3- measurement Flashcards
A characteristic of a phenomenon.
A property exists independent of its measurement
Examples: objects have mass regardless of the measurement system used to it.
Events exist in time
Fundamental properties
Measure :
quantify fundamental properties.
Examples:
pounds are a quantifiable measure of mass.
Seconds are a quantifiable measure of time.
Dimensional quantities
A single response occurs in time.
More specifically a response occurs at a certain point in time in relation to a proceeding environmental event.
Fundamental property; temporal locus
A fundamental quality of a natural phenomenon. Johnston and Penny maker, 2008
A Property
They are:
Qualitative. That is they are not measured
Properties of behavior
A dimensional quantity is a
Quantifiable, measurable aspect of a property.
The dimensional quantity associated with temporal locus is
Latency, which is the amount of time between a stimulus and response
A second fundamental property of a single response is derived from the fact that a response occupy his time.
Temporal extent
The amount of time between the beginning and the end of a response cycle.
Duration
The dimensional quantity associated with the fundamental property of temporal extent
Duration
A third fundamental property of a single response
It refers to the fact that a response can reoccur
Repeatability
The number of responses or number of cycles of the response class
Countsbility or frequency
What is the dimensional quantity associated with repeatedly
Countability
These three dimensional quantities, IIRT, rate and celeration arerelated to the combination of these two properties
Repeatability
Temporal locus
Refers to the time between two successive responses; usually the time elapsed between the end of a response cycle and the beginning of the next response cycle.
Dimensional quantity, inter response time
The ratio of the number of responses over some period of time.
Dimensional quantity rate
Fundamental datum in the study of operant behavior
Rate
______. Responding is the ratio of the number of responses over some period of time.
It’s unit of measurement is cycles per unit of time.
It is the fundamental datum in the study of operant behavior
Rate
Refers to change in behavior over time, usually increases acceleration, or decreases, deceleration in rate overtime
Dimensional quantity Celeration
Cycles/ unit of time//Unit of time
Change in rate/time OR # of responses/time/time
Celeration
The quantitative results of deliberate, planned, and usually controlled observation (Johnson and Pennypacker 2009)
Data
Datum is singular form of the term
Pivotal to the practice of ABA.
Our clinical decisions will be made by a valuation of this
Frequent, repeated allows for I’m going to valuation of our intervention(s)
Considerations;
Observer expectation
Reactivity
Complexity
Completing activities
Representative of behavior
Data
Objective
Refers only to observable
Clear
Readable and unambiguous
Allows replication, technological
Complete
Delineates the boundaries of what is and what is not an instance of behavior. Cooper do you thousand seven
Characteristics of a good response definition
Directly measures a dimensional quantity of behavior.
Most measure occurrence versus non-occurrence and thus measure a dimensional quantity, usually percent
Continuous (direct) Response measures
- Event (Frequency) recording
- Latency recording measures
- Duration recording, either per occurrence or total
- IRT Recording measures into response time
Continuous response measures
- Record time observation began
- Count the responses
- Record time observation ended
- Divide: count/unit of time
- Report as rate per unit of time
Event recording (aka frequency recording)
Use this when:
Free operant
Response has a clear beginning and end
Limitations;
Behavior that occurs for long periods of time.
Discrete trials
High rates of behavior
Event recording
Record time observation began
Record each antecedent
record each response
record time observation ended
record as responses/Antecedents/Unit of time
Event recording of restricted operants
Duration Per session:
Total amount of time individual engages in an activity
- Example: in a one hour observation total number of minutes engaged in tantrum behavior. Four minutes, five minutes each equals 20 minutes.
Duration per occurrence:
Amount of time a target behavior occupies
-. For example: 2 episodes of tantrum. First tantrum equals four minutes, second equals 10 minutes
Duration recording.
when to use:
Behavior that occurs for long periods of time.
Limitations::
not sensitive to behavior that occurs often but not for long periods of time
Unclear start and stop
Duration recording Considerations
Specify whether to start the stopwatch at the onset or at the offset of the stimulus
Specify whether or not to stop the stopwatch at the beginning or end of the response cycle.
If the latter, then you may be confounding latency and duration
Latency recording
Considerations. When to use: how much time occurs between the opportunity to respond and the response?
Limitations; will not provide information concerning accuracy of response
Latency recording
Start timing at the END of the response cycle.
