Continuous measurement and how to implement it
Continuous measurement is measurement conducted in a manner such that all instances of the response class of interest are detected during the observation period. This means that we record every instance of the target behavior.
You are viewing: Which Interval Recording Is The Most Conservative
More specifically, continuous measurement refers to the measurement of the dimensional quantities of a behavior:
• Frequency of the behavior – you count how many times a behavior occurs
• Rate of the behavior – if you divide the frequency by your observation time, you calculate the rate of the behavior
• Duration of the behavior – how long did the behavior occur for?
• Latency of the behavior – how long did it take for a behavior to begin after a SD?
• Time of the behavior – how long between behaviors?
Frequency (count)
Is a simple tally or count of the number of occurrences of a target behavior. Another way it is described is as event recording. Each time the behavior or event occurs, it is documented. For example:
• Counting the number of instances that johny requests a chip
• Counting the number of instances that billy hits his head
• Counting the number of instances that amy say hi to a peer
Rate
Is a ratio of count (frequency) over time, often expressed as count per unit of time. This helps to standardize and compare data for behavior that occurs across varying blocks of time. If you were only tracking frequency of behavior, it might be deceiving to say that Jenny requests 30 times during her session on Monday, but then only 15 times on Tuesday. Why? Well, if jenny had a session that lasted four hours on Monday and only hour on Tuesday, then the data would take on a different meaning. If we standardize our data, we can then get a more accurate picture.
When we talk about rate, we often describe behavior occurring x number of times each minute/hour/day (for example: two per minute, five per hour). In the jenny example, if you wanted to compare her requesting between Monday and Tuesday, we would convert the frequency to rate, as long as we knew how long each teaching block was. To do this, we would divide the frequency by time. For example:
• Johny requests chips five times in five minutes 5/5=1. The rate of johny’s requesting chips is one per minute.
• Billy hits his head three time in 10 minutes 3/10=0.3. The rate of Billys head hitting behavior is 0.3 per minute.
• Jenny requests items 30 times in four hours 30/4=7.5. the rate of jenny’s requesting is 7.5 per hour.
Duration
Is the amount of time in which behavior occurs. It is the basic measure of temporal extent. Another way of saying this is to record how long a behavior (or series of behavior) lasts. For example;
• The amount of time a learner engages in play with a peer
• The amount of time it takes a learner to complete the task of getting dressed
When recording duration, it is important to start your timer at the moment the behavior begins and stop your timer at the moment the behavior ends. You should have a good operational definition of the behavior so that you know exactly when to start and stop your timer.
Latency
Is the amount of time between a stimulus and a response. Worded another way, it is the amount of time that elapses between the end of a stimulus event and the beginning of a target response. In many cases, we are often looking at how long it takes an individual to response to Sd. For example, when your mom asks you a question, latency is the amount of time for when the question is asked until you begin to answer.
Inter-response time (ITR)
Is the amount of time between the end of a response and the beginning of the next response. This is different from latency, because we are now talking about the period of time that occurs between a series of behavior displayed by an individual. For example,
Let’s say that you are eating cereal. In this case, there are a series of responses that occur back to back with varying amounts of time between each one. Each bite is a response and the IRT is the period of time that occurs between removing the spoon from your mouth and then when the soon goes back into your mouth with more cereal.
How to implement discontinuous measurements
Discontinuous measurement describes any form of measurement in which some instances of the response class of interest may not be detected. This means that we do not take data on every instance of behavior, but instead only take data during predetermined times or on some specified schedule.
It is important to know that discontinuous measurement – no matter how accurate and reliable – may give you data that isn’t reflective of the whole picture.
What do we mean by this?
