3 Dimensional Quantities

RTT

1. Repeatability
   • Count
   • Rate / Frequency
   • Celeration
2. Temporal Extent
   • Duration
3. Temporal Locus
   • Response Latency
   • Interresponse Time

Repeatability

Three Types of Repeatability Measures

1. Count

2. Rate

3. Celeration

• When the behavior can be COUNTED.
• Instances of a response class occur repeatedly through time.

• Three Types of Repeatability Measures:
  • Count
  • Rate
  • Celeration

Repeatability

Three Types of Repeatability Measures

1. Count

• Add up the behaviors or items. 
• On its own count does NOT give us enough information to make decisions regarding a client's behavior intervention program.
• Most useful when observation time is constant across observations.
  • Ex: There are 12 questions on the exam.

Repeatability

Three Types of Repeatability Measures

2. Rate

AKA Frequency

• Add up the behaviors or items over time.
• Rate = Count ÷ Time
  • Ex: Ray answered 12 questions in 2 minutes
• Most popular data method used in ABA
• Rate ^aka Frequency is usually reported:
  • Per second, minute, day, week, month, year
• Report the UNIT OF TIME in your rate measure (e.g., two behaviors in one minute.)
• The unit of time must be STANDARD within the study. 
• If the units of time are NOT STANDARD, you cannot compare data successfully (e.g., two behaviors per minute, then seven behaviors in 40 seconds... how can these be compared to each other?).
• USE RATE WHEN: 
  • Recording behaviors that occur only within limited or restricted conditions (e.g., taking discrete trial data, as trials are measured by opportunities). The opposite of free operants.
  • Measuring continuous behaviors that occur for extended periods of time. 

Repeatability

Three Types of Repeatability Measures

3. Celeration

• Measure of how rates of response change over time.
• Count per unit of time ÷ time
  • Same as Frequency ÷ time
• Celeration can mean accelerate or decelerate.
  • Accelerate: When responding is faster over time.
  • Decelerate: When responding slows over time.
• A minimum of 7 measures of rate is recommended for celeration.
• Response rate is displayed on Y axis
• Time in days is displayed on X axis
• Celeration Trend Line: A straight line drawn through the graphed data representing the direction and degree of the trend.

• When DURATION of behavior can be measured.
• Every instance of behavior occurs during some amount of time.
• ONLY ONE TYPE OF TEMPORAL EXTENT MEASURE:
  • DURATION: The amount of time in which a behavior occurs.
  • USE DURATION WHEN:
    • You want to measure the amount of time of a behavior.
    • For behaviors that occur for too long a period of time or too short a period of time.
    • High rate behaviors.

1. TOTAL DURATION PER SESSION: Cumulative amount of time a person engages in the target behavior in the total session.
   • Ex: Ray engages in 3 instance of behavior; instance 1 :30 seconds, instance 2 :40 seconds, instance 3 :20 seconds → 20 + 30 + 40 = 1:30 Total Duration.
2. DURATION PER-OCCURRENCE: Duration of time that each instance of the behavior occurs.
   • Record duration of each instance 
     • instance 1 :30 seconds
     • instance 2 :40 seconds
     • instance 3 :20 seconds

• Examines when an instance of behavior occurs with respect to other events. 
• Measuring the TIME AT WHICH behavior occurs.
• LOCUS = POINT IN TIME.
• Two Types of TEMPORAL LOCUS Measures:
  1. Response Latency
  2. Interresponse Time (IRT)

Two Types of TEMPORAL LOCUS Measures:

1. Response Latency

2. Interresponse Time (IRT)

1. Response Latency: Time between onset of a stimulus and initiation of a response. 
   • Should be used when you want to measure the time between the opportunity to emit a behavior, and when the behavior is initiated.
2. Interresponse Time: Count of time that elapses between two consecutive instances of a response class.
   • Often reported as mean or median and range of IRTs per session.
   • Functionally related to rate of response. 