Stop timing at the BEGINNING of the next response cycle
IRT recording
Use this when the time between responses is of concern
Limitations:
will not provide information concerning accuracy of response
IRT recording Considerations
The Mean IRT May be estimated given the RATE, provided that the variability is not too great, or that there are no outliers
The formula is:
Recording interval/count
Helps to remember there are:
60 minutes in one hour,
3600 Seconds in one hour
Estimating mean IRT’s
Rate equals response/time
If rate = 6 per hour
That is same as 6/60 minutes
IRT = Time Units/Responses
Solve: 60/6
Estimated I RT equals 10 minutes
Estimate the mean IRT, Example
any operant whose response rate is controlled by a given opportunity to emit the response. Each discrete response occurs when an opportunity to respond exists. (aka restricted operant, controlled operate. Contrast with free operant.) duration.
Restricted Operant
Rate equals response/time If rate equals 120 per hour That is same as 120/3600 seconds IRT equals time units/responses Solve: 3600/120 Estimated IRT equals 30 seconds
Example: estimate the mean IRT
Responses are distributed across the observation.
No significant outliers
You know the rate
When to estimate mean I RT
Responses are grouped at one point of the observation interval such as the beginning or end
Significant outliers
When not to estimate Mean IRT
Percent occurrence percent correct
Trails to criterion
Discrete categorization – coding
Partial interval recording PIR
Whole interval recording WIR
Momentary time sampling MTS
PLACHECK
Discontinuous response measures
Similar to event recording of a restricted or discriminated operant. However converted into a percent
Percent correct is a special case of this
Present Occurrence
When to use this:
Interested in a proportion of correct responses
Limitations:
Not a dimensional quantity
Insufficient opportunities to respond
Percent Occurrence: Considerations
The number of consecutive opportunities to respond required to achieve a performance standard.
Record each opportunity to respond until the performance standard is vet.
Trials to criterion
- Determine what one trial will be, untied shoe.
- Decide how to report, number of trials or number of block trials.
- Record count as the measure
- Present data
Trailers to criterion steps
When to use :
Evaluate the efficacy of different teaching strategies.
Assessing learner competence.
Limitations:
behavior that is difficult to count
Trials to criterion Considerations
Responses/antecedents/unit of time
Event recording of restricted operates
A method for CLASSIFYING responses into DISCRETE categories.
Expressed as a percentage of responses for each code
Discrete categorization (Coding)
Went to use:
severity codes, Independence codes
Limitations:
not dimensional quantities
Discrete categorization
Example
Antecedent stimulus :wash your hands
Target behaviors: steps in Handwashing task analysis see below
Codes: L I equals independent; VP equals verbal prompt; GP equals gestural prompts; PP equals physical prompt; MG equals manual guidance.
- Go to sync
2 turns on Faucet
Etc
Discrete categorization
A discontinuous response measure in which a recording session is broken into short intervals of time, usually 10 to 20 seconds
A response is recorded as occurring if it occurs at any time during the interval.
-it is recorded as a non-occurrence if it does not occur at all during the interval.
Reported as a percent of intervals.
Partial interval recording
Use when there is very high frequency target behavior,
ease of data collection
Limitations:
prone to errors in estimation of actual target behavior,
length of interval
Partial interval recording
A discontinuous response measure in which a recording session is broken into short intervals of time usually 10 to 20 seconds.
A response is recorded as occurring if it occurs during the entire whole interval
-it is recorded as a non-occurrence if it occurs for less than the entire interval
Reported as a PERCENT intervals
Whole interval recording
A discontinuous response measure in which A response is recorded as occurring Only if It occurs at the point in time when the interval ends
Momentary time sampling
behavior is occurring over a long period of time.
Ease of data taking
Limitations:prone to Errors in estimation of actual target behavior
Whole interval recording
Use one continuous observation is not feasible
Prone to errors in estimation of actual target behavior
Limitations: low frequency duration
Momentary time sampling
A special case of momentary time sampling
A group of individuals is observed at the end of an interval
Count how many individuals are engaging in the target behavior
Compare with the total number of individuals
Percent of individuals engaging in behaviors
At the end of the interval observe the group and count how many of individuals are engaging in the target behavior at that moment
Divide the number engaging in the behavior by the total number of individuals
Reported as a percent of individuals engaging in behavior
PLACHECK
When there are a group of consumers
Limitations: prone to errors in exclamation of actual target behavior
PLACHECK
All indirect measurement methods are prone to errors:
WIR: underestimate
PIR: overestimate and underestimates
MTS: overestimates and underestimates
The dimensional quantity of interest
The estimated rate of the behavior
Whether to measure responses or episodes
Where to collect data
When, how often, and how long…
Who will collect the data
Resources that are available
How the data will be used
Factors to consider when selecting a response measure
If a behavior has a consistent effect on the environment it can be measured by that effect
Measuring the results of behavior is termed
Permanent product Measurement
If real time measurement is not possible this may be used to ease data collection
Permanent products
Use when behavior leave a product to be measured
Use when I’m going data collection is difficult
Limitations: real time observation maybe needed
Lose information about the behavior
Permanent product
Reliability is the consistency of measurement
In ABA, when applied to the behavior being measured, it is the coefficient of agreement between two or more independent observers
Inter-observer agreement
Usually calculated as a percentage by dividing the number of agreements by the total number of agreements plus disagreements, then multiplying by 100
Reported as percent agreement
Inter-observer agreement
Uses:
- Competence of new observers
- detecting observer drift
- validate collection methods
- increased confidence that interventions are responsible for behavior change.
