Experimental Design Flashcards
what is experimental control (functional relation)
demonstrates that the IV causes DV
DV is measured over time and treatment be repeatedly presented and extraneous variables be controlled
What is a descriptive analysis
measures the amount of a behaviour that occurs under one environmental condition or treatment variable, but does not manipulate that condition.
what is a correlational analysis
compares the amount of behavior that occur under at least two environmental conditions or treatment variables, values of treatment but does not manipulate that condition.
What is a hypothetical analysis
a reasoned guess
What is external validity in behaviour analysis
direct replication
systematic replication
study findings can be generalized to other settings
what are the two types of direct replication
intersubject - different subjects with similar characteristics are exposed to the treatment
intrasubject - same subject is used in replication
What is internal validity
extent to which the experimental controls or eliminates confounding variables
confounding variables can occur related to the setting (e.g., bootleg reinforcement, or reinforcement that is not part of and undermines an intervention), the measurement (e.g., observer drift or bias) and the independent variables (e.g., lapses in treatment integrity)
What are the threats to internal validity
Maturation - changes within the individual that occur during the experiment
Setting Confounds - - Uncontrolled aspects of the natural environment
Testing - repeated testing
Procedural Integrity - treatment not being implemented as planned - procedural drift
Loss of subjects - participants drop out (attrition)
Multiple Intervention Interference - interaction of multiple treatments
Instability - variability in behaviour
Coincidentally Intervening - coincidence that the intervention started at the same time the behavior was changing
Instrumentation - inaccurate measurement by devices or human observers
What is prediction
prediction (anticipating unknown future measurement due to steady levels of responding)
What is verification
(demonstrating that the baseline level of responding would not have changed unless the independent variable was introduced
What is replication
manipulating the independent variable repeatedly and obtaining the same outcomes
what is steady state
repeated exposure to conditions, elimination of extraneous influences, and obtaining a stable pattern o behavior before adding or changing conditions
What are the the defining features of single-subject methodology
subjects serve as their own control
repeated measurement of dependent variable over time
visual inspection of graphic data for changes in trend, level, or variability,
inter-subject replication to establish experimental control
replication across subjects, settings and materials to establish eternal validity
social importance
What is behaviour in terms of experimental methods in behaviour analysis
behaviour is an individual, continuous, phenomenon that is determined by the functional relation it holds with other events.
What is behaviour variability in the behaviour analysts
behaviour variability is the result of environmental influences and an indiviudal phenomenon, it is extrsinic to the organism, and continuous
What does the alternating treatment design control for
maturation ( changes that occur in the sbuject during the experiment)
data instability ( variability obsecures effect due to the overlapping data points)
sequence effects ( a condition run for several sessions influences he next condition) and to some degree, attrition (loss of sobjects)
What is an alternating treatment design
baseline is often run but not esential
then several session of one intervention
then several session of another intervention
conditoins are alternated or counterbalanced fashion
What is disadvantage of alternating treatment design vs. multielement designs
multiple treatment interference - so data show in the graph at a certain level may not reflect that actual level because of of multiple treatment interference since don’t know how it would be if it was implemented consistently.
unnatural nature of rapidly alternating treatments
limited capacity - only a few conditions can be compared usually 4 conditions can be compared effectively
What several threats to validity does alternating treatment/multielementnal design control for
maturation
sequence effect ( a condition run to several seasons influences the next condition)
attrition (loss of subjects)
data instability (variability obscures effects due to overlapping data points)
What is a strength of alternative treatment design
show effect quickly because of it’s alternating treatment
low participant attrition rate
but participant much experience each intervention ins a semi random fashion
allows rapid comparison
minimizes irreversibity problems
minimize sequence effects
accomdates variability in data
What is an alternative treatment design
an alternating treatment design is a multi-element design can compare treatments at the sane tune
imagine functional analysis
different data path for different conditions
How can you determine that there is a functional relation based on a graph
when there is consistent separation between the to series
no overlapping or crossing if it is experimental control is compromised.
