Surveys and interviews tell you what people say they do. Observation tells you what people actually do. The two are not always the same. A teacher may describe their lessons as highly interactive, but classroom observation might show that most talk comes from the teacher. A manager may say decisions are made collectively, but meeting observations may reveal something different.
Observation is a powerful data collection method in education, health, management, anthropology, sociology and many other fields. It can be used on its own or combined with other methods. This post explains the main types of observation, how to plan and conduct observations, how to record and analyse observational data, and how to handle the ethical and practical challenges involved.
What is observation in research?
Observation is the systematic watching, listening to and recording of behaviour, interactions, events or settings as they occur. The key word is systematic. Research observation is different from casual watching because it is guided by a research question, follows a planned procedure and produces records that can be analysed.
Observation is especially useful when:
- You want to study behaviour in its natural context.
- Participants may not be able to describe their behaviour accurately, because it is habitual or unconscious.
- There may be a gap between what people say and what they do.
- You are studying interactions, routines or processes that unfold over time.
- Participants have limited ability to report verbally, such as young children.
Types of observation
Observation methods can be described along several dimensions. Your choices on each dimension should follow from your research question.
1. Structured vs. unstructured observation
- Structured observation uses a predefined checklist or coding scheme. The observer records specific behaviours, often counting their frequency or duration. For example, recording how many times a teacher asks open versus closed questions. This produces quantitative data and is common in education and health research.
- Unstructured (or semi-structured) observation is more open. The observer records what happens in detailed field notes, focusing broadly on the research topic. This is common in qualitative and ethnographic research.
2. Participant vs. non-participant observation
- Participant observation: the researcher takes part in the activities being studied, such as working alongside staff in a workplace. This provides insider understanding but can make it harder to remain reflective and detached.
- Non-participant observation: the researcher observes without taking part, for example sitting at the back of a classroom or meeting room.
Sociologist Raymond Gold described a continuum of roles: complete participant, participant-as-observer, observer-as-participant and complete observer. Most research roles fall somewhere between the extremes.
3. Overt vs. covert observation
- Overt observation: participants know they are being observed and have agreed to it. This is the standard approach in most research and is required by most ethics committees.
- Covert observation: participants do not know they are being observed. This raises serious ethical concerns and is rarely approved, except in specific circumstances with strong justification.
4. Natural vs. controlled settings
Observation may occur in natural settings, such as classrooms, wards or public spaces, or in controlled environments such as laboratories where tasks are set by the researcher.
Advantages and limitations
Advantages
- Direct evidence of behaviour, rather than reported behaviour.
- Rich understanding of context and interaction.
- Ability to notice things participants take for granted.
- Useful for triangulating interview or survey data.
Limitations
- Observer effect (reactivity): people may behave differently when they know they are being observed, sometimes called the Hawthorne effect. This effect often decreases as participants become used to the observer.
- Observer bias: the observer's expectations may influence what they notice and record.
- Time and resources: observation can be time-consuming, especially over long periods.
- Limited access to private or sensitive settings.
- Behaviour without meaning: observation shows what people do, but not always why. This is why observation is often combined with interviews.
Planning your observation
Step 1: Link observation to your research question
Ask: what exactly do I need to observe to answer my question? Who, what, where, when and for how long?
Step 2: Choose your approach
Decide whether you need structured, unstructured or mixed observation; participant or non-participant; and how many sessions.
Step 3: Develop observation tools
For structured observation, create or adapt an observation schedule with clearly defined behaviour categories. Check whether validated observation instruments exist for your field, as many do in education and healthcare.
For unstructured observation, prepare an observation guide listing key areas of focus, such as physical setting, participants, activities, interactions, conversations, and your own reactions.
Step 4: Pilot your tools
Conduct a trial observation to check whether your categories are clear, whether you can record in real time, and how long sessions should be.
Step 5: Train observers and check reliability
If more than one person observes, train them together and check inter-observer agreement on a sample of sessions, especially for structured observation.
Step 6: Negotiate access and ethics
Obtain permission from gatekeepers and informed consent from participants. Consider people who may be present but are not participants, such as patients or members of the public, and how their privacy will be protected.
Recording observational data
Field notes
Field notes are the main record in qualitative observation. Good practice includes:
- Jotted notes during observation: brief keywords and quotes.
