Adopt a coaching role: Students choose their own questions and may feel uncertain at first, so be visibly supportive and keep expectations manageable. Walk around actively, helping groups narrow down questions and focus on feasible analyses while keeping the atmosphere curious and relaxed.

Why it works

Many students struggle to see why they need statistics or how theory connects to real data and research. Working with authentic datasets and self-chosen questions shows statistics “in real life” and makes the relevance of methods more concrete. The aim is for students to apply statistical concepts to realistic problems, practice teamwork and develop critical thinking about their own analysis choices.

Five step plan

Prepare datasets: Select several authentic datasets from literature, previous research or open sources and arrange access for students. Check that each dataset allows for simple, answerable research questions with the statistical tools covered in your course.
Brief the assignment: Explain the activity, including whether it counts as in-class (formative) or summative assessment, and include it in the study guide if graded. Make sure all students know where to find the datasets, software and any templates or rubrics.
Guide question formulation: In the tutorial, have students work in teams of 3-4 to explore a dataset and define a simple, researchable question. Move between groups to help them avoid overcomplicated questions or variable selections and to check that their question matches the available methods.
Support the analysis and poster: Each group performs the analysis using R, Python or SPSS, focusing on correct reasoning and method rather than a “right” outcome. Ask them to summarise their research question, approach, key results and interpretation in a short, clear poster or on a whiteboard or sheet of paper, so that they can explain it briefly to you or a TA during the same tutorial. (Optional: you can also do a hand-in and quick grading afterwards.)
Give feedback and repeat: Provide concise formative or pass/fail feedback using a short rubric (e.g. clarity of question, appropriateness of method, reasoning, basic reflection). Ask students to note what went well and what to improve next time. Optionally repeat the activity weekly with new datasets so students gradually tackle more complex questions and gain confidence.

Assessment

Summative: Groups hand in their work which is graded by a brief rubric, for a small part of the final grade or pass/fail. See this document for how to let students form their own groups in Canvas that you can use for quick group-grading.

Formative: Give oral feedback on the spot or use a short, but non-graded rubric. Ask students to reflect on what went well and what to improve next time.

Suggested room set-up

Small groups. An active learning space is recommended.

Tips & tricks

Watch for over-ambitious or vague questions and help students scale down to something answerable with available tests.
Emphasise that process and reasoning matter more than getting a “perfect” result; a surprising outcome can still show solid thinking.
Let students choose datasets related to their interests to boost motivation.
Allow playful or unusual questions as long as they support serious statistical reasoning.

Requirements

Access to suitable real-life datasets and statistical software (R, Python, SPSS).
Clear written assignment description and a brief rubric or grading model (e.g. in Canvas).
Lecturer or TA presence during tutorials to coach groups and listen to short poster presentations.

Key advice

Start simple with small datasets or predefined questions in early weeks, then gradually increase the freedom and complexity as students gain confidence.