Article summary with Methodological Issues in the Direct Observation of Parent-Child Interaction: Do Observational Findings Reflect the Natural Behavior of Participants? by Gardner - 2000

What is this article about?

The aim of this review is to examine the utility and validity of direct observation as a research technique in family interaction and behavior in family setting.

Direct observational research is used to gather information about typical behavior of the participants. However, participants are often required to engage in unusual settings, such as a laboratory, or to engage in certain behaviors on the researcher's demand. Furthermore, knowing that the behavior will be observed can influence the participants' behavior. Therefore, the validity and utility of direct observational techniques is often questioned.

Where does direct observational technique comes from?

To assess children’s social behavior, systematic observational techniques were first developed in the 1930s within the school setting. Hereby, psychometric issues were taken into consideration as well. Specialized on the nature of parent-child interaction, a coding system for observational technique was developed in the 1970s. Many of the systems that were developed during this time are still used.

What are the advantages and disadvantages of observational methods?

One advantage of using direct observations of parent-child interaction is they can be carried out at the participants' homes. It is unknown whether or not an unusual setting, such as laboratory circumstances, would impact parent-child interaction. Observations are limited to private settings, as public spaces can be an additional stress factor for parents.

Direct observations provide information about real behavior, such as shouts or hugs. The definition of such behavior of interest is devised by the researcher rather than the parent in order to avoid personal bias. Direct observations can also be useful in planning and evaluation interventions. Some behavior might only be observed through direct observation, rather than through self-report, as the behavior might be automatic and fast. Furthermore, observable behavior is objective, whereas self-reports reflect the perception of the participants.

On the other hand, the time-consuming nature of direct observations reflects the technique’s major drawback. This includes training observers, carrying out the observation itself, coding the interaction and ensuring inter-observer reliability. Thus, only a limited number of observational session can be carried out, which sometimes lead to unstable datasets.

How can direct observation become more valid?

Validity of direct observational data can be ensured through an observational coding system. However, the validity of those coding systems need to be ensured as well, especially for studies in which the question arises whether or not self-report would be an easier and cheaper method of research. In some studies inter-observer reliability is difficult to conduct since the participants’ behavior differ from day to day. Additionally, test-retest reliability is not always reported in studies.

Stoolmiller et al. (2000) developed a method called “censoring” to correct statistical data. They claimed that addressing reliability issues is important to avoid underestimation of data and intervention effects.

Does direct observation reflect natural behavior?

While assessing validity of direct observational behavior it is important to consider the conditions under which the data has been collected, whether or not the conditions were entirely natural. The behavior of interest has to reflect the behavior that usually occurs in problem settings. Moreover, the parental behavior during direct observations has to be similar to how they usually act.

Which factors might influence the behavior of participant’s?

The observer’s presence

The observer’s presence always influences the behavior to a certain extend. However, approaches have been developed to limit this effect. Firstly, different levels of the observer’s intrusiveness can be compared. Secondly, it is assumed that the participant’s reactivity to the observer declines after a while and when they get used to it.

Recent studies have been conducted with video recording, rather than with the traditional pen-and-paper approach. Little is yet known about the effect of video observing on the participant’s behavior.

Informing the participants about possible reactivity to being observed has not shown to be effective yet. However, reactivity can be reduced by not observing behavior the first 10 minutes. Furthermore, always being observed by the same observer can help participants to behave more natural.

There are different opinions on whether or not the observer should communicate with the participants during the observed situation. More research in different strategies to reduce reactivity is needed.

The task imposed by the observer

Parent-child interaction is often assessed by  using an artificial task in order to ensure a comparison between subjects and to increase reliability of findings.

Therefore, some participants have to interact in a setting that seems unnatural to them, which influences their parent-child interaction.

Location of the observations

Different studies on the effect of the nature and location of the study found different results.  One study found a moderate to high correlation between behavior at home and in a clinical setting, meaning that parent and child behaved in a similar way in both settings. Another study found similar behavior in some variables, but not in others. Also, the similarity of behavior differs with the age of the child. One study found poor correlation between home and laboratory setting in parent-child interaction when the child was one year old.

How can this review be concluded?

To conclude, the observer should be cautious in making assumptions about the similarity of parent-child interactions in natural settings and under laboratory circumstances. However, the participants’ reactivity to the observer does not necessarily lessen the validity of the results. The effect of video observing on the participants’ reactivity remains unclear. Furthermore, the observer needs to be cautious when generalizing the results.

More studies are needed to examine the effect of video observing on participants’ reactivity and the effect of the nature of the tasks.