Uses for IOA, (interobserver agreement
Reliability may also be assessed with regard to the treatment or intervention
It’s integrity is the degree to which an intervention is implemented as described/designated
This is often termed treatment integrity or procedural fidelity
It is vital that both the dependent variable or the behavior, we are measuring and the independent variable, or treatment, we are using are being objectively measured
IV integrity, independent variable
Two main methods are:
Total count.
Calculated by smaller over larger times 100
Percent agreement.
Calculated by agreement/agreement +disagreements x100
Yield same value as total count for direct measurement
Used primarily for indirect Measurement methods
Determining interobserver agreement
Total count IOA has a flaw. There is no guarantee that the data were collected at the same time.
This could be problematic if different treatments are in place at different times
Mean count per interval can be used to increase the confidence of event recording data
First we break the observation. Into segments
Mean count per interval IOA
Mean count per interval:
Total agreement in each interval divided by number of intervals x100
Total duration: shorter duration/longer duration times 100
Mom – Five minute tantrum
Dad – three minute tantrum
3/5×100 = 60 percent
Interval by interval:
Number of intervals agreed/number of interval agreed plus number of intervals disagreed times 100
Calculating mean account per interval
Example. 7/(7+3) = 0.7 x 100 = 70%
Scored interval
Intervals are only scored if an observer marks an occurrence
Intervals where both observer agree there was no occurrence are omitted
Unscored interval
Only intervals are score it is an observer did not mark an occurrence
Intervals we are both observer is agree there was an occurrence or omitted
Most appropriate for high frequency behaviors
In general, use more stringent methods if possible.
When using interval or time sampling consider interval by interval with scored interval or unscored interval I
Determining interobserver agreement
IOS should be at or above
IOA should be collected and scored for a minimum of
80%
33% of observations
Special class of momentary time sampling
PLACHECK (planned activity check)
A group of individuals is observed at the end of the interval
PLACHECK (Planned activity check)
Kristen is directly and openly observing staff person is. As such, their behavior may change as a result of her presents. In other words, they may react in a way that their behavior in her presence is not representative of their normal performance
Reactivity
Take care to collect data when it is
__________ Of behavior.
Example: measuring SIB at the end of acquisition trials.
Measuring behavior for increase when it is Least likely to occur
Representative
Specific responses or episodes?
What behaviors to lunch together and which to define separately
Attempts versus threats but NEVER intent
Intensity and or severity as part of the definition
Duration may be part of the definition
Latency may be part of the definition
Topographical or functional definition?
Some definitional issues – response definitions
Continuous direct response measure is directly measure a dimensional quantity of behavior
Discontinuous (indirect) response measures DO NOT
Most measure occurrence versus non-occurrence and thus Measure a dimension less quantity, usually percent
Continuous versus discontinuous measures
Clinical example
John engages in property distraction throughout the school day. John’s teacher takes data throughout the six hour school day. Aggression occurs At 9:30, 1045, 1:20 and 2:00
I RT: 75, 155, 40 = 270/3=90 minutes
Clinical example IRT
Calculate inter response times.. that is, the time at the end of one event to the beginning of the next event.
Add those times together. Divide that number by the number of times. That is the answer.
Inter response time calculation
Instead of being related to one fundamental property, the last three dimensional quantities, I RT, Rate and Celeration, Art related to the combination of two properties:
Repeatability and temporal locus
Rate # Of responses/time
Celeration
IRT. TIME/# of responses
Countability. # of responses
Duration. Time
Latency. Time
Dimensional quantities
Calculated by agreement/agreement +disagreements x100
Yield same value as total count for direct measurement
Used primarily for indirect Measurement methods
Percent agreement - IOA
