ANOVA Part 2

Question 1

What do one-way ANOVAs do?

Answer 1

Compare the means of two or more groups, and assess whether there are significant differences between the groups.

Question 2

One-way ANOVAs only assess what? and not what?

Answer 2

One-way ANOVAs can only assess main effects, and not interactions

Question 3

Variance due to the factor is known as what?

Answer 3

‘Between Variance’

Question 4

Error Variance is known as what?

Answer 4

‘Within Variance’

Question 5

A one-way ANOVA conducted on a factor with two or more levels is mathematically equivalent to what?

Answer 5

A t-test
t (squared) = F

Question 6

How do you calculate ‘Within Variance’

Answer 6

1. Calculate means for each level of your factor
2. subtract each memory score value from the corresponding mean
3. square the mean difference
4. add them all up

Question 7

How do you calculate ‘between variance’

Answer 7

1. calculate the mean of all observations (grand mean)
2. subtract the group of means frim the grand mean + square them
3. multiply the difference by the group sample size
4. add them all up

Question 8

How do you calculate the degrees of freedom for between variance

Answer 8

K - 1
(K = comparison groups)

Question 9

How do you calculate the degrees of freedom for within variance

Answer 9

n - K
(n = sample size)

Question 10

How do you carry out an F-test for a one-way ANOVA

Answer 10

1. Input the sum of squares
2. calculate the degreed of freedom
3. divide the sum of squares by the corresponding degrees of freedom
4. divide your MS-between from your MS-within
5. Compare the F-value + the F-critical

Question 11

What is the drawback to a one-way ANOVA?

Answer 11

While it can tell us that a significant difference exists, it CANNOT tell us where.
it’s an omnibus test.

Question 12

What do contrasts do?

Answer 12

They combine the group means based on the comparisons we want to make

Question 13

What is a planned comparison?

Answer 13

A scientifically sensible comparison that is decided before the experiment is conducted
* You can conduct multiple comparisons without correction provided they assess a single and scientifically sensible hypothesis

Question 14

What is a post-hoc comparison?

Answer 14

A comparison that is decided on after the experimental data has been assessed
* You should correct for multiple comparisons when they were decided on after looking at the data
* There are exceptions, but for the purpose of this course you should indicate that post-hoc comparisons are always corrected

Question 15

What are the three primary assumptions for one-way ANOVA

Answer 15

• normality
• homogeneity of variance
• independence

Question 16

How do we measure/test normality

Answer 16

Shapiro-Wilk test
- significance = not normal
- non-significant = normal

Question 17

How do we test for homogeneity of variance?

Answer 17

Levene test
- Significance = heterogeneous
- Non-significant = homogeneous
(this is for between subjects not within)