Lesson 5: Experimental Designs Flashcards
Why start with research?
Whenever you provide therapy, you are conducting an experiment
If you understand experiments, you will be able to evaluate if your services are worth the amount you charge
you can determine which components of therapy work and which do not
you can skeptically evaluate the claims of other therapists
facilitated communication
therapist doing hand over hand movement with autistic child on communication board. Can the child to it by themselves or is the therapist doing all the work? Because they’re doing hand over hand we don’t know.
single-subject design
an experimental design
the behavior of each participant is observed before and after treatment
repeated observations are made until the behavior is judged stable
why use it? focus on single person
large-group designs
an experimental design
typically 40-200 people
treatment group, control group
What are the five reasons why behavior analysts don’t use large-group designs?
1) because they are interested with helping the individual, not in getting a good statistic
2) to use large-group you need a large number of people with the same diagnosis
3) people don’t want to be assigned to the control group
4) use inferential statistics and psychologists can’t read them right. (statistically significant doesn’t mean it will for sure work)
5) single-subject designs encourage the researcher to present their individual data graphically and let the consumer decide if it was successful
Assuming that behavior changed after the intervention, if we can say that the intervention caused the behavior change, then the experiment has _________ ________.
internal validity (internal validity is good, want high internal validity)
What are the three main single-subject designs?
comparison design (A B design), reversal design (A B A), multiple baseline design (A and B start at different times for the different behaviors)
comparison design
start with baseline, introduce treatment. Not very trustworthy because you just see if there is any change with the treatment. Doesn’t necessarily mean the treatment caused the change. (NEVER say high internal validity with this one)
reversal design
baseline, go on to treatment, take off of treatment. If they get worse again, then the treatment worked. If they stay better than it’s something else, not the treatment.
multiple baseline design
Take multiple baselines of multiple behaviors and introduce the treatment to the differing baselines and different times. Do this method when you can’t reverse the treatment (e.g. being taught something)
