An evaluation of the observational methodology within psychological research.

by Research Methods & Statistics

Researchers often prefer to directly observe an individual’s behaviour rather than asking them questions about it. A preferred method allowing the researcher to do so is the naturalistic observation. This refers to scientists observing individuals in their common, natural surroundings and engaging in their everyday behaviour (Pellegrini, 1996). The researchers do not manipulate variables or interfere with things; instead they try to remain inconspicuous. The situation is not manipulated or controlled by the investigator and the situation has not been created or initiated by the investigator and therefore this research method is also referred to as an observation without intervention.

Specific to developmental psychology, this observation involves visiting a child’s home, school, nursery, public-park or playgrounds and carefully recording a child’s behaviour. It is uncommon for a researcher to intrinsically record every event that occurs instead, they test a specific hypothesis concerning one type of behaviour. An example would be aggression and researchers would focus their attention to and collect data on acts of aggression such as hitting and kicking.

Researchers may employ several differential techniques to collect data from naturalistic observations. This may involve writing down the frequencies of a certain behavior occurred in a specific amount of time, or filming a video recording of the subjects of interest. The three main methods utilized are as follow:

  • Tally chart – the observer notes down when and how many times a certain behavior occurs.
  • Observer narrative – the observer might write down notes during the observation and then refer back to these later to collect data on behavioral patterns.
  • Audio/video recordings – depending on the type of behavior being recorded the observer may record footage of the observation session.

As it is not feasible to observe every moment or aspect of a subject’s life, observational researchers often use sampling in order to gather information through naturalistic observations. The ultimate aim is to ensure the sample of data is representative of the subject’s overall behaviour. There are two main techniques in order to obtain a representative sample:

  • Time sampling – this consists of taking samples at different intervals of time. This can either be systematic or random.
  • Situational sampling – this involves observing an individual’s behaviour in a variety of settings.

One example of a naturalistic observation comes from Ginsburg and Miller (1982). They investigated sex differences in children’s risk taking behaviour due to the common belief males are more risk taking than females. Researchers wished to study this behavior in the real world rather than the laboratory and therefore observed children in the zoo. They operationally defined risk-taking and measured it in four ways:

  1. riding an elephant
  2. petting a donkey
  3. feeding animals
  4. climbing a steep wooden bridge

Researchers recorded the number of boys and girls engaging in behaviors classed as risky. They found that boys engaged in risk-taking behaviours more frequently than girls.

The main advantage of naturalistic observations is that they illustrate how an individual actually behaves in their natural environment as part of everyday life (Willems & Alexander, 1982). In other words, it allows observation of behaviour as it exactly occurs in the real world. Furthermore, this observation retains an element of ecological validity as the situation being observed exists within the natural ecology of the individual. Naturalistic observations also help to establish external validity of the findings. As the behaviour observed is real life it is easier for the researchers to generalise their findings to the general population. Additionally, they provide rich information compared to other research methods such as laboratory experiments as an individual can be observed in various social settings rather than in the artificial lab one.

Penultimately and more specific to developmental psychology, naturalistic observations can be applied to special populations such as young infants and toddlers who have not yet developed the verbal repertoire needed for other research methods including interviews and questionnaires. Finally, ethical guidelines may prohibit the manipulation of specific variables however; it may be possible to observe such conditions as they occur naturally such as reactions to traumatic stress. It would be unethical to cause such events to occur for the purposes of experimentation however, you can observe the results from events when they naturally occur.

However, there are many objections against this research method. Firstly, some behaviour types occur so rarely (such as heroic rescues) or are so socially undesirable (theft) that the observer is unlikely to witness this behaviour in the time they are conducting their observation. Secondly, in a natural setting several events usually happen at the same time and any of these events can affect how an individual behaves. This weakens the internal validity as there is concern in precisely pinpointing the causes of participant’s behaviour or of any developmental trends. A problem with the naturalistic observation is that there is no experimental control over the circumstances therefore whilst correlation cannot be shown; cause and effect cannot be established. The credit assignment problem suggests that just because variable X correlates with variable Y this does not infer that X is the cause of Y. Gentile and Anderson (2003) suggest there may be a third unmeasured variable that is the cause of the behaviour observed. There is also a problem with directionality. It is difficult to conclude which variable causes which behaviour. Instead of X causing Y, it may be that Y in fact is the cause of X.

Scientists are reluctant to rely on naturalistic observations as the presence of an observer can result in individuals behaving differently than they normally would, this refers to the issue of reactivity. For example, children may show off in front of an audience and parents may act differently with their children. Reactivity is extremely likely to occur when the participants expectations of what the experiment is about can change their responses to the demands of the situation. An example is The Hawthorne Effect (Homans, 1965). Women employees of a factory participated in a study of worker productivity and in specific, which working conditions resulted in highest levels of productivity. Productivity was measured through the number of telephone relays that were assembled each day. Researchers collected a baseline measure of productivity before changing the working conditions and measuring the change in performance. Each time the working conditions were changed, performance increased. It was proposed that such results were found as the workers knew that they were being studied and that they were aware investigators were searching for methods to increase productivity. The workers tried hard to do their best so they could give the researchers the results they were looking for.

