Observation in early childhood means watching what a child does, recording it factually, and using what you learn to plan what comes next. It sounds obvious. In practice it is the single most powerful professional tool an early years educator has, and the most commonly mangled one, buried under formats, evidence demands and the feeling that if it wasn't written down, it didn't happen.
Observation is not paperwork about children. It is attention paid to them, written down.
This is the guide we wish every new educator got in their first week: the main observation methods, a worked example of each, which method answers which question, and how to keep the whole thing light enough that you still have hands free for the children.
Why we observe at all
Three reasons, in order of importance:
- To plan. You cannot follow a child's interests if you have not noticed them. Observation is where tomorrow's provision comes from: the child who lined up every vehicle by size yesterday is telling you what to put out today.
- To assess development. Not with a stopwatch and a checklist, but by noticing patterns over time: how a child moves, talks, plays, relates. Formal frameworks (EYFS in England, Te Whāriki in New Zealand, the Bildungspläne in Germany) all build their assessment on everyday observation.
- To share with families. A specific, warm observation ("she spent twenty minutes getting the bridge to balance, and refused all help") tells a parent more than any score ever will, and it builds the partnership everything else depends on.
There is a fourth reason nobody puts in the handbook: observation is what keeps the job interesting. The educators who still love this work after twenty years are, almost without exception, the ones who never stopped being curious about what children are actually doing.
Objective vs. subjective: the one skill under all methods
Every method below stands or falls on the same skill: separating what you saw from what you think it means.
| Subjective (avoid in the record) | Objective (write this) |
|---|---|
| "Milo was aggressive today." | "Milo pushed Ben twice at the water tray, both times when Ben reached for the jug Milo was holding." |
| "Lena is a creative child." | "Lena used the pine cones as 'cakes', arranged them on plates and assigned prices: 'this one is four money.'" |
| "He can't concentrate." | "During the 20-minute circle time, Theo left the carpet four times, each time after about three minutes of sitting." |
The objective versions are not just fairer. They are more useful: Milo's record contains the trigger (possession of the jug), and Theo's contains a measurable baseline (three minutes), which is what you need to plan support and to talk to parents without putting them on the defensive.
Interpretation is not forbidden; it is a separate step. Write the facts, then add your reading, clearly marked: "Possible meaning: Milo is defending resources rather than seeking conflict. Try duplicate jugs."
The seven methods, each with an example
1. Anecdotal records
A short narrative of one meaningful moment, written after the fact. The workhorse of early years documentation.
Example: "14/11, garden, 9:40. Amara (3y4m) found a worm under the planter. She crouched, watched it for a while, then touched it with one finger and pulled her hand back laughing. She called two other children: 'Come! He's dancing!' Stayed at the planter about ten minutes."
Best for: capturing interests and milestones as they happen. Weakness: you choose what to notice, so quieter children can slip through. Fix that with a simple rota: two focus children per educator per day.
2. Running records
You sit down for five to ten minutes and write continuously everything one child does, in sequence, as it happens. The closest thing to a transcript of a child's play.
Example (excerpt): "10:02 L. takes red crayon, makes circular scribbles, left hand. 10:03 swaps to right hand, looks at K.'s paper. 10:04 says 'I'm doing a storm too.' Presses harder, paper tears, frowns, flattens the paper with both hands, continues on the untorn half."
Best for: answering a specific question about one child (pencil grip, language in free play, how a conflict actually starts). Expensive in attention, so plan it: another adult covers the room while you record.
3. Time sampling
You look at set intervals (every ten minutes, say) and note what a child is doing and with whom. The record is a grid, not a story.
Best for: patterns across a session. Does Jonas really "always" play alone, or only before lunch? Time samples turn impressions into data, and they regularly disprove the staff-room consensus about a child.
4. Event sampling
You define one behaviour (biting, crying at handover, joining group play) and record every occurrence with its context: what happened just before, who was involved, what happened after. The classic ABC structure: antecedent, behaviour, consequence.
Best for: behaviour you want to understand rather than just count. Three well-recorded events usually reveal the trigger that twenty complaints never did.
5. Checklists
A predefined list of skills or milestones, ticked when observed. Fast, comparable, and the method most prone to abuse: a checklist tells you whether, never how or why.
Best for: screening and statutory snapshots. Use them to notice gaps worth observing properly, not as the documentation itself.
6. Learning stories
A narrative written to the child, interpreting a moment through learning dispositions (courage, persistence, responsibility), with a photo and a next step, and space for the family's voice. Developed by Margaret Carr alongside New Zealand's Te Whāriki curriculum.
Best for: documentation families treasure and children revisit. They take longer to write, which is why they should be occasional and heartfelt rather than weekly and hollow. We cover them in depth, with seven full examples, in our learning stories guide.
7. Photo and video documentation
A picture of the block tower is not an observation. A picture of the block tower with the sentence "third attempt; this time she widened the base first" is one of the best observations you can make.
Best for: processes that words describe badly (movement, construction, art). Two rules: the caption carries the meaning, and consent and storage follow your data protection policy, no exceptions, no educator phones.
Which method for which question
| Your question | Reach for |
|---|---|
| "What is this child into right now?" | Anecdotal records |
| "How exactly does she approach a task?" | Running record |
| "Is my impression of this child actually true?" | Time sampling |
| "What triggers this behaviour?" | Event sampling |
| "Where are the gaps I should look at?" | Checklist, then observe properly |
| "What do I want this family to keep forever?" | Learning story |
| "How did this process unfold?" | Photo/video with a written caption |
Observing without missing the moment
The dirty secret of observation is that the recording competes with the relating. Every minute spent typing is a minute with your eyes off the room. Practical routines that keep the balance:
- Notice now, write later, but same day. A keyword on a sticky note or a ten-second voice memo preserves the moment; the full write-up can wait until nap time. Overnight, details die.
- Two focus children per day, rotating, so coverage happens by design instead of by personality.
- One good sentence beats three mediocre paragraphs. If you only have ten seconds, spend them on the child's exact words.
- Let the tool do the clerical half. This is exactly the job we built hoopooh for: you speak the observation while it happens, and it writes itself up and tags the right developmental areas, so the record costs seconds instead of an evening. Whatever tool you use, the test is the same: does it give you time back with the children, or take more?
From observation to planning: closing the cycle
The observe, assess, plan cycle is three honest questions:
- Observe: what did the child do? (The record.)
- Assess: what does it tell me about their learning? (One sentence.)
- Plan: what will I change or offer next? (One concrete thing: a material, a question, a grouping.)
The cycle fails at step three more than anywhere else. A simple team habit fixes it: at the weekly planning meeting, every educator brings one observation and the one thing they will change because of it. Twenty minutes, and your planning is suddenly built from real children instead of last year's topic web.
Ethics and consent
Children cannot consent to being documented, which puts the responsibility on us:
- Dignity first. No records of children in distress or undressed, ever, and ask yourself of every photo: would this child, at sixteen, be fine with it existing?
- Written media consent at enrolment, granular (internal documentation vs. sharing with the family vs. anything public), with opting out always possible.
- GDPR applies. Records are personal data: access-controlled, stored on compliant systems rather than private phones, deleted per your retention policy, and families can see what you hold about their child.
- Observation is for the child. The point of every record is a better experience for this child, not protection paperwork for the setting. When a format stops serving the child, change the format.
Look closer. Write less. Plan from what you actually see. That is the entire method, and it has been since long before anyone invented the paperwork.