Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
12 Cards in this Set
- Front
- Back
what 1-way ANOVA does
|
finds level-effect variance
|
|
1-way ANOVA variables
|
1DV, 1 IV (2 or more levels)
|
|
1-way ANOVA assumptions
|
homoskedasticity
normally distributed scores independence of scores |
|
ANOVA f-value
|
f<(or equal to) 1 when H0
f > 1 if H1 |
|
1-way ANOVA post-hoc - assume no effect
|
more conservative, increased type 2 error
bonferroni - estimation sheffe - very conservative sidak - more exact tukey's HSD - studentised t and studentised range distribution |
|
studentised range distribution (SRD)
|
1 value that all comparisons must exceed for sig.
2 samples from N.D. - diff. in means, repeat to find distribution of diff. 3 sample from N.D. - diff. between largest and smallest, repeat |
|
1-way ANOVA post-hoc - assume effect
|
less conservative, increased type 1 error
PLSD - least conservative, p-value for each comparison (like t-test) Neuman-Keuls - similar to Tukey's HSD but rank the means, tells you where diff's lie |
|
which 1-way ANOVA post-hoc?
|
depends on study and previous research
|
|
R^2
|
proportion of variance explained by all factors (and interactions)
|
|
eta^2
|
proportion of variance explained by a factor
|
|
adjusted r^2 (and omega^2)
|
estimate
proportion of variance explained by a factor in the population |
|
as no. levels increases
|
chance of type 1 error increases
|