Test Four Flashcards
What is a factorial design?
A design with more than one level of independent variable.
In factorial designs, what are factors and categories?
Factors are what each level of the IV is called, categories are a combination of levels.
How do you write factorial designs?
The number of integers is the number of IVs, and their value is the number of factors. (ex. 2 x 3 design)
What’s the purpose of factorial designs?
Testing more levels of IV make it more like real life–to test the generalizability of a finding.
What can factorial designs reveal? (2)
Main effects and interactions
What types of interactions are there? (2)
Crossover and spreading
What are the two types of variables in factorial designs?
Manipulated and participant variables
What do interactions show?
Interactions show moderating variables
What are the two types of groups p’s can be assigned to in factorial designs?`
Independent (between subjects) and correlated (within subjects)
How many times are p’s treated in independent groups? In corrlated groups?
1 in independent, >1 in correlated
What is mixed assignment?
Some p’s are put in an independent group, some are put in a correlated group
How does a quasi-experiment differ from a regular experiment? (3)
1) Nothing is manipulated
2) IV varied before researcher arrived
3) No random assignment
Can IVs be repeated in quasi-experiments?
Yes, but not at the will of the researcher
What threats are there to internal validity? (7)
Selection, maturation, history, regression to the mean, attrition, testing, and instrumentation
How can selection threats be avoided in quasi-experimental designs? (2)
Repeated measures and better comparison groups
How can maturation threats be avoided in QEDs?
Better comparison groups
How can testing threats be avoided in QEDs?
Good comparison groups
How can instrumentation threats be avoided in QEDs? (2)
Good comparison groups and interrupted time series
What are threats related to human subjectivity that could be present int QEDs? (3)
Observer bias, placebo effect, and experimental demand
As stated in the power points, what are two other validities that are critical for QEDs? (2)
Construct validity and statistical validity
What are three priorities you should have in mind when conducting QEDs?
Opportunity, External validity, and Ethics
How do small-N designs differ from large-N designs? (3)
1) Each p represents their own experiment; usually invokes repeated measures
2) individuals’ data presented
3) replicability determined by repeating test in another subject
When would one use small-N designs? (3)
1) special or small populations
2) research is exhausting in terms of resources
3) when single negative response would refute a theory
What are three designs that are employed for small-N designs?
Stable-baseline, repeated-baseline, and reversal
How does one conduct a stable-baseline design?
Multiple baseline measurements -> intervention -> more measurements
How does one conduct a repeated-baseline design?
Baseline measure -> Intervention -> more measurements -> new component added to baseline -> intervention -> more measurements, etc.
How does one conduct a reversal design?
Baseline measurement -> intervention -> more measurements -> removal of treatment -> more measurements
What are ethical considerations of reversal designs? (2)
1) Is it ethical to remove the treatment?
2) Is it ethical to use a treatment that isn’t empirically proven to work?
How is external validity expressed in small-N designs? (4)
1) Replicate on other similar patients
2) study in animals
3) use other techniques in larger pops
4) No expectation results will generalize to everyone
How can construct validity be measured in small-N designs?
Interrater reliability
How important is statistical validity in small-N designs?
Not as much
How can statistical validity be expressed in small-N designs? (2)
1) Effect size is measured per case
2) Graphs provide quantitative data
What must a study do to be important?
Be replicable
What are the three types of replication studies?
Direct replication, conceptual replication, and replication with extension
What is direct replication?
An exact replication of the study
What is conceptual replication?
A study that uses the same variables and studies the same questions, but those variables are operationalized differently
What is a replication with extension?
A mix of direct and conceptual replication
What are two variables that can be changed for replicability’s purposes?
Participant and situational variables
What is a meta-analysis?
An article that summarizes and reviews previously conducted studies.
What are important features of meta-analyses? (3)
1) Collect all possible examples of a particular kind of study
2) Average of all effect sizes
3) can sort groups into categories to see how effect sizes differ
What’s a problem with meta-analyses?
Findings that have been published may not reveal the entire picture.
Can you generalize to other participants?
Yes, because they’re A population, but it depends on HOW they were obtained.
Can you generalize to other settings?
Yes, and conceptual replications can provide proof for this.
What is theory-testing mode?
Testing association or causal claims to test a theory
What is generalization mode?
Goal of generalizing results from a sample to a population.
What are some results on the study of the Muller-Lyer Illusion?
Ps that grew up in a “carpentered society” were more likely to fall for the illusion.
What types of subjects are usually tested in psychology?
WEIRD subjects? (Western, Educated, Industrialized, Rich, and Democratic)
