Non-experimental Research Designs
- Case Studies
- Single-variable studies
- Multi-variable studies
Case Studies
- Detailed portraits of the experiences of an individual person or group
- External Validity is open to debate (maybe be a specific type of people but not all.
- Can inspire or generate new research
- 0 Degrees of Freedom (N-1)
Baseline Procedures
- Establish a baseline
- Transition steady state
- Reversibility (argue external validity)
- Intrasubject replication (within-subject)
- Intersubjerct replication (between subjects)
Non-experimental Research Designs
- Case Studies
- Single-variable studies
- Multi-variable studies
A-B Design
A = baseline (before)
Implemented treatment
B = treatment (after)
Comparitive data
A-B-A Design
A = Baseline
B = Treatment
A = Stop Treatment
Return toward baseline but not completely due to any lingering effects
A-B-A-B Design
A = Baseline
B = Treatment
A = Stop Treatment
B = Repeat Treatment
Shows Intrasubject replication within the same treatment
A-B-A-C
A = Baseline
B = Treatment
A = Stop Treatment
C = New Treatment
Flexible and can treatments can be modified
Punishment
Anything that decreases the way of response
For example: giving a child a cookie makes them do something less, which results in a punishment for a psychologist.
Multiple Baseline Designs
Behavior A ———-> Treatment (Ex…Say Thank You)
Behavior B ———–> Treatment (Ex…Say Please)
More than one behavior implements the changes that are supposed to effect the behavior at different times for determining which treatment was responsible for the behavior change
Multi-Element Design
Like A-B-A-C Design but does not return toward the baseline
Clinical Psychologists may provide treatment to a client which will be tested to determine if the treatment is working
Advantages of Non-experimental Research Designs
- Good for looking at an individual’s behavior
- smaller effects found are likely to be important
*Flexibility for implementing treatments or modifying their strength - No Statistics
Disadvantages of Non-Experimental Research Design
*May not be able to meet requirements (effects are reversible/baseline)
* effective counterbalancing is impossible (cannot with one person like “blue pill” vs. “red pill”
*may not be able to find small but important effects
* hard to tell how general effects are (hard to tell if how general effects are (hard to tell if it works for everyone or just Bob only)
Single-Variable Research
Population Surveys
* Censuses (everyone)
* Survey research (sample)
* Population Survey (college students)
a. cluster sampling (individuals from different groups) Example:
survey of CA college students, making sure I get students from every
college clustering them together, 100 from each major UC and 100 from
each State University.
* Epidemiological research (health)
* Public opinion research
a. marketing research
Collecting data on one variable only
Multi-Variable Research
- Correlational Methods (relationship & does not
explain why?)
a.) 3rd variable problem (confound)
i. person
ii. environment
iii. operational definitions
b.) Directionality (does A cause B or B cause A)
More than one variable together
Kind of Square
Way of measuring association between nominal and ordinal variables