It does not measure a dimensional quantity of behavior but instead is derived from direct measures of dimensional quantities. Essentially, count and time are combined in different ways to give information about the occurrence versus non-occurrence of a target behavior. But these behaviors are not continuously measured. Instead, they are considered a sample of behavior at a specific moment in time. For example,
• Checking in periodically and measuring whether timmy is playing appropriately with his toys
• Checking in every one minute to see if Julie is still sitting at the table reading
When we only check in every minute, we may have missed Julie getting up from the table and then sitting back down. Because we are not watching the whole time, we cannot count frequency or track duration (dimensional quantities). Instead, we are tracking whether the behavior occurred or did not occur on an interval, which will hopefully give us an idea of how often the behavior occurs. As long as we are consistent in our data, we can track behavioral changes.
Here are some example of discontinuous measures:
• Percentage of occurrence (% correct)
• Partial interval recording
• Whole interval recording
• Momentary time sampling
Lets learn about these and how to record them.
Percentage of opportunities or occurrence
Is a derivative measure – it comes from direct measures of dimensional quantities.
It is a ratio formed by combining the same dimensional quantities, such as count and count or duration and duration
It is frequently used in ABA to report the proportion of total correct responses
Percentage of opportunities allows us to compare data from different sessions and/or days that may differ in time or amount of opportunities.
For example, if on Monday james correctly responded to all five of 10 of the directions given to him and on Wednesday he responded to 10 of the 20 directions given to him, we can compare these data even though they do not have the same amount of data. This is because we can convert this data to a percentage, discovering that james followed directions appropriately 50% of the time on both days.
How to take percentage of opportunities data
Percentage of opportunities involves recording whether or not a response occurred (or whether or not a response occurred correctly). It usually involves a giving a checkmark or an x, a +, or -, or sometimes, yes or no.
Read more : Which Of The Following Is Not True About Technology Certifications
You track both the target response and the number of opportunities to respond. Then you convert this data to a percentage. Do this as below
– Percentage correct = [number of correct reponses (checkmarks) divided by the total number of (correct + incorrect) opportunities ] x 100.
Interval recording
The next type of discontinuous measurement we will discuss is interval recording. It is also called time sampling and it is when you record only whether a target behavior occurred or not, at predetermined times (intervals). There are three types of interval recording:
• Partial interval recording
• Whole interval recording
• Momentary time sampling
Partial interval recording
Is a method for measuring behavior in which the observation period is divided into a series of brief time intervals. The observer records whether the target behavior occurred at any time during the interval.
Partial interval recording is not really concerned with how many times the behavior occurred (frequency) during an interval or how long the behavior was present (duration), but really just that it occurred at some point during the interval.
Because of this, partial interval recording can overestimate the behavior. Lets say you are measuring nose-picking behavior across 10 consecutive intervals each lasting 30 seconds. Even if the behavior occurred only once in each 30-second interval, over 10 intervals, this would result in the behavior being reporting as occurring for 100% of the interval. You might then think that nose picking is a significant problem.
Since this type of recording can overestimate the behavior, it is best to use this measure for behavior you really wish to decrease. It will be the most conservative measure of progress. If nose picking decreased to 10% of the intervals, then that might be an acceptable outcome.
How to implement partial interval recording?
1. First, set your intervals on the timer at the time indicated in your treatment plan. (typically these are 10 to 30 seconds long, but they can vary depending on the target behavior). Usually a five minute sample of behavior is recorded at one time.
2. You can then score as a yes, if the target behavior occurs at any time during the interval, and score as a no, if the target behavior does not occur at any time during the interval.
Whole interval recording
Is another method for measuring behavior, in which the observation period is divided into a series of intervals. The observer records whether the target behavior occurred throughout the entire interval.
As opposed to partial interval recording, whole interval recording can underestimate a behavior. In the same nose picking example, the learner would have to engage in this behavior for the total duration of 30 seconds in order for the behavior to be scored as occurring in that interval. So that the learner might nose pick for 25 seconds and still the interval is scored as non-occurrence.