Orne (1962) identified a type of participant reactivity known as demand characteristics. These extraneous variables occur when people do their best to comply with what they perceive to be the demands of the observation and they try to guess what is expected from them. In contrast, some participants do the opposite of what they think is expected. Orne also identified another participant reactivity known as evaluation apprehension where by the concern experienced on behalf of the participants that they are being judged can alter their behaviour.

Weber & Cook (1972) identified four social roles of participants in observations that threaten the validity.

  1. Good Subject Role – the subject tries to validate the anticipated result (e.g. give the experiment the result they are hoping for).
  2. Faithful Subject role – participant attempts to be completely honest and faithful despite anticipating the expected result.
  3. Negativistic Subject role – the subject purposely produces responses or behaviour that is in the opposite direction of the anticipated result.
  4. The Apprehensive Subject role – participant experiences feelings of uncomfort as they were wary of being evaluated by the experimenter. As a result of this evaluation apprehension, the subject makes a conscious effort to behave in a socially desirable way in avoidance of appearing in a bad or unfavourable light.

These social roles highlight the threats to internal validity as these demand characteristics prevent observational researchers measuring what they intended to measure. Furthermore, it has been proposed that human-beings have a predisposition towards self-enhancement and maintaining positive about themselves at the cost of being unrealistic (Fiske & Taylor, 1999).

To overcome the huge flaw of such reactivity researchers have developed strategies to ensure they maintain the validity. One attempt to minimize observer influences is by videotaping participants from a concealed location such as a one way mirror as seen in Ainsworth’s Strange Situation (Ainsworth, Blehar, Waters & Wall, 1978).

Secondly, observers may spend time in the setting before conducting their observation and collecting real data so that the participants become used to the observers presence and this does not affect their behaviour.

Observational researchers may also use the single blind procedure to reduce demand characteristics. Participants are given a false account of the observational aim so that those being observed do not discover clues to the nature of the experiment. However, this can be problematic as full informed consent cannot be obtained. For example, Sherif (1935) explained to participants that they were taking part in an auto kinetic test however; they were actually participating in a social influence experiment therefore participants did not consent to the experiment. It is important to remember that if participants were aware of the true hypothesis the results would not have been valid therefore researchers deemed it appropriate to involve deception.

Expectancy on behalf of the observer may also inadvertently affect the results gained from observations. The main issue is observer expectancy where by the observers expectations alter the behaviour of the participants as demonstrated by Rosenthal (1966). A proposal to this problem is a double blind procedure where by neither the observer or subjects know what hypothesis is being tested however, this often impractical within observations. Another issue on behalf of the observer is that different researchers may draw different conclusions from the same witnessed behaviours therefore it is important that the behaviours are operationalised (clearly categorised so researchers are aware of behaviour types).

As discussed earlier, researchers may record footage of participants being observed. However, some ethical issues may arise from doing so and researchers need to ensure the confidentiality of the research participants and ensure that such data can be destroyed if the participant wishes to withdraw from the experiment.

With regards to the limitations of natural experiments, researchers may conduct structured observations in order to study unusual or undesirable behaviours that are unlikely to occur in the natural environment. A structured observation takes place in a laboratory (an artificial setting) and participants are presented with situations and exposed to settings that may trigger the behaviour of investigation and is then observed via a hidden camera or through a one-way mirror to see if the individual performs the behaviour. For example, Kuczynski (1983) got infants to promise to help him in a boring task and then left the children alone to work in a room where desirable toys were present and available. This particular research procedure enabled researchers to determine whether or not the children would break their promise.

An advantage of the structured observation is that they ensure each and every participant in the experiment is exposed to the same eliciting stimuli therefore each participant has an equal opportunity to perform the behaviour and this cannot always be done in a natural observation. However, a major drawback of this type of observation is that participants may not always behave the same way in a laboratory as they would in an everyday life setting therefore, this threatens the internal validity. Furthermore, due to the artificial and controlled environment the results are difficult to generalise to the rest of the population.

In conclusion, whilst observational studies provide a rich and detailed account of the behaviours individuals engage in (rather than say what they behave in) they may be biased by both the participants and researchers expectations which prove to be a threat to the internal validity of this research method. Furthermore, due to a lack of control in naturalistic observations it is difficult to establish cause and effect.


  1. Pellegrini, A. D. (1996). Observing children in their natural worlds: A methodological primer. Mahwah, NJ: Erlbaum.
  2. Sex differences in children’s risk-taking behavior. Ginsburg, H. J., & Miller, S. M. (1982). Child Development, 53, 426-428
  3. Willems, E. P. & Alexander, J. L. (1982). The naturalistic perspective in research. In B.B. Wolman (Ed.), Handbook of developmental psychology. Englewood Cliffs, NJ: Prentice Hall.
  4. Gentile, D.A. and Anderson, C.A. (2003). Violent video games: the newest media violence hazard. In D. A. Gentile (Ed.) Media violence and children.
  6. Orne, M.T. (1962). On the social psychology of the psychological experiment: With particular
  7. reference to demand characteristics and their implications. American Psychologist, 17, 776-783.
  9. Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York: McGraw Hill.
  12. Rosenthal R. Experimenter effects in behavioral research. New York, NY: Appleton-Century-Crofts, 1966. 464 p.
  13. Kuczynski, L. (1983). Reasoning, prohibitions, and motivations for compliance. Developmental Psychology. 19, 126-134.