Skip to main content
Our journal logo
Search

Vol. 18 No. 2 (2021)

Published: 1 September 2021

Articles

  • Editorial 18.2: Innovations and challenges in the teaching of statistics to non-specialists

    Margaret MacDougall

    Welcome to this special issue of the Journal of University Teaching and Learning Practice. Over the past 50 years, great strides have been taken in the use of technology and distance learning across the disciplines and there has been rising interest in research-informed teaching. Correspondingly, there has been a movement towards greater openness on the part of statistical educators to change their teaching practices in an effort to strive for excellence. In this special issue, a variety of approaches to teaching statistics to non-specialists are presented, all of which have a grounding in the pedagogical literature and involve innovative practices within the authors’ own institutions. With the support of a ‘ten tips’ article involving a book review, consideration is also given to maintaining quality standards. This includes recognising the place of statistics among the sciences and ensuring that making the subject accessible does not lead to compromises in student understanding of statistical inference and the conditions for use of basic statistical procedures. The papers are united in addressing challenges in the teaching of statistics to non-specialists and as such, should assist in supporting communities of practice in this teaching area. In adjusting to the varying social distancing measures presented by the Covid-19 pandemic, we have both observed and participated in new ways of building bridges with students which have unquestionably reshaped the landscape of learning and teaching. Many of us who have found these experiences rewarding will be reluctant to return to status quo as post-pandemic plans unfold, since responding to the need for creativity in distance learning provision has proved beneficial to our own personal professional development in serving the needs of our students. Others may have felt ill-prepared to reconstruct their course content due to a lack of time, training and resources. Either way, I hope that many of the ideas and recommendations provided in this special issue will prove supportive and serve as an impetus to press on in an effort to make the learning of statistics an attractive and rewarding option for non-specialists while maintaining quality standards, particularly where staff resources are limited.

    Read article: Editorial 18.2: Innovations and challenges in the teaching of statistics to non-specialists

  • What statistics instructors need to know about concept acquisition to make statistics stick

    Jochen Kruppa, Jessica Rohmann, Carolin Herrmann, Miriam Sieg, Kerstin Rubarth, Sophie Piper

    "The limits of my language are the limits of my mind. All I know is what I have words for" (Wittgenstein). When learning something completely new, we connect the unknown term to an already existing part of our knowledge. We can only build new ideas and insights upon an existing conceptual foundation. In the field of statistics, we educators frequently find ourselves met with great confusion when teaching novices. These students, entirely unfamiliar with even basic statistics, must connect the introduced statistical terms within their personal existing networks of largely non-statistical knowledge. Lecturers, on the other hand, who are well versed in statistics, have deeply internalized the content to be taught and its relevant context. The juxtaposition of the two roles may produce amusement in a lecturer upon gaining insight into the word associations made by the statistical novices. For example, a ‘logistic regression’ does not involve the ‘shipping of goods in economically difficult times,’ though this might seem entirely reasonable and intuitive to the statistics learner. Other times, these different perspectives can lead to headaches and frustration for both learners and their lecturers. In this article, we illustrate how simple statistical terms are initially connected to a student’s pre-exiting knowledge and how these associations change after completing an introductory course in applied statistics. Furthermore, we emphasize the important difference between “term”, “approach”, and “context”. Understanding this fundamental distinction may help improve the communication between the lecturer and the learner. We offer a collection of practical tools for instructors to help promote students’ conceptual understanding in a supportive, mutually-beneficial learning environment.

    Read article: What statistics instructors need to know about concept acquisition to make statistics stick

  • Revelations from story writing in business statistics: An exercise in decoding

    Collette Lemieux, Brad Quiring

    This paper describes our experiences with implementing an innovative approach to decoding our discipline, which involves writing dialogue-driven stories that explore authentic, context-rich problems within introductory business statistics. Our paper begins by introducing the decoding model created by Middendorf and Pace (2004). We then explain why we initially chose to write pedagogical stories as a meaningful way to deliver our course material, later discovering that the process also served as an alternative means of decoding our discipline. The discussion focuses on our case study, which investigates how the process of writing stories lead to significant benefits for ourselves as instructors. In particular, we connect our learning experiences to Middendorf and Pace’s (2004) work on decoding the discipline, which utilizes a seven-step process to help faculty members interrogate their teaching processes for bottleneck concepts. We present our reflections on how the process of writing stories acted as an effective alternate means of decoding the discipline.

