1 Foundations of Behavioral Measurement and Frequency

Define observable behavior

Behavior is an action or activity that an organism does. To measure it reliably, define it in terms that a person could see or hear rather than an inferred feeling, intention, or trait.

An operational definition should let different observers identify the behavior consistently. For example, replace “is disruptive” with “calls out an audible comment while another person is speaking, without being recognized.”

A measurement unit specifies what is counted or timed. For event-counting, the unit is one clearly defined occurrence, such as one call-out or one completed task. Before counting, decide what qualifies as one occurrence; for example, decide whether repeated calls separated by a pause count as one episode or several. Apply the same rule throughout each observation.

Distinguish count, frequency, and rate

A Count is the raw number of occurrences observed. Frequency, as used in this guide, is how many times the behavior occurred during a specified observation period. It reports the count in context and is not necessarily adjusted to a standard time unit.

Terminology varies across behavior-analysis sources: frequency is sometimes used for a count and sometimes for a rate. To avoid ambiguity, state the observation period explicitly and use the convention described here.

Rate expresses occurrences per unit of time. Calculate it by dividing the count by the observation time:

Use matching time units. An observation lasting $20$ minutes gives a result in occurrences per minute; convert the duration first if you need occurrences per hour. Include the time unit in the answer.

Calculate and compare rates

A learner who calls out $12$ times in $30$ minutes has a count of $12$ call-outs and a frequency of $12$ call-outs in $30$ minutes. The rate is:

The same $12$ call-outs in $60$ minutes gives a rate of $0.2$ per minute. The counts match, but the rates differ because the observation periods differ.

Other records illustrate how counts and rates can rank observations differently:

• $9$ hand raises in $15$ minutes: rate $\frac{9}{15}=0.6$ per minute.

• $18$ call-outs in $30$ minutes: rate $\frac{18}{30}=0.6$ per minute.

• $7$ comments in $10$ minutes: rate $\frac{7}{10}=0.7$ per minute.

The first two records have different counts but the same rate. The third has the smallest count but the highest rate. A count without its observation time can therefore be misleading.

For another example, a student who completes $16$ work problems in $40$ minutes has a count of $16$ problems, a frequency of $16$ problems in $40$ minutes, and a rate of $\frac{16}{40}=0.4$ problems per minute.

Select a suitable measure

Choose a measure based on the information the question requires. For a brief behavior with a clear beginning and end, such as calling out or raising a hand, count each occurrence and record the observation time. If observations differ in length, calculate rate to make the results more comparable.

A simple event count may be difficult to apply consistently when the behavior is not clearly bounded. In that case, choose a method that fits the behavior and the question instead. Across observations, use a consistent operational definition and time unit, and record the observation period alongside every count.