The flashcards below were created by user
on FreezingBlue Flashcards.
Describe the characteristics primary measures
used to evaluate performance must have.
- Administered repeatedly - can be administered continually over time
- Consistency of measurement - minmal error
- Capacity to reflect change - must show changes
- Dimensional Scale - not binary (yes, no, complete, non complete)
- Relevance of measure - should assess problem directly
- Importance of measure - should be socially important
What are ceiling and floor effects?
Change in the measure may reach and upper or lower limit and therefore must be accounted for in scale as well as behavior
What is the difference between importance and relevance?
relevance - the measure reflects the construct
Importance - The target of the intervention will make a difference and is not temporary and annoying
What are the arguments for using multiple
measures whenever possible?
single measures rarely capture the entire construct
problems are normally multifaceted
multiple perspectives are often pertinent
different methods yield different information
intervention often produce a spread affect
What is the utility of using pre-post assessments and probes throughout a study?
Pre-post assessments can be useful and informative as supplementary measures
- Probes are often used to see if the effects of and intervention has generalized.
Distinguish between reliability and validity
Reliability -consistency of measure or measurement procedure
Validity - whether the measurement assesses the domain of interest
What are the 5 points related to validity of
single measure does not convey the complete construct
- observations are limited to specific situations
- (Will behavior be the same in less structured environments)
Does the measure assess something important
the human filter can alter data
Is the change observed important
When would it be appropriate to use a frequency measure?
the Bx is discrete and take a relatively constant amount of time
What are the two ways frequency measures are used?
When the Bx is free to occur on multiple occasions
when opportunities are restricted because of restricted discrete trials.
What are the critical differences between discrete categorization and frequency measures?
- - single response performance measures are tallied
- - no real time limit
- - several Bxs included
- - limited # of opportunities to perform
What is a unique feature of discrete categorization?
bxs that form the list need not be related allowing for flexible observations
Provide an original example of a situation in
which the number of people who perform a given behavior would be of interest.
Completing all data required on a daily basis at ACE
Describe the two main versions of time-based
- Interval recording
- - bx observed for a single block divided into several small blocks of time per day
- Response Duration
- - the amount of time the response is performed, good for continuous rather than discrete bxs
What is a response-specific measure and why are these measures of use?
assessment procedures unique to the bxs under investigation
they directly assess the response or a product of the response under investigation
When would it be appropriate to use a self-report measure? What are some of the problems with self-report measures?
To address critical faucets of functioning
To evaluate perceived intervention effectiveness
- response biases - agreeableness, societal norms)
What are the advantages and disadvantages of naturalistic and contrived conditions of assessment?
- - Pro
- - Con
- Variable conditions
- hard to standardize
- - Pro
- Can study low freq bxs
- Consistent and Standardized assessment
- Minimizes confounding varaibles
- - Con
- Performance may vary in natural situations
- families may act different in natural setting
Kazdin notes that single-subject designs “bypass variability due to intersubject differences which is included in the design of
between-group experiments.” What are the advantages of this?
intersubject variability is not a part of the behavioral process of the individual
lawful affects of variables are not obscured by intersubject variability
The individual effects of the iv on the participants is lost in group designs
What does interdependence refer to and which
experimental design is it associated with?
when one phase of the experiment affects the other. (across bx, individuals, or situations)
designs without a reversal (multiple baseline)
What situations would require the use of a control or comparison group?
evaluating different levels of a variable
Separate groups are used in research when one is interested in evaluating different levels of variables that are introduced initially to the subject(s) who have not previously been exposed to any experimental phase. Can you think of another way this could
be accomplished, other than separating participants into two different
You can use the same groups just randomize the order of treatments
What are the 6 classes of results of independent variable inaccuracies that were reviewed?
cases where a difference b/t programmed and actual IV application was noted during the study prior to conclusion formation
cases where inaccuracies were noted at the conclusion fo the study but conclusion remain the same
inaccurate administration of the IV changes the conclusion causing an ineffective treatment to appear effective
cases where IV inaccuracies render effective treatments ineffective
failure to replicate was caused by faulty IV application
failure to replicate is linked to IV implementation inaccurices
How is therapist drift defined?
when the treatment agent gradually alter the treatment
Why is it problematic to only publish successful
errors in implementation of the IV are caught slower since successful studies are not scrutinized as much
Why is steady state responding not necessarily
indicative of the independent variable being applied accurately?
The fallacy of affirming the consequence
If the IV is functionally related to the DV then the DV will have a steady state
DV responding is steady thus the IV must have been applied accurately
Explain why anticipated changes in behavior when the treatment variable is applied do not necessarily indicate a robust functional relationship.
because the experimenter must prove that those changes in the DV are a product of IV manipulation and nothing else
What is the ultimate cost of lack of independent variable inaccuracy?
- accepting a weak treatment as strong
- or rejecting the strong treatment as weak
- (type1 and type II errors)
Describe the cost-effective solutions proposed
in this article.
rigorous training and spot checks of the treatment agents performance
only collecting and reporting accuracy for high risk experiments
Use IOA for IV application
Mechanical ways of tracking IV and DVs
Why would it be preferred to use partial interval recording over momentary time sampling when evaluating treatment for problem behavior?
because PIR produced higher levels of agreement with the experts.
Explain why discontinuous sampling methods may be contraindicated for small treatment effects.
because it increased the likelihood of false negative interpretations of a functional relations
Based on results, what did the authors conclude regarding use of PIR and MTS procedures?
- duration = MTS
- Prequenct = PIR