Gathering Data
Seven techniques for gathering data
| Method | Best used for |
|---|---|
| Reflective journal | Capturing the teacher’s thinking over time |
| Portfolio | Collecting student work samples that show change |
| Observations | Watching what actually happens in class |
| Interviews | Getting the depth behind a behaviour |
| Questionnaires | Reaching many students or parents quickly |
| Case studies | Going deep on one student or small group |
| Checklists | Tracking specific behaviours over many sessions |
The DO and DON’T of analysis
DO
- Design a system before you start
- Look for themes and patterns
- Share findings with a colleague
DON’T
- Let your prior assumptions guide the reading
- Censor data you have already collected
- Look at the data only once
A process for analysing qualitative data
- Write continuously about what you see, what questions emerge, what you are learning
- Look for themes, patterns, and big ideas
- Identify the points that appear most often or most powerfully
- Draw the information together
- Include support (quotes, examples, numbers) for each theme
Action research is
Process, collaboration, improvement, practice, communication.
A teacher running an action research project will at some point face the same question: what data do I actually collect, and how do I keep my own biases out of the analysis? The answer has two parts: pick the right techniques, and follow a discipline when reading what you collected.
Choosing data-gathering techniques
Action researchers have many options for gathering data. The list below is not exhaustive, but it covers most classroom projects.
Reflective journal
A teacher’s own running notes, kept consistently over the project. This is the cheapest method and often the most useful. The entries do not have to be long. Three sentences a day for two weeks beats a single 2,000 word essay at the end. The journal captures the teacher’s thinking, the small surprises, and the moments that get smoothed over by memory.
Portfolio
Samples of student work collected at different points in the project. Useful when the question concerns a skill that develops over time: writing, problem-solving, drawing. The portfolio gives you evidence the student produced, not your impression of the student.
Observations
Either you observe a colleague (or have a colleague observe you), or you observe yourself through a video recording. A short video of one lesson, even taken on a phone, is one of the most powerful tools available. Watching yourself teach reveals things memory hides.
Interviews
A conversation with one student or a small group. Useful when you want depth rather than breadth. The student’s account of why they answered a question a certain way is often more revealing than any test result.
Questionnaires
Useful when you need to reach many students, parents, or colleagues quickly. The trade-off is shallow data. A 10-question survey to 200 parents tells you something a 90-minute interview with one parent cannot, and the other way around.
Case studies
A focused study of one student, one classroom, or one event. Useful when the question is “what is really going on with this group?” rather than “what is going on across all groups?”
Checklists
A simple yes-no or count record kept over many sessions. Good for tracking specific behaviours: how often a quiet student speaks up, how many times a teaching technique is used, how many students completed a task.
The right method is the one that most directly addresses the question. A teacher who picks methods because they sound impressive ends up with data they cannot use.
Guidelines for analysing data
Three things to do, three to avoid.
DO design a system before you start
Decide before collection how you will sort and read the data. A teacher who decides on the system at the end usually picks one that confirms what they already believed. Writing the analysis plan in advance is a discipline.
DO look for themes and patterns
Read across the data, not down one entry at a time. The patterns are usually visible only at the level of the whole set. A single student journal entry is anecdote. Twenty entries with similar phrasing is data.
DO share findings with a colleague
A colleague who reads your data will ask questions you skipped. They will see patterns you missed and disagree with conclusions you drew. This is the peer examination feature of action research showing up in practice.
DON’T let your prior assumptions guide the analysis
This is the biggest threat to validity in classroom research. The teacher believes the new method works. The teacher reads the data looking for evidence that it works. The teacher finds evidence. The data did not test the belief; it confirmed it.
The fix is to look explicitly for counter-evidence. Which students did not improve? Which lessons were less effective with the new method? A teacher who can answer these questions has actually analysed the data. A teacher who cannot has only confirmed.
DON’T censor data you have already recorded
If a journal entry or a quiz result does not fit the story you are starting to tell, do not drop it. Anomalies often point at the most important finding. A teacher who throws out the inconvenient cases ends up with a tidy story and a wrong conclusion.
DON’T look through your data only once
Read the whole set, then read it again with fresh eyes a few days later. Insights show up on second and third readings that the first reading missed. Time is part of the analysis.
A process for analysing qualitative data
For data like journal entries, interview transcripts, or open-ended survey responses, a five-step process works well.
- Write continuously while you read. Note what you are seeing, what questions emerge, and what you are learning. The writing is part of the analysis, not a separate step at the end.
- Look for themes, patterns, and big ideas. What words, situations, or claims keep coming up? What seems to connect across cases?
- Identify the main points that appear most frequently and most powerfully. Frequency matters. So does intensity. A claim that appears in only one entry but is stated with strong evidence may be more important than a vague point that appears 20 times.
- Draw information together. Group the data points that belong to each theme. Look for tensions between themes.
- Include support for each theme. Quotes, numbers, examples. A theme without supporting data is not a finding. It is a hunch.
Confirmation bias: reading data looking for evidence of what you already believe
The teacher believes the new method works, looks for evidence that it works, and finds it. The fix is to look explicitly for counter-evidence: which students did not improve, which lessons did not work. A teacher who cannot identify counter-evidence has not yet analysed the data.
What action research is, in five words
The closing summary that often appears in the literature names five qualities: process, collaboration, improvement, practice, communication.
Action research is a process, not an event. It is collaborative, drawing in students, peers, and sometimes parents. It aims at improvement, not only at description. It stays close to practice, the daily work of the classroom. And it produces communication: the writing-up that lets others learn from what you tried.
A project that has all five tends to produce real change. A project missing one of them tends to drift.