Context & Scale
Students complete a survey on an issue of local/global importance. The data set is then used for students to explore in class and practice organising, analysing and visualising data. This pattern is suited to large quantitative units where students can benefit from learning multiple core concepts through a single data set.
Learning how to organise, analyse and visualise data requires students to have access to relevant, authentic data sets with an appropriate sample size. Data sets need to allow students to learn how to apply a range of different statistical methods. By asking students to complete a survey targeted to the needs of the unit, the size of the cohort allows a substantial data set to be created quickly and easily for use throughout a unit.
The literature suggests that students are “more easily convinced of the power of statistical reasoning if they see it applies to questions that are interesting and real to them” and they can “experience firsthand the many issues that arise in data collection and analysis.” (Smith, 1998, p. 2). More questions and fewer contexts are preferrable for introductory statistics courses so that connections can be made between the context and the statistical tools used (Brown, 2019).
Research has also shown that students need to find and generate their own data to gain experience in asking questions, defining problems, formulating hypotheses, designing experiments and surveys, collecting data, and analysing and communicating findings (Hogg, 1991).
Textbooks often contain many examples of data sets that students can use to learn about and apply various statistical methods. Quite often the examples used are ad hoc, and not relevant to local contexts that students can relate to. While it is possible to source authentic data sets to use in teaching statistics, finding suitable data sets is often very challenging. This is because it is difficult to find a single data set that can be used to teach multiple different statistical concepts. Further, data sets that are tied to assessment need to be changed every semester to avoid plagiarism issues, adding to the workload for the unit coordinator in an already demanding large unit.
Students are asked to complete a pre-prepared survey (created in a survey platform like Qualtrics) early in the unit on a topic relevant to the discipline, or another topic that will stimulate students’ interest related to an issue of local/global importance. The process of answering the survey draws students’ attention to the process of data collection from the outset. Once students have completed the survey, the data set generated from the responses is downloaded as an Excel spreadsheet. The data set is then used in subsequent tutorials/workshops for students to practice applying core statistical methods. That is, workshop tasks require students to organise, analyse (investigate the data using different statistical methods and tools) and visualise the data.
Create a survey that will generate data that allows students to apply the specific statistical tools that students need to learn in the unit. The survey can be set up in Qualtrics or a similar survey building tool.
Embed the survey early in the unit and make it a requirement that students complete it. For example, the survey could be introduced in the first lecture or workshop or embedded in an online module in Week 1 of the unit.
In preparation for subsequent workshops where the data will be explored, download the data set. Prepare specific tasks in the workshop/s that require students to investigate the data.
Data sets could be used in weekly workshop activities or be built into a formal assessment where students can be asked to work in groups to clean, investigate, analyse and visualise the data, and/or present their findings.
Examples of pattern in use
Example 1: Business Analytics (undergraduate)
This pattern was tested in a first-year undergraduate core unit (Quantitative Business Analysis) in the Business Analytics discipline with approximately 1000 students.
The aim was to emphasise the investigative nature of the field to students in this introductory statistics courses, as this has been shown to help develop statistical thinking (GAISE Report, 2016; Cumminsky et al., 2020).
The approach was iteratively developed and implemented. We acknowledge the unit coordinators and lecturers involved from the Business Analytics discipline, including Bern Conlon and Laurent Pauwels.
Students in the unit had previously indicated that using authentic data they could relate to would better support them in learning how to organise, analyse and visualise data. Coordinators had previously found it difficult and time consuming to incorporate authentic data sets, in particular to find data sets that were capable of accommodating all of the various statistical methods that students needed to learn and practice. In addition, to support academic integrity, data sets needed to be changed every semester when used for assessment purposes.
Students were asked to complete a pre-prepared survey on their own perceptions and uses of plastics. This topic tied in with other activities that looked at the same issue on a global and local level. The survey intentionally connected students to the issue on a personal level to stimulate their interest. The survey, created in Qualtrics, was included in students’ first online module in the unit (Week 1). Students were required to complete the module before attending their first workshop. At the end of the week, the student data generated was downloaded from Qualtrics as an Excel spreadsheet. It was then used in subsequent tutorials for students to practice cleaning, analysing and visualising the data. The benefits came from students revisiting the same data set over several weeks.
Technology / resources used
Qualtrics was used to set up the survey. The survey was then embedded directly in Canvas so that students could access it easily and complete it as part of the sequence of activities in Week 1 of the unit material.
Data was collected from student surveys and focus groups over two semesters. In the first semester almost all students (92%) indicated that completing a survey in the lecture helped them reflect on how data was collected, and 85% said that working on the data that was generated by the students themselves made them feel more engaged in the process of data analysis.
The survey revealed that the majority of students felt that the issue that the survey focused on (in this case plastics use and production) was of global and local importance. Findings showed that the more students felt plastics was an issue of importance, the more that participating in the survey helped them reflect on how data was collected, and the more they said working on that data set made them feel more engaged.
The authentic nature of the data supported student motivations and engagement as shown by the comments from a student focus group in 2020:
We can relate to it – this is actually data from BUSS1020 students; here we can apply it – they are real people, not made up statistics.
Our data base was coming from real-life communities that we’re a part of.
We went through the process of cleaning…making charts…
This is actually the most interesting thing I found in the whole thing; relevant to our cohort specifically.