Driscoll’s reflective model organises reflection around three questions: What? So what? Now what? Use them to examine one experience, explain what you learned and decide what to test next. The worked example below shows how to move beyond a description without pretending that every explanation is certain.
The University of Edinburgh Reflection Toolkit traces Driscoll’s 1994 approach to Terry Borton’s earlier questions. It also notes related adaptations, including Rolfe and colleagues’ work. The shared questions do not make the publications interchangeable. For source details rather than a worked reflection, use our Driscoll referencing guide.
Start with one incident, not your whole semester
Choose a moment you can examine closely: a decision, a misunderstanding or feedback that changed your interpretation. For the example here, the incident is a discrepancy discovered during a student laboratory presentation. The aim is to understand a checking decision, not to recount the entire project.
Original hypothetical example: the student, project, figures and feedback below are invented for teaching. This is not a real student’s account, a quotation from Driscoll or evidence that a particular intervention works. Do not submit it as your own experience.
What? Record what happened and your part in it
In our laboratory presentation, I showed a graph made from the group’s shared spreadsheet. A teammate noticed that one trial was missing. I had assumed that the spreadsheet was ready because the filename contained “final”; I had not compared it with the trial log. We paused to identify the missing row. I felt embarrassed because I had volunteered to check the graph before presenting it.
Why this paragraph is useful: it identifies a specific event and the writer’s responsibility. “One trial was missing” describes an observation. “The spreadsheet was ready” was an assumption. The paragraph does not claim that a teammate was careless or that the missing trial necessarily changed the conclusion.
So what? Explain the decision, then challenge your explanation
The filename was not evidence that every trial had been included. My check focused on labels and appearance, so it could not detect a missing row. However, I cannot yet tell whether the omission happened during data entry or when a range was copied. I need to compare the spreadsheet with the trial log before assigning a cause. I also need to recalculate the result with the omitted trial; until then, I cannot say whether the reported conclusion changes. My main learning is narrower than “we need better teamwork”: I need an explicit completeness check before approving a graph.
What makes this analysis rather than a retelling? The writer examines why the checking method failed, considers more than one explanation and states what evidence is missing. The paragraph also avoids claiming that the presentation’s interruption proves the scientific result was wrong.
If your assignment requires academic literature, connect a relevant source you have actually read to a specific part of this analysis. A citation to a reflection model explains the structure; it does not, by itself, substantiate a claim about data quality, teamwork or learning outcomes.
Now what? Specify an action and a way to review it
Before the next presentation, I will compare each plotted dataset with the trial log and record the number of trials included. I will ask a teammate to check that comparison before we label the file ready. After the rehearsal, I will note any discrepancies this check found or missed and how long it took. If a discrepancy still reaches the slides, we will examine which checking step failed and revise it. This tests a new process; it does not guarantee that future work will be error-free.
Why this is stronger than “I will check more carefully”: the plan names the task, the evidence to check, a second reviewer and a review point. Keep the proposed action proportionate to your role and the rules of your course or placement.
Three small revisions that add critical depth
- Instead of: “My group did not communicate.” Try: “We had not agreed who would compare the graph with the trial log.” The second statement identifies a process you can investigate.
- Instead of: “The missing trial ruined the result.” Try: “I need to recalculate before judging the omission’s effect.” This keeps a possible consequence separate from a verified one.
- Instead of: “Next time will go better.” Try: “At the next rehearsal, I will record what the completeness check detects and misses.” This makes follow-up possible without inventing an outcome.
A short planning sheet for your own incident
These prompts are an original planning aid for using the three-question structure, not a prescribed assessment rubric.
- What? Name one event. Record your action, the observable outcome and any assumption you made.
- So what? Examine that assumption. Identify an alternative explanation, the evidence you need and a relevant source if the task requires one.
- Now what? Choose one change within your control. Set a review point and decide what observations would make you keep or revise it.
Use your own experience, anonymise other people and follow your institution’s instructions. There is no fixed word allocation in this example: the required format and balance depend on your assessment brief.
Limits of the model and sources to check
Three headings alone do not make a reflection critical. Edinburgh’s guidance cautions that the structure can remain superficial without deeper examination. Our example therefore makes uncertainty and a testable next step visible; it is one possible application, not the only correct response.
For the original publication’s bibliographic details, PubMed records Driscoll’s “Reflective practice for practise” as Senior Nurse, 1994, 14(1), pages 47–50. That record has no abstract and is not the article’s full text. The model explanation here uses Edinburgh’s guide; do not imply you consulted the original article unless you did.
For a comparison of approaches, see our reflective models overview. For Harvard and APA source examples, return to the separate Driscoll citation guide.