Data Collection Methods
Frequency / Event and Rate Recording
These methods involve counting the number of times a behavior occurs in a specific time period. Use these methods if the behavior can be easily counted and the behavior has a clear beginning and end. Do not use these methods if the behavior is occurring at such a high rate that an accurate count is impossible (e.g., pencil tapping) or the behavior occurs for extended periods of time (e.g., 2 tantrums, but the duration of each tantrum is one hour).
Event recording is best for behaviors with a distinct beginning and end.
Event recording has been used to measure behaviors such as:
- Task initiation and protests to task demands
- Inappropriate sitting and littering
- Correct and incorrect academic responses
Teachers have used event recording to measure their own behaviors, such as:
Event recording is best for behaviors that occur with enough in between to distinguish between the end of one response and the onset of another.
- Praise statements
- Response opportunities provided to students
A frequency measure should be used only when the length of observation time is consistent from day to day (e.g., always 2 hours).
A rate measure should be used if the length of observation time varies from day to day (e.g., 60 minutes on Monday, 300 minutes on Tuesday).
Summarizing the Data:
Frequency: At the end of the observation period, total number of occurrences. For example, Anna left her seat 5 times during 7th period.
Rate: Count the number of times the behavior occurred in the time observed. Divide the count by the length of time the behavior was observed. For example, if Anna kicked a peer 30 times in a 10 minute observation, the rate would be 3 kicks per minute (30 kicks divided by 10= 3 kicks per minute).
This method documents the length of the behavior by recording the time the behavior begins and ends. Use this method if your primary concern is the length of time the student engages in the behavior and the behavior has a clear beginning and end. Do not use this method if the behavior occurs with high frequency or the behavior starts and stops rapidly.
Duration recording is appropriate for behaviors that have a distinct beginning and ending or for those that occur at such high rates that it would be difficult to get an accurate frequency count (e.g., number of taps during pencil, finger, or toe tapping).
Duration recording has been used to measure behaviors such as:
Consider collecting frequency data for the target behavior in combination with duration recording to provide a more accurate picture of behavior.
- On Task: Looking at the assignment, writing and asking questions related to the topic, using assigned materials, and following teacher directions
- Compliance to task demands
- Academic writing tasks
It can sometimes be difficult to accurately record the exact duration of the behavior. On the other hand, duration recording not only tells us how long the student engages in the behavior, but it automatically provides us with how many times the behavior occurred.
Summarizing the Data:
Duration can be summarized in two different ways:
Percentage of observation with behavior: Sum the total number of min/sec/hrs that the behavior occurred during the observation, divided the sum by the total number of min/sec/hrs of the observation, and multiply by 100.
Average Duration of Behavior: Sum the total durations and divide by the total occurrences.
During a 60 minute observation, David had 3 tantrums that lasted 3 minutes, 7 minutes, and then 5 minutes with a total duration of 15 minutes. The % of observation with behavior = 15 minutes divided by 60 minutes= .25 times 100 = Tantrums occurred during 25% of the observation. Average duration – 15 minutes divided by 3 tantrums = Average of 5 minutes per tantrum.
Interval recording documents whether a behavior occurred during a particular period. In order to determine this, an observation period is divided into brief intervals. At the end of each of these, the observer records whether or not a behavior has occurred. There are two types of interval recording: whole and partial interval. When utilizing whole-interval recording, an observer indicates whether the behavior occurred during the entire time. (Example: A student worked on an assignment during an entire thirty-second interval.) When utilizing partial-interval recording, an observer indicates whether the behavior occurred at any point during the time interval. (Example: A student worked on an assignment during fifteen seconds of a thirty-second interval. The record indicates that the behavior occurred.)
Interval recording is used when it is difficult or impractical to constantly observe behavior.Interval recording is used for continuous behaviors or for those behaviors whose onset and end are difficult to distinguish because the behaviors occur at such high rates.
Both whole and partial-interval data are reported in terms of the percentage of total intervals during which the behavior occurred (the number of intervals the behavior occurred / the number of intervals the behavior occurred + the number of intervals the behavior did not occur times 100).
- Whole-interval recording yields data on the total duration of the behavior.
- Partial-interval recording yields data on the proportion of the observations period that the behavior occurred.
Interval recoding often takes less time and effort, especially if the behavior occurs at a high frequency, because the observer records the behavior only once during the interval, regardless of how many times the behavior occurs. However, interval recording only provides an estimate of the actual number of times that a behavior occurs. If the intervals are too long (e.g., 1 hour), the results can overestimate the frequency of the behavior. The shorter the interval, the more accurate the representation of how often the behavior is occurring. For example, if aggression occurred once per hour in a 5 hour observation, a 1 hour interval recording form would conclude that the behavior occurred during 100% of intervals (1 occurrence in each 1 hour interval). However, if a ten minute interval recording for was used, the results would conclude that the same behavior occurred during only 16.7% of intervals (5 of the 10 minute intervals contained behavior out of a possible 30 intervals).
How to Collect the Data:
Identify the observation period at the times during which the behavior is most likely to occur. Typically, observations last between ten minutes and one hour, although it is more accurate and less burdensome to use shorter periods.
1. Divide the observation period into equal intervals. These intervals are usually between five and fifteen seconds long.
A. Whole interval: Record an “x”, plus sign, or check mark if the behavior occurred throughout the duration on the interval (e.g., if using ten-second intervals, the behavior must last the entire ten seconds). If the behavior did not occur for the entire interval, then record the nonoccurrence of the behavior with a minus sign or 0. (Note: No more than one behavior at a time should be observed when using whole-interval recording, due to the necessity of observing during the entire interval.
B. Partial interval: Record with an “X”, plus sign, or check mark if the behavior occurred at any point during the interval (e.g., if using ten second intervals, the behavior must occur at least once during that particular interval). If the behavior did not occur during the interval, record the nonoccurrence of the behavior with a minus or O. (Note: Multiple behaviors can be observed during partial interval recording because an observer only has to document whether a behavior occurred at all during an interval.)
2. Count the number of intervals during which the behavior occurred. Divide this number by the total number of intervals and multiply by 100 to determine the percentage of intervals during which the behavior occurred.
Keep in Mind: Interval recording provides an estimation of behavior.
- Consider using a prompt to signal the beginning and end of intervals, such as an audio recording with beeps (headphones should be used!).
- If you find that you have unmarked intervals, you may have accidentally lost track during the observation and marked the wrong interval.
Summarizing the Data:
- Whole-interval recording typically underestimates the overall duration of the behavior because if a behavior occurs-but not for the entire interval- it is not recorded or documented as occurring.
- Partial-interval recording typically overestimates the overall duration and underestimates the rate of the behavior because if a behavior occurs multiple times during an interval, it is still documented as occurring only once.
When using interval recording, the level of the behavior is reported as the percentage of intervals in which the behavior occurred. To calculate the % of intervals, count the number of intervals in which the behavior was recorded, divide the total number of intervals during the observation period and multiply by 100.
Example: Mary was out of her seat during 4 out of 10 intervals. 4/10=.40 times 100= Mary was out of her seat during 40% of intervals recorded during the observation.