BRM weeks 7&8
reversal design
ABCB-After the second phase, we implement an intervention instead of simply withdrawing and returning it to baseline. we're going to implement some condition that is intended to reverse the behavior.
AB designs
Can't show replication; or verification; can not generalize the results with one AB design to other settings, other participants, and other behavior; case study
Does the intervention seems to be working, or do adjustments need to be made? No causation can be argued
Case Study (A-B)
In using withdrawals designs, our research question
Does the intervention work? Is there a causal functional relationship between that independent and dependent variable? A way that we demonstrate that is through baseline logic
multiple baseline negatives
Id independent baselines; no within-series replication; time and resources; ethics of withholding intervention
Which of the following is true of reversal design?
It can involve the application of two or more contingencies to an alternative and incompatible behavior
Does the interventions work (a) across settings, (b) individuals, or (c) different behaviors of an individual
Multiple Baseline Design
withdrawal negatives
Not always ethical to withdraw treatment; irreversibility of iv effect on dv
Baseline logic and how we demonstrate experimental control and internal validity within our studies.
We use our baseline to compare the performance with and without the intervention in place. In applied settings, the baselines are great because it allows us to show experimentally that there is a need for the intervention.
How does multiple baseline show prediction and affirmation of the consequent
When we look at that first participant, we see a baseline phase and then we see whether or not our prediction is verified when we implement the intervention within the first intervention phase for each participant. Verification is now demonstrated with our extended baseline instead of verification within participants
Multiple probe design
Which are a variation of multiple base line design. less time and effort. We don't have to collect data continuously for all participants. And this would be a good design to use if there's a high likelihood that the behavior will not change until that intervention is implemented.
Is an intervention effect with causation arguable depending on replication?
Withdrawal Design
Baseline logic: prediction
anticipated outcomes in single case design; using trend, level and variability of data to predict the next point
multiple baseline designs are characterized by
by the staggered introduction of the intervention. They can be considered a series of AB designs with each participant or setting or behavior, with each subsequent AB design having an extended baseline and the staggered introduction of the intervention.
AB designs/ case study/ accountability
can only tell prediction and affirmation of the consequent but no replications, can't show experimental control.
withdrawal designs can tell you
can tell us that there was a change in our target variable when our intervention was implemented. It can also tell us about the nature of that change, whether it's changes in levels, trends, or variability. However, it can also tell us if the intervention was effective or that it was not effective. It lets us see if we do or do not satisfy the prediction verification replication aspect of our experimental control.
multiple baselines can tell us
can tell us that there was a change in the dependent variable when the independent variable is implemented and the nature of that change. It could also tell us which settings whose behavior changes when that intervention is implemented or replicated across those things. It could also tell us that the observed change was due to the intervention itself, and if the intervention was effective. A lack of observed change will tell us that the intervention was not effective.
Baselines
control; absence of the control we are studying; but does not mean no treatment.
multiple baseline postives
does not require withdrawal; good for irreversible bx; can document experimental control across a wide variety of conditions; optimal to assess generalization
BAB variation (withdrawal variation)
if we needed to implement our intervention immediately; does not have experimental control; problematic for research purpose; we do see replication. Can't see prediction or verification.
Stable baseline
most desirable; context to look against and make causal argument that IV is effecting DV
multiple probe design negatives
outliers; tough to talk about causal or functional relationships; tough to look at trends
Multiple baseline- baseline logic
prediction and affirmation of the consequent- by having people in baseline longer
Baseline logis: replication
repeating effects of IV; increases internal validity- less likely that a confounding variable caused change ABAB
withdrawal advantages
replication between individuals; allows for clear demonstration of functional relationship between IV & DV; demonstrates need for programming for maintenance during to baseline
Types of baseline patterns
stable; variable; ascending; descending
ABAB withdrawal
And then we have verification with our second baseline phase. We have replication with our second intervention phase. allows us to make our causal statements and demonstrate experimental control, because we have prediction, verification, and replication.
BAB Design is a type of what category of research design?
withdrawal design
ABA
Problematic ethically because we always want to end in an intervention
Does the intervention work with causal and conceptual explanations confirmed?
Reversal Design
withdrawal designs can not tell you
That the study will generalize to other participants; places; bx; we would need replications
affirmation of the consequences
The idea is that we want to see that when the intervention is put into place, the data look different than what our visual line would have been.Basically, we want to predict what the data would have looked like without the intervention in place and then when we put the intervention in place and see data that are different from that original prediction.
what are AB designs/ case study designs good for
They show differences between our predictive behavior and our behavior when the intervention is put into place. And we see differences in our baseline performances, performance relative to our intervention performance.
Multiple baselines can not tell us
tell us that we can generalize the effect that we saw beyond what we actually studied, so we would need more replications. Across other settings and behaviors and participants, we can make statements regarding generalization across a grand scale. It cannot tell us whether or not we cannot make a functional relationship. We cannot make statements about functional relationships between independent or dependent variables for each individual participant, because we have not demonstrated experimental control of each individual participant.
baseline logic
the way that we demonstrate that we have experimental control, that our internal validity is found, and that we can actually make statements about whether or not interventions were successful.
Baseline logic: verification
we're going to demonstrate that the data would have gone unchanged, if the intervention had not been put into place. The idea here is that after we've introduced the intervention we see that the data is different from our initial prediction. We are going to re-implement a baseline.