Two Derivative Measures

1. Percentage

2. Trials-to-Criterion

• Percentage: RATIO formed by combining the same dimensional quantities, such as count.

• Expresses the proportional quantity of some event in terms of the number of times the event occurred per 100 opportunities that the event could have occurred.

Two Derivative Measures

1. Percentage

2. Trials-to-Criterion

• Trials to Criterion: A measure of the number of response opportunities needed to achieve a predetermined level of performance criteria.

• A trial depends on the nature of the target behavior and the desired performance level.
• Can use count, rate, duration, and latency measures to determine trials-to-criterion data.
• Often used to compare the relative efficiency of two or more treatments.
• Useful for assessing a learner's increasing competence in acquiring a related class of concepts.

• Ex: Teaching "red"; presenting child with 'red' and 'not red' items, providing differential reinforcement for correct responses. T2C data is collected on number of items presented until concept was learned sufficiently. 

Two Definitional Measures

1. Topography

2. Magnitude

Topography

• Form or shape of response.
• Should be used when form of behavior is critical (e.g., sports, gross & fine motor activities).
• Malleable dimension of behavior = Responses of varying form are shaped and selected by their consequences.
• Topography ≠ Function; topographies can be different and have the same function OR be the same, and have different functions. 

Two Definitional Measures

1. Topography

2. Magnitude

Magnitude

• Force / intensity / severity of a behavior.
• Certain responses need to be emitted at specific levels of intensity.
  • Ex: Volume of voice differing in the library, vs. in a crowded store. 

Continuous Measurement Procedures

1. Event recording (i.e., rate, frequency count),

2. Timing (i.e., duration, IRT, latency), 

• Measurement conducted in a manner such that ALL instances of the response class of interest are detected during the observation period. 
  • Event recording (i.e., rate, frequency count), Timing (i.e., duration, IRT, latency), 
  • Ex: Taking rate data upon every occurrence of spitting behavior during a 20-minute observation. 
• Advantages:
  • Useful for behaviors that are 
    • Free operant
    • Have a discrete beginning and end
    • Require minimal displacement of the organism in time / space
    • can be emitted at any time
    • Do not require much time to complete
• Disadvantages:
  • Not useful for behaviors that
    • Occur at high rates
    • Are measured via discrete trials
    • Occur for extended periods of time
    • Are opportunity-based

Discontinuous Measurement Procedures

1. Whole interval

2. Partial interval

3. Momentary time sampling

• Measurement conducted in a manner such that SOME instances of the response class of interest may NOT be detected. 
  • Time sampling methods / interval recording methods (i.e., whole interval, partial interval, momentary time sampling).
  • Ex: Taking data via momentary time sampling after the end of a two minute interval on whether or not the child emitted spitting behavior.
• Advantages: 
  • Useful for behaviors that:
    • Occur at high rates
    • Occur for long durations of time
    • Measured via discrete trials
    • Measured via percentage
• Disadvantages:
  • Not useful for behaviors that: 
    • Are free operant
    • Recording EVERY instance is necessary
    • Require constant attention from observer

3 Procedures for Measuring Behavior

1. Event Recording

2. Timing

3. Time Sampling

Event Recording:

• Methods to record The number of times a response occurs.
• Choose An Event recording Device: Pencil and paper, wrist counters, hand tally, digital counters, masking tape, pennies, buttons, calculators, etc.
• Advantages: 
  • Fairly accurate method
  • Simple to implement; you can do this while engaging in other activities
  • Great to use with free operant behaviors
• Disadvantages: 
  • DO NOT USE EVENT RECORDING WHEN:
    • Responses occur at very high rates (e.g., hand-flapping), continuous behaviors (e.g., on task behaviors, humming etc.) and/or DTT data.

