1/45
Looks like no tags are added yet.
what is the hallmark feature of a one way anova?
1 IV with 2+ levels, 1 DV
what is an example of a one way anova?
looking at differences in prayer times among Christians, Jews, and Muslims
what is the benefit to doing a one way anova over a t-test?
t-tests would require more calculations, leading to multiple comparisons (increased risk of type 1 error)
what is a bonferroni correction?
multiply p value by the # of additional comparisons to help estimate false positives
what are the assumptions of a one way anova?
random sampling, independent data, normality of residuals, homogeneity of variance, categorical iv, continuous dv
broken down, what is the F ratio at base?
between variance / within variance (treatment variance / normal variance) (signal / noise)
on a graph, there is large signal (between groups variance) when gaps between groups is ____
big
large F ratio = ___ p-value
small
what is the main limitation of a one way anova?
it doesn't tell you which groups differ
what does this code represent: model
one way anova
An anova on 2 groups and a two-tailed independent samples t-test give what?
the exact same p-value *this will be on test*
what is the difference between a post hoc and a planned comparison?
a post hoc compares all groups and a planned comparison only compares groups stated
T/F: you only conduct a post hoc when there is significance
True
T/F: you only conduct a planned comparison when there is significance
False
what does eta squared represent (η²)
effect size (proportion of total variance in a dependent variable that is explained by an independent variable)
*how much was difference was due to our intervention*
what is a perfect account of variance?
eta = 1
what is partial η²?
proportion of variance while excluding other factors
T/F: you can have 2 groups with significant difference without the whole ANOVA having significant difference
True
what is another name for a planned comparison?
a priori
what is another name for a post hoc?
posterior
what is the standard post hoc run in R?
Tukey HSD
what does this code represent:
marginal_means
Tukey HSD post hoc
what does this code represent:
pairs(marginal_means, adjust = "none")
Fisher's LSD post hoc (no protection)
what does this code represent:
pairs(marginal_means, adjust = bonferroni)
Bonferroni post hoc (most conservative)
T/F: a planned comparison is more specific than a post hoc
true
what does this code represent:
contrast(marginal_means, method = list(c(1, -.5, -.5))
a planned comparison looking at group 1 compared to the other two groups
in planned comparison code, the numbers should always sum to what?
0
what is a factorial anova?
2+ independent variable with 2+ treatment levels and 1 DV
how many IVs and respective levels are in this anova: 2x2x4x2
4 IVs (bc 4 numbers) levels are 2, 2, 4, 2
what is a main effect?
significant effect of one independent variable ignoring other independent variables
what is an interaction?
When the effect of one IV on the DV depends on the level of another IV
give an example of an interaction
alcohol and painkillers both make you feel good but when combined can kill you
In a study looking at background music presence and task difficulty on test scores, there is no marked difference in scores for easy tests, regardless of room. however, there is a difference between medium test scores when in a silent room vs a music room. this is an _____
interaction
what happens when doing a study that has too many IVs and too many levels?
complexity kills your statistical power and it becomes way too hard to explain
T/F: you can have both main effects be significant but have no significant interaction
true
T/F: you can have multiple interactions in 1 study
true
on a factorial anova graph, lines that cross or have different slope show a what?
interaction
on a factorial anova graph, perfectly parallel lines show a what?
main effect only
what does this code represent:
model
factorial anova
what is an ancova?
analysis of covariance; anova while controlling for a covariate
what is a covariate?
thing that could interfere with data (could become confound)
give an example of an anconva
comparing Teaching Method A vs. Method B on final exam scores (dependent variable) while controlling for pre-test scores (covariate)
what are the requirements for an ancova?
continuous DV
continuous CV (usually)
categorical IV
CV correlated with DV (linear relationship needed)
CV NOT correlated with IV
DV normally distributed
homogeneity of variance
what are some applications for ancova?
1. increase test sensitivity
2. increase control (helps control for no randomization)
3. building a model
what does this code represent:
model
ancova
what does this code represent:
lines(lowess(dfCV,dfDv), lwd = 2, lty = 2
creating locally weighted regression for ancova