Since the whole interval recording can underestimate how often a behavior is occurring, we instead use this as a conservative measure for behaviors we wish to increase. For example, we want on-task behavior (attending to materials and teacher) to occur almost continuously during a class lesson, and therefore using a whole-interval recording method would be ideal. This way, the attending behavior must occur for the total interval in order for the behavior to be scored as occurring.
How to implement Whole interval recording?
1. First, you set the intervals at the time indicated in your treatment plan. (typically, the intervals are five to 15 seconds long, but they can vary depending on the target behavior). Typically, a five-minute sample of behavior is recorded at one time.
2. Then, you record yes, if the target behavior occurs for the entire interval and record no if the target behavior does not occur during the entire interval.
So what are some behavior we want to increase that would be good to measure with whole interval recording? Some examples are: appropriate play with play sets, doing homework, sitting at the table for therapy.
Momentary time sampling
Is a measurement method in which the target behavior is recorded as either occurring or not occurring at a precisely specified time. At the end of the interval, the behavior is recorded as occurring or not occurring at the exact moment in time.
Momentary time sampling is used primarily to measure continuous activity behaviors – for example, engaging in an on-task behavior with a toy or activity. It is not recommended for measuring low frequency or short duration behaviors. This is because we will likely miss the occurrence of the behavior, which will not give us an accurate sample of the behavior.
Why would we choose momentary time sample over partial or whole interval recording?
A major advantage of momentary time sampling is that the observer does not have to be attending continuously to the behavior to measure the target behavior (unlike the partial and whole). This makes it much practical measurement tool for busy, distractible environments such as large classrooms.
However, keep in mind because the behavior is observed for only a brief moment, much of the behavior will be missed.
How to implement momentary time sampling?
1. First, you set your intervals at the time indicated in the treatment plan (which can vary depending on the target behavior).
2. You then record a yes response if the target behavior is occurring at the exact end of the specified time interval. If the behavior is not occurring at the end of the specified time interval, you record a ‘no’ response.
Discontinuous measurement – reporting data
Data from partial interval recording, whole and momentary time sampling are typically reported as percentages of the total intervals in which the behavior occurred. This allows us to compare data across days/sessions and to estimate the proportion of the total observation period within which the behavior occurred.
Selecting a measurement – discontinuous or continuous?
Things to think about when deciding how to measure a target behavior include:
• What dimension of the target behavior are you interested in (rate, frequency, duration)?
• Are you interested in the proportion of the responses as it relates to other events (the percentage occurrence)?
• If you are looking at frequency, are you interested in how often the behavior occurs in some period of time (its rate)?
• If you cannot measure each instance of behavior, ask yourself: how often does the behavior occur? (discontinuous measures – PIR, WIR, MTS – wont make sense if the behavior occurs at low frequency.
• What resources are available for data collection? Do you have time to observe and record every instance of the target behavior?
Permanent product recording
The measures we discussed above all involve some direct observation in the moment.
Even though much behavior is measured in real time (by observing and recording responses), some behavior can be measured after it has occurred. When we do this, we are measuring permanent products. So what are permanent products?
Permanent products
Are the results of a behavior that produce consistent effects on the environment. They can be measured after it has occurred if the effects (products) left behind by the behavior remain unaltered until measurement.
Examples of permanent products include:
• Written compositions
• Calculations of math problems
• Chores (folding laundry, mopping the floor)
• Number of diapers in the diaper pail
Read more : Which Of The Following Is A Compute Service
Contrived permanent products
Can be natural or contrived. Much socially significant behavior has no effect on the physical environment.
• Sitting properly
• Hand flapping
However, measurement of such behaviors can be accomplished via contrived permanent products.
Some of the examples of contrived permanent products are:
• Audio recordings of students as they read out loud
• Videotaping
• Pictures of learners works
Contrived permanent products are often useful in measuring behaviors that are temporary but allow practitioners to go back and gather continuous or discontinuous measures of behavior.