3 Procedures for Measuring Behavior

1. Event Recording

2. Timing

3. Time Sampling

Timing

• Methods used to measure:
  • Duration
  • Response Latency
  • Interresponse Time

3 Procedures for Measuring Behavior

1. Event Recording

2. Timing

3. Time Sampling

Time Sampling

• Variety of methods for recording behavior during intervals or at specific moments in time. 
• Time sampling methods give us an approximation of the actual instances of behavior.

• HOW TO DESIGN AND IMPLEMENT TIME SAMPLING PROCEDURES:
  • First divide observation period into equal intervals of time. Next, record the presence or absence of behavior within or at the end of each interval. See specific instructions on how to design and implement the three forms of time sampling. 
• ADVANTAGES: 
  • Great for recording continuous and/or high rate behaviors.
• DISADVANTEGS:
  • Do NOT use when you want to record certain important but fairly infrequent behaviors.

Continuous Behaviors

vs.

Continuous Measurement Procedures

• Continuous Behaviors: Behaviors that do NOT have a clear beginning and ending (i.e., are NOT discrete) (e.g., humming, shouting, thumb sucking, body rocking).
• Counting each response or measuring its duration would be too hard.
• Use DISCONTINUOUS MEASUREMENT PROCEDURES to record continuous behaviors. 

• Continuous Measurement Procedures: Measurement procedure conducted such that ALL INSTANCE of a response are detected during observation period.
  • Ex: Event recording & timing methods.

3 Forms of Time Sampling

WPM

1. Whole interval
2. Partal interval
3. Momentary Time Sampling

• A measurement procedure in which the observer breaks a time period into intervals and measuring whether or not the behavior occurred throughout the entire interval.
• Underestimates rates of behavior

• How to Design and Implement W. I. R.:
  • Divide the observation period into a series of brief time intervals
  • At the end of each interval, record whether the target behavior occurred throughout the interval.
  • Reporting data: Always PERCENTAGE. Record the percentage of total intervals in which the target behavior occurred. 

• Advantages:
  • Best for measuring a behavior you want to INCREASE.
  • Why? To increase behavior → select conservative time sampling method → W.I.R.
• Disadvantages:
  • Not good to use to decrease behavior
    • Required to observe individual  THROUGHOUT entire interval. 

• A measurement procedure in which the observer breaks a time period into intervals and records whether or not the behavior occurred at any point within the interval.
• Overestimates rates of behavior.
• Used to represent the portion of the entire observation period that the behavior occurred. 

• How to Design and Implement P. I. R.:
  
  • Divide observation period into a series of brief time intervals (e.g., 5-10 seconds).
  • At the end of the interval record whether or not behavior occurred at any point during the interval.
  • Reporting data: Always PERCENTAGE. Record the percentage of total intervals in which the target behavior occurred. 

• Advantages:
  • Easy to measure multiple behaviors concurrently. 
  • Best for measuring behavior you want to DECREASE.
  • Why? If you want to decrease a behavior, you want to select the more conservative time sampling method, which is partial interval.
• Disadvantages:
  • Not good to use when you want to increase behavior.
  • Must observe the target behavior throughout the entire interval. 

• Breaking a time period into intervals and measuring whether or not the behavior is occurring at the end of each interval.
• Used to estimate the portion of the total observation period that the behavior occurred. 

• Advantages:
  • Do NOT need to continuously measure throughout the entire interval.
• Disadvantages:
  • Much of the behavior of interest is missed or unaccounted.
    • To avoid this problem, keep intervals short and observe the target behavior frequently. 

• A variation of Momentary Time Sampling for groups.
• Ex: A teacher observes a group of students at the end of each interval, and records the total number of students engaged in a targeted activity. 

• Measuring behavior after it has occurred by measuring the effects the behavior produced on the environment. 
• A change in the environment that lasts long enough for measurement to take place.
• Event recording, Timing, and Time Sampling can be applied to measurement by permanent product.
• Behaviors that do not have a direct effect on the environment can be measured by permanent product through video tape, photos, audio recording; these are contrived permanent products. 

