Measurement And Reliability Flashcards
- Select and __ target behaviors
Define
- Select ___ system (how we’re going to go about measuring it)
Measurement
- Collect ___
Data
- Assess ___
Reliability
Good measurement system characteristics
Objective,
reliable,
accuracy/ validity, and
sensitivity
Response (R) definitions are observable and unambiguous
Objective
Consistent (repeatable) results
Reliable
accurate (true) representation of events
Accuracy/ validity
Reflects change in R
Sensitivity
___ pertains to the specificity of the action
EX. Walking- “moving forward at slower than running pace”; what does it mean to run?
EX. saying something you don’t mean; what did you mean?
Subjectivity
Observing outcome of behavior;
Advantage: convenient; observer doesn’t have to be present
Disadvantage: don’t know what response produced product/ outcome
Product
Observing behavior as it occurs, it’s ongoing;
Advantage: flexible; accommodates wide variety of Rs
Disadvantage: requires system to account for multiple R dimensions
Process
Observable results of R
Outcome/ product
Ongoing R
Process
What R looks like; its form
Topography
Number of occurrences of R
Frequency
Amount of time taken up by R
Duration
Force of R
EX. How loud child is yelling
Magnitude/ intensity
Amount of time elapsed b/w event and R
EX. How long it takes you to hit brake when light turns red
Latency
Time elapsed b/w two Rs of same type
EX. Laughing at joke first time it’s told and not laughing when it’s told again within 10 mins.
Interresponse time (IRT)
Recording continuously; no gaps in record of behavior
EX. Count, frequency, rate (event recording)
Duration, latency, IRT (duration recording)
Continuous
Interval recording; sampling, recording only subset of R;
Divide session into equal intervals, record occurrences of R during interval, and convert to % of intervals during which R occurred
Discontinuous
In process observation, we need both continuous and discontinuous b/c we could work better in different ___
Situations
R scored if it occurs for entire/ whole interval; tends to underestimate
EX. Recording convo. For whole 10 second interval
Whole interval recording (WIR)
R scored if it occurs during any part of interval; tends to overestimate
EX. Convo start for 2 of 10 seconds in 10-second interval
Partial interval recording (PIR)
R scored if it occurs at end of interval; tends to neither underestimate nor overestimate, random errors
EX. Last word said right at last minute (the end)
Momentary time sampling (MTS)
Purpose of DeLeon and Fuqua was to increase ___
Recycling
Rationale of DeLeon and Fuqua is to increase folks’ __ on recycling;
lack of __(have to use taxpayer dollars) and utilize other types of incentives
Reliance; money
Method of DeLeon and Fuqua was that they were broken up into __ groups, and increased recycling without ___ incentives. They monitored feedback
3; monetary
Measurement used in DeLeon and Fuqua was ___ measurement (observer doesn’t have to be present). They measured the __ of recyclable material
Product; weight
The __ group did not improve recycling, but the other 2 groups did improve. This is important for applied behavior analysis (ABA)
Control
Assessing reliability and the consistency of observational code;
When behavior occurred and when it did not occur
I.e. are our records of behavior accurate or not?
Interobserver agreement
Interobserver agreement is essential b/c of the ___ of human observation, but lack of automated recording
Fallibility
What factors affect accuracy in interobserver agreement?
Bias, fatigue, difficulty recording, and drift from definition
Comparing overall session values (totals);
smaller/larger
Total agreement
Comparing entries in each interval and score as agreements or disagreements;
# of agreements/ # of intervals
EX. Did observers agree or disagree? Did it occur or not?
Interval agreements
Comparing # of instances/ seconds of behavior recorded in each interval;
# of exact agreement/ # of intervals
EX. Did they say the same exact # or not?
Exact agreement
Calculating proportion agreement interval by interval, frequency within intervals ;
Total (smaller/larger)/ # of intervals
Proportional agreement