- Expanded notes written as soon as possible afterwards, ideally the same day, filling in detail while memory is fresh.
- Separating description from interpretation: record what you saw and heard separately from what you think it means. For example, “The manager spoke for 12 of the 15 minutes” (description) versus “The manager seemed to dominate the meeting” (interpretation).
- Recording context: date, time, location, people present and seating arrangements. Simple sketches or maps can help.
- Reflexive notes: your feelings, reactions and possible influence on the situation.
Observation schedules
For structured observation, record behaviours using your schedule. Common approaches include:
- Event sampling: recording each time a behaviour occurs.
- Time sampling: recording what is happening at fixed intervals, such as every 30 seconds.
- Duration recording: recording how long a behaviour lasts.
Audio and video recording
Recording allows detailed later analysis and checking by other researchers. However, it can increase reactivity, raises additional ethical and data protection issues, and requires explicit consent.
Analysing observational data
Analysis depends on the type of data:
- Structured data can be analysed statistically, using frequencies, percentages, averages and comparisons between groups or conditions.
- Field notes can be analysed qualitatively using thematic analysis, grounded theory techniques or ethnographic interpretation. Code the notes, look for patterns, compare across sessions and settings, and pay attention to exceptions.
- Mixed approaches combine counts of behaviours with qualitative interpretation of context.
Ethical considerations
- Informed consent from participants, and permission from organisations.
- Privacy of non-participants who may be present.
- Confidentiality in field notes, using codes rather than names.
- Power and role conflicts, for example when observing colleagues or students you teach.
- What to do if you observe harm or misconduct, which should be planned with your ethics committee in advance.
- Secure storage of notes and recordings.
Example: a descriptive field note vs. an interpretive note
Here is a short fictional example from an observation of a hospital ward handover meeting.
Descriptive note: “07:02. Handover begins at the nurses' station. Six nurses present; two from the night shift, four from the day shift. Night-shift nurse A reads from a printed sheet. Day-shift nurses stand; two hold phones. At 07:06, a call bell rings; nurse C leaves and returns at 07:09. Nurse A does not pause. No questions are asked until 07:11, when nurse D asks about bed 4's medication. Handover ends 07:13.”
Interpretive/reflexive note: “Handover felt rushed; information seemed to flow one way. Nurse C missed information on beds 5–7 and did not ask for it to be repeated. Possible pattern of interruptions reducing information transfer? Check in later sessions. I noticed I felt uncomfortable standing close to the group; my presence may have made them shorter in their exchanges.”
Keeping these separate helps you, and later your readers, distinguish evidence from interpretation. Over several sessions, patterns such as interruptions and one-way communication can be compared systematically and discussed with participants in follow-up interviews.
Example: a simple structured observation schedule
For a study of student participation in seminars, a structured schedule might include the following categories, recorded in five-minute blocks:
- Tutor asks an open question.
- Tutor asks a closed question.
- Student answers a tutor question.
- Student asks a question.
- Student-to-student discussion.
- Silence longer than 10 seconds.
Each category needs a clear written definition with examples, so that different observers record the same events in the same way.
Reporting observation in your thesis
In your methodology chapter, describe the type of observation, your role as observer, the settings and number of sessions, the total time observed, the tools used (with the schedule or guide in an appendix), how notes were recorded and expanded, how you addressed reactivity and observer bias, reliability checks if applicable, and ethical procedures. In your findings, use short extracts from field notes, clearly labelled with date and setting codes, to illustrate your analysis.
Common mistakes to avoid
- Observing without a clear focus linked to the research question.
- Mixing interpretation into descriptive field notes.
- Waiting too long to write up expanded notes.
- Using unclear behaviour categories in structured observation.
- Ignoring the observer effect in the discussion.
- Drawing conclusions about motives without other evidence, such as interviews.
Final thoughts
Observation gives you direct access to behaviour and context in ways that surveys and interviews cannot. To use it well, plan carefully, choose the right type of observation for your question, develop and pilot your tools, record systematically, reflect on your own role, and handle ethics seriously. Combined with other methods, observation can add depth, credibility and a unique perspective to your research.
Are you planning to use observation in your research? Share your setting and questions in the comments below.