Definition of Measurement

Is the process of applying quantitative labels to

observed properties of events using a standard set of
rules
Designed to know what is included and excluded
Why is Measurement Necessary?
Without measurement, science is only guesswork and
opinion
Is a philosophy

ABA
Measure behavior to answer questions
Is the basis for talking about behavior
Allows you to determine if a behavior has changed
Practical Applications
Also used to measure behavior and evaluate effects of
interventions
Used before, during, and after treatment
Used to guide decision making
Used to
Discontinue ineffective treatment
Continue or increase effective treatment
Benefits of Measurement
Optimize effectiveness
Verify legitimacy of treatments
Identify and end use of pseudoscience
Accountability
Meet ethical standards
Features of Measurement
Has dimensions or distinct features that can be
measured
Three fundamental properties
Countability: behavior can be counted
Repeatability: Get the same result over and over
Temporal Extent: Duration of the Behavior
Temporal Locus: When does the behavior occur or
stop
Counting
Number of responses emitted during an observation
period
Periods can vary
Reported as frequency count

Problem
May not provide sufficient information for analysis
 Duration may be more important variable
Repeatability / Reliability
Do you get the same result over and over?
When observing or examining a behavior , should get
similar results during different observation periods.
If not, may have multiple stimuli controlling or
influencing the behavior.
Rate/Frequency
Is a ratio of count per observation period
More meaningful than count alone
Include counting time for reference
Number of Responses / 10 Minute Period
Can also examine rate of correct vs. incorrect
responses
Correct responses / Total responses
# Widgets made correctly / Time or Total Widgets
made
Is helpful in skill development
Usually reported as number / standard unit of time
Guidelines When Using Rate
Take complexity of response into account
May need to break out into different responses
Is a very useful measure for free operants
Not appropriate
For responses within discrete trials (end of a class
period)
For continuous behavior over extended period
Both can be adapted if needed.
Celeration
Is a measure of the change in rate of responding per
unit of time
Behavior studying / 15 minutes
Examine stimuli influencing the behaivor
Reported using Standard Celeration Chart
Captures behavior acceleration and deceleration
Can be used to identify stimuli that are influencing
behavior
Duration of Responding
The amount of time a behavior occurs
Total duration of session
Duration of each occurrence
Reported in standard time units
Counting plus duration measures provide different
pictures of same behavior
May be critical in changing a behavior
Response Latency
Measure of elapsed time between onset of stimulus
and initiation of response
Studying and acting out
Typically reported using mean, median, and range
Interresponse Time
Amount of time that elapses between two
consecutive instances of a response class

Direct measure of behavior

Is related to rate of responding
Reported using mean, median, and range
Derivative Measures / Percentages
Are ratios formed by combining the same
dimensional qualities
Expresses a proportional quantity
E.g., Correct to incorrect responses
Proportion of observation intervals when behavior
occurred
Considerations for Using
Percentage
Often is used incorrectly
Most accurate with divisor of 100 or more
May be misleading. Can indicate a behavior change
when an outlier observation impacts the percentage
Has limited use because has no dimensional quantity
Often sets artificial limits on behavior change
May not be practical
Trials-to-Criterion
Measure of the number of response opportunities
needed to achieve a predetermined level of
performance
Can be combined or used to determine trials-to-
criterion (e.g., rate)
Typically are calculated post facto
Used to compare effectiveness
Topography
Is the physical form or shape of a behavior
What does the behavior look like
Anger: Facial expression vs. hitting someone
Can have measurable dimension
Must be careful not to interpret.
Look at defined behvior
Malleable by consequences
Is not a fundamental quality of behavior
Magnitude
Is the force or intensity with which a response is
emitted
Important parameter for some response classes
E.g., voice volume

Is not usually a fundamental quality of behavior

Try to measure with objective systems
Automatic Measuring Meter for Loudness
Procedures for Measuring
Behavior
 Typically involve one or a combination of these
three:
1. Event recording
2. Duration
3. Time sampling methods
Event Recording
Uses procedures for detecting and recording the
number of times a behavior is observed
Devices include:
Clickers, wrist counters, digital counters, masking tape,
paper clips, match books, etc
Issues
Easy to do
Behavior must have discrete beginning and ending
Rate must not be too high
Inappropriate for behaviors with long duration
Problems
Need good observers
In applied settings (bars, restaurants, parks) may miss
behaviors
Solutions
Work in teams
 Have another team observe the first team
Recording devices – Cameras
 Allows teams to examine and recheck behavior.
Time Recording
Procedures to measure duration, response
latency, and/or interresponse time
Duration Techniques
Computer systems, stopwatch, wall clocks, tape
recorder
Response latency and interresponse time
techniques
Precise recording of duration between events of
interest
 Can use similar techniques
Whole-Interval Recording Method
Used to measure continuous behavior
Brief intervals (5-15 seconds)
Record if behavior occurs throughout the interval,
Problems:
Risk of underestimation
Reported as percentage of intervals when behavior
occurs
Partial-Interval Recording
Record if behavior occurs at any time during interval
Yes it occurs – No it does not
Multiple occurrences are scored as one
Does not capture duration

Allows recording of multiple behaviors

Reported as percentage of intervals when behavior
occurred
Momentary Time Sampling
Records whether the behavior is occurring at the end
of the interval
Does not require undivided attention
Misses much behavior
Best for continuous behavior

Reported as percentage of intervals when behavior

occurred
Planned Activity Check
At end of interval, measure number of students
engaged in target activity
Variation of momentary time sampling
Measures behavior of individuals within a group
Guidelines for Time Sampling
Use a timing device to signal beginning and end of
observation
Increase accuracy
Not distracted by watching a stopwatch
Record a response for every interval (e.g., yes or no)
Prevents losing your place with blank intervals
Can use technology to help you
Cameras etc.
Artifactual Variability
Artifact is something that appears to exist because of
the way it is examined or measured

Time sampling provides estimate of actual

occurrences
Different procedures produce different results
Differences produce variability in data
Measuring Behavior by Permanent
Product
Measures its effects on the environment
Ex post facto
All previous procedures can be applied to permanent
product measurement
Products can be a natural or contrived
Advantages of Permanent Product
Recording
Practitioner free to do other tasks
Possible measurement of otherwise inaccessible
behavior
More accurate, complete, continuous
Easier data collection (IOA, treatment integrity)
Allows for the measurement of complex behavior
Determining Appropriate Use
Is real-time measurement needed?
Are moment to moment decisions required
Can behavior be measured as a permanent product?
Each occurrence must produce same product
Product can only be produced by target behavior
Other Issues
Reactivity effects
Observing behavior can change the behavior
Usually decreases over time
What are costs to obtain and measure the permanent
product?
Availability, cost and effort of generating the product
Computer-Assisted Measurement
Data collection and analysis software combined
Multiple systems available
Sophisticated and easy to use
Laptops, hand-held computers, PDAs
Simultaneous recording of multiple behaviors across
multiple dimensions
Conclusions
Measuring behavior is important
Allows you to determine if changes are occurring
Extremely important for working in applied settings