the variable that we measure
dependent variable
alpha level typical in psychology
.05
this hypothesis predicts no difference
null
design that involves determining the extent to which two or more variables are related to one another
correlation
correlational design
The only appropriate statistic when you have more than one independent variable
ANOVA (specifically factorial ANOVA)
what is captured in the numerator of all of our statistical tests
the effect of the IV, treatment effect, differences between groups
in a factorial design, the overall effect of one IV variable on the DV, ignoring the second IV
main effect
when another variable gets confused with the independent variable (bc it is correlated with it), hurting inferences of causality
confound
when people are assigned to one and only one level of the independent variable
between-subjects design
Compares a sample mean to a population mean
One-sample t
"finding" an effect where none exists
Type 1 error
what it means when you reject the null hypothesis
you conclude that there was a significant effect
the ability to see an effect when an effect exists
power
what the design is called when you have one independent variable
one way design
Compares means from two different samples to determine if they are different from one another
Independent samples t
when the effect of one variable depends on levels of the other
an interaction
the extent to which you are certain that the IV caused changes in the DV
internal validity
a critical element of experimental design that ensures that no outside variables influence your study as it is being conducted
experimental control
when participants complete all levels of the IV
within-subjects
repeated measures
Appropriate for a pre-test/post-test design.
Paired samples t-test
a technique to combat order effects in within subjects designs
counterbalancing
readers will think they should encourage study abroad (as an IV), but people who can study abroad are different than those who can't...self-selection confound, non-equivalence confound, quasi-experimental design...
a critical element of experimental design that ensures that individual differences are controlled
random assignment
design involves a comparison between pre-existing groups (like men and women)
quasi-experimental
Appropriate when you have more than two levels of an independent variable
ANOVA
a confound in which the IV is confused with another variable or event that occurred at the same time as the IV
history confound
failing to find an effect that exists in the world
Type 2 error
what the denominator of our statistical equations captures
ERROR
the confound associated with quasi-experimental designs
non-equivalence
self-selection