Naturalistic behavioral observation is a behavioral assessment method in which an individual is observed in his or her natural environment (e.g., home, school, work), usually in a context that is most associated with a problem behavior. Typically, observations are made on a predetermined schedule by one or more observers. A time sampling interval is determined a priori (e.g., 20-second periods, 5-minute periods) and the observer records the occurrences of the target behavior and/or other relevant events during the interval. Multiple observers are often used and percentage agreement or
Behavioral Assessment Methods
Types and descriptions
Instruments and measurement devices
Behavioral rating scales
Naturalistic behavioral observation involves measurement of overt behavior in the individual's natural environment. Example: Observe a child's behavior in a classroom or at home.
Analogue behavioral observation involves the measurement of a client's overt behavior in a contrived situation that is analogous to situations the client is likely to encounter in his or her environment. Example: Code marital interactions during a marital therapy session.
Completed by individuals who are either familiar with a client's behavior or have the opportunity to directly observe a client's behavior. Example: Have a parent and teacher complete a child behavior checklist.
Systematic self-observation and recording of parameters (e.g., frequency, intensity) of targeted behaviors, environmental events, cognitions, and/or mood states. Example: Have a client diagnosed with anorexia nervosa complete an eating diary after each meal.
Measurement of physiological and motoric components of behavior problems using a variety of measurement devices, especially electromyographic, EEG, cardiovascular, and electrodermal measures. Example: Record heart rate, galvanic skin response, and blood pressure of trauma survivor while the client listens to a script recounting the traumatic event.
Behavioral Interview: A structured or semistructured interview that assesses dimensions of a client's behavior, behavior-environment interactions, behavioral contexts, and the functional relation of the behavior(s) with other controlling variables. Example: Interview inmates in a residential drug treatment facility using the Drug Lifestyle Screening Interview (see Walters, 1994).
Behavioral Questionnaire: A measurement instrument completed by the client, or individuals that know the client well, that assesses (1) behavioral dimensions and (2) functional relations of behaviors with cognitions, emotional states, and other controlling variables. Example: Have a client complete the State and Trait Food Cravings Questionnaire to identify conditions under which food cravings occur and the cues that elicit cravings.
Child Behavior Checklist Direct Observation Form; Revised Edition of the School Observation Coding System; Marital Interaction Coding System.
Child Behavior Checklist; Motivation Assessment Scale.
Self-monitoring forms are usually individually tailored to an individual's target behavior problem(s).
Polygraphs, heart rate monitors, amplifiers are available through several commercial companies.
Functional Analysis Interview Form; Drug Lifestyle Screening Interview; Alcohol Use Inventory
Motivation Assessment Scale; State and Trait Food Cravings Questionnaire other statistics are calculated (e.g., kappa) to ensure reliable data recording.
Naturalistic behavioral observation has been used as an assessment method to assess a wide array of behaviors. For example, Gulley and Northrup in 1997 observed two children diagnosed with attention deficit-hyperactivity disorder (ADHD) in their classrooms and systematically varied each child's daily dose of methylphenidate (MPH). By observing multiple, operationally specific behaviors over time (e.g., social behavior, disruptive behavior, efficiency at solving math problems, responses to comprehension problems), they were able to show the dosage level of MPH that was associated with the greatest improvement across behaviors. Teacher ratings corresponded with the academic, behavioral, and social measures for one participant, but not the other, suggesting that teacher ratings were not necessarily sensitive to changes in behavior as a function of medication dose.
Data collected using naturalistic behavioral observation can be affected by both participant-related and observer-related error variance. On the participant side, reactivity to the assessment method can change the rate of a participant's behavior and make it less likely that observed behavior reflects behavior as it naturally occurs in the environment. Reactivity effects are discussed in more detail at the end of this section.
Observer-related error variance can affect the accuracy of observations in a potentially endless number of ways. For example, the degree of observer training can affect observer reliability and accuracy Cumbersome recording forms or poorly operationalized behaviors can lead to unreliable coding. Raters can "drift" in their ratings if their understanding or application of coding rules decays over time. Other observer-related factors include observer attentional lapses during recording intervals, contamination of data if an observer is aware of another observer's recordings, errors in time-sampling parameters (e.g., frequency and duration of the time-sampling parameters are incongruent with the dimensions of the observed behavior), behavioral sampling errors (e.g., an important behavior is not included in the behavioral coding system), and observer knowledge of patient status. Frequent accuracy checks of observer ratings should be made by an independent auditor to ensure that observer agreement indices do not fall below an acceptable level (e.g., .80). If the accuracy of an observer's ratings is below the criterion, retraining should be initiated.
The validity of inferences drawn from naturalistic observation also depends on the choice of observation setting. As a general rule, observation should occur in situations where the problem behavior is most likely to occur. Because the dimensions of behavior often vary across situations and contexts, selecting the most relevant naturalistic setting for observation is an important decision.
How the data are aggregated and displayed can have a significant bearing on the validity of inferences as well. Naturalistic behavioral observation data may provide useful information when data are aggregated across many minutes but not in shorter intervals, and vice versa. In addition, how the data are interpreted af fects the validity of inferences that are made. A common interpretive strategy for observational data is to graph data and intuitively interpret the results. O'Brien in 1995 showed, however, that individuals using intuitive estimation methods often underestimate the magnitude of highly correlated variables and overestimate the magnitude of weakly correlated variables.
Was this article helpful?