Implementing permanent product recording involves recording the number or count of the items that are produced. The way these are scored depends on the behavior of interest. For example, if you are interested in the number of envelopes stuffed during a work session of one hour, you would count the total envelopes. You can even transfer this to rate data.
When contrived permanent products are used, essentially the same measurement procedures listed within this module can apply. As long as you have a clear view of each event and response during a video for example, you can apply both continuous and discontinuous measures.
Why use permanent product recording
Some benefits of permanent product recording are:
• The practitioner is free to do other things.
• It makes possible the measurement of some behaviors that occur at inconvenient or inaccessible times and places.
• The measurement can be more accurate, complete and continuous.
• It facilitates interobserver agreement – you can watch the video with multiple observers
• It enables measurement of complex behavior and multiple response classes – you can stop and rewind the tape to see if there are more than one function of behavior.
You have now learned about continuous and discontinuous measurement and how to implement it. Next, you will learn about interobserver agreement and learn how to make sure the data you collect is reliable.
Interobserver Agreement (IOA).
IOA is the amount in which two or more independent observers get the same observed results after measuring the same target behavior using the same measurement system.
Why do we need IOA?
1. It can help to know how proficient the data collectors are. If one observer consistently gets different results than the other observers, that observer may need more training.
2. It can help to detect observer drift.
a. When taking interobserver data, if two or more observers regularly get similar results, this means the operational definition for the target behavior was clear and measurable.
IOA data can help us determine that any variability in the data is not due to observer error – as the changes we observe in the data are truly a reflection of behavioral change.
IOA Formulas
The first step is to count the agreements versus disagreements for each trial or point of measurement in the data. Let’s take a pause here and learn what agreements and disagreements are:
• Agreements – the observer saw the same result (whether it means they agreed that the behavior occurred, or they agreed that the behavior did not occur)
• Disagreements – the observers do not have the same results (one observer thinks the behavior occurred and one doesn’t).
An important point to remember
Having agreement does not necessarily mean that the observers agreed that a target behavior took place. Agreement means that the observers agreed that it either did or did not occur.
After we count the agreements, we calculate it:
No. of agreements / (no. of agreements + no. of disagreements) x 100
80% is the absolute minimum acceptable IOA.
In the example, there are 7 agreements.
7/10×100=70
If its 80% that is minimal, then this data did not have an acceptable ioa. Therefore, we should either:
• Make sure the observer is properly trained – in this case, provide further instruction and feedback on how to observe and record data.
• Review the operational definition – in this case, the definition may not be clear enough and further clarification and descriptions may be necessary.
Once we have done these things, then we can record data again.
Sometimes, our ioa data may not turn out the way we expect. This may be due to any of the four following reasons:
1. Reactivity – reactivity means that the presence of the observers influences the behavior of the person being observed.
2. Observer drift – we talked about observer drift already. It means that over time the observer may unintentionally stray from the definition and or remember the definition differently than how it originally presented.
3. Complexity – complexity means that the data collection system is too difficult or onerous. An example of this is if you are measuring multiples behaviors at the same time, particularly across different measures (frequency, duration, time sampling). This can also be true for behaviors that are complex (that is, recording behavior that has a range of intensities or that merges or co-occurs with other behaviors).
4. Expectancy – expectancy is a bias that happens because of preconceived ideas about either the target behavior or the person being observed or both.
These issues can happen with any behavior plan that requires IOA data. Here are some strategies to help reduce these problems.
What are two ways in which interobserver reliability can improve? There are two strategies that can help improve interobserver reliability:
1. Choose observers who have experience taking data but are naïve to the study itself. This will reduce the issue of expectancy and observer drift. They may be more likely to adhere to the operational definition, but will not have any expectations about what they should be looking for.
2. Analyze the data regularly (preferably daily). This will have to make sure everyone is on the same page and reporting data correctly. This can also correct any errors early on before poor habits are developed.
Source: https://t-tees.com
Category: WHICH
