Prepare a clear presentation and useful feedback
Unit 14 · Wednesday 2 December 2026 · 12:30–16:00
Gavin Simpson
Aarhus University
Start from your Unit 13 work
Use your revised report, main figure and rehearsal notes.
Complete student project folder · Presentation brief
Date/time and formal submission arrangements TBA.
Student-facing preparation deck. Students can read it before the final class using existing work. It is not an additional substantial assignment. Show only relevant slides in class if the preparation was already covered in Unit 13.
Name the result before showing it
Outcome + comparison + units + time + population.
“What difference are we estimating, for whom and when?”
Tell us what the data actually represent.
Use the group’s prespecified target. For supplied cases explicitly label synthetic teaching data. Do not let the question become a collection of every contrast in a model output table.
Draw the design
Show how researchers assigned treatments to animals or pens. If they only observed existing conditions, show which farms or groups they selected.
Label repeated measurements separately from replication.
Explain why your model reflects that structure.
Ask students to make the design visible with a diagram rather than only a formula. For observational data, show the sampled units and say no treatment was assigned. Keep paper context separate from the hypothetical supplied design.
Show one main figure
Label axes, units, which group is subtracted from which and uncertainty.
Explain the practical magnitude and remaining uncertainty.
A mean interval does not describe every individual outcome.
Invite students to point to the estimate and explain what an interval covers. A broad interval can still be informative about what is and is not established. Avoid converting it into an unsupported equivalence or causal claim.
Make checks consequential
What did the model check reveal?
In your alternative analysis, which plausible assumption did you change?
What changed in the estimate, uncertainty or conclusion?
Students use one justified comparison already developed in class. No new compulsory complex modelling. A reassuring plot does not prove every assumption; mention the most consequential remaining one.
Explain the next experiment
Before collecting new outcomes, choose a difference that would matter and say whether researchers would assign treatments to animals, pens or pasture groups.
Compare a few candidate sizes under two assumptions .
State what the simulation omits and what feasibility needs checking.
Prospective planning targets are not selected from realised estimates alone. Use simglm to simulate the planned experiment and its stated analysis. Report fitting failures and uncertainty caused by using a limited number of simulated experiments; explain simplifications. Explain a different intervention question where appropriate; students are proposing, not conducting, animal experiments.
Share eight minutes
A
Question and design
2:00
B
Model, checks and main figure
3:00
C
Follow-up and conclusion
3:00
Use the six-slide outline .
Speaking allocation is an example. Each member should understand the whole project. Rehearse once with a timer and simplify crowded slides. Do not reward quantity of output over explanation.
Ask a question that affects the conclusion
“What would change if this assumption were different?”
“Does that sample size count animals, pens or pasture groups, and what treatment would each receive?”
“What does the new experiment let us conclude?”
Feedback should identify a real consequence, not demand unnecessary models or more outcomes. Ask for clarification when the design or target was not recoverable. Use neutral wording and acknowledge evidence that was clearly explained.
Listen, record, revise
Peer sheet : evidence → question → feasible revision.
Final checklist : science, planning and communication.
Keep an offline copy of your figure, report and slides.
These materials support formative learning. Inability to project slides can be handled with a report or printed figure. Students document code assistance and any previously accessed solutions; individual questions test their reasoning. Any final examination requirements must be announced separately.
Each member explains a decision
Justify your model from the question and design.
Explain one alternative and its consequences.
Answer a directed question in your own words.
Disclose assistance and previous solution access.
Use at least one directed question for each student within the five-minute discussion. Award feedback for defensible reasoning and checks, not agreement with a reference formula. Students acknowledge assistance and earlier worked solutions, distinguishing supplied choices from their own reasoning.
