The method–context requirement

An observation is a research method in which behaviour is watched and recorded systematically. In this practical, observation must be used in the context of human development: patterns of continuity and change in behaviour across the lifespan. The developmental behaviour under investigation must be visible or recordable in some other way. An observer cannot directly see what a participant thinks or feels.

The minimum sample is one participant. That is a minimum, not a recommended maximum, so the design can include more participants. Repeatedly observing the same person does not increase the sample size; the sample still consists of one participant.

A suitable write-up clearly explains:

• the psychological question and the developmental behaviour being investigated
• how the behaviour was made specific enough to recognize and record consistently
• the observation type and why it was appropriate
• how the participant or participants were sampled
• the setting, procedure and method of recording behaviour
• how ethical considerations were addressed
• what patterns were found
• how bias, external variables, validity, reliability and the sample affect the conclusions

Classifying an observation

Observation designs are classified along three separate dimensions. For example, one study could be naturalistic, overt and non-participant. These labels describe different choices rather than competing names for a single choice.

Naturalistic or controlled

A naturalistic observation is an observation carried out in the participant's usual environment without the researcher arranging the setting. Because the behaviour is more likely to reflect everyday life, this can support external validity. There is less control, though. Unexpected events or differences in the surroundings may influence behaviour.

A controlled observation is an observation carried out in a setting or situation arranged by the researcher, allowing conditions to remain relatively consistent. Recording tends to be more orderly, and researchers can compare participants or occasions more easily. However, behaviour in the arranged situation may be less natural.

Overt or covert

An overt observation is an observation in which participants know they are being observed for research. It supports informed participation and makes withdrawal easier to respect. At the same time, knowing that they are being watched may change participants' behaviour.

A covert observation is an observation in which participants do not know at the time that they are being observed for research. This may reduce behavioural changes caused by awareness of the observation, but it raises serious questions about consent, privacy, deception and the possibility of withdrawal. A public setting does not automatically remove the researcher's ethical responsibility.

Participant or non-participant

A participant observation is an observation in which the researcher joins the activity or group being studied while recording behaviour. Taking part may give the researcher close access to the context and help them understand what actions mean to those involved. It can also influence events, make consistent recording harder and increase the effect of the researcher's perspective.

A non-participant observation is an observation in which the researcher records behaviour without joining the activity under study. This allows the observer to focus on recording and may reduce direct interference. Even so, a visible observer can still affect behaviour.

Selecting observation rather than another method

The psychological question should guide the choice of method. Observation is particularly suitable when the aim is to investigate what people do, rather than what they say they do. It records behaviour in context, but on its own it cannot reveal a participant's private reasoning or establish why the behaviour happened.

The main alternatives produce different kinds of evidence:

An observation can include careful comparisons without automatically becoming an experiment. The researcher's actual actions identify the method, not the vocabulary chosen for the report.

Recording behaviour and measurement

Before recording starts, the behaviour must be described in observable terms. A broad label such as “positive interaction” allows too much interpretation. Instead, the procedure should state which visible actions count and how to decide where one action ends and the next begins. This improves consistency while keeping the measure connected to the intended developmental characteristic.

Measurement is the process of representing a psychological characteristic through systematically recorded evidence. During an observation, this may involve recording whether a behaviour occurs, how often it occurs, how long it lasts or which predefined category describes it best. The question should determine the recording choice—not convenience alone.

Predetermined categories can improve consistency, although they may leave out important behaviour. Open descriptive notes retain context and unexpected detail, but they also give the observer greater freedom to select and interpret events. A well-justified procedure needs to balance these concerns.

Bias and reflexivity

Bias is a systematic influence that pushes evidence or conclusions away from an accurate representation of what is being studied. A single isolated mistake is not bias.

Researcher bias is bias produced when a researcher's expectations, decisions or social position systematically influence data collection, interpretation or reporting. During an observation, this could affect which actions the researcher notices, how an ambiguous action is categorized or which incidents receive emphasis in the conclusion.

Participant bias is bias produced when participants systematically alter their behaviour or responses because of their expectations about the research or concern about how they will be judged. It is especially relevant to overt observation, though covert designs bring greater ethical difficulties.

Reflexivity is a research process in which researchers examine how their own assumptions, identity, relationships and decisions may have influenced the research. A reflexive account identifies plausible sources of influence and explains how they were checked. Simply claiming objectivity isn't enough.