    Read article: Revelations from story writing in business statistics: An exercise in decoding

  • When one size does fit all: Simultaneous delivery of statistics teaching to multiple audiences

    Rachel Hilliam, Karen Vines

    In 2019, The Open University celebrated its 50th anniversary. For most of its fifty years it has been delivering statistics education to both specialists and non-specialists, simultaneously and at scale. A single module must deliver statistics teaching across multiple qualifications in many different disciplines which means it is not possible to tailor module material and examples for each individual qualification. This is contrary to the way many service teaching modules are designed. For example, the introductory statistics module at The Open University delivers basic statistics and probability to around 1500 students a year. By structuring the material around topics to which everyone can relate; money, education and health, the module is able to engage students across a range of different qualifications from mathematics to economics, business studies, computing and natural sciences. Feedback from students has been overwhelmingly positive. Students are provided with printed material, which is supplemented with a variety of interactive elements, which will be described in this paper. Online forums provide a place for peer support. Small group tutorials, both face to face and online, are provided by tutors. Each tutor will typically support 20 students and provide feedback to those students on their assessable work. The students in an individual tutor’s group could be studying one of many different qualifications. All the module material, including all the interactive elements, are embedded in the university’s virtual learning environment which has made it possible to monitor how students engage with the material. This knowledge, and qualitative feedback from students, has provided support for the idea that by teaching through a narrative approach, with examples based on scenarios which have a broad interest to everyone, it is possible to engage all students with statistics regardless of their qualification. By taking this same narrative approach a new module is being created which will be studied concurrently by students in the final stages of their qualifications in mathematics, statistics, data science and economics. This paper will outline how the module will utilize Jupyter notebooks to deliver practical modelling with real data sets chosen to be of interest to all the different cohorts of students. During the last 50 years The Open University has always had to deliver modules at scale to students across many disciplines. With the increased use of data across qualifications, and the need for people to upskill in data analysis, it is essential to explore ways in which statistical training can be delivered at scale.

    Read article: When one size does fit all: Simultaneous delivery of statistics teaching to multiple audiences

  • A case study in the e-assessment of statistics for non-specialists

    Iain Weir, Rhys Gwynllyw, Karen Henderson

    The need to be able to choose, perform and interpret the results from appropriate statistical tests is ubiquitous in many STEM disciplines and beyond. In this case study, we illustrate a learning and assessment approach that has been used successfully for first year Business School students taking a module on Business Decision Making for Marketing and Events. The objective of the module is to give students a grounding in statistics, including learning to use the SPSS package, and to apply the techniques learnt to a large real-world data set. Due to the large number of students and their diverse abilities the learning and teaching strategy is centred on using formative key skill e-assessments to encourage students to self-learn. On accessing each e-assessment, students are presented with a randomly generated data set, which they are required to import into SPSS to then appropriately analyse and report on. Marking and feedback occurs automatically on submission and repeated use ensures that a student thoroughly learns a key skill and covers various analysis outcome scenarios, for instance significant or not significant test outcomes. Each key skill e-assessment has multiple embedded links to comprehensive Help pages that provide SPSS ‘how-to’ information or output interpretation. Access to these means that staff in the PC sessions can concentrate on giving higher-level support as opposed to merely helping with the mechanics of producing SPSS output. The e-assessments were created using Dewis, which is a fully algorithmic open-source e-assessment system. Dewis’ ability to communicate with the R programming language facilitates the creation of authentic e-assessments, generating bespoke student data and providing answers that match SPSS screen output. The assessment of students’ statistical competency is by in-class e-exams, which are summative and sat under exam conditions. The fact that the formative and summative e-assessments are marked immediately by Dewis represents a huge saving in marking time compared to human marking and allows students to have instant feedback on their work. Repeated use of the formative e-assessments allows students to independently practice and check their understanding. In this case study, we find that those students that did practice achieved significantly higher marks in the e-exams. We are now in our third year of delivery and the module has developed further due to stronger connections between the business and statistics strands of the module.

    Read article: A case study in the e-assessment of statistics for non-specialists

  • Online seminars in statistics for doctoral students: A case study

    Christine Davies

    The Doctorate in Professional Practice (DProf) programme at the University of Wales Trinity Saint David (UWTSD) provides a means for practitioners in a range of professions to undertake research, and achieve doctoral status, whilst maintaining their professional roles. The programme is generic, covering a wide range of professional backgrounds, though most are allied to the social sciences. Students on the programmes are generally mature individuals, senior in their fields, from across the world. They have significant expertise in their professional fields, but often have limited understanding of quantitative research processes, including the use of statistics. This deficit is addressed via a module in Quantitative Research delivered via a series of seven online seminars utilising meeting/webinar applications. Though the use of online technology has become commonplace since the advent of COVID19, its use as a means of delivering a series of discursive, problem-orientated statistics seminars for small groups of students is innovative. Also, learners are required to undertake specific reading and viewing of videos before each seminar so that they become acquainted with key points in advance. During seminars, students discuss specific topics, participate in problem-solving, and learn how to use software for statistical tests. Screen-sharing permits real-time observation of the use of software. At the end of the seminar series, students are provided with at least one individual online tutorial so that specific issues can be addressed. This Case Study examined learners’ views of the Quantitative Research module by means of dialogic interviews. Thematic analysis of interview transcripts indicated that the use of an online seminar series to deliver the module was viewed positively by learners. They particularly appreciated the social aspects of the seminars which helped to remove fears and encourage participation. Understanding was enhanced by viewing short videos before and after the seminars, and by activities during the sessions. This innovative way of teaching statistics appears to have been effective in equipping DProf candidates with the statistical knowledge and skills required to undertake their Part 2 research.