• Advantages:
  • Practitioner is free to do other tasks
  • Makes inconveniently timed behaviors easier to observe
  • Measurement can be more complete, accurate and continuous
  • Facilitates IOA
  • Enables data collection for behaviors with multiple response classes
  • reduced potential reactivity
•  Disadvantages
  • Responses accountable for producing a particular outcome may vary.

• Recording equipment, in the presence of which a person may act differently.

1. Validity
2. Accuracy
3. Reliability

3 Indicators of Trustworthy Measurement

1. Validity
2. Accuracy
3. Reliability

• 3 Elements of Validity:
  1. Directly MEASURING socially significant TARGET BEHAVIOR.
  2. Measuring DIMENSION of the target behavior relevant to the question or concern about the behavior.
  3. Ensuring that data are representative of the behavior's occurrence under conditions during times that are most relevant to the concern about the behavior. 
• Measurement has VALIDITY when it yields data that are directly relevant to the phenomenon measured and to the reasons for measuring it. 

3 Indicators of Trustworthy Measurement

1. Validity
2. Accuracy
3. Reliability

Threats to Validity

1. Indirect Measurement
   • Secondhand or filtered information
   • Measuring proxy or stand in for behavior of interest.
   • Direct Measures of behavior will always have more validity than Indirect Measures.
2. Measuring the Wrong Dimension of the Target Behavior
   • More threatening to validity than indirect measurement
     • Ex: Measuring frequency when you should be measuring duration
3. Measurement Artifacts
   • Discontinuous measurement: measurement in which some instances of the response class are NOT detected
   • Poorly scheduled measurement periods: Recording data on  the behavior during a time that does not properly depict behavior.
   • Insensitive / limiting measurement scales: A scale that imposes an artificial floor or ceiling on performance 

3 Indicators of Trustworthy Measurement

1. Validity
2. Accuracy
3. Reliability

• The extent to which the OBSERVED VALUE matches the TRUE VALUE of an event.
• True value requires special or extraordinary precautions that ensure that all possible sources of error have been removed. 
• If measurement is not valid, accuracy is moot.

Establish TRUE VALUE for accuracy:

• Measurement system must be different than observed value
• Use IOA

3 Indicators of Trustworthy Measurement

1. Validity
2. Accuracy
3. Reliability

• The extent to which a measurement procedure yields the same value when brought into REPEATED contact with the same state of nature
• Same results repeatedly.
  • Ex: Bathroom scale measures at 185, then 5 minutes later 185, then 20 minutes later 185.
• Reliability ≠ Valid. 
  • Ex: Scale at home measures 185 3x. Scale at Dr.'s office and gym reads 175 3x. 

• Observers unknowingly alter the way they measure a behavior. 
• Unintended changes in measurement systems may cause measurement error.
• Occurs when observers have a shift in interpretation of behavioral definition. 
• Target behavior definitions change over time. 

• How to minimize:
  • Retrain observers and provide feedback on accuracy & reliability of measurement

• Nonrandom measurement error.
• Errors in measurement likely to be in 1 direction.
• Data that over/underestimates true value of event.
• Observer Reactivity: Measurement error resulting from an observer's knowledge that others are evaluating the data s/he reports. An observer can be influenced by how s/he anticipates another observer will record data. 

How to minimize:

• Monitor and record data unobtrusively. Use videotaping for second observer to observe the same session.

• The degree to which two or more independent observers report the same values after measuring the same event. 

• 3 Requisites for obtaining valid IOA:
  • Observers must:
    1. Use SAME MEASUREMENT system
    2. Measure SAME EVENT
    3. Be INDEPENDENT

Interobserver Agreement

Benefits

1. Determines competence of new observers. Detects observer drift
2. Increases / decreases confidence that the definition of the target behavior and measurement systems were clear.
3. Increases confidence that variability in data are not due to observer variability → changes / variability in data reflect changes in behavior. 

