PSY248 Lecture Notes - Lecture 1: Concurrent Validity, Intraclass Correlation, Informed Consent
2 Jun 2018
School
Department
Course
Professor
Revision of Design and Stats material: PSY248
Research question → what we want to find out; human behaviour
• Will always have some sort of theoretical explanation; address scientific
theories
o Must be falsifiable = scientific evidence
o Theory-driven research = best
Experiment vs. quasi-experiment → research hypothesis vs. null hypothesis
• Null = introduction of any variable will have no effect on another variable
• Research hypothesis = proposes that the introduction of one variable will lead
to change in some other measured variable
• Testing hypothesis through studies
o Experiments → you have the potential to randomly allocate
observations to conditions
o Quasi-experiments → unable to randomly allocate conditions
o Correlational studies (no manipulation/IV) – looking at two naturally
occurring variables and seeing its relation → continuous vs. invariant
Between group design vs. repeated measures
• Within-subjects design (aka repeated measures) – measuring same thing on
day 1 and day 2 (comparison of one group)
• Between-group design – one group gets one type, other group gets a different
type of practice
• Only BETWEEN-GROUPS design is possible in quasi-experiments – it
cannot work within-subjects (i.e. you cannot say you are male on one day and
female on the next day)
Independent vs. dependent variables:
• IV = manipulation (e.g. practice schedule)
• DV = measuring (e.g. skill level of a skill)
• You do not want your extraneous variable to become a potentially
confounding variable
• Extraneous variables = always going to be there, but we want to ensure that
these variables do not become a potential confound
Confounding variables:
• Validity = measuring what I claim to measure when designing a study
• Internal validity of experiments → isolation of relationship between IV and
DV (isolate through random allocation/random order)
• Alternatives to random allocation includes COUNTERBALANCE
• Counterbalance means that every participant will do an equal task
• External validity → refers to whether we can generalize from a sample to the
population from which the sample was drawn
• Population and ecological external validity
o Ecological validity = generalizing to neighbouring populations
find more resources at oneclass.com
find more resources at oneclass.com
