Week 1 Flashcards
(57 cards)
Definition of Measurement
Measurement is the process of applying quantitative labels to describe and differentiate objects and natural events
the process of applying quantitative labels to describe and differentiate objects and natural events
Definition of Measurement
Three steps of Measurement
- identifying the behavior to be measured
- defining the behavior in observable terms
- selecting an appropriate observation and data-recording method
Concept of measurement includes
characteristics of what is being measured
quality and appropriateness of the measurement tools
technical skill of the measurer
how the measurers obtained are used
Frequent measurement helps practitioners:
optimize their effectiveness
verify the legitimacy of treatments
identify and terminate treatments based on pseudoscience, fad, fashion, or ideology
be accountable to clients, consumers, employers, and society
achieve ethical standards
Three fundamental and measurable dimensional quantities of behavior
- repeatability: instances of a behavior can occur repeatedly through time (behavior can be counted)
- temporal extent: every instance of behavior occurs during some amount of time (duration of behavior)
- Temporal locus: every instance of behavior occurs at a certain point in time with respect to other events (when behavior occurs)
Measures based on Repeatability (3)
Count: simple tally of number of occurrences of a behavior
Rate: number of responses per unit of time
Celeration: how rates of response change over time
Repeatability: Count
counting may not provide enough information to allow behavior analysts to make useful program decisions or analyses
Observation period, or counting time, should always be noted
e. g. data showing that Katie wrote correct answers to 5, 10, and 15 math problems over three consecutive math class periods suggests improving, but if it was over 5 minutes, 20 minutes, and 60 minutes, this changes the improvement level.
- this is why we need to include counting time
Repeatability: Rate
One of the most widely used measures in behavior analysis
Converting count to rate makes it more meaningful - e.g. Katie's math problems have a rate of 1.0, .5, and .25 per minute over three consecutive class periods
Often reported as a count per 30 seconds, per minute, per hour, per day
Six guidelines to interpret rate data:
- Reference the counting time
- calculate correct and incorrect rates of response when assessing skill development
- consider the varied complexity of responses
- use rate to measure free operant
- do not use rate to measure behaviors that occur within discrete trials
- do not use rate to measure continuous behavior that occur for extended periods
- Reference the counting time
guideline to interpret rate data
2 methods to calculate rate of response:
- total observation time (most used)
- IRT
Must include duration of the observation time
- not doing that leads to faulty data
- e.g. running 1 mile at a rate of 7 minutes per mile is different than running a marathon at a rate of 7 minutes per mile
- calculate correct and incorrect rates of response when assessing skill development
(guideline to interpret rate data)
when they can get the answers right or wrong, need to have a rate of response for each behavior
- important because if you just have rate of right, then you could have illusory data because wrong could be increasing in rate as well.
- consider the varied complexity of responses
guideline to interpret rate data
Many important behaviors are composites of two or more component behaviors, and different situations call for varied sequences of combinations of the component behaviors
One method for measuring rate of responding (that’s complex) is to count the operations required to achieve a correct response
- e.g. measuring students math calculation performance by counting which step they completed correctly in the sequence. students given 20 multiplication or division questions. for 55 x 5 = ? if they got it right it would be four correct responses because finding the missing factor requires four operations. 4 x ? = 164. There are multiple ways to solve this. Can use division (four steps) or multiplication (two steps) so get the average of three to count for that one. Count the number of operations completed correctly and incorrectly per set of 20 problems and reported correct and incorrect rates of response
- use rate to measure free operant
guideline to interpret rate data
free operant = behaviors that have discrete onsets and offsets, do not depend on SD, involve minimal displacement of the organism in time and space, and can be emitted over a wide range of responses rates
-e.g. bar press by rats and key pecks by pigeons
rate of response is a preferred measurement for free operants because it is sensitive to changes in behavior values and because it offers clarity and precision by defining a count per unit of time
- do not use rate to measure behaviors that occur within discrete trials
(guideline to interpret rate data)
not appropriate for behaviors that occur within limited or restricted situations (DTT).
e.g. responding to a flashcard shown by a teacher, answering a questions when asked
there is a presentation of an SD and because of this measures such as percentage of response opportunities in which a response was emitted or trials-to-criterion should be used. Not rate.
- do not use rate to measure continuous behavior that occur for extended periods
(guideline to interpret rate data)
e.g. playing a playground game or being on task during classroom activity
Dont use rate. Measure whether the behaviors are on or off at any given time, yielding data on duration or estimates of duration obtained by interval recording
Repeatability: Celeration
measure of how rates of response change over time
should use acceleration for increasing rates and deceleration for decreasing rates of response
celeration = rate/per unit of time
Measures based on Temporal Extent (1)
duration: the amount of time from the onset to the end point of a response
the amount of time from the onset to the end point of a response
duration, Measures based on Temporal Extent
Temporal Extent: Duration
important when measuring the amount of time a person engages in a target behavior
appropriate measure for behaviors that occur at very high rates or for an extended time
- e.g. how long a tantrum occurs, or a student that cant stay on task more than 30 seconds.
Two kinds of duration measures:
- total duration per session
- duration per occurrence
- total duration per session
a measure of the cumulative time a person engages in the target behavior
2 procedures to measure and report:
- the cumulative amount of time a target behavior occurs during an observation period.
- the amount of time a person spends completing a specific task, without specifying a minimum or maximum observation period
- total duration per session measure and report:
1. the cumulative amount of time a target behavior occurs during an observation period.
> e.g. measuring child playing alone time, stop timer once behavior stops, but don’t reset timer, when behavior starts again, start timer again
if the duration remains constant, total duration per session data can be reported in standard units of time (e.g. 6 min of play)
if the duration varies, MUST be converted to a percentage of total time observed (e.g. 6 min of play in 10 min session = 60%)
- duration per occurrence
a measure of the duration of time that each instance of the target behavior occurs
e.g. sound-recording device automatically records (a) the number of times noisy disruptions by children on a school bus exceeded a specified sound threshold and (b) the duration in seconds that each disruption remained above that threshold. tally the mean duration per occurrence of noisy disruptions as one measure for evaluating the intervention’s effect
Selecting and Combing Measures of Count and Duration
Count measures:
- repeatability
Duration measures:
- temporal extent
e. g. teacher can count each instance of out-of-seat behavior, but because out-of-seat behavior could last any amount of time, could also use total duration recording
- because of the relevance of temporal extent in this case, duration is better than count
Buuuuuut duration per occurrence measures BOTH duration and repeatability so this is preferable to total duration
