Unit 3 Flashcards
Discontinuous Recording Methods are
Methods in which all instances of the target behavior are not captured during the observation period. Occurrences are known/planned to be missed.
When are discontinuous recording methods appropriate?
When continuous measures are challenging
Interval Recording
Discontinuous measures in which the observation period is divided into smaller time periods, and criteria for occurrence or nonoccurrence are established
Interval recording is also known as
Time sampling
Data in interval recording are typically reported as
Percentage of intervals
Types of interval recording
Partial-Interval Recording (PIR)
Whole-Interval Recording (WIR)
Momentary Time Sampling (MTS)
Planned activity check (PLACHECK)
Partial-Interval Recording
A form of interval recording in which occurrence is noted if the behavior occurs at any time during the interval
PIR Advantages
-Easier to record than rate/frequency with high-rate responses and responses without a clear start and end.
-High IOA
PIR Disadvantages
Underestimates frequency of high-rate behavior
Overestimates duration of responding
When to use PIR
For moderately high frequency behavior
For behavior without clear start/stop
For behaviors targeted for decrease
Whole-Interval Recording
A form of interval recording in which occurrence is only noted if the behavior occurs for the entire interval
WIR advantages
-May be easier to record than duration per occurrence or total duration (can stop observing as soon as the behavior stops occurring)
WIR Disadvantages
Underestimates duration of responding
WIR should only be used for
Behaviors targeted for increase
When to Use WIR
-For behavior with significant duration
-Continuous measurement of duration may not be feasible
-For behaviors targeted for increase
Momentary Time Sampling
A form of interval recording in which occurrence/nonoccurrence of behavior is observed and recorded at the end of the interval
MTS Advantages
-Very easy to record
-May be most accurate representation of duration (doesn’t have a bias towards over or underreporting)
MTS Disadvantages
May not be representative (doesn’t capture behavior during the interval) - shortest of snapshots
When to use MTS
-When continuous observation is not feasible
-When observer has other responsibilities
-For behavior targeted either for increase or decrease
Planned Activity Check (PLACHECK)
A variation of momentary time sampling where the number of members of a group engaging in the target behavior is noted at the end of the interval
When to use PLACHECK
-When data are needed on the behavior of individuals within a group
-When continuous observation is not feasible
Bias in measurement
A tendency to consistently over or underreport events
Discrete trial probes
A discontinuous recording procedure in which a sample of the behavior is collected during discrete trials when treatment is not yet, or is no longer, in effect
Discrete Trial Probes: Advantages
-Easier to collect than continuous (all trial) data
-May correlate with data from more sessions (three or more)
Discrete Trial Probes: Disadvantages
-Not continuous; limited representation
-May lead to premature determination of mastery
Discrete trial probes are also known as
First-trial probes
Derivative Measures
Forms of data obtained from direct measures of behavior
Two derivative measures of behavior
-Percentage
- Trials-to-criterion
Percent Correct
The ratio of correct responses out of the total number of responses as a fraction of 100
Percent Occurrence
The ratio of target responses out of the total number of opportunities to respond
Advantages of Percentage
-Puts response dimensions in context
-Not just frequency, but frequency out of opportunities
Disadvantages of percentages
-Lose reference to behavioral dimensions
-Percent and percent correct are not dimensional quantities
Discrete Categorization (Coding)
A data collection method for classifying responses - codes for behavior
Discrete categorization is reported as
Percent of responses for each code
Discrete Categorization (Coding) Advantages
-Allows collection of data on qualitative measures
-Can measure topography, magnitude, intensity, etc
Discrete Categorization (Coding) Disadvantages
-Often nominal (naming) rather than quantitative data
-May not graph easily - need to be careful with interpretation
Trials-to-Criterion
A measure of the number of opportunities required until responding reaches a targeted level of performance
Trials-to-criterion is a measure of
Learning efficiency
Trials-to-Criterion can be used to
Assess learner competency or to evaluate the efficiency of different teaching strategies
Permanent Products
A change in the environment produced by behavior
Permanent products are sometimes known as
Response products
Permanent Products: Advantages
-Don’t have to be there when the behavior happens (measurement may be easier and more accessible)
-May be more accurate, complete, continuous (allows for multiple people to observe same results and more than one time)
Permanent Product Disadvantages: Lacks _____; Cannot always ________ and ________; Use of recording can _____
-Lacks correspondence between products and behaviors
-Cannot always be sure of the response that produced the product
-Cannot always be sure who engaged in the response
-Use of recording can cause people to behave differently (reactivity)
Amenable to Measurement by Permanent Products - Two Rules
1 - Each occurrence of behavior produces the same product
2 - Permanent product is ONLY produced by the target behavior
Low-Tech Measurement Tools
-Pencil and paper data sheets
-Frequency counters
High-Tech Measurement Tools
Using machines/computers to record data
Advantages of Automated Measurement
Do what they’re designed to do, don’t get distracted
Disadvantage of automated measurement
May not account for variations in topography, intensity, etc.