    Read article: Online seminars in statistics for doctoral students: A case study

  • Teaching research methods to undergraduate dental students

    Sam Leary, Andy Ness

    The provision of undergraduate statistics teaching varies across UK dental schools, but some challenges are faced in all schools. These include the need to (i) demonstrate the relevance of statistics to dentistry, (ii) address the phenomenon of statistical anxiety likely to be experienced, and (iii) deliver the teaching with a limited number of statistically trained staff. Our objective was to design a research methods course that combined epidemiology and statistics teaching for undergraduates at Bristol Dental School (University of Bristol) that was clinically relevant, focused on concepts and interpretation rather than calculation, and was sustainable, using new technology to enhance learning. The research methods course was introduced in 2008, extensively developed over the next two years based on student and tutor feedback, and has run with only minor updates ever since. The aim of the course is to introduce year 2 dental students to the skills needed to practice evidence-based dentistry, i.e. understand and critically appraise published research. Basic epidemiological concepts, different types of study design, summarising and interpreting data, and choosing appropriate statistical analyses are covered. The course is introduced by a face-to-face lecture. This emphasises the relevance to future careers, and pre-empts the feelings of statistical anxiety by presenting evidence that exam results for this course are not associated with having achieved an A-level in mathematics. The rest of the course is delivered using the flipped classroom approach. Didactic teaching is in the form of nine e-lectures, each lasting 20-25 minutes. These are split into chapters to allow easy navigation, and include pop-up questions. Small group (up to 10 students) structured tutorials (one per e-lecture) are used to reinforce the material covered in the e-lectures, drawing on real clinical examples from research publications. They are interactive, and also include e-voting quizzes to allow the tutors and students to gauge the level of understanding that is being achieved. At the end of the course there is a revision session and written assessment, which must be passed before students can progress to year 3 of the dental degree. Approximately 70 students take the course each year, split into eight groups, and tutorials are based on pre-prepared materials. Tutoring on this course provides an attractive opportunity to gain teaching experience, with only a modest investment in time. Therefore, it has always been possible to recruit enough statisticians or epidemiologists in the Faculty of Health Sciences, University of Bristol to act as tutors. At least 89% of the students have passed the exam on first sitting each year, and all students have passed their re-sit exam. Marks ranged from 50-86% in 2019, and similar ranges were seen in previous years. Student feedback is consistently high, with virtually all students rating all components of the course at least satisfactory, and high percentages rating them good or excellent (e-lectures and tutorials were rated to be good or excellent by 72% and 95% of the students respectively in 2019). External examiners have been very complimentary, and only ever requested minor changes. Hence it has been possible to run a sustainable research methods course that engages students’ interest and produces excellent learning outcomes.

    Read article: Teaching research methods to undergraduate dental students

  • Ten tips for statistical educators in response to a constructive review of the book How to Analyze Data

    Margaret MacDougall

    This paper comprises a review of the Pocket Study Skills book How to Analyze Data which I was invited by MacMillan International Higher Education to provide as a contribution to the current special issue. This includes both a preliminary review and an extension of this review to provide ten tips for educators to enhance the statistical learning of non-specialists in statistics. The tips are intended to provide a constructive approach to the review process through safeguarding educators, including dissertation supervisors, and learners, from being misled by some of the shortcomings that I have identified in this book. It is in this sense that I refer to these tips as “remedial tips”. As such, these tips should not be regarded as exhaustive in their own right in reflecting the statistical learning needs of non-specialists. Nevertheless, one of the important responsibilities professionally trained statisticians face in the teaching of statistics to non-specialists is that of empowering learners to stem the tide of misuse of statistics within their own disciplines, and one of the corresponding challenges is that of identifying published material that is potentially unhelpful and even counterproductive in achieving this goal. Thus, it is also my intention that these tips will be of broader pedagogical value as a handy reference, including for those uninitiated teachers of statistics to non-specialists who may be tempted to reach out for a pocket guide to support them in balancing the demands of teaching and research in an effort to make statistics less mystifying to their students.

    Read article: Ten tips for statistical educators in response to a constructive review of the book How to Analyze Data