ANOVA between + within

Question 1

between group vs within groups variance

what do they contain

Answer 1

Between
- effect of independent variable
- random error

within
- random error (individual diff)

Question 2

H0 and H1 predictions for ANOVA

Answer 2

H0 = SS between/SSwithin equal to 1
H1 = SSbetween/SSwithin greater than 1

Question 3

When to use t-test vs ANOVA

Answer 3

t-test 1 or 2
ANOVA = 3 groups

Question 4

familywise error is

Answer 4

probability of committing at least 1 type1 error across stat tests
1-(1-alpha)n

to the power of n (number of comparisons)

Question 5

Type 1 error

Answer 5

false positive
accepting that 5% of time, incorrectly reject the null hypothesis

Question 6

ANOVA is

Answer 6

analysis of variance
- ratio of variance accounted for by independent variable

Question 7

Between-subjects ANOVA assumptions

Answer 7

1. dependent variable interval/ratio
2. population normally distributed
3. equal variance of samples - homogeneity (levene’s non-significant is good)
4. independence of observation

Question 8

between ANOVA steps

Answer 8

1. sums of squares (grand mean, group mean)
   - ss total
   - ss between
   - ss within
2. degrees of freedom - df total N-1, dfbetween k-1, df within (number in each group-1) added
   - k =number of groups
3. mean squares
4. f ratio
5. p value

sstotal-within = betwee
ssbetween+within = total

ss total-within=between

Question 9

F-table

need df 1, df 2

Answer 9

df1 across top row is dfbetween
df2 down the side column is dfwithin

Question 10

ANOVA cohen’s d

both within and between

Answer 10

n2
- small .01
- medium .06
- large .14

Question 11

ANOVA within
ERROR PARSING IS

Answer 11

remove individual differences variance from SSwithin
- more accurate estimate of error
- f-ratio bigger
- greater chance of rejecting null

Question 12

ANOVA within
effect size writing

Answer 12

0.79
79% of variance in DV is accounted for by IV
removed/partialled out individual differences

Question 13

ANOVA Within assumptions

Answer 13

1. DV interval or ratio scale
2. populations norm distributed
3. independent of observations in each IV
4. sphericity - variances are equivalent
   - Mauchleys test need it non-significant

Question 14

ANOVA

which groups are different tho?

at least one group/condition is different from the other

Answer 14

post-hoc tests
- bonferroni (t-tests with smaller aplha = alpha/number of comparisons)
- holm’s stepdown until retained