Manual (human) measurement advantages
More flexible/responsive - able to detect variation in topography, intensity
Manual (human) measurement disadvantages
Humans can get distracted
Guidelines for selecting a response measure
-Choose continuous measures, if possible
-Compromise with discontinuous measures, if needed.
Evaluating Measurement
Effective and Analytic
Effective
Have we made a change?
Did behavior change?
Did the change matter?
Analytic (Evaluating Measurement)
Confidence that the change in behavior was due to the independent variable.
Three essential factors
Validity, accuracy, reliability
Measurement validity
Correspondence of a measure to behavior, legitimately and directly measured.
Measurement accuracy
Comparison of a measure with a known standard
Measurement accuracy determines
Whether the observed value matches the true value
Measurement Reliability
Consistency of measurement
Rank the terms by order of importance: Reliability, validity, accuracy
Validity, accuracy, reliability
Considerations in Selecting Observation Methods
-Training
-Feedback
-Contingencies
-Ensure accuracy
-Prevent observer drift
Troubleshooting issues: Failure to collect data
-Make as easy as possible
-Practice to fluency
-Set expectations before or use contract
-Transmit data more frequently
-Reinforce and shape
Troubleshooting issues: Losing data sheets
-Electronic
-Brightly covered folders/binders
-Central location or near where behaviors occur
-Make portable if possible
Troubleshooting issues: Data are being estimated
-Support the efforts
-Stress the importance of honesty - it’s better to have no data than to have inaccurate data
-Shaping
Interobserver Agreement (IOA)
A comparison of two or more independent observers’ recording of the same events
IOA as a Measure of Reliability
Might be considered a measure of inter-rater reliability.
IOA is not a measure of
Validity or accuracy
IOA is more so a measure of
Believability
Uses of IOA
-Determine competence of new observers
-Detect observer drift
-Increase confidence in definitions and measurement procedures
-Confirma that variability in data reflects behavior and not observers
General standard of IOA
80%
IOA is reported as
Percent agreement
Two general approaches to calculating IOA
1 - Smaller number divided by the larger number
2 - Number of agreements divided by the number of agreements plus disagreements
Total Count IOA - Measurement and formula
-Rate, frequency
- Smaller number/larger number
Mean Count-per-Interval IOA
Int 1 IOA + Int 2 IOA…./N intervals
Most stringent form of count per interval IOA
Exact Count
Exact Count-per-interval IOA formula
of intervals with 100% agreement/N intervals
Total Duration IOA formula
Shorter duration/longer duration
Mean Duration-per-occurrence IOA formula
Dur 1 IOA + Dur 2 IOA…/# of responses with duration
Types of Count IOA
Total count, mean-per-interval, exact count-per-interval
Types of duration IOA
Total Duration, Mean Duration-per-occurrence
Interval-by-Interval IOA formula
intervals agreed/#agreed+#disagreed
Scored-interval IOA is also known as
Occurrence IOA
Scored-interval IOA only looks at
Agreement on occurrence of behavior (scored)
Unscored-interval IOA is also known as
Nonoccurrence IOA
Unscored-interval IOA only looks at
Agreement on nonoccurrence of behavior (unscored)
IOA data should be collected
-Across phases/conditions
-For all observers
-For 20-33% of all observations
