A. White, & L. Guyot (Eds.), Looking back, looking forward. Proceedings of the Tenth International Conference on Teaching Statistics (ICOTS-10), Kyoto, Japan. Voorburg, The Netherlands: International Statistical Institute.
Pasquier, P., Rahwan, I., Dignum, F., & Sonenberg, L. (2007). Argumentation and persuasion in the cognitive coherence theory: preliminary report. Argumentation and Persuasion in the Cognitive Coherence Theory (193–210). Springer Berlin Heidelberg.
Smith, G. (2018). Beware the AI delusion. Retrieved from https://www.fastcompany.com/90247482/beware-the-ai-delusion?partner=feedburner&utm_source=feedburner&utm_medium=feed&utm_campaign=feedburner+fastcompany&utm_content=feedburner
Weihs, C. and Ickstadt, K. (2018). Data Science: the impacts of statistics. International Journal of Data Science and Analytics . https://doi.org/10.1007/s41060-018-0102-5
World Economic Forum (2018). Machines Will Do More Tasks Than Humans by 2025 but Robot Revolution Will Still Create 58 Million Net New Jobs in Next Five Years. Retrieved from http://reports.weforum.org/future-of-jobs-2018/press-releases/