PSY 1004 Lecture Notes - Lecture 7: Analysis Of Variance, Dependent And Independent Variables, Nonprobability Sampling
○Main Effects: effect of one IV, averaged over the effects of everything else (ex:
Main effect of A)
■Post Hoc Test may need to be done
○Interactions: effect of one IV differs across levels of another IV (ex: AxB
interaction)
○Example: Does facial expression (smiling, neutral) and attractiveness (attractive,
unattractive) affect judgment of guilt in others?
■Main Effects – look at
●Column means – vertical
●Row means – horizontal
■Interaction – look at cell means
●8 Possible Outcomes for 2X2 Factorial
○Interactions and Main Effects:
■It is possible to have an interaction with no main effect and visa versa
■There are 8 prototypical patterns for a 2X2 factorial.
■Example: Studying the degree of attraction in couples as a function of
distance
(near, far) and love
(true love, not true love)
●LOOK AT GRAPHS AT THE END OF SLIDE
Topic 6
You can have an interaction without a main effect, vice versa, they are not dependent on each
other
Sampling
●Population: Entire collection of individuals considered
●Sample: Subset of population
●Statistics: The study of interpreting empirical (observable) data
○Organize data
○Summarize data
○Interpret results
●Parameter: value from population
●Statistics: value from sample
●Ideal is to select a representative sample of the population
●Law of large numbers: A large sample will probably be more representative than a
small sample
○25 to 30 individuals for each group/treatment
Document Summary
Main effects : effect of one iv, averaged over the effects of everything else (ex: Post hoc test may need to be done. Interactions: effect of one iv differs across levels of another iv (ex: axb. It is possible to have an interaction with no main effect and visa versa. There are 8 prototypical patterns for a 2x2 factorial. Example: studying the degree of attraction in couples as a function of distance (near, far) and love (true love, not true love) Look at graphs at the end of slide. You can have an interaction without a main effect, vice versa, they are not dependent on each other. Statistics: the study of interpreting empirical (observable) data. Ideal is to select a representative sample of the population. Law of large numbers: a large sample will probably be more representative than a small sample. 25 to 30 individuals for each group/treatment by the researcher.