Introduction
Direct observation involves looking and recording of target behavior. Many researches rely on the use of direct observation despite the fact that issues have been raised with regard to the accuracy of the observers. It has been observed that the data that is provided is not always commensurate with the widespread use of observation methodologies. There are myriad factors that affect accuracy of data that is derived from direct observation.
These factors include reactivity, observer’s drift, the procedure of recording used, location of the observation, reliability, expectancy and feedback, and the characteristics of subjects, observers, and settings. This critical writing paper will do a critical review of observation as a research method (Repp, Nieminen, Olinger, & Brusca, 1988, p.1).
Reactivity
When an observer records the behavior of interest, the subject becomes aware of their presence. It is assumed that the subject will behave in the same way as when the observer was not present. Naturally, reactivity is likely to occur as the subjects under investigation react to the presence of the observer by changing their behaviors. In the process the behaviors may increase, decrease, become more variable or remain not affected at all.
As a matter of fact, some subjects under investigation will do their best to present their best sides. There is a likelihood that only socially accepted behaviors will be exhibited by the subjects. Nevertheless, other subjects are likely to exhibit opposite reaction. Reactivity impacts study’s internal validity because its effects cannot be separated from those of experimental variables.
Reactivity also affects external validity of a study and in the process raises concerns with regard to whether the study can be applied to non-reactive situations and pretest sensitization. Researches that have been conducted have illuminated reactivity at the expense of the effects of observation on the behavior of observers.
The latter is called observer reactivity. Observer reactivity is very common in circumstances where teachers and their students are involved. When teachers act as observers, they tend to give more prompts to the subjects under study. Others may increase their rate of instructing and that of giving positive feedback (Repp et al, 1988, p.31).
Observer’s drift
Other than reactivity another factor that affects accuracy of observers is the observer’s rift. Observer’s drift comes about as a result of the shift of the observer from the original response definition. The behavior of subjects under observation gets inconsistently recorded. Observers drift is quite common among between-group and within-group designs. This calls for considerable concern. Drift makes the data collected incomparable across conditions because precise responses cannot be quantified.
Observer drift is also a pointer to the distinction that is there between observer agreement and observer accuracy. Observer agreement results from comparison of a number of observers while observer accuracy is occasioned by comparison of scores with already established criteria. There are instances when high observer agreement is given prominence at the expense of accuracy (Repp et al, 1988, p.15).
Location of the observation
It is common knowledge that data derived from direct observation are normally collected in situ. In fact to reduce the effects of obtrusiveness of observation some data collectors do prefer use of audio or videotapes. These devices in the initial stages of their use may occasion some elements of reactivity but later the effects of reactivity may be ephemeral.
This is supported by findings in the studies where there were no systematic changes in families who were being observed by use of audiotape to see if they changed their positive and negative response over time. After a series of observations, there were no un-lasting changes (Repp et al, 1988, p.1).
Reliability
Reliability in direct observation denotes inter-observer agreement. Observers are allowed to evaluate themselves after they have undergone some kind of training. However, caution should be taken to ensure that the observers do not reach high agreement scores to the detriment of accuracy.
Experiments must therefore do calculations for themselves. Awareness of observers during reliability checks impacts both observer accuracy and the reliability of the direct observation study. Observers tend to be more accurate when they have a perception that somebody is monitoring them. Inter-observer agreement is higher when observers harbor a feeling that reliability is being assessed. This makes reactivity a reality in this aspect.
High frequency behaviors affect how inter-observer agreement is calculated. It is known to inflate percentage agreement on occurrence. Low frequency behaviors on other hand inflate percentage agreement on non occurrence. Existence of relationship between rate of behavior and formulas used in calculating reliability has created a scenario where there is no standard of acceptable agreement levels (Repp et al, 1988, p.1).
Observer’s expectancy
Biasness may result when observers already know what they expect from the subject like sex, behavior of peers, or purposes of intervention. Biasness can weaken conclusions that are drawn from independent variables. For example if respondents are told that a given intervention can reduce a certain stereotypic behavior, the subjects are likely to report that the target behavior has reduced after administration of that intervention (Repp et al, 1988, p.1).
Conclusion
Behavioral observations also have problems related to calculation and inference. Researches that involve behavioral observations really depend on accuracy of observers.
Observers in the field must therefore have some form of calculation of agreement. Accuracy of the observers is always tested by comparing their coding of video or audio tapes with an established criterion. Factors that affect accuracy of observers include reactivity, observers drift, and location of the observation, reliability, and expectancy of feedback among some other factors.
Reference List
Repp, A.C., Nieminen, G.S, Olinger, E., and Brusca, R., (1988). Direct Observation: Factors Affecting the Accuracy of Observers. NJ: Prentice Hall