Researchers can check unconscious bias by agreeing on precise behavioural descriptions beforehand and keeping a record of coding decisions. Another observer might examine the same behaviour, while interpretations that lack support in the records should be reconsidered. These checks cannot make a researcher perspective-free, but they can make its influence easier to spot.

External variables and causality

An external variable is a factor outside the relationship of primary interest that may influence the observed behaviour or its measurement. In a developmental observation, possible alternative explanations include differences in setting, time, nearby people, familiarity with the activity or the observer's presence.

Causality is a relationship in which a change in one factor produces a change in another. Observation can identify behaviour, context and association, but it usually cannot isolate a cause. Two participants might behave differently in a way that makes age or developmental stage seem relevant. Other differences, however, may explain the result. A controlled observation can reduce some alternatives, but control alone does not support a causal claim if factors weren't manipulated and competing explanations remain.

Ethical considerations and responsibility

Ethical considerations are principles used to protect the dignity, rights and welfare of people affected by research. They shape which observational designs are acceptable, what researchers may record and how they may store or report the evidence.

In an overt observation, participants should receive understandable information about taking part and be free to choose whether to participate. They should know how to withdraw and must be protected from unnecessary distress. Records must not reveal identities without appropriate permission. Extra care is needed when a participant may have limited capacity to give informed agreement. Permission from someone else does not remove the researcher's duty to respect that participant's welfare and willingness to take part.

Covert observation needs particularly strong justification because advance consent may be absent and withdrawal may be difficult. Researchers must ask whether people could reasonably expect privacy and whether the question could instead be answered overtly. They should also consider whether the observation could expose sensitive behaviour, and whether a later explanation is possible without causing further harm.

Responsibility is the obligation to consider and manage the consequences of psychological research for participants, researchers and wider groups. An observation isn't justified simply because it is methodologically interesting; its value must outweigh the likely harm or intrusion.

Validity and reliability

Validity is the extent to which evidence and conclusions accurately represent what the research claims to investigate. Its different forms raise different questions about an observation:

• Internal validity is the extent to which a study supports its conclusion about what produced the observed result, rather than an explanation based on uncontrolled influences. External variables and participant awareness can weaken it.
• External validity is the extent to which findings apply beyond the particular participants, setting and time studied. A naturalistic setting may help, but a sample of one sharply limits claims about a population.
• Content validity is the extent to which a measure adequately covers the range of behaviour making up the characteristic being investigated. Categories that are too narrow may miss important aspects of development.
• Face validity is the extent to which a measure appears, on inspection, to assess what it claims to assess. It provides a useful initial check, though appearance alone is weak evidence.
• Construct validity is the extent to which a measure genuinely represents the theoretical psychological characteristic it is intended to represent. A visible action should not automatically count as a complete measure of a complex developmental characteristic.

Reliability is the extent to which a research procedure produces consistent results when applied in a comparable way. An observational record becomes more reliable when behavioural categories are clear enough for different observers to apply similarly and for researchers to repeat the same procedure. Agreement between observers supports reliability, but does not prove validity. Two observers may consistently apply a poor category.

Sampling and the minimum sample

A sample is the participant or group of participants from whom research evidence is collected. One participant meets the minimum for this practical, but the selection of that person still matters.

Sampling decisions are connected to ethics and access. A theoretically strong technique may be impossible when recruiting or observing the relevant population would be intrusive. Convenience alone does not justify including someone.

Generalizability, transferability and credibility

Generalizability is the extent to which findings from a sample can be applied to a wider population or to other settings and times. A judgment about it should take account of sample size, sampling technique and the similarity between the sample and target population. Variation within the population also matters, as do the realism of the setting and the stability of the observed pattern across occasions. With one participant, conclusions should normally stay closely tied to that case.

Transferability is the extent to which findings developed in one context may be meaningfully applied to another context after their similarities and differences have been considered. The researcher does not claim that a finding applies everywhere. Instead, they provide enough contextual detail for others to judge whether it might fit another population or setting.

Credibility is the extent to which a research account is well supported by the evidence and presents a convincing representation of the participants and context studied. Clear records and transparent category decisions can strengthen it. So can comparing interpretations, examining bias reflexively and describing the observation context in detail.

Generalization and transfer involve different judgments. Generalization asks how far findings apply across a wider population or range of situations. Transferability asks whether the insight fits a particular new context. Careful observation alone guarantees neither.