Direct Observation

Description of the Strategy

Practitioners and researchers keep track of the outcomes of behavior modification and cognitive behavior therapy using direct and indirect recording methods. Indirect methods involve using surveys, interviews, and analyses of existing information as indications of target behaviors. Direct methods include examination of permanent products (e.g., tests or writing samples) and direct observation, which is described in this entry.

Direct observation means watching an individual and recording his or her performance. When an informal observation is completed, a narrative summary of the performance is prepared, which is referred to as an anecdotal record. In a formal observation, the observer decides ahead of time who will be observed doing what when, how the behavior will be counted, and where the observation will be conducted. The behavior should be described in operational and measurable terms by emphasizing its dimensions. A recording schedule, metric, and location are selected and applied. After the observation has been completed, the counts are tallied and reported.

Defining Behavior. Behaviors must be defined in operational and measurable terms so that any observer can consistently determine when they are occurring and keep a record of it.

• Well-defined behavior example: Mandy says “No” when the teacher gives instructions.
• Poorly defined behavior example: Mandy is noncompliant.

Identifying Dimensions of Behavior. Identifying the dimension of behavior not only helps in the selection of appropriate recording methods but also allows the specific area(s) of concern to be pinpointed. For example, is the primary concern how often the behavior occurs (frequency or rate), how long it lasts (duration), how much time elapses between the antecedent stimulus and the onset of the behavior (latency), or how the behavior is described (topography, magnitude, or locus)?

• Frequency is the number of times the individual engages in a target behavior.
• Rate is the frequency of behavior expressed as a ratio to time.
• Duration is how long an individual engages in a target behavior.
• Latency is the length of time between instructions to begin a behavior and the actual start of a behavior.
• Topography is what the behavior looks like (i.e., the motor components of the behavior).
• Magnitude, or force, is the intensity or severity of the behavior.
• Locus is where the behavior occurs.

Recording Behavior. Direct observations can be recorded in different ways, depending on the behavior dimensions identified.

• Event recording measures the number of times a behavior occurs during a preset observation period. For example, event recording is used to document the number of times a student raises his or her hand during a 30-minute observation period.
• Interval recording measures the number of time blocks in which a behavior occurs, and it can be conducted in two basic ways. Both methods begin by dividing the observation period into a specific number of blocks or intervals of time (typically 30 seconds to 3 minutes). In the whole-interval method, the observer records only if the behavior occurs during the entire observation interval. In the partial-interval method, the observer records if the behavior occurs during any part of the observation interval. For example, a 10-minute observation period can be divided into 30-second intervals. If the individual engages in the target behavior for an entire 30-second interval, an observer using the whole-interval method would mark that interval on the recording sheet. An observer using the partial-interval method would mark that interval if the target behavior occurs at least once at any time during the 30-second interval. The result is the percentage of intervals. In general, an observer would use the whole-interval method when the goal is to increase a target behavior and the partial-interval method when the goal is to decrease a behavior.
• Momentary time sampling also requires the observer to divide the observation period into equal time intervals. In this method, the observer records only if the behavior is occurring at the exact moment the interval ends. For example, the observer may set a timer to ring every 2 minutes during a 30-minute observation period and would record only if the target behavior is occurring at the moment the timer rings. The result is the percentage of intervals.
• Duration recording measures the length of time over which a behavior or response occurs. For example, recording the time a behavior starts and stops and then calculating the total time the behavior occurred (e.g., 48 minutes of a 60-minute observation period). By recording the duration of each behavioral occurrence, a cumulative duration or an average duration per event can be calculated.
• Latency recording measures the time between a request (or triggering antecedent event) for behavior and an actual response. For example, the observer can record the time between asking a student to perform a behavior and the time the student starts doing it. If several latency measures are obtained during an observation session, an average latency score can be calculated.

Reporting Behavior. Collecting and reporting data on an ongoing basis helps practitioners and researchers make decisions. Immediately following an observation, the observer should convert the data collected into a usable, numerical form: a frequency count, a percentage of intervals or time, a rate, or a length of [Page 1268]time. Converting observational data allows comparisons across days, settings, or individuals.

...