How to tell a graph is an alternating treatment design vs.
alternating treatment design, intervention occurs at different times will have a staggered pattern,
but if it’s graphed daily it will it will line up varticially.
with simultaneous treatment the design will always line up vertically because the interventions are implemented at the same time
What is alternating treatment (multielement, multiple schedules)
rapid alternation between two or more treatments
what is simultaneous treatment (concurrent schedules)
two or more concurrently operating contingencies; in effect subject “chooses” thd intervention
What is multiple baseline
staggered implementation of the intervention in a step-wise fashion across subjects, settings or behaviors
what is withdrawal
alternation between baseline and a particular intervention
multiple treatments/ multiple treatment design
a reversal design that evaluates the effects of two or more independent variables relative to each other or to baseline conditions
when to use a changing criterion design
changing criteria is usually a rate
not the best for evaluating skill acquisition program.
not for comparing
not for behaviour that has little tolerance ( severe head banging)
how to tell if the changing criterion has experimental control
several rapid changes in behaviour corresponding to criterion changes
different length of conditions
different size criterion changes
a phase contrary to the general trend
achievement of stability for each criterion
if criterion fall above the line for some criteria and blow the for other, experimental control is compromised
if there are large increases in the criterion and the behviour is able to meet it
What is the procedure for implement a changing criterion design?
establish stable baseline behaviour
change values of an IV systematically in small steps
change criterion only after the behaviour has stabilized
the size of the criterion should be large enough to show practical effects but small enough to achieve the effect
How to demonstrate a functional relationship in multiple baseline design
a behaviour change only with the onset of the intervention
a replication of behaviour change across at least two behaviours, persons or setting (but usually three or more)
staggered implementation of the intervention.
What is the steps for using a multiple baseline design
behaviours should be of different response classes (independent)
stable responding to be achieved for all target behaviours before intervening
intervene on the most stable behavior first
use only one intervention
provide differences in the length of the multiple baselines
intervene on the next target when stable responding has occurred on the current target
What is the multiple probe procedure
collect one to three probe measures for each step in the task
if the data remain low and stable for all three probes, implement the intervention for step one until criterion is attained
collect three probe measures for all three steps. Repeat #1 through #3 until criterion has been achieved on all steps and a final probe condition has been implemented
What assessment strategy does the multiple pobe design employ
multiple baseline assessment strategy of the acquisition of behavioural performances.
it consists of an initial assessment of a learner’s performance prior to training
measuring the learner’s performance on every skill before training that skills
measuring the learner’s performance on every skill after mastery performance is achieve on each skill
Are continuous probe necessary
unlikely that student is going to acquire a skill inthe absence of trianing
multiple probe decreasing only requires that probes be conducted periodically.
An advantage of alternating treatment design over ABAB reversal design
can evaluate two or more interventions
each intervention serves as the control for the other so you can skip the baseline phase and start the intervention right away
What is bootleg reinforcement
can be avoided by using reinforcers that are unavailable elsewhere
What is a risk of multiple baseline design for self-injury
the extended baseline needed in a multiple baseline is a risk for self=injury
When to use a BAB design
provides a reasonable means of determining experimental control without having to delay treatment because it begins with the intervention phase
So use when the target can cause injury to the client or others
when an intervention is in already in place
when time is limited
What are comparative studies
focus on competing interventions:
innovative vs. established intervention
original vs. refined intervention
comparisons to understand interactions with contextual variables
does not compare two studies that are novel
What is a parametric analysis
determines the effective range of values of an independent variable ( dosages)
What are threats to internal validity for comparative analysis
vulnerable to same type of threats as other studies
(ex. multi-treatment interferences, non-reversibility of effects, incorrect attrition of the effect to only one treatment)
What is a component analysis
helps determine which component is effect - it systematically withdraw components on at a time.
what is a multiple treatments design?
multiple treatment design compares an intervention to a baseline condition or other interventions or combination of interventions.
only adjacent conditions could be compared
a minimum of two comparison of adjacent conditions is required to demonstrate experimental control
how to increase internal/external validity of component analysis
each treatment condition should have a stable pattern of responding prior to introducing a new condition
external validity, the experimental effect should be prleicated with other pariticpants