Methods For Calculating IOA

1. Timing / Duration (2)
2. Time Sampling / Interval Recording (3)
3. Event Recording (4)

1. Timing / Duration (2)
   
   1. Total Duration
   2. Mean Duration
2. Time Sampling / Interval Recording (3):
   1. Interval-by-Interval
   2. Scored Interval
   3. Unsecured Interval
3. Event Recording (4):
   1. Total Count
   2. Mean Count-per-Interval
   3. Exact Count-per-Interval
   4. Trial-by-Trial

Methods For Calculating IOA

1. Timing / Duration (2)

1. Total Duration
2. Mean Duration-Per-Occurence

1. Total Duration IOA

• Computed by dividing the shorter of the 2 durations reported by the 2 observers by the longer duration and multiplying by 100.
• Formula:  

                             Shorter Duration              × 100

Longer Duration 

2. Mean Duration Per Occurrence IOA

• A more conservative and usually more meaningful assessment of IOA than total duration data. 
• Formula can also be used for the mean latency-per-response IOA or mean IRT-per-response IOA.
• Formula:

Duration Behavior1 + Duration Behavior2 +Duration BehaviorN

 N Behaviors with Duration IOA

Methods For Calculating IOA

2. Time Sampling / Interval Recording (3)

1. Interval-By-Interval
2. Scored Interval
3. Unsecured Interval

1. Interval-By-Interval IOA:

• Primary observer's data for each interval is matched to the secondary observer's data for the same interval.
• Likely to overestimate the actual agreement measuring behaviors that occur at very high or very low rates. 
• Subject to random or accidental agreement between observers.
• Formula: 

# of intervals both recorders are in agreement x100 

Total # of intervals

2. Scored Interval IOA:

• Only uses intervals in which both observer's scored an occurrence of the behavior to calculate the IOA.
• Minimizes the effects of chance agreements for interval data on behaviors that occur at very high or very low rates because it ignores the intervals in which measure y chance is highly likely.
• Recommended for behaviors that occur at frequencies of approximately 30% of intervals or fewer to avoid overinflated and possibly misleading IOA measures. 
• Formula:

# of intervals all recorded occurrence  x100

# of intervals at least one recorded occurrence

3. Unsecured Interval IOA:

• Minimizes the effects of chance agreements for interval data on behaviors that occur at very high or very low rates.
• Only considers intervals in which either or both observers recorded a non-occurrence of behavior.
• Formula:

# of intervals all recorded non-occurrence  x100

# of intervals at least one recorded non-occurrence

Methods For Calculating IOA

3. Event Recording (4)

1. Total Count
2. Mean Count-Per-Interval
3. Exact Count-Per-Interval
4. Trial-By-Trial 

1. Total Count IOA:

• Simplest Method for event recording
• Percentage of agreement between the Toal number of repossess recorded by 2 observers
• Calculated by dividing the smaller of the counts by the larger count and multiplying by 100.
• Overestimates the extent of actual agreement.
• Formula:

Smaller #    x100

Larger #        ,       

2. Mean Count-Per-Interval IOA:

• Dividing the observation period into a series of smaller counting times.
• Observers record the number of occurrences of the behavior within each interval.
• Calculate the agreement between the count of the 2 observers within each interval.
• Use the agreements per interval as the basis for calculating the IOA for the total observation periods.
• Formula:

Interval 1 IOA + Interval 2 IOA + Interval N IOA

Total # of Intervals

3. Exact Count-Per-Interval IOA:

• The percentage of intervals in which all observers recorded the same count.
• The most strict event recording IOA method.
• Formula:

# of Intervals of 100% Agreement    x100

Total # of Intervals

4. Trial-By-Trial IOA:

• The agreement between two or more observers who measured the occurrence or non-occurrence of discrete trial behaviors for which the count for each trial, or response opportunity, can only be 0 or 1.
• Can be calculated by comparing the observers' total counts by comparing their counts on a trial-by-trial basis.
• Formula:

# of Trials of Agreement    x100

Total # of Trials